Evolution of Early Earth's Atmosphere, Hydrosphere, and Biosphere, Constraints From Ore Deposits

Downloaded from memoirs.gsapubs.
org on February 23, 2010
Preface
INTRODUCTION
This volume was stimulated by a Pardee Symposium titled “Evolution of the Early Atmosphere, Hydro-
sphere, and Biosphere: Constraints from Ore Deposits,” which we convened in 2002 at the national meeting
of the Geological Society of America. The history of Earth’s early atmosphere, hydrosphere, and biosphere,
from Hadean through Archean and into Proterozoic time, is one of the enduring puzzles in the geological
sciences. When did the oceans appear, and did they remain liquid throughout Earth’s history? What was
the composition of the early atmosphere, and how did it affect climate? How did the atmosphere and ocean
compositions change through time, and why? When, where, and how did life emerge on Earth? When did
cyanobacteria, sulfate reducers, methanogens, and eukarya appear, and how did they affect their geologic
environments? How did changes in the atmosphere, hydrosphere, and biosphere affect the lithosphere, and
vice versa?
One of the most prominent aspects of this puzzle is when Earth’s atmosphere became oxic. Controversy
today focuses on two possibilities: that the atmosphere has been oxic since early Archean time (ca. 3.8 Ga)
or that it gained oxygen between about 2.3 and 2.1 Ga, an event recently termed the Great Oxidation Event
(GOE). The availability of free oxygen would have had a strong effect on the (bio)geochemical cycles of
elements that exist in more than one oxidation state in nature. Of the major rock-forming elements, only
iron does this. In contrast, multiple oxidation states are common in nature for many of the trace elements
that are concentrated in some ore deposits, such as manganese, molybdenum, uranium, and vanadium. The
atmospheric concentrations of CO2 and CH4 are also of concern as potential greenhouse gases to resolve the
faint young sun problem, as the sources and products of biological activities, and as the source of acid rain for
weathering the early continents. Finally, the concentrations of various forms of sulfur species in the oceans,
especially H2S and SO42–, are of interest because they are linked to the atmospheric pO2 history, the evolu-
tion of a variety of sulfur-utilizing microbes (e.g., sulfate-reducing bacteria and sulfide-oxidizing bacteria),
and the origins of a variety of mineral deposits, including banded iron-formations, volcanogenic massive
sulfide, sedimentary exhalative, and Mississippi Valley–type (MVT) deposits. The abundance ratios of many
of the elements in ore deposits respond to the oxygen-carbon-sulfur geochemistry of the atmosphere and
oceans, making ore deposits particularly good indicators of their geochemical environment.
Historically, the first ore deposit types linked to atmosphere-hydrosphere compositions were uranium
paleoplacers and banded iron formations. Later work has expanded the list to include other types of uranium
deposits, sedimentary manganese deposits, laterites, and sedimentary exhalative and MVT lead-zinc depos-
its. Papers in this volume deal with most of these deposit types and include new research results, as well
as summaries of the current opinions on how they relate to proposed histories of Earth’s early atmosphere,
hydrosphere, and biosphere.
SUMMARY OF THE VOLUME
This volume includes papers based on some of the presentations at the symposium, as well as additional
papers. It starts with a section dealing with the biosphere and the origin of life. Here, Russell and Hall show
that low-temperature hydrothermal systems could have constituted simple reactors in which H2 and CO2
formed acetate, complex organic molecules and even cells (i.e., the emergence of life) using a froth-like
substrate of iron-sulfides. Next, Grassineau et al. use sulfur and carbon isotope geochemistry of Archean
v
Downloaded from memoirs.gsapubs.org on February 23, 2010
vi Preface
rocks and ore deposits to trace the metabolic evolution of life from 3.8 to 2.7 Ga. The second section focuses
on development of the early continents and begins with the study by Thiart and de Wit demonstrating that
extraction of ore elements from the mantle has become less efficient through time. This is followed by a
study by Ishihara et al. showing that the range in oxidation states of Archean granitoid intrusions is essen-
tially the same as that of younger granitoids, and a final study by Kerrich et al. indicating that atmospheric
N2 has been drawn into ore deposits and other parts of the lithosphere through geologic time by N-fixing
microorganisms.
The third section of the volume, which focuses on uranium deposits and their relation to atmospheric
evolution, starts with two contrasting views of the famous Mesoarchean Witwatersrand Supergroup in South
Africa that contains Earth’s largest gold-uranium deposits. These deposits are hosted by quartz-pebble con-
glomerates, and controversy centers on whether they formed as paleoplacers. In the first paper, Minter shows
that the sedimentary setting of gold, pyrite, and uraninite favors a detrital origin and their deposition under
a reducing atmosphere. In the second paper, Law and Phillips describe geologic and geochemical features
indicating that the gold, uranium, and sulfur were deposited by hydrothermal fluids after deposition of the
conglomerates and conclude that these deposits do not provide evidence for a reducing Archean atmosphere.
In a third paper, Yamaguchi and Ohmoto argue against a paleoplacer origin for pyrite in the slightly younger
Paleoproterozoic Huronian Supergroup in Canada, which also contains quartz-pebble conglomerates that
host only uraninite and pyrite, but no gold. In the final paper, Gauthier-Lafaye shows that the well-known
Oklo sandstone-type uranium deposits of Gabon, which formed at ca. 2.1 Ga, are best explained if they
formed when oxygenated water was available to transport dissolved uranium.
The fourth section of the volume concerns lead-zinc and manganese deposits and the information that
they provide about the evolution of sulfur in seawater. The first two papers by Lyons et al. and Kesler and
Reich show that the temporal history of sedimentary exhalative (sedex) and MVT lead-zinc deposits, respec-
tively, is best explained by an ocean that lacked widespread, abundant sulfate until at least middle Proterozoic
time. Liu et al. show that Neoproterozoic Mn-carbonate deposits in China resulted from low levels of sulfate
in the ocean caused by an anomalous influx of Fe from lateritic soils rather than from a “snowball Earth.”
The final section of the volume deals with banded iron formation (BIF) deposits and the insights that
they provide about the geochemistry of iron, oxygen, sulfur, and carbon and the nature of organisms in the
oceans. The paper by Raiswell deals with the relative importance of weathering, diagenetic, and exhalative
processes as sources of iron. Brown discusses the problem of depositing iron and concludes from experi-
ments and energy considerations that microbes could have accounted for large volumes of iron carbonate and
oxide precipitates, and Gutzmer et al. show that high-grade hematite BIF deposits formed under conditions
similar to those of today. In the final two papers, Kato et al. and Ohmoto et al. suggest that BIFs formed
throughout geologic history by the mixing of locally discharged hydrothermal fluids with ambient seawater
and show that this is consistent with an ocean that has been oxygen- and sulfate rich since ca. 3.8 Ga.
REMAINING CONTROVERSIES
It will be apparent to even the most casual reader that authors who have contributed to this volume hold
a wide range of views about the composition of the early Earth atmosphere, hydrosphere, and biosphere and
the constraints offered by ore deposits, and we regard this as one of the major contributions of the volume.
Our sincere hope is that the range of views propounded in this volume will help students and researchers to
realize the nature of the remaining controversies and the sorts of data that are required to resolve them. For
example, why do Lyons et al. and Kesler and Reich conclude that the early ocean was low in sulfate whereas
Ohmoto et al. conclude that it was high in sulfate? What sort of evidence do they have, how did they interpret
this evidence, and what sorts of new studies are necessary to test their assumptions and conclusions? What
new methods can be used to resolve these controversies?
Our understanding of early Earth conditions has evolved tremendously over the past few decades, and
we expect that this will continue in the future, spurred in part by studies inspired by this volume. As this
research continues, three points are apparent. First, in the final analysis, our conclusions are based largely on
the rock record, and that record is complex. Mineralogical and textural features of undisturbed rocks are dif-
ficult enough to understand, but things get much harder with older rocks because in general, they have been
subjected to higher degrees of metamorphism and/or to longer periods of weathering. As geochemical stud-
ies become more and more sophisticated, it is increasingly important that the utmost care be taken to assure
Preface vii
that samples truly represent the event or feature of interest and to document the textures and features that
prove it. Second, the abundance of Precambrian rocks, especially those formed in near-surface and active
tectonic settings, decreases with increasing geologic age because most of them were eroded or destroyed by
later tectonic events, but also perhaps because the earlier continents were smaller. Therefore, more accurate
reconstructions of paleogeographic and tectonic settings of the major Archean and Proterozoic terrains are
necessary in order to relate the temporal change in the preserved number/size of a specific type of ore deposit
to the chemical and biological evolution of early Earth. Third, there is growing evidence that Earth’s ore
deposit history rarely reflects a single dominant control and instead is result of complex interaction among
many factors. These can be classified in many ways, but are perhaps best viewed as the tectonic, hydrologic,
and biologic cycles. Our growing understanding of earth system science shows how intimately these cycles
are related and it comes as no surprise that ore deposits reflect this interplay. Hopefully, when the successor
to this volume is compiled, continuing research will have brought us closer to an understanding of the con-
straints that ore deposits provide on early Earth’s atmosphere, hydrosphere, and biosphere.
ACKNOWLEDGMENTS
The symposium that stimulated this volume was supported generously by the Geological Society of
America, the Society of Economic Geologists and the NASA Astrobiology Institute. We are grateful to all of
these for their interest in this important subject.
Stephen E. Kesler
Hiroshi Ohmoto
Contents
Preface .........................................................................................................................................................v
Origin of Life
1. The onset and early evolution of life ...................................................................................................1

M.J. Russell and A.J. Hall
2. Early life signatures in sulfur and carbon isotopes from Isua, Barberton,
Wabigoon (Steep Rock), and Belingwe Greenstone Belts (3.8 to 2.7 Ga) .......................................33
N.V. Grassineau, P. Abell, P.W.U. Appel, D. Lowry, and E.G. Nisbet
Evolution of the Early Continents
3. Fingerprinting the metal endowment of early continental crust to test for secular changes
in global mineralization .....................................................................................................................53
C. Thiart and M.J. de Wit
4. Discovery of the oldest oxidized granitoids in the Kaapvaal Craton and its implications
for the redox evolution of early Earth...............................................................................................67
S. Ishihara, H. Ohmoto, C.R. Anhaeusser, A. Imai, and L.J. Robb
5. Secular variations of N-isotopes in terrestrial reservoirs and ore deposits .....................................81

R. Kerrich, Y. Jia, C. Manikyamba, and S.M. Naqvi
Evolution of Uranium Deposits and Atmospheric Evolution
6. The sedimentary setting of Witwatersrand placer mineral deposits

in an Archean atmosphere...............................................................................................................105
W.E.L. Minter
7. Witwatersrand gold-pyrite-uraninite deposits do not support a reducing

Archean atmosphere ........................................................................................................................121
J. Law and N. Phillips
8. Evidence from sulfur isotope and trace elements in pyrites for their multiple
post-depositional processes in uranium ores at the Stanleigh mine, Elliot Lake,
Ontario, Canada...............................................................................................................................143
K.E. Yamaguchi and H. Ohmoto
9. Time constraint for the occurrence of uranium deposits and natural nuclear fission
reactors in the Paleoproterozoic Franceville Basin (Gabon).........................................................157
F. Gauthier-Lafaye
iii
iv Contents
Evolution of Seawater and Basinal Sulfur Geochemistry
10. Proterozoic sedimentary exhalative (SEDEX) deposits and links to evolving

global ocean chemistry ....................................................................................................................169
T.W. Lyons, A.M. Gellatly, P.J. McGoldrick, and L.C. Kah
11. Precambrian Mississippi Valley–type deposits: Relation to changes in composition

of the hydrosphere and atmosphere ................................................................................................185
S.E. Kesler and M.H. Reich
12. Superheavy S isotopes from glacier-associated sediments of the Neoproterozoic

of south China: Oceanic anoxia or sulfate limitation? ..................................................................205
Liu Tie-bing, J.B. Maynard, and J. Alten
Evolution of Seawater Iron and Oxygen Geochemistry and Banded Iron Formations
13. An evaluation of diagenetic recycling as a source of iron for banded iron formations ...............223
R. Raiswell
14. Microbially mediated iron mobilization and deposition in iron formations since
the early Precambrian .....................................................................................................................239
D.A. Brown
15. Oxygen isotope composition of hematite and genesis of high-grade BIF-hosted iron ores .........257
J. Gutzmer, J. Mukhopadhyay, N.J. Beukes, A. Pack, K. Hayashi, and Z.D. Sharp
16. Rare-earth elements in Precambrian banded iron formations: Secular changes

of Ce and Eu anomalies and evolution of atmospheric oxygen .....................................................269
Y. Kato, K.E. Yamaguchi, and H. Ohmoto
17. Chemical and biological evolution of early Earth: Constraints from banded
iron formations ................................................................................................................................291
H. Ohmoto, Y. Watanabe, K.E. Yamaguchi, H. Naraoka, M. Haruna, T. Kakegawa, K. Hayashi,
and Y. Kato
Index ........................................................................................................................................................333
Geological Society of America

Memoir 198
2006
The onset and early evolution of life

Michael J. Russell*
Scottish Universities Environmental Research Centre, Rankine Avenue, East Kilbride, Glasgow G75 0QF, Scotland, and
Géosciences, LGGA, Université de Grenoble 1, 1381, rue de la Piscine, 38400 St. Martin d’Heres, France
Allan J. Hall*
Department of Archaeology, University of Glasgow, G12 8QQ, Scotland
ABSTRACT
The tension between CO2 dissolved at relatively high atmospheric pressure in

the Hadean ocean, and H2 generated as ocean water oxidized ferrous iron during
convection in the oceanic crust, was resolved by the onset of life. We suggest that
this chemosynthetic life emerged within hydrothermal mounds produced by alkaline
solutions of moderate temperature in the relative safety of the deep ocean floor. Exo-
thermic reaction between hydrothermal H2, HCOO - and CH3S - with CO2 was cata-
lyzed in inorganic membranes near the mound’s surface by mackinawite (FeS) nano-
crysts and “ready-made” clusters corresponding to the greigite (Fe5NiS8) structure.
Such clusters were precursors to the active centers (e.g., the C-cluster, Fe4NiS5) of a
metalloenzyme that today catalyzes acetate synthesis, viz., the bifunctional dehydro-
genase enzyme (ACS/CODH). The water, and some of the acetate (H3C.COO -), pro-
duced in this way were exhaled into the ocean together as fluid waste. Glycine (+H3N.
CH2.COO -) and other amino acids, as well as tiny quantities of RNA, generated in
the same milieu were trapped within tiny iron sulfide cavities.
Energy from the acetate reaction, augmented by a proton gradient operat-
ing through the membrane, was spent polymerizing glycine and other amino acids
into short peptides upon the phosphorylated mineral surface. In turn these peptides
sequestered, and thereby protected, the catalytically and electrochemically active
pyrophosphate and iron/nickel sulfide clusters, from dissolution or crystallization.
Intervention of RNA as a polymerizing agent for amino acids also led to an adven-
titious, though crude, process of regulating metabolism—a process that was also to
provide genetic information to offspring. The fluxes of energy and nutrient available
in the hydrothermal mound—commensurate with the requirements of life—encour-
aged differentiation of the first microbes into two separate domains. At the bifurca-
tion the Bacteria were to specialize in acetogenesis and the Archaea into methanogen-
esis. Representatives of both these domains left the mound by way of the ocean floor
and crust to colonize the deep biosphere.
Once life had emerged and evolved to the extent of being able to reduce nitro-
gen for use in peptides and nucleic acids, light could have been used directly as an
energy source for biosynthesis. Certain bacteria may have been able to do this, where
protected from hard UV by a thin coating of chemical sediment produced at a sub-
aerial hot spring operating in an obducted and uplifted portion of the deep biosphere.
*[email protected]; [email protected]
Russell, M.J., and Hall, A.J., 2006, The onset and early evolution of life, in Kesler, S.E., and Ohmoto, H., eds., Evolution of Early Earth’s Atmosphere, Hydro-
sphere, and Biosphere—Constraints from Ore Deposits: Geological Society of America Memoir 198, p. 1–32, doi: 10.1130/2006.1198(01). For permission to copy,
contact [email protected]. ©2006 Geological Society of America. All rights reserved.
1
2 M.J. Russell and A.J. Hall
Embedded in fresh manganiferous exhalites, early photosynthetic bacteria could fur-

ther protect themselves from radiation by adsorbing manganese on the membrane.
Organization of the manganese with calcium, within a membrane protein, happened
to result in a CaMn3O4 cluster. In Mn(IV) mode this structure could oxidize two mol-
ecules of water, evolve waste oxygen, and gain four electrons and four protons in
the process to fix CO2 for biosynthesis. All these biosynthetic pathways had probably
evolved before 3.7 Ga, though the reduced nature of the planet prevented a buildup
of free atmospheric oxygen until the early Proterozoic.
Keywords: CODH/ACS, greigite, origin of life, oxygenic photosynthesis, ranciéite.
By autogeny we understand the origin of a most simple organic individual in an inorganic formative fluid, that is, in a fluid which
contains the fundamental substances for the composition of the organism dissolved in simple and loose combinations (for example,
carbonic acid, ammonia, binary salts, etc.).
—Ernst Haeckel 1892, p. 414
INTRODUCTION fides, mackinawite and greigite—sulfides which accommodate

that effective and common catalytic metal, nickel. The primary
In a posthumous paper published in 1952, Goldschmidt pre- carbon dioxide, which composed a proportion of the atmosphere/
sented three principles to be adhered to in origin-of-life studies, ocean system (the volatisphere), is fixed by reaction with acti-
principles derived from his geochemical and mineralogical expe- vated hydrothermal H2 emanating from the highly reduced Earth
rience: to provide the basic organic molecules of life. Hydrogen, as a
carrier and donor of high-energy electrons, is the first fuel of life.
The first principle is that an a-biotic environment, poor in elementary And as soon as organic molecules are generated they can inhibit
oxygen, is suitable for the preservation and accumulation of ... organic
molecules. crystal growth. Indeed growth of inorganic clusters may be
The second principle is the collection, concentration and ordering of arrested at a very early stage by certain charged or polar organic
such molecules on free rectilinear planes, crystal faces, of minerals, giving molecules (Rickard et al., 2001). In some cases these clusters can
them possibilities for further mutual interaction between themselves and the
“basement” crystal. act as catalysts for further organic synthesis.
The third principle is the hypothesis that carbon dioxide (and its nearest Although the potential energy available for reaction between
derivatives) may be the primary material. the highly reduced Earth and its moderately oxidized volati-
sphere is substantial, the kinetic barriers are formidable (Shock,
These principles, consistent with early geological interpreta- 1992). When hydrothermal solutions first titrated with a sterile
tions of the moderate redox state of the early atmosphere, stand in prebiotic ocean, much of the thermal energy was effectively dis-
stark contrast to the assumptions of Oparin (1924, 1938), Haldane persed. Not so the chemical energy. Here we attempt to show
(1929), Urey (1952), Oró (1961), Deamer (1985), Joyce (1989), how a hydrothermal mound at moderate temperature focused and
Miller (1992), and Bada (2004) that life originated from the catalyzed the reaction between the main molecules fundamental
plethora of organic molecules supposedly delivered from space to life, H2 and CO2, and then fractionated, concentrated, and con-
or generated in a putative reduced atmosphere. Goldschmidt, tained the longer charged products. In so doing we rely on the
like the evolutionist Haeckel (1892) before him, inferred life to thermodynamic calculations and kinetic considerations of Shock
have emerged autogenically (i.e., from the simplest of inorganic and his collaborators in predicting the likely products of the ear-
substances), whereas Oparin assumed, and his followers assume liest metabolism (Shock, 1990, 1992; Amend and Shock, 1998;
still, a plasmogenic or organotrophic inheritance to explain their Shock and Schulte 1998; Shock et al., 1998). We also appreciate
RNA and lipid worlds (Bada, 2004). the operations of Geochemist’s Workbench in the presentation
In this contribution we develop Goldschmidt’s autogenic of relative stability fields in Pourbaix (Eh/pH) diagrams inter-
principles to show that evolutionary steps may be traced, though pretable by both geochemists and biochemists (Bethke, 1996).
uncharted in places, mechanistically from aqueous geochemistry Even these geochemical considerations must be given geologi-
and mineralogy, through chemosynthetic biochemistry to oxy- cal context. The emergence of life must be seen as a geological
genic photosynthesis. The abiotic environment we favor for the issue, as the first stage in the “evolution of species” and not some
accumulation and preservation of organic molecules is within separate conception to be examined merely by dismantling a bac-
FeS microcavities in a submarine hydrothermal mound. The terial cell and looking for the oldest bits (Leduc, 1911). And the
basement crystals collecting, concentrating, ordering, and pro- geochemical reactions must be translatable to an early plausible
moting mutual further interactions are the metastable iron sul- biochemistry. For example, reactions that make organic polymers
The onset and early evolution of life 3
only at temperatures above 200 ºC have little direct bearing on or serpentinite and protons are returned in its stead (Seyfried and
the problem. Life exists as the energy trap and catalyst for reac- Bischoff, 1981; Douville et al., 2002):
tions between reduced and oxidized components. Like a mineral
exploration geologist searching for a lithochemical aureole to an 2MgSiO3(enstatite) + Mg2+ + 3H2O →
orebody as an indicator of “spent” ore solutions (Russell, 1974), Mg3Si2O5(OH)4 (serpentine) + 2H+ (1)
in order to comprehend the process of emergence we have to
examine life’s waste products, the entropic sinks. Amongst these At the same time any sulfate is scrubbed out as magnesium
are the fluid wastes, water, acetate, methane, hydrogen sulfide, hydroxide sulfate hydrate is produced within the crust, further
and oxygen. These waste products continued to be dispersed, contributing to the low pH of high-temperature hydrothermal
then and now, by convection and advection, albeit at differing solutions (Bischoff and Seyfried, 1978; Janecky and Seyfried,
rates. The solid mineral wastes, especially the sulfides, are gener- 1983):
ally deposited close to the source.
So what were the geological and geochemical conditions in 2Mg2+ + 3H2O + SO42- → Mg(OH)2:MgSO4:H2O + 2H+ (2)
the Hadean that gave rise to life and its waste products?
In the near absence of sulfate (equation 2) but with springs
INITIAL CONDITIONS exhaling into an acidulous ocean there would have been no anhy-
drite and sulfide chimneys, and no black smokers in the Hadean.
As soon as the first ocean condensed and cooled around 4.4 Free sulfide concentrations in the hot fluids were relatively low.
Ga Earth was primed for life (Wilde et al., 2001; Russell and Hall, The little there was would have reacted with zinc to produce sta-
1997). But where on Earth could life have begun? Conditions ble ZnS and Zn2S3- clusters (Walker and Brimblecombe, 1985;
were anything but equable. The temperature of the oceans fluctu- Luther et al., 1999). Minor to trace quantities of other “biophile”
ated wildly. Large meteorites that partially vaporized the ocean elements such as Mn, Zn, Ni, Co, Mo, Se, and W (Goldschmidt,
increased atmospheric pressure so that the remaining water might 1937) would also have been delivered, with iron, to the Hadean
have reached 300 ºC or so. In lulls in the bombardment, high CO2 ocean through these high temperature springs at oceanic spread-
pressures could induce a 100 ºC greenhouse (Kasting, 1993), yet ing centers (Von Damm, 1990; Hemley et al., 1992). The iron
meteorite-induced dust clouds might, on occasion, have masked concentrations contributed in this way probably approached 20
radiation from the weak young sun. If so, a short-lived icehouse mmol. We base this estimate on analyses of the Rainbow hydro-
could have ensued (Nisbet and Sleep, 2001). However, condi- thermal system operating in ultramafic rock at the slow spreading
tions in this “water world” were generally tempestuous; the ocean Mid-Atlantic Ridge (Douville et al., 2002) (Table 1). Ocean floor
surface was no place to organize the first cell, and shorelines, spreading in the Hadean was unlikely to have been fast because
where they existed, would have suffered continual storms and of the inhibiting effects of a 30-km-thick crust produced by the
huge tides, a response to the closer moon rapidly orbiting about voluminous melting of a very hot and dry mantle, not to men-
an Earth whose day lasted a mere five hours. Darwin’s “warm
little pond,” if it was not swamped, would have been subject to
deleterious hard UV at eight times the present flux (Canuto et al.,
1982; Bahcall et al., 2001; Abe and Ooe, 2001). The only safe TABLE 1. COMPARISON BETWEEN HIGH AND MODERATE
place to be was on the ocean floor. TEMPERATURE SUBMARINE SPRINGS
Parameter Juan da Fuca Rainbow Lost City Eyjafjordur
Can we imagine a window of opportunity for life to onset? T °C 224 °C 365 °C 40°–90 °C 71.4 °C
After all, there was no shortage of the appropriate energies, both pH 3.2 2.8 d11 10.03 (at 24 °C)
H2 na 13 d15 na
physical and chemical. Hydrothermal convection currents within H2S 3.5 1.0 0.064 0.01
the thick, fractured, and permeable Hadean crust focused a chem- Fe 18.74 24 na 0.00014
ical disequilibrium between reduced iron (in ferrous minerals and Mn
Mg
3.58
0
2.25
0
na
9–19
0.0000018
0.01
as minor native iron) and the relatively oxidized volatisphere. Ca 96.4 67 22 0.061
Although there was chemical potential for reaction between the Na
K
796
51.6
553
20
482
na
3.4
4.2
H2, continuously emanating from Earth’s interior and CO2 in the SiO2 23.3 6.9 na 1.6
atmosphere and ocean, an electrochemical potential between CO2 G4.46 na na 0.57
SO4 0 (0) 5.9–12.9 0.2
redox couples H+/H2 (the hydrothermal state) and Fe3+/Fe2+ (the Cl 1087 380 548 1.26
state in the ocean) also obtained. Co na 0.013 na na
Ni na 0.003 na na
A geological phenomenon to excite an earth scientist imag- Zn 0.9 0.16 na na
ining the onset of life is the black smoker (Corliss et al., 1981). Mo na 0.000002 na na
Duration yr >1000 >1000 >30,000 ~11,000
Such a very hot spring, emanating at close to the critical point of Note: Data from Von Damm, (1990), Douville et al., (2002), Kelley et al.,
seawater, is acidic as a result of the serpentinization of pyroxenes (2001, 2005), Früh-Green et al. (2003), and Marteinsson et al., (2001).
Elemental and molecular concentrations are in millimoles. Note that most
at high temperature. In these conditions Mg2+ from ocean water submarine springs probably last for at least 100,000 years (e.g., Lalou et al.,
in the convective downdrafts is precipitated as brucite (Mg(OH)2) 1993; Früh-Green et al., 2003).
tion the additional tens of kilometers of oceanic volcanic plateaus if any atmospheric weathering and runoff because, though there
produced from mantle plumes (Arndt and Chauvel, 1990; Arndt, probably were continents, radioactive heating made them plastic
1999; Russell and Arndt, 2005) (Fig. 1). And hydrostatic pres- and they rarely emerged above the surface of the relatively deep
sures, and thereby temperatures of the hydrothermal fluid, may ocean (Sandiford and McLaren, 2002; Russell and Arndt, 2005).
have been higher in the Hadean because of the deep penetration However, the iron contributions from the magmatically driven hot
of fluids to the margins of the crystallizing magma chambers springs were diluted by the iron-free submarine alkaline springs
under a deeper ocean. As a consequence, iron concentrations also exhaling from the ridge flanks and on the deep ocean floor. There
may have been correspondingly even higher (Von Damm, 2000; were also quantities of ferric oxyhydroxide (FeOOH) particles
Bounama et al., 2001; Allen and Seyfried, 2003). (denoted by FeIII) produced by photolysis (Cairns-Smith et al.,
Although it has been pointed out that temperatures of such 1992; Russell and Hall, 2002) (Fig. 2A):
hot acidic springs at spreading centers in the Hadean were so
high as to destroy organic molecules (Miller and Bada, 1988), 2Fe2+ + 2H+ + light → 2FeIII↓ + H2↑ (3)
they would have provided phosphate, as well as the trace ele-
ments that were to help energize and catalyze life, directly to the Taking account of dilution, photo-oxidation and precipitation, we
ocean (Kakegawa et al., 2002). speculate that the carbonic ocean carried up to 10 mmol of fer-
Given these hydrothermal contributions, what was the state rous iron.
of ocean chemistry? As atmospheric CO2 of mainly volcanic der- The hydrothermal convection cells feeding the alkaline
ivation was at a pressure of anywhere between 0.2 and 10 bars, springs were partly driven by exothermic serpentinization, a pro-
oceanic pH probably varied between 5 and 6. It could have been cess that begins slowly at around 85 ºC (Martin and Fyfe, 1970;
higher following major meteorite impacts when large quantities Wenner and Taylor, 1971). Flow in fractures within the oceanic
of rock dust were raised to the atmosphere (Nisbet and Sleep, crust was facilitated partly by geodynamic stresses and partly by
2001). Rainfall was likely to have been high, but there was little the tidal stresses induced by the close and rapidly orbiting moon
Figure 1. Cross-section of mantle convection cell for Earth at >4 Ga (Smith, 1981; Campbell et al., 1989; Davies, 1992; Karsten et al., 1996;
Foley et al., 2003; Russell and Arndt, 2005). Chemosynthetic life emerged at a warm alkaline seepage and expanded into the surrounding sedi-
ments and crust, and was conveyed by ocean floor spreading toward a constructive margin produced largely by obduction. Once uplifted at the
margin, a proportion of cells invaded sediments in the photic zone where, at a sulfurous spring, some evolved to exploit solar photons. Oxygenic
photosynthesis was a further evolutionary development.
Figure 2. The focusing of solar energy to produce (A) photolytic iron oxidation and the potentiation of chemosynthetic life (Cairns-Smith et al.,
1992); (B) reduction of ferredoxin and the onset of photo-induced non-cyclic electron transport (Blankenship, 2002); (C) photo-oxidation of
Ca-Mn bicarbonate and generation of a precursor to the water oxidizing complex (Anbar and Holland, 1992; Dismukes et al., 2001; Russell and
Hall, 2001, 2002); and (D) oxygenic photosynthesis through reduction of Mn4IV (Blankenship, 2002). Iron and manganese are exhaled from hot
springs at ocean floor spreading centers and at island chains.
(Gaffey, 1997). Nevertheless serpentinization and carbonation generated on the oxidation of the vestiges of native iron in the
may have blocked most fractures at temperatures above ~115 ºC Hadean crust (Righter et al., 1997).
(Wenner and Taylor, 1971).
Although it is well understood that high temperature springs
have a pH of between 2 and 3, a consequence of the loss and fixa-
tion of Mg2+ and the concomitant release of two protons (equa-
tions 1 and 2) (Von Damm, 1990; Allen and Seyfried, 2003), less
well known is the fact that magnesium is rendered more soluble
during the exothermic serpentinization of olivine and pyroxene
below 200 °C (Fig. 3) (Macleod et al., 1994; Palandri and Reed,
2004). Because of this, hydroxyl rather than H+ is generated as a
byproduct of the serpentinization:
3MgFeSiO3 + 7H2O → 3SiO2 + Fe3O4 +

3Mg2+ + 6OH- + 4H2↑ (4)
Thus these moderate temperature springs are buffered to a pH

of 10–11 by the precipitation of brucite [Mg(OH)2] and their
temperature may be controlled at ~115 °C (Wenner and Taylor,
1971).
Hydrogen is produced in both the low and high temperature Figure 3. Photomicrograph showing typical texture of a millimetric
systems on the oxidation, by water, of ferrous-iron-bearing min- grain of serpentinized olivine as found in diverse rock types. The pro-
erals to magnetite (equation 4). Hydrogen will also have been cess only takes place below 350 °C (Allen and Seyfried, 2003).
3Fe0 + 4H2O → Fe3O4 + 4H2↑ (5) be produced in the low temperature alkaline system are ammo-
nia (NH3), methane thiol (CH3S-), hydrosulfide (HS-), formate
Furthermore, reaction with the nickel-iron alloys produced dur- (HCOO-), and minor cyanide (CN-) (Muller, 1995; 995; Russell
ing serpentinization will have released hydrogen to the hydro- and Hall, 1997; Shock et al., 1998; McCollom and Seewald,
thermal fluid (Krishnarao, 1964). Other reduced entities likely to 2003). These molecules provide most of the basic nutrient and
Figure 4. Pyrite botryoids at a 350 Ma fossil warm spring at the Tynagh mine, Ireland: (A) The top surface (field of view measures 2 cm across).
(B) Cross-section through pyrite botryoids, revealing bubbles. (C) A natural chemical garden from Tynagh comprising pyrite spires embedded
in barite (photographed on centimetric background). (D) Photograph by Martin Beinhorn of a sulfide structure produced as 250 mmol of Na2S is
injected into a 25 mmol FeCl2 solution (field of view measures 4 cm across). (Pyrite is presumed to have replaced iron monosulfide membranes
in cases A to C [Banks, 1985; Russell, 1988].)
energy requirements of life. On meeting and mixing with the The large edifices at the Lost City spring are composed
Hadean ocean, hydrothermal mounds would be formed that seem mainly of carbonate and brucite [Mg(OH)2], though Kelley et al.,
to us the likely hatcheries of life (Russell and Hall, 1997). (2001) deduce a preoxidation sulfide concentration of ~5 mmol
kg–1 in the hydrothermal solution. The former presence of a sim-
EVIDENCE FOR A LOW TEMPERATURE MOUND ilar <100 °C alkaline spring at a transform fault in the Indian
Ocean is also indicated by deposits of finely layered hydrated
We first assumed (Russell et al., 1988, 1989) that hydrother- magnesium silicate (sepiolite) mixed with poorly crystalline Fe-
mal precipitates at an alkaline spring or seepage would result in Mn hydroxides (Bonatti et al., 1983). Any original carbonate has
a porous mound somewhat comparable to the submarine exhala- been redissolved. Another rather similar deposit, though precipi-
tive sulfide constructs at the Tynagh and Silvermines orebodies tated from fresh water, has been discovered off the north coast
in Ireland (Fig. 4A–C) (Boyce et al., 1983; Banks, 1985; Banks of Iceland (Marteinsson et al., 2001; Geptner et al., 2002) (Table
and Russell, 1992; Samson and Russell, 1987). These were the 1). Here cones of Mg-rich clay (saponite) tens of meters high are
deposits that first excited our interests in life’s emergence (Rus- forming where warm (72 °C) alkaline (pH 10) submarine spring
sell et al., 1988, 1989, 1994). Although the iron sulfide mounds waters exhale into a fjord. Although the cones do offer the kind
at Tynagh and Silvermines resulted from acidic solutions, origi- of porous morphology we expected, no sulfides are recorded
nally at ~250 °C and emanating into ~60 °C alkaline brine pools (Geptner et al., 2002; Martin and Russell, 2003). Some of the
in faulted basins at the bottom of the Mississippian sea ca. 350 differences between our expectations and the modern submarine
Ma, we reasoned that the chemical reactions would just as well springs can be ascribed to contrasting conditions in the Hadean
result in similar precipitates if the two solutions were inverted when the crust was more reduced, Fe2+ concentrations in the
(Russell et al., 1989). Observations of the kind of alkaline mod- ocean were high, and O2 was negligible or absent.
erate temperature hydrothermal spring we envisaged (Russell et In the light of these present-day discoveries how then might
al., 1989, 1998; Shock, 1992) have been made in 1.5 m.y. old we refine our model for the emergence of life? Where the con-
ultramafic oceanic crust, 15 km from the Mid-Atlantic Ridge at vective up-drafts were vigorous, the alkaline spring waters (pH
the so-called “Lost City” field (Kelley et al., 2001; Früh-Green 10–11, ~100 °C) would have exhaled directly into the acidulous
et al., 2003) (Table 1). Nevertheless, although the solutions here Hadean ocean (pH ~5.5) (Russell et al., 1989; Shock, 1992;
are alkaline as we expected, and carried some H2, the mounds Macleod et al., 1994; Russell, 2003). We have assumed that at
contrasted significantly with our predictions (Russell et al., 1994; times when solar radiation was masked, this ocean was ~20 °C or
Russell and Hall, 1997). less (Fig. 5). Precipitation at the exhalative center was rapid, but
Figure 5. Model environment for the

emergence of life at a submarine seep-
age on the ocean floor (after Russell and
Hall, 1997).
as the fracture conduits in the mounds became fouled with car- We originally imagined iron sulfide structures to have pre-
bonate, and with gels and microcrysts composed of silica, sepio- cipitated spontaneously at the interface of the hot alkaline seep-
lite, saponite, brucite, green rust, and iron sulfide, the fluid egress age waters containing millimoles of HS- and H2 with the cool
became diffuse and seepages replaced springs. Such restraint in carbonic ocean water bearing millimoles of Fe2+ and particulate
the seepage mound favored—depending on solutes and local FeOOH (Figs. 6A, 6B). Our attempts to reproduce similar struc-
pH—development of siderite, ferrous hydroxide, and/or iron tures in the lab were relatively successful, though rather high
(nickel) monosulfides. What might the structure of these precipi- concentrations of HS- (250 mmol) were required to produce
tates have looked like? bubbles (Fig. 4D) (Russell, 1988; Russell et al., 1989). Depend-
Figure 6. (A) A Pourbaix (Eh/pH) diagram illustrating the stabilities of siderite, mackinawite (as FeS), protoferredoxin, greigite, pyrite, green
rust, and hematite, produced for activities of H2S(aq) = 10−3, and Fe2+ = 10−6, using GWB (Bethke, 1996). The inset shows notional phase rela-
tions emphasizing the intermediate oxidation state of the FeS component of membrane protoferredoxins and is positioned to indicate the Eh-pH
conditions pertaining to alkaline hydrothermal fluid as it enters Hadean seawater. Note that the pH boundary of monophosphate/polyphosphate
intersects this redox position (and see Fig. 16). (B) An Eh-pH diagram computed for modern atmospheric CO2. At higher pCO2 the siderite field
would expand as indicated by arrows, at the expense of FeS. Such a release of HS- from pyrrhotite accumulations in the crust to the hydrothermal
solution would, on meeting Fe2+ within the growing mound, reprecipitate as FeS (Hall et al., 1994). The Fe2+/Fe(OH)3 boundary is projected
(dashed line to show its approximate position at very low pCO2. Calculated using GWB (Bethke, 1996). (C) An Eh/pH diagram illustrating the
electrochemical energy potentiating the onset of life and the first microbe. (D) The electrochemical energy available to modern iron reducing
bacteria (Zachara et al., 2002) compared with that available from the reduction of native sulfur. After Russell and Hall (1997).
ing on the pH, precipitation of mackinawite [(Fe>Ni)1+xS] and Aided by the electrochemical gradients obtaining near the
green rust [FeII4FeIII2(OH)12.CO32–.2H2O] produced chemical gar- mound’s surface, one of the effects of the inorganic membrane
dens comparable to those found at Tynagh (Fig. 4C) (Russell, would have been to split hydrogen into electrons, protons, and
1988). Greigite [Fe3S4] also occurred, as did violarite [Fe2Ni4S8] transient activated hydrogen atoms. We imagine electrons trans-
if Ni2+:Fe2+ ratios were high (Russell, 1988; Russell et al., 1998). ferring from one semiconducting nanocrystal to the next through
Mackinawite nanocrysts were probably the main components of the membrane as they were drawn toward external FeIII and/or
the chemical garden and of the membranous walls to individual HCO3- within the membrane (Fig. 8). To maintain charge balance
compartments. We suggest bubbles like these, or at least micro- the protons were forced to follow by rotational/translational dif-
cavities within a mackinawite precipitate, acted as the original fusion of water/hydronium molecules that adhered to the crystal-
catalytic culture chambers for early metabolism and embryonic lite surfaces (da Silva and Williams, 1991, p. 103). As we shall
life. Described thus, the hydrothermal mound begins to take on see, the addition of further protons to the exterior of the FeS com-
the attributes of a natural catalytic flow reactor and fractionation partments would have augmented the natural protonmotive force
column and we now examine it in this light (Russell et al., 2003; acting on the membrane.
Stone and Goldstein, 2004; Russell and Martin, 2004). Divalent metal ions can also invade the sulfur layer (Fig. 7),
and nickel and minor cobalt as well as other metals can replace
MACKINAWITE—MEMBRANE MINERAL, iron in the metallic layer. Also mackinawite is potentially a major
PREBIOTIC CATALYST, AND ELECTRON TRANSFER temporary sink for many trace metals in anoxic conditions, even
AGENT calcium (Morse and Arakaki, 1993). Morse and Arakaki (1993)
demonstrate that the surface affinity of mackinawite during
Mackinawite provided the inorganic structure and reaction adsorption of Mn2+, Cr3+, Co2+, Ni2+, Zn2+, Cd2+ and Cu2+ is in
surfaces of the first membrane. At the molecular level macki- the order of their decreasing solubility as sulfides. When taken
nawite (Fe1+xS) comprises layers of offset Fe2S2 rhombs (Wol- in conjunction with their high surface-to-volume ratios, macki-
thers et al., 2003). At the atomic level it can be seen that the iron nawite nanocrysts have excellent catalytic properties (Cody,
layers in mackinawite are semiconducting (Vaughan and Ridout, 2004; Cody et al., 2004). And as we would expect of a catalyst,
1971) (Fig. 7), yet across the layers the van der Waals bonding mackinawite is a highly reactive mineral prone to oxidation. In
of the sulfurs confers an insulating capacity to the mineral. The an anaerobic environment it can be oxidized in two ways. Oxi-
particle sizes of the mackinawite precipitates are bimodal—one dized by Fe3+, it converts to greigite by loss of electrons from, and
is 2 × 2 × 1.5 nm (at pH 8), the other 7 × 7 × 3 nm (at pH 6) reorganization and even some dissolution of, the iron (Krupp,
(Wolthers et al., 2003). 1994) (Figs. 7, 9A). So we expect greigite to be a minor phase,
Figure 7. Structure of mackinawite Fe1+xS.

Mackinawite consists of an assemblage of
[2Fe2S] rhombs (Wolthers et al., 2003)
arranged in such a way that it acts as a
semiconductor in the bc plane and an in-
sulator through the c axis. Comprising the
membrane, mackinawite nanocrysts may
have acted as the first electron transfer
agents from the interior of the protocells
to FeIII, the exterior electron acceptor (Fig.
8) (Russell and Hall, 1997; Russell et al.,
1998; cf. Ferris et al., 1992).
Figure 8. Supposed emergence of chemi-

osmosis driven by reduction of FeIII on
the exterior of the FeS membrane. Elec-
trons are conducted through mackinawite
nanocrysts from H2 oxidized on the in-
terior (cf. Ferris et al., 1992). Protons
track electrons through aqueous films by
rotational/translational diffusion of H3O+/
H2O molecules adhering to the crystallite
surfaces (da Silva and Williams, 1991, p.
103) to conserve charge balance. Else-
where mackinawite may act as an insu-
lator (Fig. 7). The membrane potential is
augmented by protons in the acidulous
ocean—an ambient protonmotive force.
Figure 9. Structural relatedness of (A) greigite Fe5NiS8; (B) the thiocubane [Fe4S4] unit in protoferredoxins and ferredoxins; (C–E) the [Fe4NiS5]
open cuboidal complexes in CO-dehydrogenase; and (F) the twinned center to nitrogenase. Affine sulfur sub-lattices, cubic close-packed in A,
III/3+
are distorted in the metalloenzyme centers. The presence or absence of Fe , Mo, Ni, and organic ligands may dictate which of these entities
formed in the first cells. RS denotes a link to the protein through the sulfur of cysteine, HN involves a link through a nitrogen ligand of histidine
to the same protein, while hc is homocitrate (-OOC.CH2.COH.COO-.CH2.COO-). Structure (A) from Vaughan and Craig (1978), Krupp, 1994;
(B) Hall et al. (1971), Beinert et al. (1997); (C) Drennan et al. (2001); (D) Dobbek et al. (2001); (E) Doukov et al. (2002), Darnault et al. (2003);
Svetlitchnyi et al. (2004); (F) Helz et al. (1996), Einsle et al. (2002), Seefeldt et al. (2004).
Figure 11. The hydrothermal mound as an acetate generator. The de-

tailed cross-section of the surface illustrates the sites where organic
ions are produced, retained, react, and self-organize to emerge as pro-
tolife (from Russell and Martin, 2004).
Figure 10. Atomic structure of pyrite drawn to show the ferrous iron
ligated to six sulfur-pairs, Fe 2+ (S2-2 )6 . The complex structure makes
the mineral difficult to nucleate, and difficult to reduce (Finklea et al., Fresh sulfide nanoparticles acted as sites of adsorption, absorp-
1976) unlike the iron sulfide clusters of metal enzymes (cf. Fig. 9). tion, and catalysis. Ferric iron particles attracted to the outside
of the membrane may have oxidized some of the mackinawite
to greigite and even pyrite. Cyanide would have fractionated
from the formaldehyde, the former adsorbed on the pyrite, the
particularly toward the outer margins of the membrane (see next latter upon the mackinawite (Leja, 1982; Rickard et al., 2001).
section). Further oxidation converts it to the relatively inert pyrite In places membranes and barriers composed of mackinawite and
(Fig. 10). In this latter case it is the sulfur that is oxidized (to minor greigite acted as solid phases for further chemical inter-
S22-). Rickard et al. (2001) have shown that this stage of oxidation actions between the reactive solutes (Russell et al., 1994, 2003;
is prevented by formaldehyde, significant because, as we shall Schoofs et al., 2000). What might these reactions have been?
see below, greigite has a structural affinity to ancient metallo- The reaction expected to have released the most energy, i.e.,
enzymes (Fig. 9). with the greatest thermodynamic drive, was the production of
methane and water from the carbon dioxide and the hydrogen
HYDROTHERMAL MOUND AS REACTOR AND (Amend and Shock, 2001):
ACETATE WASTE GENERATOR
CO2 + 4H2 → CH4 + 2H2O (7)
Acting as a natural flow reactor and fractionation column,
the hydrothermal mound stood vertically and was composed However, there is a major kinetic barrier that faces this reaction,
essentially of brucite, clay, minor sulfides, green rust, and ephem- which takes place spontaneously only above 500 °C (Shock et
eral carbonate (Fig. 11). Hydrothermal fluid entered through the al., 1998). The reaction to produce acetate dissipates less energy,
permeable and porous base. The bubbles and pores stemmed the but the kinetic barrier is also lower, though not low enough for H2
flow of “electron-rich” molecules such as H2, NH3, HCOO-, CN-, to react spontaneously with the CO2 (Shock et al., 1998; Amend
CH3S-, and HS-. The H2, HS-, HCOO-, and NH3 were the most and Shock, 2001). Shock et al. (1998) calculate that carbon in
concentrated at 10 or so mmol. This hydrothermal solution mixed metastable equilibrium states obtained by mixing hydrothermal
with about 100 mmol of HCO3– and several millimoles of Fe2+ fluids with anoxygenic seawater below 110 °C would be mainly
and HP2O73-/HPO32- in ocean water that was entrained through as acetate, with subsidiary propanate and dodecanoate (Fig. 12).
the sides of the mound. The HS- reacted with Fe2+ and FeII (i.e., Therefore, we expect the hydrothermal mound and its compart-
fixed ferrous iron) to precipitate nickeliferous mackinawite: ments to have catalyzed the production of acetate (CH3COO-)
and water from CO2 (as bicarbonate, Fig. 13) and H2 below 50 °C,
Fe(Ni)2+ + HS- → Fe(Ni)S + H+ (6) degrading energy in the process (Russell and Martin, 2004):
thioacetate (CH3COSCH3) in the presence of an FeS/NiS slurry

optimally at pH 6.4:
CO + 2CH3SH → CH3.COSCH3 + H2S (9)
Methane thiol has been synthesized from CO2 at 100 °C in the

presence of FeS and H2S (Heinen and Lauwers, 1996, 1997).
At the same time, the FeS was oxidized to pyrite, as might be
expected from the “pyrite-pulled” model of Wächtershäuser,
1988). The yield with respect to H2S was ~0.25%. In theory
methane thiol activities would rise a thousandfold when gener-
ated from H2 and CO (rather than CO2) in the crust or the hydro-
thermal mound (Schulte and Rogers, 2004). These conditions
would have been met within the mackinawitic membrane that
Figure 12. Mole percent distribution of the three most “metastable”
carboxylic acids at low to moderate temperatures theoretically ob- separated ocean water at about pH 5.5 from the more alkaline
tained through the mixing of hydrothermal fluids with anoxygenic sea- hydrothermal fluids that contained CO and organic sulfides, i.e.,
water. They are acetate (CH3.COO-), propanate (CH3.CH2.COO-), and the thiols (Russell and Hall, 1997). Hydrolysis would have liber-
dodecanoate (CH3.(CH2)10.COO-). (Minor species have been neglected ated a proportion of the acetate while the methane thiol catalyst
in this redrawing of part of Figure 5.7 in Shock et al., 1998.) was returned to the solution:
CH3.COSCH3 + OH- → CH3.COO- + CH3SH (10)
The acetate reaction, augmented by the proton gradient

operating across the membrane, may also have been coupled to
pyrophosphate generation. (Currently this process is driven by a
sodium gradient.) The surviving methyl acetate became involved
in further biosynthetic reactions such as the generation of amino
acids and lipids. Can we generalize the first evolutionary step?
The notional hydrothermal reactor works to the rule of the
second law of thermodynamics. For the most part, we might
expect that geochemicals far from equilibrium would react irre-
versibly, degrading energy in the process (Shock et al., 1998).
Minerals such as mackinawite form exothermically on reaction
between hydrothermal HS- and the Fe2+ in the entrained ocean
water (equation 6), so we can consider the onset of life as simi-
larly exothermic overall and suggest a simplified equation as a
demonstration:
{407H2+10NH3+HS-}hydrothermal + {210CO2+H2PO4-
+ Fe,Ni,Co,Zn,Mo2+}ocean
Figure 13. Pourbaix diagram using Geochemist’s Workbench ACT2
(Bethke, 1996) showing the acetate and carbonate fields (thick lines) → {C70H129O65N10P(Fe,Ni,Co,Zn,Mo)S}proto-life
with respect to dissolved and solid iron phases (dashed lines). Condi-
tions: 100 °C, p = 1.013 bars, CO2(g) log f-3, Fe2+ log a-20, fields + {70H3C.COOH+219H2O}waste (11)
extended below stability field of water (dotted lines).
The ratio of waste acetate plus water to “proto-life” in this con-
ceptual reaction is high. Most of the output from the reacting
monomers elutes to the ocean and entropy thus increases, but
4HCO3- + 8H2 + 2H+ → 2CH3.COO- + 8H2O (8) the iron sulfide botryoids, bubbles, and pores could act, albeit
inefficiently at first, as tiny electrically powered compartments
There is some experimental evidence to support the conclusions or “tureens” in which a “warm organic soup” could have been
of Shock et al. (1998). Reacting 100 mmol CH3SH (methane synthesized, constrained and concentrated to a critical mass that
thiol) and 4.5 mmol CO at 100 °C at normal pressure, Huber and encouraged further interactions (Russell et al., 1988, 1994; Braun
Wächtershäuser (1997) produced micromolar amounts of methyl and Libchaber, 2004). Thus, within the compartments processes
were reversible, and though entropy was exported, it decreased The amino acids were mostly adsorbed within compartments in
within the compartments themselves (Prigogine, 1978). the mound. With the addition of formaldehyde (HCHO), minor
This approach does correspond to what is understood of early concentrations of RNA would also have been produced from
life from microbiology: many microbes, including those near these components (Ferris and Hagan, 1984).
the base of the evolutionary tree, can gain energy by generating Apart from the fluid wastes, mainly water and acetate, there
acetate using the enzyme carbon monoxide dehydrogenase with is also the solid waste product to consider.
acetyl-coenzyme-A (which is also an organic sulfide or thiol, co-
A.SH), through the acetyl-coenzyme-A pathway, that is in part PYRITE—A SOLID WASTE PRODUCT
homologous with the Huber-Wächtershäuser reaction (equation
9) (Schink, 1997; Peretó et al., 1999; Amend and Shock, 2001; When mackinawite or greigite is oxidized to pyrite, for
Russell and Martin, 2004): example during the generation of methane thiol, the rhombic
[Fe2S2] building block is lost; instead, the ferrous iron is ligated
2CO2 + 8[H] + co-A.SH → co-A.S.OC.CH3 + 3H2O (12) to six partially oxidized sulfur-pairs (S22-) (Fig. 10). The resulting
Fe(S22-)6 is a complex structure which makes the mineral difficult
A portion of the acetate and the energy released in these exergonic to nucleate and very difficult to reduce back to FeS (Finklea et al.,
reactions would have gone to waste. But waste, the generation of 1976). Wächtershäuser (1988) has argued for the formation of
entropy, is life’s raison d’être. We might think of the mound as pyrite on oxidation of FeS by hydrogen sulfide, with the genera-
optimizing the generation of acetate over time while side reaction of hydrogen, as the primeval energy source for the origin of a
tions, including many involving the activated thioacetate (CH3. (surface) metabolist. Although we consider hydrogen to be deliv-
COS-), synthesized the more complex molecules that interacted ered to the mound in the alkaline solution, the Wächtershäuser
to produce life. This non-vivocentric view is now examined in reaction works in acidic and neutral conditions (Taylor et al.,
the context of a notional reactor that produces the acetate and 1979; Drobner et al., 1990; Rickard, 1997). Moreover, the hydro-
water (Figs. 11–13). phobic pyrite precipitate recorded at the gas-water boundary in
Bubbles comprising the iron sulfide membrane could have the experiments of Heinen and Lauwers (1996) demonstrates that
been hydraulically inflated over warm seepages, where they this irreversible (i.e., non-catalytic) oxidation can drive reduc-
encapsulated the reduced alkaline hydrothermal solution (Fig. tions of CO2 as suggested by Wächtershäuser (1988). Although
4D) (Russell et al., 1989, 1993). As the bubbles became dis- this reaction produces less energy in alkaline conditions, were
tended they weakened, failed, and daughter bubbles were gener- pyrite to have been formed it could have taken no further part in
ated above the punctures (Fig. 4B). Thus the redox and pH front protometabolism and must, therefore, be considered as a waste
remained at the growing surface of the mound (Figs. 4D, 5, 11). product. Nevertheless, its hydrophobic surfaces would have ren-
Bubbles farthest from the feeder veins would have been disad- dered it a surficial trap for organic molecules and cyanide gleaned
vantaged unless the structure of their membranes particularly from the hydrothermal fluid (Leja, 1982; Russell et al., 1988).
disposed them to supporting an osmotic pressure. This osmotic
pressure would have been induced by the generation of abiotic ENERGY FOR POLYMERIZATION
charged organic molecules. Contiguous compartments (Fig. 5),
generated by budding of the iron monosulfide membrane (Fig. Although we note empirically that the thermal gradient
4B), would have contained fluid mixes at slightly different Eh responsible for convection lies between 115° and <20 °C, the
and pH conditions and therefore would have harbored different electrochemical potential to drive biosynthesis and polymeriza-
reactants and products, as energy cascaded from one chemical tion could be subscribed by the H+/H2 couple (effectively the
and electrochemical level to another. These possibilities have hydrogen electrode) representing the hydrothermal emanations,
been elegantly considered for other types of inorganic mem- as against the Fe3+/Fe2+ couple representing the photolytic ferric
brane by Cairns-Smith (1982, p. 327 and 351–356). As discussed iron (representing an initially dispersed positive electrode). The-
below, organic synthesis would have been catalyzed by the iron oretically the hydrothermal hydrogen could have reduced ferric
(nickel) monosulfide, which, unlike fine metal and oxide/hydrox- oxyhydroxide (signified as FeIII) to ferrous ions, and produced
ide catalysts, cannot be poisoned by sulfidation. protons in a reversal of equation 3 in which the redox potential
Although acetate and water were the main fluid products, the was influenced by light energy:
hydrothermal NH3 and minor CN- would have reacted with bicar-
bonate on mineral surfaces and produced amino acids, especially H2 + 2FeIII → 2H+ + 2Fe2+ (14)
glycine (+H3N.CH2.COO-) (Hennet et al., 1992):
Pourbaix diagrams idealize the thermodynamic potentials on
proton activity (pH) and electron activity (Eh) coordinates (Fig.
3CO2 + 5H2 + NH3 + HCN →
6). Oxidation of H2 on one side of the membrane by reduction,
2(+H3N.CH2.COO-) + 2H2O (13) through electron transfer, of FeIII on the other side of the mem-
brane, theoretically generates a potential of 770 mV (Fig. 6C)
(Russell et al., 2003). Other species such a Fe(OH)3 have lower

redox potentials, for example Fe(OH)3/Fe(OH)2 = 270 mV (Gar-
rels and Christ, 1965, p. 183). Even this potential is more than the
250 mV (−11.5 kcal per mole) required to drive polymerization,
especially when the pH gradient is taken into account (Thauer et
al., 1977; and see Kell, 1988, Van Walraven et al., 1997). This
chemiosmotic proton potential is augmented by proton potential
of the acidulous ocean (pH ~5.5 at ~20 °C) acting across the inor-
ganic membrane on the alkaline hydrothermal solution (pH ~10.5
at ~100 °C), a potential that results in a further 300 mV (Fig. 6C).
Geobacter metallireducens, as well as many other bacteria in the
lowest branches of the evolutionary tree, can reduce FeIII using
much the same electrochemical potential (Fig. 6D, 14) (Liu et al.,
1997; Vargas et al., 1998; Zachara et al., 2002).
In sum, the overall “protonic” potential approximates 1 V,
enough geochemical energy to have engendered metabolism
across an inorganic membrane on early Earth (Russell et al.,
2003). Using a “Beutner rig” (Beutner, 1913, Fig. 1), we have
Figure 15. Plot of mV developed across an FeS membrane with time.
found that a spontaneously generated FeS membrane, formed Preliminary results with conditions as defined in the inset. Schematic
on reaction between 10 mmol solutions of FeSO4 and Na2S (at illustration of the notional photoelectrochemical cell assumed to have
pH 4.12 and 10.0 respectively), creates a tension of ~600 mV, obtained on the Hadean Earth and ocean, and of a preliminary experi-
which is held for several hours (Fig. 15). Filtness et al. (2003), mental emf/pmf.
using more fastidious conditions obtain a total tension closer to
700 mV, differences maintained for more than 4 h. That the pH-
dependent boundary between the mono- and di-phosphate fields
intersects the (primarily) Eh-dependent iron sulfide fields dem-
onstrates how “proticity” (a proton current) could have driven
the condensation of inorganic phosphate (Pi) to pyrophosphate
(PPi) (Fig. 16). So in theory the redox reaction outlined in equa-
tion 3 could have been simply coupled through the membrane to
Figure 16. GWB diagram (Bethke, 1996) illustrates how high energy
Figure 14. Evolutionary tree (after Woese et al., 1990; Stetter, 1996; phosphoanhydride, stable at lower pH and water activity and higher
Martin and Russell, 2003). The last common community (the LCC) T, forms from low energy monophosphate (HOP2O63-) and can drive
occupied the mound in which life had emerged. Many prokaryotes in dehydration polymerizations on its hydration (cf. the ATP4-/ADP3-/or
the lowest branches of the tree can use FeIII as an electron acceptor AMP2- energy cycle of life). In oxidative phosphorylation the dehy-
(Liu et al., 1997; Vargas et al., 1998; Kashefi et al., 2002). Note that drating power of ATP is renewed by pmf (acidification), whereas in
methanogens are found only in the Archaeal domain and that oxygenic substrate level phosphorylation ATP is renewed by removal of H+ and
photosynthesis is a property only of the cyanobacteria. OH- by a NAD-associated enzyme.
this dehydration or condensation of monophosphate (Russell and (2003, chapter 4, Fig. 5) finds between mackinawite and arsenate.
Hall, 2002): Whatever the mechanism, these first amino acid polymers or pep-
tides probably consisted mainly of glycine with occasional alanine,
H+ + 2HPO42- → HP2O73- + H2O (15) aspartate, serine, and valine (Hennet et al., 1992). The last four
amino acids have particular stereochemistries; i.e., they can be right-
In biochemical energetics the process is known as “oxidative handed (dextro) or left-handed (levo). Polymerization of a racemic
phosphorylation,” a reference to the fact that protons are initially mix of dextro and levo amino acids would result in a heterochiral
driven to the outside of the membrane to maintain their balance with peptide; i.e., it would have had a mixed chirality. A lack of chiral-
electrons flowing outward to an electron acceptor before returning ity is thought to be a major problem for theorists of the emergence
to recharge the phosphate (Mitchell, 1967) (Fig. 8). The protonmo- of life (Cairns-Smith, 1982). We show later that a lack of chiral-
tive force is the power behind metabolism and is therefore indis- ity in the first peptides, far from being a problem, was a positive
pensable to life, so the fact that this ambient force is a feature of the advantage with regard to self-organization of the constituents of the
redox and alkaline-to-acid interface separated by a semiconducting first compartments, though we must emphasize that this unregulated
inorganic membrane is a compelling aspect of the hypothesis. Nev- polymerization of amino acids, while a necessary step, was an evo-
ertheless, the “ATPase” responsible for the conversion in today’s lutionary dead end. However, before we examine the unregulated
organisms is a complex rotating turbo-motor (Elston et al., 1998). peptides we need to take a careful look at the structure of another
We assume, after Baltscheffsky et al. (1999), that the original pro- key metal sulfide mineral precipitated in the reactor, viz., greigite.
cess used a static prototype H+-PPase, basically an enriched domain
of phosphate upon mineral surfaces within the membrane (Fig. 8). GREIGITE—PRE-ENZYME AND ELECTRON
What may be surprising is the assumption that dehydrating DOCKING SITE
reactions can take place in hydrothermal conditions. What to con-
sider is the fact that anions, generated and concentrated in the mem- The structure of greigite tends to an inverse spinel, written
brane, would have competed successfully for a place on a growing notionally as SFe3+S[Fe42.75+S4]SNi2+S (Fig. 9A). In fact the electrons
mineral surface against the minor negative charges on the oxygens are delocalized, so there is a certain amount of valence resonance. If
of polar water molecules. Given the centrality of acetate, the first further nickel is introduced, the inverse spinel verges toward the true
organic phosphate may have been acetyl phosphate (CH3.CO.PO42- thiospinel violarite (as SNi2+S[Fe23+Ni23+S4]SNi2+S). We assume that
), perhaps produced by phosphorylation of the acetyl thioester (from some of the hydrogen dissolved under high pressure in the hydro-
equation 9) (de Duve, 1991; Russell et al., 1994): thermal fluid was adsorbed on, or absorbed within, the [Fe4S4] cube.
Once there, hydrogen was split to an electron (reducing the iron) and
CH3COSCH3 + HPO42- → CH3SH + CH3COPO42- (16) a proton (protonating one of the four sulfides), leaving a hydrogen
radical. This nascent hydrogen [H•] was highly active and might
Reacting acetyl phosphate with inorganic phosphate (HPO42-) in have attacked a bicarbonate ion, the ultimate electron acceptor, at a
the presence of FeII minerals, de Zwart et al. (2004) have gen- tetragonally coordinated nickel site (Fig. 9A). Because of the likely
erated pyrophosphate with a yield of 25% at ~40 °C. FeS was contribution of greigite (NiFe5S8) to inorganic membranes devel-
found to strongly retard hydrolysis of, and thus preserve, the oped at alkaline submarine seepages (Russell, 1988), and its simi-
pyrophosphate: larity to the structure of the active centers to the most ancient pro-
teins, the ferredoxins, we have suggested that molecular constructs
HPO42- + CH3.CO.PO42- → HP2O73- + CH3.COO- (17) of the mineral were, at a later stage of emergence, incorporated by
peptides as the first electron transfer agents, redox enzymes and
The resulting pyrophosphate (PPi) would have had the energy synthases (Russell and Hall, 1997, 2002; Milner-White and Russell,
to polymerize amino acids (Romero et al., 1991; Baltscheffsky 2005). They are the ready-made, modular mineral clusters (Beinert
et al., 1999): et al., 1997) that, when combined with other metals and/or organic
structures, constitute components of what Baymann et al. (2003)
have termed “the redox protein construction kit.” The kit is partly
+
H3N.CH2.COO- + HP2O73- + +H3N.CH2.COO- →
based on the (inverse) spinel or greigite structure, which contains an
H3N.CH2.CO.NH.CH2.COO- + 2HPO42- + H+
+
(18) [Fe4S4]~2+ cage or cubane in which the electrons are delocalized and
in which the iron atoms have a nominal positive charge of 2.5 and
The monophosphate might then have been repolymerized with pro- have the tendency to switch valence.
tons as shown in equation 15.
Thus, before the advent of the ribosome, random amino acid FROM CATALYSTS TO ENZYMES
polymerization could have been driven by pyrophosphate bonded
through to sulfide on the surface of mackinawite in a -Fe-S-O-P- We have seen that iron sulfides and iron nickel sulfides have
motif as suggested by EXAFS (Pattrick, 2001, personal commun.). the catalytic propensity to produce some of the simple molecu-
Such a conformation is congruent with the relationship Wolthers lar modules of life from inorganic constituents. Examples of
organic products are methane thiol, the simple amino acids, and 1999) (equation 18). Alternatively, where sulfide concentrations
acetate. As acetate is the likely first major product of the hydro- were high, the more reactive thiocarboxyl group (-CSO-) might
thermal reactor or mound, it is instructive to consider how mod- have been attacked by the amino group with the loss of a thiol
ern homoacetogens (i.e., acetogenic bacteria that use only inor- (RSH) (Wächtershäuser, 1992). There are also experimentally
ganic nutrients and fuel) synthesize acetate (Müller, 2003). The inspired suggestions for peptide formation (Ferris et al., 1996;
overall reaction of the acetyl-coenzyme-A pathway is shown in Huber and Wächtershäuser, 1998; Huber et al., 2003; Leman et
equation 12. The key enzyme of the pathway is carbon monox- al., 2004), but a mechanism for prebiotic polymerization has yet
ide dehydrogenase (CODH) (Peretó et al., 1999; Dobbek et al., to be agreed.
2001). As a single unit or homodimer, the enzyme employs five
[Fe4S4] clusters including a unique [Fe4NiS5] cluster where CO2 PEPTIDE NESTS FOR SULFIDES AND PHOSPHATE
is reduced to CO (Figs. 9B, 9D). A more complex form of the
enzyme possesses additional FeS clusters and reduces not only Potentially the two structures involved in energy transfer
CO2 to CO, but also condenses Ni-bound methyl and CO, yield- are the metal sulfide clusters and the phosphates. The early sul-
ing a Ni-bound acetyl moiety that is transferred to co-ASH in the fide clusters are likely to have been sequestered by thiolate (e.g.,
acetyl-coenzyme-A synthase reaction (Fig. 9E) (Lindahl, 2002). [Fe4S4][CH3S]42-) in aqueous solution (Bonomi et al., 1985). Thus
The formula of the [Fe4NiS5] C-cluster in the enzyme is also both the phosphates (e.g., HP2O73-) and the thiolated iron(nickel)
comparable to that of greigite [Fe5NiS8] (Fig. 9C–E). Indeed, the sulfide clusters were anionic. As the nitrogens of the amino
[Fe4S4] cube of greigite is also congruent, or nearly so, with the groups carry a δ+ charge even when part of a peptide, the inor-
cubanes in the most ancient proteins, the ferredoxins (Eck and ganic anions would have been drawn to these δ+ charges on the
Dayhoff, 1966) (Fig. 9B). That there is some play in the placing peptide chain (Fig. 17). At the same time the chain would have
of the Fe and Ni ions and of their charge in greigite is also echoed bowed to satisfy the negatively charged clusters. Encased in such
by the way Ni and Fe are variously but characteristically sited in
the dehydrogenases (redox enzymes) and synthases (biosynthesis
enzymes) (Fig. 9C–E). We have considered the component build-
ing blocks of the greigite structure to be the likely “ready-made”
molecules co-opted by early life before they could be interred in
sulfides or oxides (Russell and Hall, 1997; Russell and Martin,
2004; and see Bonomi et al., 1985).
THE ORGANIC TAKEOVER
Recalling our inorganic predilection, we use the way metals

are coordinated in minerals and their natural cluster precursors as
a heuristic device to inform us as to the likely, or at least possible,
coordination chemistries adopted by early biology. Metals now
constitute the active centers of ~50% of all protein types (Jerni-
gan et al., 1994). This percentage will be optimal for the present-
day geochemical environment where metals are harder to come
by. Mono-, di- and tri-phosphates are also vital to coenzymes.
At the onset of life amino acid polymers—the peptides—would
have sequestered and protected the ubiquitous active inorganic
centers in what was the first stage in an organic takeover. The
peptides are the mediators of the takeover.
Amino acids are zwitterions, amphoteric molecules that
have a negative and a positive charge (“zwitter” is German for
hermaphrodite) as well as an organic side chain R—except for
glycine (+H3N.CH2.COO-), which only caries a hydrogen atom.
In normal conditions amino acids have a negatively charged car-
boxyl group (-COO-) and a positively charged amino group (-
NH3+) e.g., +H3N.HCR.COO-. Before the advent of RNA, amino
acids could have been polymerized by the loss of the constituents Figure 17. Ball and stick sketches of early nests of short peptides con-
stituting alternating glycines with other abiotic racemic amino acids
of water, OH- from the carboxyl of one amino acid and of H+ from binding (A) an Fe3S4(RS3-)3 to produce a protoferredoxin, and (B) an
the next. The reaction may have been driven by pyrophosphate inorganic phosphate in a primitive P-loop. R represents side chains.
hydration where water activity was low (Baltscheffsky et al., After Milner-White and Russell (2005).
“peptide nests” the active centers were partially protected from

dissolution, or nucleation and crystallization, and would have
remained active (Milner-White and Russell, 2005). The cosseting
of active centers in this way would have been optimal when the
peptide was composed either of glycine, the one achiral amino
acid, or of amino acids with random stereochemistries. In other
words, for efficient self-organization of protoenzymes (i.e., the
ferredoxins, dehydrogenases, and synthases) and the proto-coen-
zymes (e.g., PPi) it is better for the peptide to have been hetero-
chiral and have consisted of racemic amino acid residues. Chiral
peptides would tend to have formed helices or sheets rather than
nests. The nest configuration was a likely early outcome because
the easiest amino acid to make abiotically is glycine and because
the other abiotic amino acids would have been racemic (Hennet
et al., 1992; Huber and Wächtershäuser, 2003).
We have seen how electrons, carried first as constituents of
the hydrogen molecule in hydrothermal solution, seek out elec-
tron acceptors. These are ferric iron and the more recalcitrant car-
bon dioxide or bicarbonate (Figs. 6D, 13, 18). Electron transfer
agents and catalysts are required for the redox reactions. Later
in evolution sulfate, nitrogen, and nitrate also came to be used
as electron acceptors (Fig. 18). Depending on what oxidation or Figure 19. Ferredoxin in Peptococcus aerogenes as a comparison to
reduction takes place, electrons had to be transported singly, two a nested [4Fe-4S] cubane shown in Figure 17A. This is a typical fer-
at a time, three at a time, or even four at a time. Simple and mul- redoxin of the type that probably formed by the twinning of a single
cubane ligated through cysteines. This ferredoxin has a total of 54
tiple twinning to dimers and multidimers of the [4Fe-4S]-bearing
simple amino acids (redrawn form Adman et al., 1973). Inset shown is
ferredoxin, results of gene duplications, facilitates multiple elec- the mechanism of the transfer of a solvent proton to the buried redox
tron flow (Adman et al., 1973; Steigerwald et al., 1990) (Fig. 19). center via the side chain of aspartate (Chen et al., 2000). (A—alanine,
But these complex proteins could only have formed when pep- Y—tyrosine, V—valine, I—isoleucine, N—asparagine, D—aspartate,
tides were relatively long and had particular and regulated amino S—serine, C—cysteine, G—glycine, K—lysine, P—proline, E—glu-
tamate, Q—glutamine.)
acid sequences to facilitate their folding and the sequestration of
the inorganic active clusters.
COENZYMES WITH ORGANIC RING COMPONENTS
One of the first requirements for organic molecules dur-

ing early evolution was to cooperate with the redox tasks of the
ready-made metal-bearing clusters and their associated peptides.
Carbon/nitrogen rings containing three conjugated double bonds
and unpaired and delocalized electrons are the coenzymes that
carry out some of these processes (Pullman, 1972). Furthermore
they have the advantage of being bonded to aliphatic compounds
involved in structuring the cell. All the nucleic acid bases and the
three aromatic amino acids have such rings as, or on, their side
chains (Fig. 20). In terms of the organic takeover, the ferredox-
ins originally involved in the oxidation of glucose were largely
replaced by nicotinamide adenine dinucleotide (NAD) (Daniel
and Danson, 1995). And, when iron is in short supply, flavodox-
ins can substitute as electron carriers for ferredoxins.
Although the heterocyclic rings do not grow to mineral pro-
portions, some of them polymerize to macrocyclic compounds.
Figure 18. Redox potential/pH vectors between the cellular fluid of These compounds can sequester a variety of single metal ions
prokaryotes and that of the living environment represented by electron coordinated through four nitrogen atoms (Eschenmoser, 1988).
acceptors. After Russell and Hall (1997). Different metal ions can bestow remarkably different properties
Figure 20. Table of the “universal” genetic

code. Side chains of amino acids are shown
to be mainly distinctive for three-letter co-
dons, which consist of First Base, Middle
Base, and Auxiliary Base. Hence, AAG
and AAA are codons for lysine but the only
codon for methionine is AUG (Fig. 21).
Concentrations indicate amino acids that
have been obtained in “prebiotic” synthe-
ses (Hennet et al., 1992; and see Marshall,
1994) and are therefore assumed to be the
commonest on the early Earth. Ornithine,
not analyzed for in Hennet et al. (1992), has
been tentatively assigned to the arginine co-
dons because of the similar nature of their
side chains and because arginine has not
been recorded in abiotic experiments. The
four starred amino acids have been shown
to attach to RNA strands that contain their
codons. (From Russell et al., 2003 and ref-
erences therein.)
on these organometallic compounds. Cobalt and nickel corri- many others, both bases and a variety of sugars can be formed at
noids are involved in biosynthesis; magnesium (and in acid solu- low temperature by the condensation of hydrogen cyanide and
tions, zinc) produces the chlorophylls used in photosynthesis; phosphoglycerate, respectively.
and iron produces the heme groups used for electron transfer and The carbon nitrogen ring compounds constituting some of
in oxygen chemistry (Eschenmoser, 1988). Pratt (1993) consid- the nucleic acid bases may have formed by the condensation of
ers the structures based on the macrocyclic corrin ligand to date HCN on a sulfide surface such as pyrite (Leja, 1982; cf. Sowerby
back four billion years. They certainly must be older than 3.8 Ga, and Heckl, 1998). Adenine (HCN)5, the most common of the
for photosynthesis employs a myriad of such structural variants bases, vital to energy storage as well as one of the components of
(Blankenship, 2002). RNA, may have formed this way (Oró and Kimball, 1961). And
guanine might also have been synthesized in hydrothermal con-
THE PARTICULAR PROBLEMS OF RNA SYNTHESIS ditions, but at low yields (Ferris et al., 1978). Uracil is a hydro-
lysis product of HCN oligomers (Voet and Schwartz, 1982). But
No special place in this “unintentional” world is given to how or whether unstable cytosine formed at this early stage is
RNA in our decentralized system beyond its being metaboli- not known.
cally useful and therefore a surviving molecule. Although an The synthesis of ribose phosphate, the particular pentose
unstable entity, once formed RNA would have been less mutable sugar attached to the bases, is also difficult to understand but
when secured upon a mineral surface, especially in the presence may have been the stable end product, assembled upon a macki-
of highly reduced fluids. Nevertheless, the synthesis of nucleic nawite surface, from a reaction between phosphorylated chiral
acids, composed of a phosphorylated ribose sugar attached to one glyceraldehyde (GA3P) and the achiral dihydroxy acetone phos-
of four possible bases, is a problem more daunting than that of phate (DHAP), themselves derived by condensation of formal-
the amino and carboxylic acids. As a start we note that pyrophos- dehyde adsorbed from the alkaline fluids upon FeS (Quayle and
phate, introduced through volcanoes to the early oceans, would Ferenci, 1978; Schulte and Shock 1995; Pontes-Buarques et al.,
have remained in solution in the relatively acidic ocean, although 2001; Rickard et al., 2001; Russell et al., 2003; Ricardo et al.,
some would have been precipitated as vivianite (Fe2(PO4)2.8H2O) 2004). If so a feedback or autocatalytic cycle may have been
and as a condensed pyrophosphate on mixing with moderate- initiated, with a contribution from activated hydrogen that acted
temperature alkaline fluids at the hydrothermal mound (Rouse as another sink for carbon dioxide (Russell et al., 2003). Phos-
et al., 1988; Yamagata et al., 1991; Russell and Hall, 1997; de phogluconate formed in this way decomposes exothermically to
Zwart et al., 2004). There are plausible hydrothermal sources of GA3P and DHAP again. Reaction between the ribose phosphate
NH3 and HCN to explain the synthesis of some of the bases of adhering to the mackinawite surface and the bases would then
RNA (Schulte and Shock, 1995; Shock et al., 1998). As noted by produce RNA.
Of course it may be that the particular nucleic acids used before polymers might have been packed together on a mineral surface,
life had fully developed were different, and/or that there were two if only ephemerally, takes some of the mystery out of how life
rather than four bases (e.g., Reader and Joyce, 2002), perhaps works or worked—how, for example, nucleic acid polymers may
adenine and uracil (Jimenez-Sanchez, 1995). We are forced to step have first adventitiously coded for peptides and proteins. Look-
over this period for lack of knowledge. Anyway, apart from the eas- ing at the origin of the code from a mineralogical perspective
ily synthesized adenine, only small concentrations of the rest were leads us to follow Woese (1967) in his view that genetic informa-
needed because, as we shall see, RNA polymers were the “molds” tion was first transferred directly by selection through a some-
that may have produced a myriad of peptide “casts.” what indiscriminate “codon-amino acid pairing,” which relied
upon the affinity of the shape and charge of the codon (a triplet of
THE ORIGIN OF THE CODE AND THE FIRST CODED three nucleic acids) to the shape and charge of the amino acid and
POLYMERS OF PARTICULAR CHIRALITY especially of its side chain (Woese et al., 1966: Woese, 1967, p.
174–175). Thus, what is known as the peptidyl transferase reac-
Although the analogue of the acetyl-coenzyme-A pathway tion of an RNA molecule probably evolved via direct translation
provides a sink for carbon dioxide in the early atmosphere and on a protoribosome (Woese, 1967). This relationship happened
ocean, and it is possible to envisage how such a growing system to provide a rudimentary but direct coding to the polymerizing
would bud and reproduce, this is insufficient for replication and amino acid sequence.
evolution. For this a code was required, probably reliant on a Developing this idea, Mellersh (1993) emphasized that RNA
replicating and evolving RNA, to generate and order functionally triplets would only offer a cleftlike (tridentate) conformation to
useful peptides. In living cells amino acids are sequenced and attract amino acids when adhering to a solid phase. For such a
polymerized in a process centered on the ribosome, composed solid phase we favor mackinawite (Russell et al., 2003). We have
essentially of RNA, that ratchets along messenger RNA (Ban et already noted that phosphates may coat a mackinawite surface and
al., 2000). Such a complex process must have evolved from a there act as a random polymerizing agent for peptide formation.
simpler system. A hint of what this was is provided by the shape The phosphates of RNA also could have been bonded through to
of the RNA codons and the characteristics of the amino acid side sulfide on the surface of mackinawite in the same way (Fig. 21).
chains (Fig. 21). First and foremost we should see the affinity between the RNA
Geologists are less familiar with the structures of the car- triplets as offering a mechanism of polymerization more efficient
bon-, nitrogen-, and oxygen-bearing main-chain polymers and than the chance condensation of amino acids on a simple phos-
their various and characteristic side chains that constitute the phorylated mineral surface. The rows of RNA triplets could have
amino and nucleic acids than they are with the internal structure gripped and juxtaposed amino acid monomers in such a way as to
of minerals. Nevertheless, the way the side chains of the organic offer the carboxyl group of one to the amino group of the next for
Figure 21. Sketch to show how an amino acid may have been gripped by an RNA triplet (AUG) and offered its nucleophilic amino group to the
electrophilic carbon of the thiocarboxyl group of an adjacent amino acid to dimerize (Mellersh, 1993; cf. Muth et al., 2000, fig. 2). A nitrogen
atom acts as the basic binding site for the amino group on adenine (A), and oxygen on the next ribose (R) acts as the binding site of the matching
thiocarboxyl group. A peptide chain built incrementally in this way would be released by acid-base catalysis (Muth et al., 2000). There would
have been a tendency for the AUG triplet to act as the codon for methionine as shown here, if available in the FeS membrane (Mellersh, 1993;
Russell et al., 2003). If not, another hydrophobic amino acid at relatively high concentration was likely to have occupied the site.
bonding (Mellersh, 1993). At the same time the affinity between The chance stereochemistry of the short RNA polymer
the clefts of RNA and the side chains of the amino acids would would determine whether it catalyzed the polymerization of D-
happen to effect crude selection by codon-amino acid pairing as or L-amino acids into peptides (Mellersh, 1993). To achieve a
envisaged by Woese (1967). For example, clefts in which ura- low-energy state, as with mineral growth, we might expect RNA
cil was the central base would tend to attract only amino acids to tend to lengthen while preserving either left or right chiralities,
with hydrophobic side chains such as that of methionine (NH3+. i.e., a favorable packing arrangement (Joyce et al., 1984). Were
CH2.(CH2)2.SH.CH3.CSO-), whereas those in which adenine was a monomer with the opposite stereochemistry to be added to a
central would show affinity for the hydrophilic charged or polar growing chain, growth would be thwarted (Sandars, 2003). That
amino acids (Figs. 20, 21). filter would have been sufficient to tip the holochirality scale
We assume amino and nucleic acids would have occupied because, despite the presence of a racemic mixture of amino
the iron sulfide cells, though the former would have vastly out- acids in the microcavities, only amino acids of the same α-carbon
weighed the latter, and the crude coding of, or translation to, configuration (similar stereochemistry) would preferentially have
peptides would have been a feed-forward process. RNAs on the ended up in peptides, to yield a population of distinctly handed
surface of nanocrystals may themselves have replicated, where peptides. Some of these peptides would eventually feed back in a
water activity was low, by Watson-Crick hydrogen bonding in hypercyclic manner to favor the syntheses of their “stereochemi-
which A bonded to U and G bonded to C (Béland and Allen, cally appropriate” polymerizing template.
1994). The “protoribosome” would then have operated as a repli- Eventually the more robust but less reactive DNA (deoxy-
case, via the replication of triplets (the “triplicase” of Poole et al., ribonucleic acid) molecules took over from RNA and thence
1999). The first cycle of replication provided an antisense codon, survived. Braun and Libchaber (2004) have demonstrated that
so that a GCC triplet (the codon for alanine and the most likely secondary convection and thermophoresis driven by temperature
first triplet) produced the antisense codon “read” in the oppo- gradients within microcavities in a hydrothermal mound could
site direction as GGC (glycine) (Trifonov, 2000). Therefore, as have concentrated, elongated, and driven the replication of DNA.
touched on above, if the initial triplet coded for a hydrophilic It remains to be seen if RNA could be elongated and replicated
amino acid (with adenine occupying the central position), then by the same process.
its opposite would have coded for a hydrophobic residue (with Although the codon-amino acid affinity concept explains
uracil as the central base) (Béland and Allen, 1994; Konecny et why the chiralities of polynucleotides and peptides in life are
al., 1995). opposite, a relationship not required by Crick’s (1968) frozen
Point (single base) mutations on these triplets would have accident hypothesis, it does not explain why, on our planet, dex-
tended to code for the amino acids synthesized by Hennet at al. tro-DNA and RNA code for levo-proteins. As the energetic dif-
(1991), i.e., glycine (GGC), alanine (GCC), aspartate (GAC), and ferences between right- and left-handed chiral molecules, even
serine (UCC). More astonishing is the fact that these amino acids for thiosubstituted DNA analogues, are negligible (MacDermott
constitute the common sequence in that group of metalloproteins et al., 1992) and would have been “lost in the noise” within a
acknowledged to have the longest pedigree, the ferredoxins (Eck natural hydrothermal reactor, we can only conclude that both chi-
and Dayhoff, 1966; Trifonov et al., 2001; Russell et al., 2003). ral forms emerged separately, but that at some unknown stage the
This system of direct coding would have been relatively robust, present pairing survived, either by fortuitously stealing a march
in that mutations not involving the central RNA monomer would on the other, perhaps through the chance development of a better
have attracted amino acids with similar side chains and thereby ribosome, or later in a chirality war between the rival prokaryotes
similar properties. within the biosphere.
However, during the organic takeover the protoribosome
would have required another surface in place of mackinawite. A HYDROPHOBIC ORGANIC MEMBRANE
This might have been supplied by a peptide sequence rich in pos-
itively charged side chains. Such a peptide would have attracted As abiotic lipids of equal carbon chain lengths are unlikely
the phosphates of RNA that they might polymerize and still offer to have been delivered to, and survived on, early Earth in quan-
the triplet clefts. Lysine, arginine, and ornithine would have tities large enough to have allowed continued reproduction
been equally useful in such a peptide (Fig. 20). Mellersh and (Cairns-Smith, 1982), we have been left to consider FeS bounded
Wilkinson (2000) have demonstrated that poly-adenosine, which “cells” as the hatcheries of life. Short noncoded peptides gen-
includes the clefts AAA expected to have affinity for lysine, does erated in hydrothermal conditions (Russell et al., 1994; Huber
stereoselectively bind L-lysine from dilute aqueous solution of and Wächtershäuser, 1998, 2003; Ferris et al., 1996) could have
L-amino acids (Mellersh and Wilkinson 2000). Moreover, about played an important role in improving membrane characteris-
half the amount of L-arginine and L-ornithine also was found tics. These peptides and other polymers produced in the mound
to bind with poly-adenosine. As adenosine was likely the most would have coated the inside of inorganic compartments and par-
common of the nucleic acids, and lysine and probably ornithine tially plugged pore spaces. Excess organic sulfides and nitrides
can be made abiotically, then we have the makings of a feedback also could have been “entropy driven” into this, the first organic
cycle that involves the transfer of information. membrane (Cole et al., 1994). We imagine these polymers to
have cohered to form proteinaceous membranes and walls to the full reduction of carbon dioxide (Amend and Shock, 2001).
protocells—organic structures that offered several advantages. Moreover, the catalytic/enzymatic machinery required is similar
Eventually, genetically controlled proteinaceous cell envelopes (Thauer, 1998; Fontecilla-Camps and Ragsdale, 1999).
composed substantially of hydrophobic heterochiral peptides, for Before their release from the mound the first cells would
example, would have the advantage of including metal clusters have responded to environmental differences deterministically
such as the [Fe4S4] centers within their structure as stabilizers and in a manner more comparable to an ecosystem than to an indi-
electron transfer agents. At a later stage lipids would have been viduated cell. Compartments near the edge of the hydrothermal
generated from acetyl-coenzyme-A by continued addition of C2 mound, those most distant from the “inorganic formative fluid”
components, perhaps until the stable C12 fatty acid (dodecanoate) (cf. Haeckel, 1892), lie in the steepest chemical, electrochemical,
was realized (Fig. 12) (Shock et al., 1998). Once this happened, and thermal gradients but at lower temperature. It is here that H+,
lipids would have more efficiently denied protons an uncon- FeOOH, PPi, and CO2 were most concentrated, and this was where
trolled short circuit back into the cellular interiors. we assume the onset of life to have taken place at around 40 °C
(Fig. 14). At this early stage, mineral circumscribed compart-
EARLY EVOLUTION AND THE COMMON ments in the more restricted environment below this distal group
ANCESTRAL COMMUNITY and closer to the hydrothermal fluid were hotter, and though the
gradients would have been lower and the immediate environ-
The universal ancestor of life probably comprised a comment would have been more depleted in the constituents men-
munity of single-celled organisms still housed within its hydro- tioned above, concentrations of CO, H2, thiols, and the abiotic
thermal hatchery that possessed all of the attributes common to amino acids would have been higher. In this proximal zone ele-
all Bacteria and Archaea: the genetic code; the ribosome; DNA; mental sulfur polymers, generated in the atmosphere by photol-
a supporting core and intermediate metabolism needed to supply ysis of H2S and SO2 (Pavlov and Kasting, 2002) and sedimented
the constituents of its reproduction; replication; compartmenta- within the mounds, could stand in as an electron acceptor in
tion from the environment; redox chemistry; and the use of a pro- place of ferric iron, albeit at a lower potential (Fig. 6D) (Stet-
ton gradient. This last common community (the LCC of Woese, ter and Gaag, 1983). Alternatively, if cells deeper in the mound
1998; Macalady and Banfield, 2003) existed in the hydrothermal were beyond the reach of external electron acceptors, another
mound at the dawn of the biochemical revolution where genes way of ridding the system of excess reductant was through the
and proteins were diversifying into a myriad of functions, where discharging of electrons to the oceanic and atmospheric sink in
metal sulfide catalysts were being replaced by proteins, where CH4 (de Duve, 1991).
new pathways and cofactors were being invented to augment Evolution in the mound extended beyond mere optimization
and substitute their mineral and RNA precursors, where FeS was of the acetate and methane reactions (Martin and Russell, 2003).
being incorporated into proteins as Fe(Ni)S clusters, an imprint A next step was adaptation that exploited the reduced carbon and
of which would be reflected in the FeS centers of ancient protein, energy to be found in waste products and dead cells:
and where biochemistry started to diversify into the forms that
were both possible and useful (Eck and Dayhoff, 1966; Hall et CH3.COO- + 8FeIII + 4H2O → 8Fe2+ + 2HCO3- + 9H+ (19)
al., 1971; Martin and Russell, 2003) (Figs. 14, 19).
From the standpoint of protein structure, this age of inven- The prior use of FeIII as an electron acceptor during autotrophic
tion would have witnessed the origin of basic building blocks biosynthesis (equation 14) provided a means of such respiration
of biochemical function that (i) are conserved at the level of 3D (oxidative metabolism) (Vargas et al., 1998). Other potential
structure among Bacteria and Archaea and (ii) are recognizable electron acceptors were photolytic S0 and MnIV (Figs. 6D, 18)
as functional modules in various electron transporting proteins (Nealson and Stahl, 1997; Bahcall et al., 2001; Baymann et al.,
(Beinert et al., 1997; Baymann et al., 2003). From the standpoint 2003).
of amino acid sequence complexification, this age of invention We conclude that the last common ancestral community
would have been a phase of molecular evolution where proteins occupied the very hydrothermal hatchery in which life first
were diversifying and improving their efficiency (Baymann et emerged. The proto-Bacteria were initially suited to low to mod-
al., 2003). erate temperatures, and the proto-Archaea originally evolved to
It has been argued that the first acetogens were the forerun- withstand the shock of relatively high temperatures (i.e., ~60 °C)
ners of the Bacteria (Fig. 14) (Russell and Martin, 2004). We (Fig. 14). But the propensity to live well above 40 °C was passed
suggest that a minority of these cells, derived from those that back to the nascent Bacteria through genetic transfer. A period of
emerged at around 40 °C, exploited the potential offered at higher high ambient temperature, caused either by a meteorite impact or
temperature deeper in the mound where the kinetic energy was by a carbon dioxide greenhouse (Kasting and Ackerman, 1986;
greater and the activation energy required for reduction, through Kasting and Brown, 1998; Nisbet and Sleep, 2001) could explain
acetate (equation 8), all the way to methane (equation 7), was why the last common community may have been thermophilic,
lower. These first so-called methanogens may have evolved perhaps living at 50–60 °C (Gaucher et al., 2003 but see Brochier
while still in the mound, as there is even more energy to be had in and Philippe, 2002).
DIFFERENTIATION INTO TWO DOMAINS toxic intermediates produced by thermodegradation at high tem-
perature. Srere (1987) points out that the clustering of the func-
Although the thermodynamic drives would be lower in tionally related genes responsible for these complexes would also
proximal compartments of the mound (Schoonen et al., 1999), have conferred evolutionary advantage when returned to meso-
the lower kinetic barriers to reductions and condensations at thermal conditions.
these higher temperatures would encourage reaction. In these Let us recall the importance of a regulated dynamic system.
hotter compartments at first there would have been no RNA We could show that the hydrothermal system was both thermo-
regulation of peptides, and no replication or evolution. This is stat and chemostat. This took the onus off the living system to be
because RNA is unstable at high temperature (Poole et al., 1999). a thermostat but it would have needed its own control system for
But early biochemical evolution in the outermost compartments governing the internal state of the protocells. Such a regulatory
could have produced some “thermotolerant” DNA from RNA power is known as homeostasis. It probably emerged as newly
that could have invaded the metabolic husks of those below and generated protons, driven out of the cell by electron transfer, kept
co-opted this poorer, “thermochallenging” environment (Fig. 22) the cell neutral to alkaline. This process was augmented by the
(Forterre, 1995, 2002). Glansdorf has argued that cotranslation oxidative formation of disulfides such as the amino acid dimer
of functionally related proteins from integrated anabolic genes cystine, from the monomer cysteine, by protons (Russell et al.,
“facilitated the formation of multienzyme complexes” that chan- 1994):
neled thermolabile substrates that could invade hotter environ-
ments (Glansdorff, 1999, p. 432). Here inherently thermolabile 2RS- + 2H+ ⇔ H2 + RSSR (20)
proteins acted to stabilize and protect the whole ensemble (For-
terre, 1995). Operons—linear sequences of genes transcribable At the point of differentiation or bifurcation into the two pro-
as a single unit together with a regulatory operator—emerged as karyotic domains, we see the precursors to the Bacteria occu-
a response to these increasing temperatures (Glansdorff, 1999). pying the broad front of the growing hydrothermal mound at
Such operons would have facilitated the production of multien- its interface with the ocean. The precursors to the Archaea, the
zyme complexes capable of reducing the deleterious effects of sturdy but slowly evolving second domain of the prokaryotes,
Figure 22. Chemosynthetic life emerges at a warm alkaline seepage and at (A) differentiates into the precursors of the Bacteria and Archaea,
and expands downward into the surrounding sediments and crust (Martin and Russell, 2003; Russell and Martin, 2004). From here a proportion
is conveyed by ocean floor spreading toward a constructive margin (B) produced by obduction. Once uplifted at the margin, some of the cells
happen to invade sediments in the photic zone where, at a sulfurous spring, some evolve to exploit solar photons. Numbers 1–3 relate to life’s
emergence, and 4 marks the point of differentiation of the Archaea from the Bacteria. Roman numerals V and VI mark evolutionary stages of
the Archaea, and 5 and 6 indicate stages of evolution of the Bacteria in the deep biosphere. Photon energy was first mastered by the green sulfur
bacteria (7), followed by the heliobacteria (8). These photosynthesizing bacteria had appeared at least by the early Archaean (Westall et al.,
2001). Oxygenic photosynthesis (9) is a further development that may have evolved at a manganiferous hot spring by 3.75 Ga. (Various scales.)
(From Russell and Arndt, 2005.)
bring up the rear (Figs. 14, 22) (Woese et al., 1990; Russell and ous MoS42- and FeS (Russell et al., 1994; Helz et al., 1996). This
Hall, 2002). Both the proto-Bacteria and the proto-Archaea lived “Siamese-like twin” possibly contains a nitrogen atom in the cen-
up to the opposite edges of reproductive viability, the former at tral site (X in Fig. 9F) (Einsle et al., 2002; Smith, 2002). Exactly
a distance from fuel and challenged by kinetics at low tempera- where and how nitrogen is reduced to ammonia via the nitride
ture, the latter subject to the dangers of pyrolysis. Certainly the N3- on this metal sulfide center is as yet unresolved (Seefeldt et
Archaea appear to be the conservative cousins of the Bacteria, as al., 2004).
though they have had to hoard resources internally and defend
themselves against untoward perturbations. In the anaerobic OBDUCTION AND PHOTOSYNTHESIS
environment that obtained in the mound at the dawn of life, some
of the proto-Archaea probably lived off redox reactions that Bac- We have noted that conditions for the earliest cells in the open
teria have never mastered, relying on organic sulfides in a series sea were inhospitable and periodically impossible. How then do
of electron donations and generating methane waste (an electron we explain the emergence of photosynthetic organisms in the full
carrier) from carbon dioxide without a metallic electron acceptor glare of hard UV from the young sun (Canuto et al., 1982)? The
(Schäfer et al., 1999; Amend and Shock, 2001). first step was for organisms to have approached the photic zone
Even then, the proto-Bacteria and the proto-Archaea found with the “safety of numbers.” Because of the particular geometry
it advantageous to live syntrophically both within and across of Hadean oceanic crust, buoyant sediment and hydrated basaltic
domains. Cells would have interacted with their neighbors by crust piled up over the subducting, delaminated, eclogitized lower
swapping genes, providing some of the nutrients, and removing parts of the slab (Russell and Arndt, 2005). There were no deep
some of the waste. Unfortunates that were entrained within the ocean trenches. Obduction and uplift of oceanic sediments, and of
hydrothermal solution and dispersed to the relative desert of the the hydrated basalt beneath, passively transported some bacterial
uncertain ocean could not have survived such vicissitudes and colonies into shallow water and into the photic zone on the mar-
dilution of nutrient (e.g., Bjerrum and Canfield, 2002). The only gins of volcanic chains (cf. Margolis et al., 1978) (Fig. 22). Cells
safe migration route was down onto the ocean floor and into the were protected from deleterious solar radiation beneath a mineral
warm sediments and permeable basalts below, where the essen- coating (Cockell and Knowland, 1999). Opportunistic protection
tials—H2 and CO2—were assured. by superposed minerals and mineral excretions are well-known
Thus we conclude that the most significant of all cellular survival gambits (Phoenix et al., 2001). In these conditions some
differentiations, that between the Bacteria and the Archaea, prob- Bacteria near the surface augmented their protective shield by
ably took place before the mound was evacuated (Koga et al., developing a UV pigment protector from a ring of organic bases.
1998; Martin and Russell, 2003), although up till this time of Pigments comprising macrocyclic aromatic rings probably date
divergence, genes were shared in what may be called a cellular back to at least 4 Ga (Pratt, 1993). Single ions of Fe, Mg, Zn, Co,
cooperative (Fig. 14). This differentiation of the precursors of the and Ni could have been sequestered in variants of what is known
Archaea from those of the Bacteria was expedited by entropy and as the corrin ring, itself comprising four C/N rings (Eschenmoser,
the random changes in genes it caused. It eventually produced the 1988). Pigments developed for photoprotection could then have
two mutually exclusive genotypes (Wicken, 1987). been adapted as electron transfer agents, as photosynthetic reac-
tion centers and antenna proteins (Fig. 2B) (Mulkidjanian and
ESTABLISHMENT OF THE DEEP BIOSPHERE Junge, 1997; Allen, 2004).
Sediments in and overlying an obducted pile are likely, in
As the proto-Archaea and the proto-Bacteria began to colo- places, to have been subjected to hydrothermal H2S of magmatic
nize their surrounds, they eventually found themselves expand- or metasomatic derivation (Fig. 22). The first photosynthesist
ing into the sediments and volcanic horizons at the base of the may have been a precursor of the green sulfur bacteria. As the
mound (Fig. 22). Here conditions were not so different excepting name suggests, like many “primitive” bacteria, they relied on
the much lower flux of nutrient and fuel (Wolin, 1982). Thus the hydrogen sulfide as an electron donor (Baymann et al., 2001;
deep biosphere was inaugurated (Parkes et al., 1990, 1994; Ped- Blankenship, 2002). In these conditions, a reaction center (RC1)
ersen 1993). Once in the deep biosphere SO42– and N2 are likely was developed that could generate elemental sulfur as waste, and
to have joined FeIII, S0, and MnIV as electron acceptors (Fig. 18). gain electrons for biosynthesis in the process:
Although this probably happened quickly (Pinti, 2002; Shen and
Buick, 2004; Raymond et al., 2004) the age of these innovations H2S + CO2 + light → (CH2O)life + H2O + 2S0 (21)
is not known. We take the view that once chemical energy poten-
tials are available then their exploitation is relatively rapid. Yet at As we might expect of relatively gradualistic evolution, the
these early stages mineral-like clusters would again have played green sulfur bacteria retain a reliance on iron sulfide clusters as
a critical role in evolution. The metal center responsible for N2 electron transfer agents (Vermaas, 1994; Blankenship, 2002).
reduction is comparable to a greigite twin along the sulfur plane. An evolutionary variant of the fermenting bacteria, a photosyn-
excepting the presence of one proximal Mo atom comparable thetic precursor of the heliobacteria, could fix carbon dioxide by
to a naturally occurring MoFe3S4 cluster produced from aque- the concomitant oxidation of organic waste and detritus. At the
same time they used a similar reaction center (RC1) (Baymann

et al., 2001; Blankenship, 2002). These two reaction centers
hybridized to form what is known as photosystem 2 (PS2), the
photosystem invariably employed by all cyanobacteria and the
chloroplasts derived from them in plants (Michel and Deisen-
hofer, 1988; Allen, 2005). PS2 works in conjunction with the
first reaction center (RC1), which evolved into photosystem 1
(Baymann et al., 2001). Photosystem 2 is capable, in conjunction
with photosystem 1, of oxidizing two molecules of H2O (cf. H2S
in the green sulfur bacteria and (CH2O)x in the halobacteria) for
the generation of every one molecule of O2, while gaining four
electrons and four protons for the fixation of carbon from CO2
for biosynthesis in the process (Hansson and Wydrzynski, 1990;
Allen, 2005):
2H2O + light + PS2oxidized → 4H+ + 4e- + PS2excited + O2 (22)
To generalize:
H2O + CO2 + light → (CH2O)life + O2 (23)

Figure 23. Mn and Fe plots (GWB) of species and phases in simi-
Reaction 22 invariably involves the oxygen evolving center lar conditions. Activities of Mn2+ and Fe2+ = 10−6; fugacity of CO2
(OEC) in PS2. The OEC is a [CaMn4]-structure consisting of a = 1 (~3000 times the present atmosphere). Pyrolusite is suppressed
trigonal pyramid with calcium at the apex, 3.4Å from each of the to favor hydrated and mixed valence oxides and hydroxides such as
three manganese atoms at its base: a distal manganese lies in the birnessite [(Na,K,Ca)(Mg,Mn)Mn6O14.5H2O)] and the thermodynami-
same plane as the other three and 3.3Å from its nearest neighbor cally uncharacterized ranciéite [CaMn4O9.3H2O]. In theory, clusters
of birnessite and ranciéite have the potential to oxidize water to the
(Loll et al., 2005). One of the manganese atoms constituting the peroxide at low pH.
base of the pyramid is also 3.3Å from another, whereas the other
two are 2.7Å apart. To explain such an extraordinary innovation
Russell and Hall (2001, 2002) assumed that anoxygenic photo-
synthesizers in subaerial hot spring carbonate pools adsorbed to take up a shape within the membrane that would have facili-
exhalative manganese precipitates onto their membranes, per- tated the oxidation of water (Figs. 2D, 23, 24). Invagination and
haps while using photolytic MnIV as an electron acceptor (Fig. evolution of the protein led to the emergence of the cyanobac-
2C) (Burns and Burns, 1979; Myers and Nealson, 1988; Chafetz teria, organisms that were eventually to change the face of the
et al., 1998). In reduced form a manganese coating could protect planet to its present blue-green cast (Dismukes et al., 2001).
the cells from hard UV injury (cf. Daly et al., 2004). Photo-oxi- Unless one considers the Banded Iron Formations to be a
dation would generate ranciéite (the calcium-bearing birnessite, good indicator (Holm, 1987; Dymek and Klein, 1988), the impact
CaMn4O9.3H2O) (Anbar and Holland, 1992; Russell and Hall, of oxygenic photosynthesis is not discernable until the early
2001; 2002, fig. 6). Thus, a cluster of ranciéite may have con- Proterozoic (Bekker et al., 2004). However, Rosing (1999) and
tributed the “ready-made” [CaMn4] structure that was co-opted Rosing and Frei (2004), remarking on the apparently high rates
by one of the reaction centers, though if so the Mn-Mn distances of organic production, advance morphological, as well as stable
of 2.9Å characteristic of the cluster constrained in ranciéite must and radiogenic isotopic evidence, to favor an origin of oxygenic
have been modified to a conformation more typical of hollandite photosynthesis prior to 3.7 Ga. And there is independent sulfur
where two Mn-Mn distances are 2.7Å and two are 3.3Å (Sauer isotope evidence for the presence of trace atmospheric oxygen
and Yachandra, 2004; Loll et al., 2005). The manganese cluster, by 3.5 Ga (Ohmoto et al., 1993; Shen and Buick, 2004). Because
once sequestered in this reaction center and stabilized by the cal- we can see no reason why, given the appropriate conditions, the
cium atom, could have oxidized two water molecules, passing emergence of oxygenic photosynthesists from anoxygenic pho-
four protons and four electrons from the hydrogen to the cell for tosynthetic bacteria should have taken any longer than the emer-
biosynthesis while the oxygen went to waste (Fig. 23) (Russell gence of life, these inferences, surprising to some (e.g., Blank,
and Hall, 2001). Without the calcium atom the structure would be 2004), seem entirely reasonable. The long lag time between the
irreversibly reduced to the cubane moiety of the spinel hausman- photobacterial production of oxygen waste and its appearance
nite (Russell and Hall, 2002, fig. 6). As it is the structure readily as an atmospheric gas in the early Proterozoic (Farquhar et al.,
reverts to the oxidized form. Thus we conclude that a minimum 2000) must then be explained by the reductive capacity of the
of genetic control was required for a hydrated [CaMn4] complex planet, particularly of reduced iron, organic detritus, and biogenic
Figure 24. Electron pathways of oxygenic

photosynthesis for cyanobacteria for pH =
7 though photosystem 2 (PS2) is displaced
for clarity. In non-cyclic photophosphory-
lation NADP+ is reduced by electrons
from PS1 to make NADPH; PS2 provides
an electron to replenish PS1 whereas dis-
sociation of water to evolve oxygen pro-
vides the electron to replenish PS2. In
cyclic photophosphorylation only PS1 is
used and NADP+ does not receive an elec-
tron; protons are transferred out of the cell,
generating a pmf that produces ATP. Pho-
tosynthetic production of NADPH, ATP
and protons contribute to the synthesis of
carbohydrate from CO2. Light provides
an alternative source of redox energy for
H2 provision of chemical energy (Fig. 2)
although FeS remains at the core of the
energy-transfer system (cf. Blankenship,
2002, Figure 11.7).
methane (Lécuyer and Ricard, 1999; Kasting, 2001; Catling et ing temperature though the kinetics become ever more sluggish
al., 2001; Bjerrum and Canfield, 2002; Bekker et al., 2004). (Shock, 1992). The main products of the reaction are likely to
have been acetate and water (Huber and Wächtershäuser, 1997).
CONCLUSIONS Other products such as glycine (amino acetate), other amino
acids, and traces of nucleic acids constituted the first organic
The onset of life was not a haphazard affair but the meta- molecules (Russell and Hall, 1997, 2002). At this moderate tem-
stable evolutionary outcome of focused reactions between hydro- perature the acetate reaction had to be catalyzed (Shock et al.,
thermal hydrogen and bicarbonate in the ocean, with inputs from 1998). Mackinawite (FeS) and greigite (NiFe5S8) nanocrysts
trace metals, phosphate, and ammonia. Considering the fragility comprising the first membranes are significant in this respect,
of some organic polymers, especially RNA, and the theoretical and plausible evolutionary steps may be imagined between these
and experimental support for pyrophosphate and organic syn- and the fully fledged enzymes with [NiFe4S5] centers involved in
thesis at moderate temperatures, ~40 °C is a likely temperature the generation of acetate to this day (Russell and Martin, 2004).
for life’s emergence (Moulton et al., 2000). Life would have Energy from the acetate reaction, augmented by a natural pro-
emerged as soon as such a temperature was realized—a state tonmotive force, the consequence of the pH and redox gradients
probably reached, at least intermittently, by 4.3 Ga (Nisbet and acting across the semiconducting and semipermeable membrane,
Sleep, 2001). Convection cells operating within the oceanic crust was coupled to the formation of pyrophosphate. In turn the
in the relative tranquility of ridge flanks or in the deep ocean floor energy in the phosphate bond drove polymerization. Small quan-
would have produced hydrothermal mounds—mounds that also tities of amino acids, metal-bearing clusters, and eventually RNA
would have acted as chemical fractionation reactors. The free precursors, self-organized to become involved in the more effi-
energy of reaction between H2 and CO2 increases with decreas- cient generation of peptides and acetate waste, a thermodynamic
imperative. RNA genetic regulators eventually evolved to the the oxidation of ferrous iron kept oxygen concentrations vanish-
state where they could be passed on to offspring and be shared ingly low until the early Proterozoic (Lécuyer and Ricard, 1999;
with their neighbors (Mellersh, 1993; Hanczyc et al., 2003; Koo- Catling et al., 2001).
nin and Martin, 2005). Amyloidal peptides and other polymers Although compartmentalized, the biosphere overall is auto-
eventually took over from iron sulfides as membrane and cell trophic. Bernal (1960) put it well: “Life, geologically speaking,
wall constituents. consists of the interference with secondary lithosphere-atmo-
Pressurized microflow and circulation reactors could be used sphere reactions so as to produce a small but ever-renewed stock
in parallel and series to test these aspects of the model (Russell et of organic molecules.” Energized mainly by the sun’s rays, the
al., 2003; Braun and Libchaber, 2004). two essential contributors to this stock are (i) hydrogen, gained
The first gene-swapping cellular cooperative would have for the most part from water through photosynthesis with a minor
emerged at moderate temperature where chemical and electro- component released during the hydrous oxidation of ferrous iron,
chemical gradients were high near the surface of the mound. and (ii) carbon dioxide released during volcanism, metamor-
These acetogenic cells were the ancestors of the Bacteria, organ- phism, subduction, and reoxidation or fermentation of biogenic
isms that evolved while still within the mound to use electron detritus. Of course there are all kinds of continuing interactions
acceptors other than CO2 and FeIII, perhaps in the order MnIV between the biosphere and lithosphere. However, we can con-
and S0 (Baymann et al., 2001). But free energy was also to be clude that the autogenic view of the emergence of life posed here
had deeper within the mound where temperatures were higher, sits well with the way the biosphere operates, and operated, right
though gradients may have been lower. Here a small number of from its inception.
cells may have been able to withstand higher temperatures where
the methane reaction was favored. Eventually specializing in ACKNOWLEDGMENTS
generating methane waste, these methanogenic cells differenti-
ated from the parent population (the last common community) We thank Bill Martin, Anthony Mellersh, Everett Shock,
to become the forerunners of the Archaea, the second domain of Wolfgang Nitschke, James Milner-White, John Allen, Nick Arndt,
prokaryotic life (Martin and Russell, 2003). Both the Bacteria Jacques Meyer, and Minik Rosing for discussions, although not
and the Archaea expanded into environments offering compara- all of the views expressed here may coincide with theirs.
ble conditions. Although nutrient supply was at a premium, only
the oceanic sediments and crust could support these early vulner- REFERENCES CITED
able communities. Living syntrophically, the prokaryotes inau-
gurated the deep biosphere, well out of the way of the early and Abe, M., and Ooe, M., 2001, Tidal history of the Earth-Moon dynamical system
before Cambrian age: Journal of the Geodetic Society of Japan, v. 47,
late bombardments (Parkes et al., 1990). Here they developed the p. 514–520.
capacity to use SO42- and N2 as electron acceptors. Adman, E.T., Sieker, L.C., and Jensen, L.H., 1973, The structure of a bacterial
Convection remained a driving force for evolution. At a large ferredoxin: Journal of Biological Chemistry, v. 248, p. 3987–3996.
Allen, J.F., 2004, Cytochrome b6f: Structure for signaling and vectorial
scale, plate tectonics drove portions of the deep biosphere into metabolism: Trends in Plant Science, v. 9, p. 130–137, doi: 10.1016/
the photic zone on ocean island chains. Supplied with nutrients j.tplants.2004.01.009.
(e.g., H2 and H2S) and trace metals from subaerial hot springs Allen, J.F., 2005, A redox switch hypothesis for the origin of two light reac-
tions in photosynthesis: FEBS Letters, v. 579, p. 963–968, doi: 10.1016/
(Ca, Fe, Mn, Ni, Co) and protected by chemical and detrital sedi- j.febslet.2005.01.015.
ment, some of the Bacteria adapted the power of photons to drive Allen, D.A., and Seyfried, W.E., 2003, Compositional controls on vent fluids
electron transfer and thereby metabolism. The transition ele- from ultramafic-hosted hydrothermal systems at mid-ocean ridges: An
experimental study at 400°C, 500 bars: Geochimica et Cosmochimica
ments could be photo-oxidized, only to be re-reduced by organic Acta, v. 67, p. 1531–1542, doi: 10.1016/S0016-7037(02)01173-0.
molecules. Accumulation of ambient manganese within certain Amend, J.P., and Shock, E.L., 1998, Energetics of amino acid synthesis in
photosynthetic bacteria is likely to have protected them from hydrothermal ecosystems: Science, v. 281, p. 1659–1662, doi: 10.1126/
science.281.5383.1659.
UV injury (cf. Daly et al., 2004). Once within the cell, a rancié- Amend, J.P., and Shock, E.L., 2001, Energetics of overall metabolic reactions of
ite cluster [CaMn4O9.3H2O] was co-opted to act as the oxygen thermophilic and hyperthermophilic Archaea and Bacteria: FEMS Micro-
evolving center [CaMn4] ± 2H2O) which, indirectly excited by biology Reviews, v. 25, p. 175–243, doi: 10.1016/S0168-6445(00)00062-
0.
photons, extracted the hydrogen (as protons and electrons) from Anbar, A.D., and Holland, H.D., 1992, The photochemistry of manganese and
water. The protons and electrons were used in the reduction of the origin of banded iron formations: Geochimica et Cosmochimica Acta,
CO2 for biosynthesis while oxygen was emitted as waste. There v. 56, p. 2595–2603, doi: 10.1016/0016-7037(92)90346-K.
Arndt, N.T., 1999, Why was flood volcanism on submerged continental plat-
is no compelling reason to assume life required billions of years forms so common in the Precambrian?: Precambrian Research, v. 97,
to configure the process of oxygenic photosynthesis and several p. 155–164, doi: 10.1016/S0301-9268(99)00030-3.
reasons to suppose that this biosynthetic pathway had emerged Arndt, N.T., and Chauvel, C., 1990, Crust of the Hadean Earth: Bulletin of the
Geological Society of Denmark, v. 39, p. 145–151.
before our stratigraphic record began ca. 3.75 Ga (Rosing and Bada, J.L., 2004, How life began on Earth: A status report: Earth and Planetary
Frei, 2004). It can therefore be inferred that the core metabolic Science Letters, v. 226, p. 1–15, doi: 10.1016/S0012-821X(04)00470-4.
cycles were in place by that time (Pace, 2002). Reaction with Bahcall, J.N., Pinsonneault, M.H., and Basu, S., 2001, Solar models: Current
epoch and time dependences, neutrinos, and helioseismological proper-
biogenic methane, the use of oxygen as an electron acceptor, and ties: Astrophysical Journal, v. 555, p. 990–1012, doi: 10.1086/321493.
Baltscheffsky, M., Schultz, A., and Baltscheffsky, H., 1999, H+-PPases: A Campbell, I.H., Griffiths, R.W., and Hill, R.I., 1989, Melting in an Archaean
tightly membrane-bound family: FEBS Letters, v. 457, p. 527–533, doi: mantle plume: Heads it’s basalts, tails it’s komatiites: Nature, v. 339,
10.1016/S0014-5793(99)90617-8. p. 697–699, doi: 10.1038/339697a0.
Ban, N., Nissen, P., Hansen, J., Moore, P.B., and Steitz, T.A., 2000, The com- Canuto, V.M., Levine, J.S., Augustsson, T.R., and Imhoff, C.L., 1982, UV radia-
plete atomic structure of the large ribosomal subunit at 2.4 Å resolution: tion from the young Sun and oxygen and ozone levels in the prebiological
Science, v. 289, p. 905–920, doi: 10.1126/science.289.5481.905. palaeoatmosphere: Nature, v. 296, p. 816–820, doi: 10.1038/296816a0.
Banks, D.A., 1985, A fossil hydrothermal worm assemblage from the Catling, D.C., Zahnle, K.J., and McKay, C.P., 2001, Biogenic methane, hydro-
Tynagh lead-zinc deposit in Ireland: Nature, v. 313, p. 128–131, doi: gen escape, and the irreversible oxidation of early Earth: Science, v. 293,
10.1038/313128a0. p. 839–843, doi: 10.1126/science.1061976.
Banks, D.A., and Russell, M.J., 1992, Fluid mixing during ore deposition at Chafetz, H.S., Akdim, B., Julia, R., and Reid, A., 1998, Mn- and Fe-rich black
the Tynagh base-metal deposit, Ireland: European Journal of Mineralogy, travertine shrubs: Bacterially (and nanobacterially) induced precipitates:
v. 4, p. 921–931. Journal of Sedimentary Research, v. 68, p. 404–412.
Baymann, F., Brugna, M., Mühlenhoff, U., and Nitschke, W., 2001, Daddy, Chen, K., Hirst, J., Camba, R., Bonagura, C.A., Stout, C.D., Burgess, B.K., and
where did (PS)I come from?: Biochimica et Biophysica Acta, v. 1507, Armstrong, F.A., 2000, Atomically defined mechanism for proton trans-
p. 291–310. fer to a buried redox centre in a protein: Nature, v. 405, p. 814–817, doi:
Baymann, F., Lebrun, E., Brugna, M., Schoepp-Cothenet, B., Giudici-Orticoni, 10.1038/35015610.
M.T., and Nitschke, W., 2003, The redox protein construction kit: pre- Cockell, C.S., and Knowland, J., 1999, UV radiation screening compounds: Bio-
last universal common ancestor evolution of energy-conserving enzymes: logical Reviews, v. 74, p. 311–345, doi: 10.1017/S0006323199005356.
Philosophical Transactions of the Royal Society of London, ser. B, v. 358, Cody, G.D., 2004, Transition metal sulfides and the origin of metabolism:
p. 267–274, doi: 10.1098/rstb.2002.1184. Annual Review of Earth and Planetary Sciences, v. 32, p. 569–599, doi:
Beinert, H., Holm, R.H., and Münck, E., 1997, Iron-sulfur clusters: Nature’s 10.1146/annurev.earth.32.101802.120225.
modular, multipurpose structures: Science, v. 277, p. 653–659, doi: Cody, G.D., Boctor, N.Z., Brandes, J.A., Filley, T.R., Hazen, R.M., and Yoder,
10.1126/science.277.5326.653. H.S., 2004, Assaying the catalytic potential of transition metal sulfides
Bekker, A., Holland, H.D., Wang, P.-L., Stein, H.J., Hannah, J.L., Coetzee, L.L., for abiotic carbon fixation: Geochimica et Cosmochimica Acta, v. 68,
and Beukes, N.J., 2004, Dating the rise of atmospheric oxygen: Nature, p. 2185–2196, doi: 10.1016/j.gca.2003.11.020.
v. 427, p. 117–120, doi: 10.1038/nature02260. Cole, W.J., Kaschke, M., Sherringham, J.A., Curry, G.B., Turner, D., and Russell,
Béland, P., and Allen, T.F.H., 1994, The origin and evolution of the genetic M.J., 1994, Can amino acids be synthesised by H2S in anoxic lakes?: Marine
code: Journal of Theoretical Biology, v. 170, p. 359–365, doi: 10.1006/ Chemistry, v. 45, p. 243–256, doi: 10.1016/0304-4203(94)90007-8.
jtbi.1994.1198. Corliss, J.B., Baross, J.A., and Hoffman, S.E., 1981, An hypothesis concerning the
Bernal, J.D., 1960. The problem of stages in biopoesis, in Florkin, M. ed., relationship between submarine hot springs and the origin of life on Earth:
Aspects of the origin of life: New York, Pergamon Press, p. 30–45. Proceedings, 26th International Geological Congress, Geology of Oceans
Bethke, C., 1996, Geochemical reaction modeling: Oxford, UK, Oxford Uni- Symposium, Paris, July 7–17, 1980, Oceanologica Acta, no. SP, p. 59–69.
versity Press, 397 p. Crick, F.H.C., 1968, The origin of the genetic code: Journal of Molecular Biol-
Beutner, R., 1913, New electric properties of a semipermeable membrane of ogy, v. 38, p. 367–379, doi: 10.1016/0022-2836(68)90392-6.
copper ferrocyanide: Journal of Physical Chemistry, v. 17, p. 344–360, da Silva, J.J.R.F., and Williams, R.J.P., 1991, The biological chemistry of the
doi: 10.1021/j150139a006. elements: Oxford, UK, Clarendon Press, 561 p.
Bischoff, J.L., and Seyfried, W.E., 1978, Hydrothermal chemistry of seawater Daly, M.J., Gaidamakova, E.K., Matrosova, V.Y., Vasilenko, A., Zhai, M.,
from 25° to 350°C: American Journal of Science, v. 278, p. 838–860. Venkateswaran, A., Hess, M., Omelchenko, M.V., Kostandarithes, H.M.,
Bjerrum, C.J., and Canfield, D.E., 2002, Ocean productivity before about 1.9 Makarova, K.S., Wackett, L.P., Fredrickson, J.K., and Ghosall, D., 2004,
Gyr ago limited by phosphorus adsorption onto iron oxides: Nature, Accumulation of Mn(II) in Deinococcus radiodurans facilitates gamma-
v. 417, p. 159–162, doi: 10.1038/417159a. radiation resistance: Science, v. 306, p. 1025–1028, doi: 10.1126/sci-
Blank, C.E., 2004, Evolutionary timing of the origins of mesophilic sulphate ence.1103185.
reduction and oxygenic photosynthesis: A phylogenomic dating approach: Daniel, R.M., and Danson, M.J., 1995, Did primitive microorganisms use non-
Geobiology, v. 2, p. 1–20, doi: 10.1111/j.1472-4677.2004.00020.x. heme iron proteins in place of NAD/P?: Journal of Molecular Evolution,
Blankenship, R.E., 2002, Molecular mechanisms of photosynthesis: Oxford, v. 40, p. 559–563, doi: 10.1007/BF00160501.
Blackwell Science. Darnault, C., Volbeda, A., Kim, E.J. Legrand, P., Vernède, X., Lindahl, P.A.,
Bonatti, E., Simmons, E.C., Breger, D., Hamlyn, P.R., and Lawrence, J., 1983, and Fontecilla-Camps, J.C., 2003, Ni-Zn-[Fe4-S4] and Ni-Ni-[Fe4-S4]
Ultramafic rock/seawater interaction in the oceanic crust: Mg-silicate clusters in closed and open α subunits of acetyl-CoA synthase/carbon
(sepiolite) deposit from the Indian Ocean floor: Earth and Planetary Sci- monoxide dehydrogenase: Nature Structural Biology, v. 10, p. 271–279,
ence Letters, v. 62, p. 229–238, doi: 10.1016/0012-821X(83)90086-9. doi: 10.10.1038/nsb912.
Bonomi, F., Werth, M.T., and Kurtz, D.M., 1985, Assembly of FenSn (SR)2- Davies, G.F., 1992, On the emergence of plate tectonics: Geology, v. 20, p. 963–
(n=2,4) in aqueous media from iron salts, thiols and sulfur, sulfide, thio- 966, doi: 10.1130/0091-7613(1992)020<0963:OTEOPT>2.3.CO;2.
sulfide plus rhodonase: Inorganic Chemistry, v. 24, p. 4331–4335, doi: Deamer, D.W., 1985, Boundary structures are formed by organic components
10.1021/ic00219a026. of the Murchison carbonaceous chondrite: Nature, v. 317, p. 792–794,
Bounama, C., Franck, S., and von Bloh, W., 2001, The fate of the Earth’s ocean: doi: 10.1038/317792a0.
Hydrology and Earth System Sciences, v. 5, p. 569–575. de Duve, C., 1991, Blueprint for a cell: The nature and origin of life: Burling-
Boyce, A.J., Coleman, M.L., and Russell, M.J., 1983, Formation of fossil ton, North Carolina, Neil Patterson Publishers, 275 p.
hydrothermal chimneys and mounds from Silvermines, Ireland: Nature, de Zwart, I.I., Meade, S.J., and Pratt, A.J., 2004, Biomimetic phosphoryl trans-
v. 306, p. 545–550, doi: 10.1038/306545a0. fer catalysed by iron(II)-mineral precipitates: Geochimica et Cosmochi-
Braun, D., and Libchaber, A., 2004, Thermal force approach to molecular mica Acta, v. 68, p. 4093–4098, doi: 10.1016/j.gca.2004.01.028.
evolution: Physical Biology, v. 1, p. 1–8 (DOI: 10.1088/1478–3967/1/1/ Dismukes, G.C., Klimov, V.V., Baranov, S.V., Kozlov, Yu.N., DasGupta, J.,
P01), 8 p. and Tyryshkin, A., 2001, The origin of atmospheric oxygen on Earth:
Brochier, C., and Philippe, H., 2002, A non-hyperthermophilic ancestor for bac- The innovation of oxygenic photosynthesis: Proceedings of the National
teria: Nature, v. 417, p. 244, doi: 10.1038/417244a. Academy of Sciences of the United States of America, v. 98, p. 2170–
Burns, R.G., and Burns, V.M., 1979, Manganese oxides, in Burns, R.G., ed., 2175, doi: 10.1073/pnas.061514798.
Marine minerals: Washington, D.C., Mineralogical Society of America, Dobbek, H., Svetlitchnyi, V., Gremer, L., Huber, R., and Meyer, O., 2001, Crys-
Reviews in Mineralogy, v. 6, p. 1–46. tal structure of a carbon monoxide dehydrogenase reveals a [Ni-4Fe-5S]
Cairns-Smith, A.G., 1982, Genetic takeover and the mineral origins of life: cluster: Science, v. 293, p. 1281–1285, doi: 10.1126/science.1061500.
Cambridge, UK, Cambridge University Press, 477 p. Doukov, T.I., Iverson, T.M., Seravalli, J., Ragsdale, S.W., and Drennan, C.L.,
Cairns-Smith, A.G., Hall, A.J., and Russell, M.J., 1992, Mineral theories of the 2002, A Ni-Fe-Cu center in a bifunctional carbon monoxide dehydroge-
origin of life and an iron sulphide example: Origins of Life and Evolution nase/acetyl-CoA synthase: Science, v. 298, p. 567–572, doi: 10.1126/sci-
of the Biosphere, v. 22, p. 161–180, doi: 10.1007/BF01808023. ence.1075843.
Douville, E., Charlou, J.L., Oelkers, E.H., Bienvenu, P., Colon, C.F.J., Donval, Geptner, A., Kristmannsdöttir, H., Kristjánsson, J.K., and Marteinsson, V.Th.,
J.P., Fouquet, Y., Prieur, D., and Appriou, P., 2002, The rainbow vent fluids 2002, Biogenic saponite from an active submarine hot spring, Iceland:
(36°14′N, MAR): The influence of ultramafic rocks and phase separation Clays and Clay Minerals, v. 50, p. 174–185, doi: 10.1346/00098600276
on trace metal content in Mid-Atlantic Ridge hydrothermal fluids: Chemi- 0832775.
cal Geology, v. 184, p. 37–48, doi: 10.1016/S0009-2541(01)00351-5. Glansdorff, N., 1999, On the origin of operons and their possible role in evolu-
Drennan, C.L., Heo, J., Sintchak, M.D., Schreiter, E., and Ludden, P.W., 2001, tion toward thermophily: Journal of Molecular Evolution, v. 49, p. 432–
Life on carbon monoxide: X-ray structure of Rhodospirillium rubrum 438.
Ni-Fe-S carbon monoxide dehydrogenase: Proceedings of the National Goldschmidt, V.M., 1937, The principles of distribution of chemical elements
Academy of Sciences of the United States of America, v. 98, p. 11973– in minerals and rocks: Journal of the Chemical Society, v. 1937, p. 655–
11978, doi: 10.1073/pnas.211429998. 673.
Drobner, E., Huber, H., Wächtershäuser, G., Rose, D., and Stetter, K.O., 1990, Goldschmidt, V.M., 1952, Geochemical aspects of the origin of complex
Pyrite formation linked with hydrogen evolution under anaerobic condi- organic molecules on Earth, as precursors to organic life: New Biology,
tions: Nature, v. 346, p. 742–744, doi: 10.1038/346742a0. v. 12, p. 97–105.
Dymek, R.F., and Klein, C., 1988, Chemistry, petrology and origin of banded Haeckel, E., 1892, The history of creation, Volume 1 (4th edition), translated by
iron-formation lithologies from the 3800 Ma Isua Supracrustal Belt: West E.R. Lankester: London, Kegan Paul, Trench, Trübner and Co., 422 p.
Greenland: Precambrian Research, v. 39, p. 247–302, doi: 10.1016/0301- Haldane, J.B.S., 1929, The origin of life: Rationalist Annual, v. 3, p. 3–10.
9268(88)90022-8. Hall, D.O., Cammack, R., and Rao, K.K., 1971, Role for ferredoxins in the
Eck, R.V., and Dayhoff, M.O., 1966, Evolution of the structure of ferredoxin origin of life and biological evolution: Nature, v. 233, p. 136–138, doi:
based on living relics of primitive amino acid sequences: Science, v. 152, 10.1038/233136a0.
p. 363–366. Hall, A.J., Boyce, A.J., and Fallick, A.E., 1994, A sulphur isotope study of iron
Einsle, O., Tezcan, F.A., Andrade, S.L.A., Schmid, B., Yoshida, M., Howard, sulphide in the late Precambrian Dalradian Ardrishaig Phyllite Formation,
J.B., and Rees, D.C., 2002, Nitrogenase MoFe-protein at 1.16Å resolu- Knapdale, Argyll: Scottish Journal of Geology, v. 30, p. 63–71.
tion: A central ligand in the FeMo-cofactor: Science, v. 297, p. 1696– Hanczyc, M.M., Fujikawa, S.M., and Szostak, J.W., 2003, Experimental mod-
1700, doi: 10.1126/science.1073877. els of primitive cellular compartments: Encapsulation, growth and divi-
Elston, T., Wang, H., and Oster, G., 1998, Energy transduction in ATP synthase: sion: Science, v. 302, p. 618–622, doi: 10.1126/science.1089904.
Nature, v. 391, p. 510–513, doi: 10.1038/35185. Hansson, Ö., and Wydrzynski, T., 1990, Current perceptions of Photosystem II:
Eschenmoser, A., 1988, Vitamin B12: Experiments concerning the origin of its Photosynthesis Research, v. 23, p. 131–162, doi: 10.1007/BF00035006.
molecular structure: Angewandte Chemie International Edition in Eng- Heinen, W., and Lauwers, A.M., 1996, Organic sulfur compounds resulting
lish, v. 27, p. 5–39, doi: 10.1002/anie.198800051. from the interaction of iron sulfide, hydrogen sulfide and carbon dioxide
Farquhar, J., Bao, H., and Thiemens, M.H., 2000, Atmospheric influence of in an anaerobic aqueous environment: Origins of Life and Evolution of
Earth’s earliest sulfur cycle: Science, v. 289, p. 756–758, doi: 10.1126/ the Biosphere, v. 26, p. 131–150, doi: 10.1007/BF01809852.
science.289.5480.756. Heinen, W., and Lauwers, A.M., 1997, The iron-sulfur world and the origins
Ferris, F.G., Jack, T.R., and Bramhill, B.J., 1992, Corrosion products associ- of life: Abiotic synthesis from metallic iron, H2S and CO2: A comparison
ated with attached bacteria at an oil field water injection plant: Canadian of the thiol generating FeS/HCl(H2S)/CO2-system and its Fe0/H2S/CO2-
Journal of Microbiology, v. 38, p. 1320–1324. counterpart: Proceedings Koninklijke Nederlandse Akademie van Weten-
Ferris, J.P., and Hagan, W.J., 1984, HCN and chemical evolution: The pos- schappen, Amsterdam, v. 100, p. 11–25.
sible role of cyano compounds in prebiotic synthesis: Tetrahedron, v. 40, Helz, G.R., Miller, C.V., Charnock, J.M., Mosselmans, J.F.W., Pattrick, R.A.D.,
p. 1093–1120, doi: 10.1016/S0040-4020(01)99315-9. Garner, C.D., and Vaughan, D.J., 1996, Mechanism of molybdenum
Ferris, J.P., Hill, A.R., Liu, R., and Orgel, L.E., 1996, Synthesis of long pre- removal from the sea and its concentration in black shales: EXAFS evi-
biotic oligomers on mineral surfaces: Nature, v. 381, p. 59–61, doi: dence: Geochimica et Cosmochimica Acta, v. 60, p. 3631–3642.
10.1038/381059a0. Hemley, J.J., Cygan, G.L., Fein, J.B., Robinson, G.R., and D’Angelo, W.M.,
Ferris, J.P., Joshi, P.C., Edelson, E.H., and Lawless, J.G., 1978, HCN: A plau- 1992, Hydrothermal ore-forming processes in the light of studies in rock-
sible source of purines, pyrimidines and amino acids on the primitive buffered systems: I. Iron-copper-zinc-lead sulfide solubility relations:
Earth: Journal of Molecular Evolution, v. 11, p. 293–311, doi: 10.1007/ Economic Geology and the Bulletin of the Society of Economic Geolo-
BF01733839. gists, v. 87, p. 1–22.
Filtness, M.J., Butler, I.B., and Rickard, D., 2003, The origin of life: The prop- Hennet, R.J.-C., Holm, N.G., and Engel, M.H., 1992, Abiotic synthesis of
erties of iron sulphide membranes: Transactions of the Institution of Min- amino acids under hydrothermal conditions and the origin of life: A per-
ing and Metallurgy, Applied Earth Science, v. 112B, p. 171–172. petual phenomenon?: Die Naturwissenschaften, v. 79, p. 361–365, doi:
Finklea, S., Cathey, S., and Amma, E.L., 1976, Investigation of the bonding 10.1007/BF01140180.
mechanism in pyrite using the Mössbauer effect: Acta Crystallographica, Holm, N.G., 1987, Possible biological origin of banded iron formations from
v. A32, p. 529–537. hydrothermal solutions: Origins of Life and Evolution of the Biosphere,
Foley, S.F., Buhre, S., and Jacob, D.E., 2003, Evolution of the Archaean crust v. 17, p. 229–250.
by delamination and shallow subduction: Nature, v. 421, p. 249–252. Huber, C., and Wächtershäuser, G., 1997, Activated acetic acid by carbon fixa-
Fontecilla-Camps, J.C., and Ragsdale, S.W., 1999, Nickel-iron-sulfur active tion on (Fe,Ni)S under primordial conditions: Science, v. 276, p. 245–247,
sites: Hydrogenase and CO dehydrogenase: Advances in Inorganic Chem- doi: 10.1126/science.276.5310.245.
istry, v. 47, p. 283–333. Huber, C., and Wächtershäuser, G., 1998, Peptides by activation of amino acids
Forterre, P., 1995, Thermoreduction, a hypothesis for the origin of prokaryotes: on (Fe,Ni)S surfaces: Implications for the origin of life: Science, v. 281,
C.R: Academy of Science, v. 318, p. 415–422. p. 670–672, doi: 10.1126/science.281.5377.670.
Forterre, P., 2002, A hot story from comparative genomics: reverse gyrase is Huber, C., and Wächtershäuser, G., 2003, Primordial reductive amination
the only hyperthermophile-specific protein: Trends in Genetics, v. 18, revisited: Tetrahedron Letters, v. 44, p. 1695–1697, doi: 10.1016/S0040-
p. 236–237, doi: 10.1016/S0168-9525(02)02650-1. 4039(02)02863-0.
Früh-Green, G.L., Kelley, D.D., Bernascono, S.M., Karson, J.A.. Ludwig, Huber, C., Eisenreich, W., Hecht, S., and Wächtershäuser, G., 2003, A possible
K.A., Butterfield, D.A., Boschi, C., and Proskurowski, G., 2003, 30,000 primordial peptide cycle: Science, v. 301, p. 938–940, doi: 10.1126/sci-
years of hydrothermal activity at the Lost City vent field: Science, v. 301, ence.1086501.
p. 495–498, doi: 10.1126/science.1085582. Janecky, D.R., and Seyfried, W.E., 1983, The solubility of magnesium-hydrox-
Gaffey, M.J., 1997, The early solar system: Origins of Life and Evolution of the ide-sulfate-hydrate in seawater at elevated temperatures and pressures:
Biosphere, v. 27, p. 185–203, doi: 10.1023/A:1006578315384. American Journal of Science, v. 283, p. 831–860.
Garrels, R.M., and Christ, C.L., 1965, Minerals, solutions and equilibria: New Jernigan, R., Raghunathan, G., and Bahar, I., 1994, Characterization of interac-
York, Harper and Row, 450 p. tions and metal ion binding sites in proteins: Current Opinions in Struc-
Gaucher, E.A., Thomson, J.M., Burgan, M., and Benner, S.A., 2003, Inferring tural Biology, v. 4, p. 256–263, doi: 10.1016/S0959-440X(94)90317-4.
the palaeoenvironment of ancient bacteria on the basis of resurrected pro- Jimenez-Sanchez, A., 1995, On the origin and evolution of the genetic code:
teins: Nature, v. 425, p. 285–288, doi: 10.1038/nature01977. Journal of Molecular Evolution, v. 41, p. 712–716.
Joyce, G.F., 1989, RNA evolution and the origins of life: Nature, v. 338, p. 217– Loll, B, Kern, J., Saenger, W., Zouni, A., and Biesiadka, J., 2005, Towards com-
224, doi: 10.1038/338217a0. plete cofactor arrangement in the 3.0Å resolution structure of photosys-
Joyce, G.F., Visser, G.M., van Boeckel, C.A.A., van Boom, J.H., Orgel, L.E., tem II: Nature, v. 438, p. 1040–1044, doi: 10.1038/nature04224.
and Westrenen, J., 1984, Chiral selection in poly(C)-directed synthesis of Luther, G.W., 2004, Kinetics of the reactions of water, hydroxide ion and sul-
oligo(G): Nature, v. 310, p. 602–604, doi: 10.1038/310602a0. fide species with CO2, OCS and CS: Frontier molecular orbital consider-
Kakegawa, T., Noda, M., and Nannri, H., 2002, Geochemical cycles of bio- ations: Aqueous Geochemistry, v. 10, p. 81–97.
essential elements on the early Earth and their relationships to the origin Luther, G.W., Theberge, S.M., and Rickard, D.T., 1999, Evidence for aque-
of life: Resource Geology, v. 52, p. 83–89. ous clusters as intermediates during zinc sulfide formation: Geochi-
Karsten, J.L., Klein, E.M., and Sherman, S.B., 1996, Subduction zone geo- mica et Cosmochimica Acta, v. 63, p. 3159–3169, doi: 10.1016/S0016-
chemical characteristics in ocean ridge basalts from the southern Chile 7037(99)00243-4.
ridge: Implications of modern subduction systems for the Archean: Macalady, J., and Banfield, J.F., 2003, Molecular geomicrobiology: Genes and
Lithos, v. 37, p. 143–161, doi: 10.1016/0024-4937(95)00034-8. geochemical cycling: Earth and Planetary Science Letters, v. 209, p. 1–
Kashefi, K., Tor, J.M., Holmes, D.E., Gaw Van Praagh, C.V., Reysenbach, A.- 17, doi: 10.1016/S0012-821X(02)01010-5.
L., and Lovley, D.R., 2002, Geoglobus ahangari, gen. nov., sp., nov., a MacDermott, A.J., Tranter, G.E., and Trainor, S.J., 1992, The search for large
novel hyperthermophile capable of oxidizing organic acids and growing parity-violating energy differences finds fruit in thiosubstituted DNA ana-
autotrophically on hydrogen with Fe (III) serving as the sole electron logues: Chemical Physics Letters, v. 194, p. 152–156, doi: 10.1016/0009-
accepter: International Journal of Systematic and Evolutionary Microbi- 2614(92)85525-F.
ology, v. 52, p. 719–728, doi: 10.1099/ijs.0.01953-0. Macleod, G., McKeown, C., Hall, A.J., and Russell, M.J., 1994, Hydrothermal
Kasting, J.F., 1993, Earth’s early atmosphere: Science, v. 259, p. 920–926. and oceanic pH conditions of possible relevance to the origin of life: Ori-
Kasting, J.F., 2001, The rise of atmospheric oxygen: Science, v. 293, p. 819– gins of Life and Evolution of the Biosphere, v. 24, p. 19–41, doi: 10.1007/
820, doi: 10.1126/science.1063811. BF01582037.
Kasting, J.F., and Ackerman, T.P., 1986, Climatic consequences of very high Margolis, S.V., Ku, T.L., Glasby, G.P., Fein, C.D., and Audley-Charles, M.G.,
carbon dioxide levels in the earth’s early atmosphere: Science, v. 234, 1978, Fossil manganese nodules from Timor: Geochemical and radio-
p. 1383–1385. chemical evidence for deep-sea origin: Chemical Geology, v. 21, p. 185–
Kasting, J.F., and Brown, L.L., 1998, The early atmosphere as a source of bio- 198, doi: 10.1016/0009-2541(78)90044-X.
genic compounds, in Brack, A., ed., The molecular origins of life: Cam- Marshall, W.L., 1994, Hydrothermal synthesis of amino acids: Geochi-
bridge, UK, Cambridge University Press, p. 35–56. mica et Cosmochimica Acta, v. 58, p. 2099–2106, doi: 10.1016/0016-
Kell, D.B., 1988, Protonmotive energy-transducing mechanisms: some physi- 7037(94)90288-7.
cal principles and experimental approaches, in Anthony, C. ed., Bacterial Marteinsson, V.Th., Kristjánsson, J.K., Kristmannsdöttir, H., et al., 2001,
energy transduction: London, Academic Press, p. 429–490. Discovery of giant submarine smectite cones on the seafloor in Eyjaf-
Kelley, D.S., Karson, J.A., Blackman, D.K., et al., 2001, An off-axis hydro- jordur, Northern Iceland, and a novel thermal microbial habitat: Applied
thermal vent field near the Mid-Atlantic Ridge at 30° N: Nature, v. 412, and Environmental Microbiology, v. 67, p. 827–833, doi: 10.1128/
p. 145–149, doi: 10.1038/35084000. AEM.67.2.827-833.2001.
Kelley, D.S., Karson, J.A., Früh-Green, G.L., et al., 2005, A serpentinite-hosted Martin, B., and Fyfe, W.S., 1970, Some experimental and theoretical observa-
ecosystem: The Lost City hydrothermal field: Science, v. 307, p. 1428– tions on the kinetics of hydration reactions with particular reference to
1434, doi: 10.1126/science.1102556. serpentinization: Chemical Geology, v. 6, p. 185–202.
Koga, Y., Kyuragi, T., Nishihara, M., and Sone, N., 1998, Did archaeal and bac- Martin, W., and Russell, M.J., 2003, On the origin of cells: An hypothesis for
terial cells arise independently from noncellular precursors? A hypothesis the evolutionary transitions from abiotic geochemistry to chemoautotro-
stating that the advent of membrane phospholipid with enantiomeric glyc- phic prokaryotes, and from prokaryotes to nucleated cells: Philosophical
erophosphate backbones caused the separation of the two lines of descent: Transactions of the Royal Society of London, ser. B, v. 358, p. 27–85, doi:
Journal of Molecular Evolution, v. 46, p. 54–63. 10.1098/rstb.2002.1183.
Konecny, J., Schöniger, M., and Hofacker, G.L., 1995, Complementary coding
McCollom, T., and Seewald, J.S., 2003, Experimental constraints on the hydro-
conforms to the primeval comma-less code: Journal of Theoretical Biol-
thermal reactivity of organic acids and acid anions: I. Formic acid and
ogy, v. 173, p. 263–270, doi: 10.1006/jtbi.1995.0061.
formate: Geochimica et Cosmochimica Acta, v. 67, p. 3625–3644, doi:
Koonin, E.V., and Martin, W., 2005, On the origin of genomes and cells within
10.1016/S0016-7037(03)00136-4.
inorganic compartments: Trends in Genetics, v. 21, p. 647–654.
Mellersh, A.R., 1993, A model for the prebiotic synthesis of peptides which
Krishnarao, J.S.R., 1964, Native nickel-iron alloy, its mode of occurrence, dis-
throws light on the origin of the genetic code and the observed chirality
tribution and origin: Economic Geology and the Bulletin of the Society of
of life: Origins of Life and Evolution of the Biosphere, v. 23, p. 261–274,
Economic Geologists, v. 59, p. 443–448.
doi: 10.1007/BF01581903.
Krupp, R.E., 1994, Phase relations and phase transformations between low
Mellersh, A.R., and Wilkinson, A.-S., 2000, RNA bound to a solid phase can
temperature iron sulfides mackinawite, greigite and smythite: European
Journal of Mineralogy, v. 6, p. 389–396. select an amino acid and facilitate subsequent amide bond formation: Ori-
Lalou, C., Reys, J.-L., Brichet, E., Arnold, M., Thompson, G., Fouquet, Y., and gins of Life and Evolution of the Biosphere, v. 30, p. 3–7, doi: 10.1023/
Rona, P.A., 1993, New age data for Mid-Atlantic Ridge hydrothermal A:1006620421068.
sites: TAG and Snakepit chronology revisited: Journal of Geophysical Michel, H., and Deisenhofer, J., 1988, Relevance of the photosynthetic reaction
Research, v. 98, p. 9705–9713. center from purple bacteria to the structure of photosystem II: Biochemis-
Lécuyer, C., and Ricard, Y., 1999, Long-term fluxes and budget of ferric iron for the try, v. 27, p. 1–7, doi: 10.1021/bi00401a001.
redox states of the Earth’s mantle and atmosphere: Earth and Planetary Sci- Miller, S.L., 1992, The prebiotic synthesis of organic compounds as a step
ence Letters, v. 165, p. 197–211, doi: 10.1016/S0012-821X(98)00267-2. toward the origin of life, in Schopf, J.W., ed., Major events in the history
Leduc, S., 1911, The mechanism of life: London, Rebman, 172 p. of life: Boston, Jones and Bartlett, p. 1–28.
Leja, J., 1982, Surface chemistry of froth flotation: New York, Plenum Press, Miller, S.L., and Bada, J.L., 1988, Submarine hot springs and the origin of life:
640 p. Nature, v. 334, p. 609–611, doi: 10.1038/334609a0.
Leman, L., Orgel, L., and Ghadiri, M.R., 2004, Carbonyl sulfide-mediated pre- Milner-White, E.J., and Russell, M.J., 2005, Nests as sites for phosphates and
biotic formation of peptides: Science, v. 306, p. 283–286, doi: 10.1126/ iron-sulfur thiolates in the first membranes: 3 to 6 residue anion-binding
science.1102722. motifs: Origins of Life and Evolution of the Biosphere, v. 35, p. 19–27,
Lindahl, P.A., 2002, The Ni-containing carbon monoxide dehydrogenase fam- doi: 10.1007/s11084-005-4582-7.
ily: Light at the end of the tunnel?: Biochemistry, v. 41, p. 2097–2105, Mitchell, P., 1967, Proton-translocation phosphorylation in mitochondria, chlo-
doi: 10.1021/bi015932+. roplasts and bacteria: Natural fuel cells and solar cells: FASEB, v. 26,
Liu, S.V., Zhou, J., Zhang, C., Cole, D.R., Gajdarziska-Josifovska, M., and p. 1370–1379.
Phelps, T.J., 1997, Thermophilic Fe(III)-reducing bacteria from the deep Morse, J.W., and Arakaki, T., 1993, Adsorption and coprecipitation of diva-
subsurface: The evolutionary implications: Science, v. 277, p. 1106–1109, lent metals with mackinawite (FeS): Geochimica et Cosmochimica Acta,
doi: 10.1126/science.277.5329.1106. v. 57, p. 3635–3640, doi: 10.1016/0016-7037(93)90145-M.
Moulton, V., Gardner, P.P., Pointon, R.F., Creamer, L.K., Jameson, G.B., and Poole, A.M., Jeffares, D.C., and Penny, D., 1999, Early evolution: prokary-
Penny, D., 2000, RNA folding argues against a hot origin of life: Journal otes, the new kids on the block: BioEssays, v. 21, p. 880–889, doi:
of Molecular Evolution, v. 51, p. 416–421. 10.1002/(SICI)1521-1878(199910)21:10<880::AID-BIES11>3.0.CO;2-P.
Mulkidjanian, A.Y., and Junge, W., 1997, On the origin of photosynthesis as Pratt, J.M., 1993, Nature’s design and use of catalysts based on Co and the
inferred from sequence analysis—A primordial UV-protector as com- macrocyclic corrin ligand: 4 × 109 years of coordination chemistry: Pure
mon ancestor of reaction centers and antenna proteins: Photosynthesis and Applied Chemistry, v. 65, p. 1513–1520.
Research, v. 51, p. 27–42, doi: 10.1023/A:1005726809084. Prigogine, I., 1978, Time, structure, and fluctuations: Science, v. 201, p. 777–785.
Muller, A.W.J., 1995, Were the first organisms heat engines? A new model for Pullman, B., 1972, Electronic factors in biochemical evolution, in Ponnamperuma,
biogenesis and the early evolution of biological energy conservation: C., ed., Exobiology: North Holland Publishing Company, p. 136–169.
Progress in Biophysics and Molecular Biology, v. 63, p. 193–231, doi: Quayle, R.J., and Ferenci, T., 1978, Evolutionary aspects of autotrophy: Micro-
10.1016/0079-6107(95)00004-7. biological Reviews, v. 42, p. 251–273.
Müller, V., 2003, Energy conservation in acetogenic bacteria: Applied and Raymond, J., Siefert, J.L., Staples, C.R., and Blankenship, R.E., 2004, The
Environmental Microbiology, v. 69, p. 6345–6353, doi: 10.1128/ natural history of nitrogen fixation: Molecular Biology and Evolution,
AEM.69.11.6345-6353.2003. v. 21, p. 541–554.
Muth, G.W., Orteleva-Donnelly, L., and Strobel, S.A., 2000, A single adenosine Reader, J.S., and Joyce, G.F., 2002, A ribozyme composed of only two different
with a neutral pKa in the ribosomal peptidyl transferase center: Science, nucleotides: Nature, v. 420, p. 841–844, doi: 10.1038/nature01185.
v. 289, p. 947–950, doi: 10.1126/science.289.5481.947. Ricardo, A., Carrigan, M.A., Olcott, A.N., and Benner, S.A., 2004, Borate
Myers, C.R., and Nealson, K.H., 1988, Bacterial manganese reduction and minerals stabilize ribose: Science, v. 303, p. 196, doi: 10.1126/sci-
growth with manganese oxide as the sole electron acceptor: Science, ence.1092464.
v. 240, p. 1319–1321. Rickard, D., 1997, Kinetics of pyrite formation by the H2S oxidation of iron(II)
Nealson, K.H., and Stahl, D.A., 1997, Microorganisms and biogeochemi- monosulfide in aqueous solutions between 25° and 125°C: The rate equa-
cal cycles: What can we learn from layered microbial communities, in tion: Geochimica et Cosmochimica Acta, v. 61, p. 115–134, doi: 10.1016/
Banfield, J.F., and Nealson, K.H., eds., Geomicrobiology: Interactions S0016-7037(96)00321-3.
between microbes and minerals: Washington, D.C., Mineralogical Soci- Rickard, D., Butler, I.B., and Olroyd, A., 2001, A novel iron sulphide switch
ety of America, Reviews in Mineralogy, v. 35, p. 5–34. and its implications for earth and planetary science: Earth and Planetary
Nisbet, E.G., and Sleep, N.H., 2001, The habitat and nature of early life: Nature, Science Letters, v. 189, p. 85–91, doi: 10.1016/S0012-821X(01)00352-1.
v. 409, p. 1083–1091, doi: 10.1038/35059210. Righter, K., Drake, M.J., and Yaxley, G., 1997, Prediction of siderophile ele-
Ohmoto, H., Kakegawa, T., and Lowe, D.R., 1993, 3.4-billion-year-old bio- ment metal-silicate partition coefficients to 20 GPa and 2,800°C: The
genic pyrites from Barberton, South Africa: Sulfur isotope evidence: Sci- effects of pressure, temperature, oxygen fugacity, and silicate and metal-
ence, v. 262, p. 555–557. lic melt compositions: Physics of the Earth and Planetary Interiors, v. 100,
Oparin, A.I., 1924, Proiskhozhdenie Zhizny: Moscow, Rabochiî. p. 115–134, doi: 10.1016/S0031-9201(96)03235-9.
Oparin, A.I., 1938, Origin of life: New York, Dover. Romero, I., Gómez-Priego, A., and Celis, H., 1991, A membrane-bound pyro-
Oró, L., 1961, Comets and the formation of biochemical compounds on the phosphatase from respiratory membranes of Rhodospirillium rubrum:
primitive earth: Nature, v. 190, p. 389–390. Journal of General Microbiology, v. 137, p. 2611–2616.
Oró, J., and Kimball, A.P., 1961, Synthesis of purines under possible primi- Rosing, M.T., 1999, 13C-depleted carbon microparticles in >3700-Ma sea-floor
tive Earth conditions. I. Adenine from hydrogen cyanide: Archives of sedimentary rocks from West Greenland: Science, v. 283, p. 674–676,
Biochemistry and Biophysics, v. 94, p. 217–227, doi: 10.1016/0003- doi: 10.1126/science.283.5402.674.
9861(61)90033-9. Rosing, M.T., and Frei, R., 2004, U-rich Archaean sea-floor sediments from
Pace, N.R., 2002, The large scale topology of the tree of life [abs.]: Astrobiol- Greenland—Indications of >3700 Ma oxygenic photosynthesis: Earth
ogy, v. 2, p. 484. and Planetary Science Letters, v. 217, p. 237–244, doi: 10.1016/S0012-
Palandri, J.L., and Reed, M.H., 2004, Geochemical models of metasomatism 821X(03)00609-5.
in ultramafic systems: Serpentinization, rodingitization, and sea floor car- Rouse, R.C., Peacor, D.R., and Freed, R.L., 1988, Pyrophosphate groups in the
bonate chimney precipitation: Geochimica et Cosmochimica Acta, v. 68, structure of canaphite, Ca2Na2PO7.4H2O: The first occurrence of a con-
p. 1115–1133, doi: 10.1016/j.gca.2003.08.006. densed phosphate mineral: American Mineralogist, v. 73, p. 168–171.
Parkes, R.J., Cragg, B.A., Fry, J.C., Herbert, R.A., and Wimpenny, J.W.T., Russell, M.J., 1974, Manganese halo surrounding the Tynagh ore deposit, Ire-
1990, Bacterial biomass and activity in deep sediment layers from the land: A preliminary note: Transactions of the Institution of Mining and
Peru margin: Philosophical Transactions of the Royal Society of London, Metallurgy, section B, Applied Earth Science, v. 83, p. B65–B66.
ser. A, v. 331, p. 139–153. Russell, M.J., 1988, Chimneys, chemical gardens and feldspar horizons ± pyr-
Parkes, R.J., Cragg, B.A., Bale, S.J., Getliff, J.M., Goodman, K., Rochelle, rhotine in some SEDEX deposits: aspects of alkaline environments of depo-
P.A., Fry, J.C., Weightman, A.J., and Harvey, S.M., 1994, Deep bacterial sition, in Zachrisson, E., ed., Proceedings of the Seventh IAGOD Sympo-
biosphere in Pacific Ocean sediments: Nature, v. 371, p. 410–413, doi: sium, Stuttgart, Schweizerbartsche Verlagsbuchhandlung, p. 183–193.
10.1038/371410a0. Russell, M.J., 2003, On the importance of being alkaline: Science, v. 302,
Pavlov, A.A., and Kasting, J.F., 2002, Mass-independent fractionation of sulfur p. 580–581, doi: 10.1126/science.1091765.
isotopes in Archean sediments: Strong evidence for an anoxic Archean Russell, M.J., and Arndt, N.T., 2005, Geodynamic and metabolic cycles in the
atmosphere: Astrobiology, v. 2, p. 27–41, doi: 10.1089/15311070275362 Hadean: Biogeosciences, v. 2, p. 97–111, doi: 1726-4189/bg/2005-2-97.
1321. Russell, M.J., and Hall, A.J., 1997, The emergence of life from iron monosul-
Pedersen, K., 1993, The deep subterranean biosphere: Earth Science Reviews, phide bubbles at a submarine hydrothermal redox and pH front: Journal
v. 34, p. 243–260, doi: 10.1016/0012-8252(93)90058-F. of the Geological Society of London, v. 154, p. 377–402.
Peretó, J.G., Velasco, A.M., Becerra, A., and Lazcano, A., 1999, Comparative Russell, M.J., and Hall, A.J., 2001, The onset of life and the dawn of oxygenic
biochemistry of CO2 fixation and the evolution of autotrophy: Interna- photosynthesis: Respective roles of cubane core structures [Fe4S4]2+ and
tional Microbiology, v. 2, p. 3–10. transient [Mn4O4]4+[OCaO]2: Sixth International Conference on Carbon
Phoenix, V.R., Konhauser, K.O., Adams, D.G., and Bottrell, S.H., 2001, Role of Dioxide Utilization, September 9–14, 2001, Breckenridge, Colorado,
biomineralization as an ultraviolet shield: Implications for Archean life: Abstracts, p. 49.
Geology, v. 29, p. 823–826, doi: 10.1130/0091-7613(2001)029<0823: Russell, M.J., and Hall, A.J., 2002, From geochemistry to biochemistry: chemi-
ROBAAU>2.0.CO;2. osmotic coupling and transition element clusters in the onset of life and
Pinti, D.L., 2002, The isotopic record of Archean nitrogen and the early evolu- photosynthesis: The Geochemical News, no. 113/October, p. 6 – 12.
tion of the early Earth: Trends in Geochemistry, v. 2, p. 117. Russell, M.J., and Martin, W., 2004, The rocky roots of the acetyl-CoA path-
Pontes-Buarques, M., Tessis, A.C., Bonapace, J.A.P., Monte, M.B.M., Cortés- way: Trends in Biochemical Sciences, v. 29, p. 358–363, doi: 10.1016/
Lopez, G., de Souza-Barros, F., and Vieyra, A., 2001, Modulation of ade- j.tibs.2004.05.007.
nosine 5′-monophosphate adsorption onto aqueous resident pyrite: Poten- Russell, M.J., Hall, A.J., and Turner, D., 1989, In vitro growth of iron sulphide
tial mechanisms for prebiotic reactions: Origins of Life and Evolution of chimneys: possible culture chambers for origin-of-life experiments: Terra
the Biosphere, v. 31, p. 343–362, doi: 10.1023/A:1011805332303. Nova, v. 1, p. 238–241.
Russell, M.J., Daniel, R.M., and Hall, A.J., 1993, On the emergence of life via Shock, E.L., McCollom, T., and Schulte, M.D., 1998, The emergence of
catalytic iron sulphide membranes: Terra Nova, v. 5, p. 343–347. metabolism from within hydrothermal systems, in Wiegel, J., and Adams,
Russell, M.J., Daia, D.E., and Hall, A.J., 1998, The emergence of life from FeS M.W.W., eds., Thermophiles: The keys to molecular evolution and the
bubbles at alkaline hot springs in an acid ocean, in Wiegel, J., and Adams, origin of life: Washington, Taylor and Francis, p. 59–76.
M.W.W., eds., Thermophiles: The keys to molecular evolution and the Smith, J.V., 1981, The first 800 million years of the Earth’s history: Philosophical
origin of life: Washington, Taylor and Francis, p. 77–126. Transactions of the Royal Society of London, ser. A, v. 301, p. 401–422.
Russell, M.J., Hall, A.J., and Mellersh, A.R., 2003, On the dissipation of ther- Smith, B.E., 2002, Nitrogenase reveals its inner secrets: Science, v. 297,
mal and chemical energies on the early Earth: The onsets of hydrothermal p. 1654–1655, doi: 10.1126/science.1076659.
convection, chemiosmosis, genetically regulated metabolism and oxy- Sowerby, S.J., and Heckl, W.M., 1998, The role of self-assembled monolay-
genic photosynthesis, in Ikan, R. ed., Natural and laboratory-simulated ers of the purine and pyridine bases in the emergence of life: Origins
thermal geochemical processes: Dordrecht, Kluwer Academic Publishers, of Life and Evolution of the Biosphere, v. 28, p. 283–310, doi: 10.1023/
p. 325–388. A:1006570726326.
Russell, M.J., Hall, A.J., Cairns-Smith, A.G., and Braterman, P.S., 1988, Submarine Srere, P., 1987, Complexes of sequential metabolic enzymes: Annual
hot springs and the origin of life: -correspondence- Nature, v. 336, p. 117. Review of Biochemistry, v. 56, p. 89–124, doi: 10.1146/annurev.
Russell, M.J., Daniel, R.M., Hall, A.J., and Sherringham, J., 1994, A hydrother- bi.56.070187.000513.
mally precipitated catalytic iron sulphide membrane as a first step toward Steigerwald, V.J., Beckler, G.S., and Reeve, J.N., 1990, Conservation of
life: Journal of Molecular Evolution, v. 39, p. 231–243, doi: 10.1007/ hydrogenase and polyferredoxin structures in the hyperthermophilic
BF00160147. Archaebacterium Methanothermus fervidus: Journal of Bacteriology,
Samson, I.M., and Russell, M.J., 1987, Genesis of the Silvermines zinc-lead- v. 172, p. 4715–4718.
barite deposit, Ireland: fluid inclusion and stable isotope evidence: Eco- Stetter, K.O., 1996, Hyperthermophilic prokaryotes: FEMS Microbiology
nomic Geology and the Bulletin of the Society of Economic Geologists, Reviews, v. 18, p. 149–158, doi: 10.1016/0168-6445(96)00008-3.
v. 82, p. 371–394. Stetter, K.O., and Gaag, G., 1983, Reduction of molecular sulphur by methano-
Sandars, P.G.H., 2003, A toy model for the generation of homochirality during genic bacteria: Nature, v. 305, p. 309–311, doi: 10.1038/305309a0.
polymerization: Origins of Life and Evolution of the Biosphere, v. 33, Stone, D.A., and Goldstein, R.E., 2004, Tubular precipitation and redox gra-
p. 575–587. dients on a bubbling template: Proceedings of the National Academy of
Sandiford, M., and McLaren, S., 2002, Tectonic feedback and the ordering Sciences of the United States of America, v. 101, p. 11537–11541, doi:
of heat producing elements within the continental lithosphere: Earth 10.1073/pnas.0404544101.
and Planetary Science Letters, v. 204, p. 133–150, doi: 10.1016/S0012- Svetlitchnyi, V., Dobbek, H., Meyer-Klaucke, W., Meins, T., Thiele, B., Römer,
821X(02)00958-5. P., Huber, R., and Meyer, O., 2004, A functional Ni-Ni-[4Fe4S] cluster in
Sauer, K., and Yachandra, V.K., 2004, The water-oxidation complex in pho- the monomeric acetyl-CoA synthase from Carboxydothermus hydrogeno-
tosynthesis: Biochimica et Biophysica Acta – Bioenergetics, v. 1655, p. formans: Proceedings of the National Academy of Sciences of the United
140–148. States of America, v. 101, p. 446–451, doi: 10.1073/pnas.0304262101.
Schäfer, G., Engelhard, M., and Müller, V., 1999, Bioenergetics of the Archaea: Taylor, P., Rummery, T.E., and Owen, D.G., 1979, Reactions of iron mono-
Microbiology and Molecular Biology Reviews, v. 63, p. 570–620. sulfide solids with aqueous hydrogen sulfide up to 160ºC: Journal of
Schink, B., 1997, Energetics of syntrophic cooperation in methanogenic degra- Inorganic Nuclear Chemistry, v. 41, p. 1683–1687, doi: 10.1016/0022-
dation: Microbiology and Molecular Biology Reviews, v. 61, p. 262–280. 1902(79)80106-2.
Schoofs, S., Trompert, R.A., and Hansen, U., 2000, The formation and evolution Thauer, R.K., 1998, Biochemistry of methanogenesis: A tribute to Marjory Ste-
of layered structures in porous media: effects of porosity and mechanical phenson: Microbiology, v. 144, p. 2377–2406.
dispersion: Physics of the Earth and Planetary Interiors, v. 118, p. 205– Thauer, R.K., Jungermann, K., and Decker, K., 1977, Energy conservation
225, doi: 10.1016/S0031-9201(99)00148-X. in chemotrophic anaerobic bacteria: Bacteriological Reviews, v. 41,
Schoonen, M.A.A., Xu, Y., and Bebie, J., 1999, Energetics and kinetics of the p. 100–180.
prebiotic synthesis of simple organic and amino acids with the FeS-H2/ Trifonov, E.N., 2000, Consensus temporal order of amino acids and evolu-
FeS2 redox couple as reductant: Origins of Life and Evolution of the Bio- tion of the triplet code: Gene, v. 261, p. 139–151, doi: 10.1016/S0378-
sphere, v. 29, p. 5–32, doi: 10.1023/A:1006558802113. 1119(00)00476-5.
Schulte, M.D., and Rogers, K.L., 2004, Thiols in hydrothermal solution: Stan- Trifonov, E.N., Kirzhner, A., Kirzhner, V.M., and Berezovsky, I.N., 2001,
dard partial molar properties and their role in the organic geochemistry of Distinct stages of protein evolution as suggested by protein sequence
hydrothermal environments: Geochimica et Cosmochimica Acta, v. 68, analysis: Journal of Molecular Evolution, v. 53, p. 394–401, doi: 10.1007/
p. 1087–1097, doi: 10.1016/j.gca.2003.06.001. s002390010229.
Schulte, M.D., and Shock, E.L., 1995, Thermodynamics of Strecker synthesis Urey, H.C., 1952, On the early chemical history of the earth and the origin
in hydrothermal systems: Origins of Life and Evolution of the Biosphere, of life: Proceedings of the National Academy of Sciences of the United
v. 25, p. 161–173, doi: 10.1007/BF01581580. States of America, v. 38, p. 351–363.
Seefeldt, L.C., Dance, I.G., and Dennis, R.D., 2004, Substrate interactions with Van Walraven, H.S., Hollander, E.E., Scholts, M.J.C., and Kraayenhof, R.,
nitrogenase: Fe versus Mo: Biochemistry, v. 43, p. 1401–1409. 1997, The H+/ATP ratio of the ATP synthetase from the cyanobacterium
Sen, S., Igarashi, R., Smith, A., Johnson, M.K., Seefeldt, L.C., and Peters, J.W., Synechoccus 6716 varies with growth temperature and light intensity:
2004, A conformational mimic of the MgATP-bound “on state” of the Biochimica et Biophysica Acta, v. 1318, p. 217–224.
nitrogenase iron protein: Biochemistry, v. 43, p. 1787–1797. Vargas, M., Kashefi, K., Blunt-Harris, E.L., and Lovley, D.R., 1998, Microbial
Seyfried, W.M., and Bischoff, J.L., 1981, Experimental seawater-basalt interac- evidence for Fe(III) reduction on early Earth: Nature, v. 395, p. 65–67,
tion at 300°C, 500 bars: Chemical exchange, secondary mineral formation, doi: 10.1038/25720.
and implications for transport of heavy metals: Geochimica et Cosmochi- Vaughan, D.J., and Ridout, M.S., 1971, Mössbauer studies of some sulfide
mica Acta, v. 45, p. 135–147, doi: 10.1016/0016-7037(81)90157-5. minerals: Journal Inorganic Nuclear Chemistry, v. 33, p. 741–747, doi:
Shen, Y., and Buick, R., 2004, The antiquity of microbial sulfate reduc- 10.1016/0022-1902(71)80472-4.
tion: Earth-Science Reviews, v. 64, p. 243–272, doi: 10.1016/S0012- Vaughan, D.J., and Craig, J.R., 1978, Mineral chemistry of natural sulfides:
8252(03)00054-0. Cambridge, UK, Cambridge University Press, 493 p.
Shock, E.L., 1990, Geochemical constraints on the origin of organic compounds Vermaas, W.F.J., 1994, Evolution of heliobacteria: Implications for photosyn-
in hydrothermal systems: Origins of Life and Evolution of the Biosphere, thetic reaction center complexes: Photosynthesis Research, v. 41, p. 285–
v. 20, p. 331–367, doi: 10.1007/BF01808115. 294, doi: 10.1007/BF02184169.
Shock, E.L., 1992, Chemical environments of submarine hydrothermal sys- Voet, A.B., and Schwartz, A.W., 1982, Uracil synthesis via HCN oligomeriza-
tems: Origins of Life and Evolution of the Biosphere, v. 22, p. 67–107, tion: Origins of Life and Evolution of the Biosphere, v. 12, p. 45–49, doi:
doi: 10.1007/BF01808019. 10.1007/BF00926910.
Shock, E.L., and Schulte, M.D., 1998, Organic synthesis during fluid mixing Von Damm, K.L., 1990, Sea floor hydrothermal activity: black smoker chemis-
in hydrothermal systems: Journal of Geophysical Research, v. 103E, try and chimneys: Annual Review of Earth and Planetary Sciences, v. 18,
p. 28513–28527, doi: 10.1029/98JE02142. p. 173–204, doi: 10.1146/annurev.ea.18.050190.001133.
Von Damm, K.L., 2000, Chemistry of hydrothermal vent fluids from 9°-10°N, Woese, C.R., 1998, The universal ancestor: Proceedings of the National Acad-
East Pacific Rise: “Time zero,” the immediate post eruptive period: Jour- emy of Sciences of the United States of America, v. 95, p. 6854–6859,
nal of Geophysical Research, 105B, 11,203–11,222. doi: 10.1073/pnas.95.12.6854.
Wächtershäuser, G., 1988, Pyrite formation, the first energy source for life: A Woese, C.R., Dugre, D.H., Saxinger, W.C., and Dugre, S.A., 1966, The molecu-
hypothesis: Systematic and Applied Microbiology, v. 10, p. 207–210. lar basis for the genetic code: Proceedings of the National Academy of
Wächtershäuser, G., 1992, Groundworks for an evolutionary biochemistry: The Sciences of the United States of America, v. 55, p. 966–974.
iron-sulphur world: Progress in Biophysics and Molecular Biology, v. 58, Woese, C.R., Kandler, O., and Wheelis, M.L., 1990, Towards a natural system
p. 85–201, doi: 10.1016/0079-6107(92)90022-X. of organisms: proposal for the domains Archaea, Bacteria, and Eucarya:
Walker, J.C.G., and Brimblecombe, P., 1985, Iron and sulfur in the pre-biologi- Proceedings of the National Academy of Sciences of the United States of
cal ocean: Precambrian Research, v. 28, p. 205–222, doi: 10.1016/0301- America, v. 87, p. 4576–4579.
9268(85)90031-2. Wolin, M.J., 1982, Hydrogen transfer in microbial communities, in Bull, A.T.,
Wenner, D.B., and Taylor, H.P., 1971, Temperatures of serpentinization of ultra- and Slater, J.H. eds., Microbial interactions and communities: London,
mafic rocks based on 18O/16O fractionation between coexisting serpentine Academic Press, p. 323–356.
and magnetite: Contributions to Mineralogy and Petrology, v. 32, p. 165– Wolthers, M., 2003, Geochemistry and environmental mineralogy of the iron-
185, doi: 10.1007/BF00643332. sulphur-arsenic system: Geologica Ultraiectina, Mededelingen van de
Westall, F., de Witt, M.J., Dann, J., Van der Gaast, S., de Ronde, C., and Faculteit Aardwetenschappen Universiteit Utrecht, no. 225, 185 p.
Gerneke, R., 2001, Early Archean fossil bacteria and biofilms in hydro- Wolthers, M., Van der Gaast, S.J., and Rickard, D., 2003, The structure of disor-
thermally influenced sediments from the Barberton Greenstone Belt, dered mackinawite: American Mineralogist, v. 88, p. 2007–2015.
South Africa: Precambrian Research, v. 106, p. 93–116, doi: 10.1016/ Yamagata, Y., Wanatabe, H., Saitoh, M., and Namba, T., 1991, Volcanic produc-
S0301-9268(00)00127-3. tion of polyphosphates and its relevance to prebiotic evolution: Nature,
Wicken, J.S., 1987, Evolution, information and thermodynamics: Extending the v. 352, p. 516–519, doi: 10.1038/352516a0.
Darwinian program: New York, Oxford University Press, 243 p. Zachara, J.M., Kukkadapu, R.K., Frederickson, J.M., Gorby, Y.A., and Smith,
Wilde, S.A., Valley, J.W., Peck, W.H., and Graham, C.M., 2001, Evidence from S.C., 2002, Biomineralization of poorly crystalline Fe(III) oxides by dis-
detrital zircons for the existence of continental crust and oceans on the similatory metal reducing bacteria (DMRB): Geomicrobiology Journal,
Earth 4.4 Gyr ago: Nature, v. 409, p. 175–178, doi: 10.1038/35051550. v. 19, p. 179–207, doi: 10.1080/01490450252864271.
Woese, C.R., 1967, The genetic code: The molecular basis for genetic expres-
sion: New York, Harper and Row, 200 p. MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA


Memoir 198
2006
Early life signatures in sulfur and carbon isotopes

from Isua, Barberton, Wabigoon (Steep Rock), and Belingwe
Greenstone Belts (3.8 to 2.7 Ga)
N.V. Grassineau*
Department of Geology, Royal Holloway, University of London, Egham, Surrey TW20 0EX, UK
P. Abell†
Department of Chemistry, University of Rhode Island, Kingston, Rhode Island 02881, USA
P.W.U. Appel
Geological Survey of Denmark and Greenland (GEUS), Oester Voldgade 10, 1350 Copenhagen, Denmark
D. Lowry
E.G. Nisbet
Department of Geology, Royal Holloway, University of London, Egham, Surrey TW20 0EX, UK
ABSTRACT
Carbon and sulfur isotopes have been measured on samples from four Archean
greenstone belts dating from 3.8 Ga to 2.7 Ga, in order to trace metabolic changes as
life evolved over this one-billion-year period. In the Isua Greenstone Belt (3.8 Ga),
Greenland, δ34S in sulfide minerals from sedimentary sequences range from –3.8‰
to +3.4‰. δ13Cred measured in BIFs, turbidites and conglomerates vary from –29.6‰
to –14.7‰; this range permits us to hypothesize the presence of hyperthermophilic
and chemotrophic species in transient settings, or possibly pelagic photoautotrophic
microbes, or both. In the Barberton Greenstone Belt, South Africa, sulfide minerals
show δ34S values from +1.5‰ to +5.6‰. Black shales have δ13Cred values from –32.4‰
to –5.7‰, suggesting that oxygenic photosynthetic and sulfate-reducing bacteria were
present by ca. 3.24 Ga. The δ13Cred measured in the stromatolites of Steep Rock Lake
(3.0 Ga), Ontario, Canada, are from –30.6‰ to –21.6‰, giving clear evidence for occu-
pation of a shallow water environment by cyanobacteria. The wide isotopic ranges for
δ34S in sulfides from –21.1‰ to +16.7‰ and for δ13Cred in carbon-rich cherts and black
shales from –43.4‰ to –7.2‰ in the Belingwe Greenstone Belt, Zimbabwe, indicate
that photosynthetic microbial mat communities were well established at 2.7 Ga. In
these well-preserved Late Archean formations, modern-style biological sulfur and car-
bon cycles may have been in operation. The δ34S and δ13Cred ranges, respectively 37‰
and 36‰, indicate a great variety of biological processes interacting with each other.
Keywords: early life, sulfur and carbon isotopes, Archean, Isua, Belingwe, Barberton,
Steep Rock Lake.
*[email protected]
†
Deceased.
Grassineau, N.V., Abell, P., Appel, P.W.U., Lowry, D., and Nisbet, E.G., 2006, Early life signatures in sulfur and carbon isotopes from Isua, Barberton, Wabigoon
(Steep Rock), and Belingwe Greenstone Belts (3.8 to 2.7 Ga), in Kesler, S.E., and Ohmoto, H., eds., Evolution of Early Earth’s Atmosphere, Hydrosphere, and
Biosphere—Constraints from Ore Deposits: Geological Society of America Memoir 198, p. 33–52, doi: 10.1130/2006.1198(02). For permission to copy, contact
[email protected]. ©2006 Geological Society of America. All rights reserved.
33
34 N.V. Grassineau et al.
INTRODUCTION: USE OF STABLE ISOTOPES IN this methane produce δ13C down to –80‰ (Coleman et al., 1981).
STUDYING EARLY LIFE An additional process is anoxygenic photosynthesis with fraction-
ations that are usually smaller (as Chloroflexus: –20‰ to –10‰
The best record of early life is in sedimentary rocks from [e.g., van der Meer et al., 2001]; or Rubisco II [e.g., Robinson et
Archean greenstone belts. This study uses isotopic analyses of al., 2003; Schidlowski, 2001]). Conversely, the inorganic carbon
carbon- and sulfide-rich deposits from four Archean belts to seek byproduct of the photosynthesis has a narrow range represented
evidence for the onset of major metabolic processes. The Isua by most Phanerozoic marine carbonates (Table 1).
Greenstone Belt (ca. 3.8 Ga) in West Greenland and the Belin- In the Archean, carbon degassing from the mantle entered
gwe Greenstone Belt (2.7 Ga) in Zimbabwe are the two main the atmosphere and thence the ocean. Carbon extracted by sedi-
areas studied. Additional δ13C and δ34S data were obtained from mentation from the Late Archean ocean/atmosphere system was
the Barberton Greenstone Belt in South Africa and Steep Rock fractionated between a relatively small 13C-depleted organic car-
Lake succession in northwestern Ontario, Canada. bon portion (δ13C ~ −28‰), and a larger residual reservoir of
Metabolic processes produce distinctive isotopic frac- 13
C-enriched inorganic carbon (δ13C ~ 0‰) (e.g., Schidlowski et
tionations when selecting carbon and sulfur from chemical or al., 1975; Abell et al., 1985a, 1985b). Studies of Archean biology
organic substrates accessible to the organisms. The recogniz- (e.g., Hayes et al., 1983; Schidlowski et al., 1983) have recognized
able δ34S and δ13C signatures in the sediments after burial, both this distinct carbon isotopic fractionation as marking the onset
in residual organic matter and in associated minerals, are par- of the global-scale processing of carbon by cyanobacteria, now
ticularly useful in efforts to understand the nature and extent of known to be via Rubisco I. The rapid and successful production
microbial activity (e.g., Nisbet and Fowler, 1999; Grassineau of cyanobacteria then induced the start of carbon management
et al., 2001a, 2002). These isotopic fingerprints are among the by oxygenic photosynthesis that may have been sudden in geo-
very few biological signals remaining in Archean rocks. Other logical terms. As O2 is the process by-product, aerobic or at least
evidence for early microbial phylogeny comes from molecular microaerobic environments of deposition would be expected in
studies (Woese, 1987; Pace, 1997; review in Nisbet and Sleep, rocks that showed the characteristic isotopic signature. Although
2001). In this study, stable isotopes have been used to infer the an aerobic facies, organic δ13C signature, or δ13Ccarb signature of
presence of prokaryotic communities in relics of microbial mats 0‰ may not be evidence of cyanobacteria individually, the pres-
from Archean rocks. ence of all three is diagnostic. The start date of photosynthetic
Carbon-bearing compounds in the modern earth system have processes in the Archean is very controversial, and more isotopic
very different δ13C values (Table 1). Biological fractionations investigation is needed.
produce wide δ13C ranges, with two distinctive signatures domi- Anoxygenic photosynthesis predates the oxygenic process
nating modern organic carbon. The first, from –28‰ to −22‰, in standard phylogenetic models (see discussion in Nisbet and
is due to the fractionation by the Rubisco I enzyme during pho- Sleep, 2001). Therefore, there may be an interval in the geologi-
tosynthesis (Sirevåg et al., 1977; Pierson, 1994), in aerobic and cal record where the δ13C fractionation between organic matter
microaerobic environments. This characteristic carbon isotopic and carbonate was less than –28‰. However, it is still in debate
fractionation occurs when carbon supply is abundant (Erez et whether large-scale carbonate deposition occurred before oxy-
al., 1998) and taken from seawater in exchange with a CO2-rich genic photosynthesis. Most geologists concur that methanogen-
atmosphere. The second is for methanogenic archaea that pro- esis dates back to the Archean (Hayes, 1994; Nisbet and Sleep,
duce metabolic methane cycling, generating more fractionated 2001), and may well predate photosynthesis, though Cavalier-
δ13C, with a range of values mainly between –40‰ and –30‰. Smith (2002) has argued that methanogens did not appear until
Methanotrophs or sulfate reducers that gain organic matter from the late Proterozoic and hence a very “light methane” signature
would not be expected. Thus the “geological” consensus that
methanogenesis is ancient needs testing. Primary δ13C of marine
carbonate and organic carbon in the Archean are still not well
TABLE 1. MAIN ISOTOPIC RESERVOIRS WITHIN THE EARTH SYSTEM constrained, but the wide δ13Cred range so far obtained (–52‰ to
Reservoirs and domains G34S‰CDT G13C‰PDB –13‰; Strauss and Moore, 1992) suggests that biological pro-
Seawater Sulfate + 21.0 ± 0.2‰† Carbonate +0.6 ± 1.6‰
Atmosphere –30 to +30‰† CO2 –8‰† cesses were operational. A key target is to seek evidence for oxy-
CH4 –47‰‡ genic photosynthesis (δ13Ccarb ~ 0‰; δ13Cred ~ –28‰) or anaero-
Mantle + 0.3 ±0.5‰§ CO2 –8 to –5‰#
CH4 –33 to –15‰# bic conditions within the Archean.
Igneous rocks –11 to +9‰† Graphite –33 to –6‰# Two isotopically distinct primary sulfur reservoirs domi-
Modern sediments –60 to +20‰†† –30 to –10‰†
Petroleum –8 to +32‰† –34 to –18‰‡ nate today: homogeneous seawater sulfate, and sulfide derived
Natural gas –8 to +32‰† CH4 –80 to –14‰‡ from the mantle (Table 1). δ34S of sulfides in modern sediments
Coal –30 to +24‰† –35 to –15‰†
†
In Faure (1986). ranges from –60‰ to +20‰ (Hoefs, 1997). Significant isotopic
‡
§
Stevens (1988). fractionations can be caused by inorganic processes, especially
e.g. Sakai et al. (1984).
#
Taylor (1986). in hydrothermal fluids, but biological activities produce even
††
Hoefs (1997). greater fractionation (e.g., Schidlowski et al., 1983), particularly
Isua, Barberton, Wabigoon (Steep Rock), and Belingwe Greenstone Belts 35
in repetitive recycled reactions (e.g., Canfield and Teske, 1996). and Hayes, 1978, for carbon; Grassineau et al., 2001b, for sulfur).
The resulting fractionation between seawater sulfate and precipi- This high-resolution technique measures local isotopic fraction-
tated biogenic sulfides in sediments has been up to 80‰ since ations in samples at sub-millimeter scale, which is necessary to
the beginning of the Phanerozoic, generating 34S-enrichment in detect specific activities of the biological communities, otherwise
seawater sulfate, which has averaged between +10‰ to +30‰ at undetectable in larger samples.
different periods of Earth’s history. A precision of ± 0.1‰ for δ34S was obtained on hand-picked
δ34S of Archean seawater sulfate remains unknown, as sul- sulfide minerals as small as 0.8 mg for pyrite. The six standards
fate is very rare in most early sediments, but is estimated to be analyzed, including NBS123, NBS127, and IAEA-S3, cover a
close to 0‰–2‰ (Ohmoto, 1992). Then oceans contained mod- range from –31.6‰ to +20.3‰. The samples for reduced car-
est amounts of sulfate (e.g., Veizer et al., 1989; Canfield and bon analysis were first treated in 20% HCl at 120 °C for 12 h.
Teske, 1996), and Habicht and Canfield (1996) suggested that A precision better than ± 0.1‰ in δ13C was obtained on hand-
high SO4 concentrations in seawater did not occur before the picked samples of 0.07 mg for pure carbon, to 30 mg for whole
rise of atmospheric oxygen in post-Archean time. On the other rock with 0.1 wt% C. The standards measured, including NBS21
hand, Ohmoto et al. (1993) interpreted the 9‰ δ34S range for and IAEA-CO9, cover a range from –47.1‰ to +3.3‰. Blank
3.4–3.2 Ga pyrite in one specimen (South Africa) as evidence for contamination from tin capsules is <34 ppm C, as measured in
microbial reduction of seawater sulfate, suggesting that seawater the laboratory, but all samples with less than 200 ppm C, where
already had a high SO4 content. Shen et al. (2001) and Shen and the blank isotopic effect might be significant for interpreting the
Buick (2004) reached the same conclusion with values of +3‰ results, have been rejected from the data set.
to +9‰ in 3.47 Ga sulfate samples from northwestern Australia. Pure carbonate samples (0.5 mg) were measured using an
Thus there is currently no consensus. Isocarb automated carousel connected to a PRISM mass spec-
The evolution of the sulfur cycle is also disputed, as there are trometer. Impure carbonates were analyzed using a modified
few constraints on the timing of the first appearance of key parts Micromass Multiflow connected to an Isoprime mass spectrom-
of the cycle. Biochemical evidence from the “standard” model eter. The system requires as little as 200 μg of pure carbonate or
of microbial phylogenetic evolution (Woese, 1987; Pace, 1997) up to 50 mg of whole rock powder with 0.5% carbonate. Internal
indicates strongly, though circumstantially, the great antiquity of precision is better than ±0.07‰ for δ13C and ±0.10‰ for δ18O for
both S-oxidation and S-reduction processes. Many biochemical both systems. The standards used are NBS19 limestone and an
processes rely on a S-containing enzyme, and thus at least some internal laboratory calcite.
S-utilizing bacteria are Archean (e.g., Londry and Des Marais,
2003). Ohmoto and Felder (1987) suggested that microbial sul- EARLY ARCHEAN: THE ISUA GREENSTONE BELT,
fate reduction was established by 3.5 Ga, and further examples WEST GREENLAND
have confirmed this (Barberton, South Africa, at 3.4 Ga [Ohmoto
et al., 1993; Kakegawa and Ohmoto, 1999]; North Pole, north- The Isua Greenstone belt (IGB), West Greenland, is a vol-
western Australia, at 3.47 Ga [Shen and Buick, 2004]). However, cano-sedimentary relic exposed in an arcuate belt surrounded
some authors consider that though sulfate reduction may be old, and locally intruded by a variety of tonalitic gneisses. The belt
some of the sulfur cycle evolved only in the Proterozoic (Blank, is dominated by thick sequences of mafic pillow lavas interca-
2004), and that the complete cycle with important S-isotope frac- lated with numerous beds of iron formations (Appel et al., 1998;
tionation only appeared ca. 0.86–1.0 Ga (Canfield and Teske, Fedo et al., 2001; Myers, 2001; Polat et al., 2003). Its minimum
1996). Hitherto, the δ34S range of Archean sulfide minerals was age constraint of 3.7 Ga comes from U-Pb of zircons in gra-
considered to be less than 10‰ (Cameron, 1982; Habicht and nitic sheets crosscutting the belt (Nutman et al., 1997). Sulfides
Canfield, 1996), and up to 13‰ in mineral deposits (Ohmoto, in the IGB are mainly pyrrhotite and pyrite. They occur mainly
1992). However, Grassineau et al. (2001a, 2002) reported a as thin stratabound layers and disseminations throughout the
wider range of 37‰ at 2.7 Ga, suggesting that the sulfur cycle rocks. Pyrite is most abundant in BIF and metacherts, often in
was already well on its way to full operation in the Late Archean association with fuchsite. Pyrrhotite dominates in pillow lavas,
at 2.7 Ga, with evidence of sulfate-reduction, sulfur-oxidation, often associated with chalcopyrite in discordant quartz veins and
and possibly disproportionation. Furthermore Shen et al. (2001) veinlets (Appel, 1979). The IGB furthermore hosts sedimentary
found a range of 16‰ at 3.47 Ga. This demonstrates that the sul- sequences, comprising mica schists or garnet mica schists, with
fur cycle is of great antiquity (Grassineau et al., 2001a). or without staurolite as well as chemical sediments such as chert
and iron formation. Rollinson (2002) observed five different met-
ANALYTICAL TECHNIQUES amorphic and structural domains, based on garnet study (Fig. 1).
The southwestern zone has undergone two main metamorphic
δ13C for reduced carbon and δ34S were analyzed with a events (Rollinson, 2002) at 3.74 Ga (Frei and Rosing, 2001: Pb/
VG/Fisons/Micromass “Isochrom-EA” system, consisting of an Pb on metabasalts) and 2.84 Ga (Frei et al., 1999: Pb/Pb on mag-
elemental analyzer (EA1500 Series 2) online to an Optima mass netite). The latest episode was at high metamorphic grade with
spectrometer operating in He continuous flow mode (Matthews widespread metasomatic overprinting (Rose et al., 1996; Frei et
Figure 1. Map of Isua Greenstone Belt (3.7–3.8 Ga) showing the five structural domains defined by various authors and compiled by Rollinson
(2002) (modified from Rollinson, 2002). Domain I underwent low-strain metamorphism at 3.69 Ga. Domains II and V recorded two high-
grade metamorphic events both at 3.74 Ga. Domains III and IV underwent the same metamorphisms with higher intensity, and a late event at
ca. 2.8 Ga. However, Domain IV preserved some low-strain areas. The localities studied are shown.
al., 1999, Rollinson, 2002). Conversely, the northeast zone has Sedimentary Formations Sampled
experienced only one event of lower intensity, which took place
at 3.69 Ga (Frei et al., 1999 [Pb-Pb on magnetite and tremolite]) Sampling of the sedimentary rocks of the Isua Greenstone
and no deformation thereafter (Moorbath and Kamber, 1998). Belt focused on the regions that had the lowest metamorphic
The latter authors suggest that this age might represent the depo- grade and were affected only by the early 3.74–3.69 Ga event, as
sitional period. This low-strain area has well-preserved primary these are most likely to preserve evidence of biological activity.
sedimentary and igneous features (Appel et al., 1998; Rollinson, Such areas are much rarer in the southwest than in the northeast
2002). Several carbonate alteration events took place in the belt. (Rollinson, 2002). Eight formations were sampled in this study
The earliest occurred during the deposition of the sediments and (Fig. 1). Six of the sampled units are banded-iron formations
extrusion of lava flows (J.S. Myers, personal commun.). Later (BIF), three from the western and three from the northeastern part
carbonate alteration especially affected the western part of the of the belt. The northeastern BIF samples are actinolite-bearing
IGB from 3.7 to 2.8 Ga (Moorbath and Whitehouse, 1996; Rose metachert and iron formation consisting of magnetite-rich bands
et al., 1996; Frei and Rosing, 2001). The first author carried out alternating with grunerite-rich bands, or magnetite-rich bands
fieldwork in the IGB in 1998 and 1999, as part of the Isua Multi- alternating with quartz-bands. The largest example is the Iron
disciplinary Research Project. Mountain Formation, a chert-BIF of alternating layers of quartz
and magnetite with local amphibole-rich bands (Frei et al., 1999;

Myers, 2001). The BIF samples from the west, including one of
carbonate facies, are far more strongly deformed by higher-grade
metamorphism than those from the northeast.
Another group of samples are metasedimentary rocks from
the northwest that were formed by turbidity currents (Bouma
sequence; Nutman, 1986). They consist of a series of medium-
grained quartzites grading into fine-grained metapelites contain-
ing numerous small carbon particles (Rosing, 1999). In addition,
a metaconglomerate was sampled in the northeast. The poorly
sorted polymict assemblage consists mainly of quartz pebbles
(metachert) with few pebbles of amygdaloidal volcanic rocks
and BIF clasts (Fedo, 2000).
Previous Stable Isotope Studies at Isua
Monster et al. (1979) found a narrow range of −1.0‰ to

+2.6‰ for sulfides in the IGB. More recently, Strauss (2003)
gives a variation from −3.0‰ to +1.0‰, and Mojzsis et al.
(2003) reported values of −0.9‰ to +2.2‰ on two pyrite grains
(Fig. 2). These values around 0‰ are in the range of magmatic
hydrothermal sulfides and provide no positive evidence that sul-
fur-utilizing bacteria were active during deposition.
Previous studies have obtained a wide range for δ13Cred
(reduced carbon) from –28‰ to –6‰ (Fig. 3) (Perry and Ahmad,
1977; Schidlowski et al., 1979; Hayes et al., 1983; Naraoka et al.
1996; Rosing, 1999; Ueno et al., 2002; van Zuilen et al., 2002),
suggesting biological fractionation. For example, Rosing (1999)
proposed a biogenic origin for the graphite grains in the Bouma
sequence, with δ13Cred of –19‰ close to the original depositional
isotopic signature.
New Results Obtained for the Isua Belt
Sulfur Isotope Results and Discussion

The sulfide minerals analyzed from the sedimentary
sequences (Table 2, Figs. 2A, 2B) are mainly disseminated
pyrite, with only a few secondary pyrrhotites from the carbonate
facies BIF. The δ34S obtained for the pyrites are from −3.8‰ to
+3.4‰ (n = 19) (Grassineau et al., 2000). This 7.2‰ range is
larger than previously measured in Isua. The extent of the range
found in this study suggests that more diverse processes oper-
ated than hitherto thought. This result is particularly significant
because the largest range of values is seen in the northeastern
sulfide-rich BIF and Iron Mountain, which have the lowest met-
amorphic grade in the belt (Rollinson, 2002).
The δ34S has different ranges for each BIF. In the Iron
Mountain, the values are only positive (+3.1‰ to +3.4‰; n = 3).
In contrast, the two other northeastern BIFs show mainly nega-
tive values (−3.8‰ to +1.1‰; n = 9). The metaconglomerate
with an average of +0.4 ± 0.5‰ (n = 4) is very homogeneous Figure 2. Distributions of δ34S obtained for sulfide minerals from the
IGB. (A, B) δ34S for the sulfides in the metasediments, with a range of
and close to 0‰, similar to the values of Monster et al. (1979). 7.2‰. (C) δ34S in the sulfides from igneous rocks. (D, E) δ34S for the
The only western BIF analyzed for δ34S gives an average of −0.7 gold and fuschite deposits. Results from previous studies are repre-
± 0.1‰ (n = 3). sented by thick horizontal bars.
TABLE 2. ISUA GREENSTONE BELT RESULTS (3.8–3.7 Ga)

Lithology G34S‰ G13Cred‰ Carbon wt% G13Ccarb‰
East
Iron Mountain Metachert
462-401Na +3.2
462-401Nb +3.4
462-401Nc +3.1
Sulfide-rich BIF
462-439a –2.5
462-440a –1.1 –22.7 0.02
462-440b –1.9 –29.6 0.03
462-440c –27.4 0.02
462-441a –29.4 0.02
462-441b –26.6 0.02
462-442a –1.6 –23.4 0.02
462-443a –0.0 –22.6 0.03
462-443b –24.3 0.02
465-722a –3.8
465-723a –1.2
465-724a +1.1
Magnetite-rich BIF
462-423a –1.2
Polymict conglomerate
465-726a +0.3
465-726b +0.9
465-727a +0.7 –8.6 0.12
465-727b –0.4 –8.7 0.09
465-728a –27.2 0.03
465-728b –23.4 0.02
465-728c –24.6 0.02
465-729a –28.6 0.03
West
Bouma sequence
466-855b –14.7 0.12
466-857a –16.0 0.52
466-857b –14.9 0.44
466-859a –14.9 0.10
466-859b –18.4 0.09
466-863a –17.7 0.12
466-863b –17.9 0.14
Carbonate-rich BIF
462-497a –3.1
462-499c –2.3
462-00a –3.0
462-01 –3.9
462-08a –0.8 –16.1 0.07 –0.4
462-08ab –0.6 –20.1 0.05 –1.0
462-09a –0.7 –6.5 0.11 –2.2
462-09b –1.7
Other western BIFs
466-876a –12.0 0.02
466-876c –11.3 0.03
466-888a –10.1 0.18
Hydrothermal deposits (mainly in SW Isua)
Gold deposits –5.4 to +0.6‰
†
N = 25 (–2.5 ± 1.8‰)
Fuchsite deposits –10.0 to +3.0‰
N = 21 (–1.1 ± 3.8‰)†
†
Averages are in parenthesis with1 s.d.
In order to provide a basis for comparison, pyrites from

igneous formations and a slightly later Tarssartôq dyke (3.47 Ga)
have been analyzed. The δ34S ranges obtained are between –0.5
and +1.0‰ (n = 9) and would suggest that high-temperature fluids
circulating in the belt had a signature of +0.4 ± 0.5‰ (Fig. 2C).
This interpretation is supported by pyrite from metabasalts that
show a narrow range of +0.5 ± 0.1‰ (n = 2).
Figure 3. Distributions of δ13C obtained for the IGB. (A, B) δ13Cred for Further comparison has been made with sulfides from two
the metasediments, compared with values in non-sedimentary rocks
(C). (D) δ13Ccarb from the western carbonate facies BIF and the second- 3.7–3.8 Ga mineral deposits hosted by the western area of the
ary metacarbonates in the east of the belt. Results from previous stud- IGB (Pb-Pb on galena; Frei and Rosing, 2001). These formed
ies are represented by thick horizontal bars. from low to moderate temperature hydrothermal fluids (Figs. 2D,
2E). δ34S from disseminated galena and sphalerite associated phism to greenschist facies, in which case sulfide would not have
with gold mineralization in a tonalite sheet varies from –5.4‰ survived with a primary isotopic signature. Rollinson (2002), on
to +0.6‰ (n = 25) (Grassineau and Appel, 2000). More remark- the other hand, determined a temperature for the metamorphism
ably, the range of –10.0‰ to +3.0‰ (n = 21) obtained for dis- not higher than 520 °C by analyzing metapelites. So it is likely
seminated sulfides associated with quartz-fuchsite deposits is the that this fairly moderate event did not re-homogenize δ34S; hence
widest yet found in Isua. Such wide isotopic fractionation is most the isotopic variation remained completely or partially preserved
likely created by sulfate-sulfide reactions (Ohmoto, 1992). in the sulfides.
Western Samples. The carbonate-facies BIF is dominated by The sulfide-rich BIF contains alternating quartz and magne-
siderite, with secondary pyrrhotite in veins and along fractures. tite bands, but with up to 5% of pyrites in some zones, mostly in
The low to moderate oxygen fugacity and high pH, constrained quartz but also within magnetite. Some pyrite crystals are agglom-
by the presence of magnetite, kept the isotopic variation rela- erated in clusters, suggesting that they are not re-crystallized by
tively small for the hydrothermal fluids generated during meta- a post-metamorphic event. The wider range of δ34S and the lower
morphism (Ohmoto, 1986). If there was any δ34S signature left by metamorphic grade make these BIFs the most likely to preserve
organisms then it has been overprinted by the 2.84 Ga metamor- a primary bacterial signature, and it is possible that a much wider
phic event and re-homogenized as pyrrhotite. This is shown by range produced by bacterial sulfate reduction has been narrowed
the narrow range of values. Two main primary sources for sulfur by metamorphic homogenization. It cannot be ruled out though
can be suggested: the meta–pillow lavas and sulfur derived from that the range was produced by hydrothermal magmatic sulfur, or
seawater during the deposition of the BIF on the seafloor. The that there was a mixing between biogenic sulfur and hydrother-
homogenized δ34S value of –0.7‰ obtained for the BIF is lighter mal sulfur (+0.5‰) during metamorphism. However, it would
than the magmatic sources (+0.5‰). Although the shift in values be difficult to fractionate inorganically the hydrothermal source
of 1.2‰ could be explained entirely by hydrothermal processes to significantly lower δ34S without involving a second source of
remobilizing magmatic sulfur, it is possible that another source lighter sulfur, except at extremely low pH, or at higher fO in the
2
caused this shift during re-homogenization, possibly primary sul- presence of sulfates (Rye and Ohmoto, 1974). Sulfate though has
fur in the BIF. so far not been found in these particular samples.
Northeastern Samples. In contrast, the better preserved and
less deformed metasediments in the northeast are more likely to Carbon Isotopic Results and Discussion
preserve the original δ34S signatures. They show a larger δ34S δ13C of reduced carbon from the sedimentary sequences has
range of 7.2‰ (n = 26), but values are different for each forma- been measured after acid treatment (Table 2, Figs. 3A, 3B). The
tion. In the polymict metaconglomerate, δ34S is homogenous δ13Cred range is from −29.6‰ to –6.5‰ (n = 27) (Grassineau
(+0.4 ± 0.5‰; n = 4). The pyrites are from the matrix and the et al., 2000). The carbon content of the Isua rocks is between
clasts and show similar δ34S. This is expected, as the clasts con- 0.02 and 0.52 wt% (n = 27). The carbon was not visible in hand
sist of mafic volcanic rocks and metacherts (Fedo, 2000). The specimen and the samples analyzed were whole rock chips.
detrital pyrites probably originated as magmatic sulfur, at ~0‰. Where known from the literature the carbon is referred to as
There is no sign of a biological signature, which should have graphite; otherwise it is called reduced carbon. In addition, car-
given more isotopic variation. bonates were measured with δ13Ccarb of –2.2 ± 1.1‰. The two
The three samples analyzed from the Iron Mountain metach- northeastern formations analyzed are the sulfide-rich BIF and
ert are very homogeneous (δ34S of +3.2 ± 0.2‰). These δ34S val- the polymict conglomerate with δ13Cred as light as –29.6‰ (0.03
ues are heavier than those of the five other BIFs. The pyrites are wt% C) (n = 8). With one exception, the values are lighter than
idiomorphic and located close to or in veins, indicating that the –22.6‰.
secondary sulfides were introduced by hydrothermal fluids that The δ13Cred range in the western formations is from –20.1‰
may have been produced by the early metamorphism. The sulfur to –6.5‰ (n = 13). The results obtained from the dark layers of
in these fluids could have originated from different sources: (1) the Bouma sequence are from −18.4‰ to –14.7‰ (n = 7) with
sulfur already present in situ in the form of barite (barite traces carbon contents from 0.09 to 0.52 wt%. The values are simi-
exist in the metachert) reduced by interaction with hydrothermal lar to those of Rosing (1999) (−19.1‰ to –11.4‰). The two
fluids (e.g., Ohmoto and Godhaber, 1997); 2) sulfur from volca- small BIFs associated with ultramafic rocks have homogeneous
nic formations in the belt. This view is supported by Lepland et δ13Cred of –11.1‰ ± 0.8‰ (n = 3). The carbonate-facies BIF
al. (2002) who interpreted REE patterns for apatites in the Iron shows the entire isotopic range obtained in the west, indicat-
Mountain as a result of a pervasive fluid from mixed sources. ing that more than one process occurred. Carbonates from this
Sulfide- and magnetite-rich BIFs from the northeast have a formation, mainly siderite, give δ13Ccarb values from –3.9‰ to
relatively wide and mainly negative range of δ34S (4.9‰; from –0.4‰ (n = 8).
–3.8‰ to +1.1‰; n = 9) (Fig. 2A). This heterogeneity stands out To differentiate between carbon in the sediments and the
compared to the results of the two previously mentioned north- other formations, we measured 18 reduced carbon samples of
eastern formations. Komiya et al. (1999) analyzed metabasites igneous rocks and graphites from hydrothermally affected units
and suggested that Domain I had undergone retrograde metamor- in the northeast part of the belt (Fig. 3C). Carbon contents range
from 0.02–3.5 wt%. Nine graphite samples from amphibolite carbon cannot be ruled out. Rosing (1999), who obtained the
facies give a wide range of δ13Cred from –20.3‰ to –10.2‰ least modified δ13C at –19.1‰, inferred that this value represents
(−14.8 ± 5.4‰). These values fall roughly within the range planktonic-like organisms. Though by no means proving the case
already reported by previous workers (Fig. 3). δ13Cred was also for biogenicity, the evidence suggests that the most depleted δ13C
measured in five metacarbonate samples, with a narrow spread (here –18.4‰) might record biological activity at 3.8 Ga.
of –6.3 ± 1.3‰. Finally two ultramafic rocks were analyzed with Northeastern samples. The two northeastern sedimentary
values of –8.5 and –5.8‰, and two volcano-felsic rocks, with formations show lighter values. With the exception of two sam-
–24.2 and –8.6‰. The δ13Cred range (close to 20‰) is large in ples, all δ13Cred are more depleted than –22‰ (Fig. 3A). This is
these formations, but two main sources can be pointed out. The consistent with the lower metamorphic grade in this area, imply-
heavier values between –8.6 and –5.8‰ are most likely of mag- ing that more original signatures might have been preserved. The
matic origin, whereas the lighter range of –14.6‰ to –10.2‰ sulfide-rich BIF values average –25.7 ± 2.7‰, with one δ13Cred at
may record the circulation of high-grade metamorphic fluids –29.6‰, the lightest found in this study of Isua. These samples
during the early event (Perry and Ahmad, 1977; Naraoka et al., consist mainly of alternating quartz and magnetite bands. There
1996; van Zuilen et al., 2002, 2003). These authors suggested is no visible carbon in the quartz-layers in thin sections, so it is
that the siderite decomposed to graphite at the high temperatures likely associated with the magnetite. The range of 6‰ in samples
reached by the regional metamorphism. This graphite would be with low reduced carbon content does not suggest a high-tem-
13
C-enriched during isotopic re-equilibration with the siderite perature fluid source (as with δ34S, a fluid would have left a more
(Ueno et al., 2002). homogenized δ13C range for reduced carbon). Thus the preferred
The δ13Ccarb range of metacarbonates for the northeastern part interpretation is that some of this reduced carbon had an organic
is wider (from –5.7‰ to +1.2‰) than for the western samples origin. The δ13Cred distribution possibly records two carbon
(from –3.9‰ to –1.0‰) (Fig. 3D). The homogenization of δ13C sources, syn-sedimentary organic carbon and a small component
in the western carbonate facies BIF is consistent with the higher of inorganic CO2 from post-depositional hydrothermal fluids
grade; in fact, most of the carbonates here were most likely remo- circulating through the belt. If so, because metamorphism can
bilized during the 2.84 Ga metasomatic event (Rose et al., 1996). enrich the graphite in 13C by up to 20‰ at 450 °C (Schidlowski,
The wider variation shown in the northeastern carbonates reflects 2001), the primary signature of the organic component in these
the lower grade of the Domain I metamorphism, where the isoto- BIFs may have been even lighter than –29‰. In this case, the
pic overprint was less effective. biological environment could have been anaerobic.
Western Samples. Comparison between the results shows Fedo (2000) described the polymict conglomerate as detrital
that δ13Cred is generally lighter in the sedimentary rocks than in with a mafic origin for some of the clasts. δ34S in this study (+0.4
the ultramafic and mafic rocks, with a wider range of 23.1‰. ± 0.5‰) supports this interpretation, and the two δ13Cred at –8.6‰
This is not the case for the two small western BIFs; the value of are also similar to the results obtained for metabasalts (Figs. 3A,
–11.1 ± 0.8‰ indicates that the high-grade metamorphism has 3C). However, four values of –28.6‰ to –23.4‰ indicate a sec-
overprinted the δ13C, with a re-equilibration toward the general- ond carbon source. They are close to the sulfide-rich BIF values
ized graphite value of –12‰ obtained in other parts of the belt. and may record an organic origin. Could traces of life be found
Despite the proximity of the three western BIFs, the meta- in a conglomerate? Preservation is possible and the metamorphic
morphic imprint is less obvious in the carbonate facies BIF. In grade that affected the metaconglomerate was low enough not
contrast to δ34S, the δ13Cred range for this BIF is wide, 13.6‰, to cause significant shift in primary signatures. The view that
indicating that the re-homogenization was not as efficient as in the original environment hosted life and preserved an organic
the other western BIFs, and that some areas of the formation signature is debatable. The shallow subaqueous setting for the
may have partially retained the primary value. The lightest δ13C, conglomerate deposition suggested by Fedo (2000) could host
at –20.1‰, is similar to the original value suggested by Rosing life. Therefore the value of –28‰ might be close to the primary
(1999) for kerogen in the Bouma sequence nearby. Though meta- signature of the organic activity. It is interesting to compare this
morphic overprinting is likely, −20.1‰ might be primary too: if with the isotopic signature of oxygenic photosynthesis, though
so, it could represent the burial of organic matter within the BIF. this is no more than speculation in the absence of other support-
On the other hand, δ13Cred at –6.5‰ from the same BIF appears ing evidence.
to be related to the late (2.84 Ga) high-grade metamorphism. The In summary, metamorphism has caused widespread over-
carbonate in this rock is mainly siderite, which locally may have printing of the original δ13C signatures and especially the δ34S
decomposed to graphite during the high temperatures reached values of the IGB. However, some areas have better preserved
during the 2.84 Ga event. the primary isotopic compositions. This is the case for the west-
The last western formation studied is the sedimentary out- ern turbiditic rocks, and more particularly in the northeastern
crop interpreted as turbidites displaying a Bouma sequence. The part, which was subjected to slightly less strain and lower-grade
δ13Cred range measured in this study (−18.4‰ to –14.7‰; n = metamorphism (Rollinson, 2002, 2003). The results presented
7) might represent a primary biological signature, but the pos- here are not inconsistent with the hypothesis that there is a record
sibility of mixing with the –12‰ high-temperature metamorphic of planktonic life in the western part (Rosing, 1999). From the
TABLE 3. FIG TREE GROUP AND STEEP ROCK LAKE RESULTS

Lithology G34S‰ n G13Cred‰ Carbon wt% N G13Ccarb‰ N
Fig Tree Group (3.24 Ga) +1.5 to +5.6‰ 15 –32.4 to –5.7‰ 0.02–0.25% 18 –4.9 to –4.4‰ 6
S- and C-rich Black shales (+2.3 ± 1.0‰) (–22.4 ± 9.1‰) (-4.7 ± 0.2‰)
Steep Rock Lake (3.0 Ga) +5.0‰ 1 –30.6 to –21.6‰ 0.22–2.29% 32 +0.1 to +2.9‰ 56†
Stromatolites (–25.4 ± 2.0‰) (+2.0 ± 0.6‰)
Note: Averages are in parenthesis with 1 s.d.
†
Including 9 analyses by P. Abell (by conventional line technique).
results obtained in the eastern BIFs and metaconglomerate, bio- –2.0‰ (de Ronde et al., 1992). All these authors came to the
genic processes, possibly including methanogenesis and photo- conclusion that photosynthesis involving marine organisms was
synthesis (Rosing and Frei, 2004), might have been in operation occurring at the time.
at 3.7 Ga.
New Isotopic Results and Discussion of the Fig Tree Group
MIDDLE ARCHEAN: THE BARBERTON
GREENSTONE BELT The newly analyzed samples consist of three 10–16-cm-
long cores of sulfide-rich carbonaceous shales. Similar ranges to
The Barberton Greenstone Belt, located in the Kaapvaal those previously obtained have been found on samples analyzed
Craton in South Africa, is one of the best-preserved mid-Archean at millimeter scale, with δ34S in pyrites and pyrrhotites ranging
successions. The regional metamorphic grade is low to moder- from +1.0‰ to +5.6‰ (n = 15) (Table 3). The sulfide minerals
ate, much lower than in the IGB, and though there are many are an assemblage of agglomerated fine grains, as seen in Fig-
shear zones and décollement horizons, the sequences of rocks ure 4. δ13Cred varies from –32.4‰ to –5.7‰ (n = 18) in samples
generally have experienced only low strain (Viljoen and Viljoen, with carbon contents up to 0.25 wt%, with no direct correlation
1969). Among the components of the belt are the Onverwacht between the C content and δ13C. Of the samples analyzed, 73%
Group (mainly mafic and ultramafic formations) and two clastic
and chemical sedimentary sequences, the Moodies and Fig Tree
Groups (Paris, 1987).
This study is based on the Fig Tree Group, which consists of
ferruginous cherts, greywackes, shales, BIFs, and pelites, depos-
ited in a submarine setting (Paris, 1987). The material, collected
by J. Kramers and C. Siebert (Bern, Switzerland), comes from a
horizontal borehole composed of greywackes and shales, drilled
from the Fairview mine in an ESE direction out of the orebody
toward the Sheba mine, in the central part of the belt (Siebert,
2003; Kramers et al., 2004). The Fig Tree Group has an age
between 3.26 and 3.22 Ga from the measurements of zircons in
tuff layers (Lowe and Byerly, 1999).
Previous Stable Isotope Investigations
Many carbon and sulfur isotope measurements have been

made on the sedimentary groups. Sulfides in the Fig Tree Group
show δ34S ranges of −0.4‰ to +4.0‰ (Strauss and Moore, 1992),
+1.2‰ to +3.9‰ (de Ronde et al., 1992), and –0.9‰ to +4.4‰
(Kakegawa and Ohmoto, 1999). An average of +3.4 ± 0.2‰ for
barite (Strauss and Moore, 1992) suggests that mid-Archean sea-
water sulfate was not highly fractionated. Studies in the Moodies
and Onverwacht Groups show similar ranges (e.g., Ohmoto et al.
1993: –3.0‰ to +8.6‰ in pyrites).
Carbon-rich shales in the Fig Tree Group contain up to 14.0
wt% of organic carbon with δ13Cred from –29.5‰ to –9.5‰ (de
Figure 4. Detailed isotopic analyses of 279B dark shale sample from
Ronde and Ebbesen, 1996). Other studies give ranges of –32.8‰ the Fig Tree Group (3.24 Ga). The core has a central 5-cm-thick layer
to –24.3‰ (Hayes et al., 1983), or –35.4‰ to –27.0‰ (Strauss rich in sulfides and secondary carbonate. At each end, δ13Cred is homo-
and Moore, 1992). The δ13C in carbonates are from –4.5‰ to geneous around –31‰.
are lighter than –20‰. Pervasive calcite in bands has a very nar- LOWER MIDDLE ARCHEAN: THE STEEP ROCK
row δ13Ccarb range of –4.9‰ to –4.4‰ (n = 6). GROUP
A detailed section of one of the samples, presented in Fig-
ure 4, illustrates the isotopic variations obtained. Core 279B has The Steep Rock Group from the Wabigoon Greenstone Belt
thin dark laminations, and a central sulfide- and carbonate-rich in northwestern Ontario, Canada, contains an Archean carbonate
layer ~5 cm thick. There is a wide range of δ13Cred, from –31.9‰ platform, the Mosher Carbonate Formation, which is 500 m thick
to –7.8‰, but at both ends of the core δ13Cred is homogeneous, and made of laminated carbonate, mainly limestone, in part stro-
with –31.4 ± 0.4‰; the δ13C tends to become heavier toward the matolitic (Wilks and Nisbet, 1985, 1988; Kusky and Hudleston,
center, reaching –7.8‰ inside the “central” layer. This layer has 1999). Regional metamorphic grade is broadly lower greenschist.
peripheral 5-mm-thick carbonate-rich bands with very homoge- The age of the Steep Rock Group is not fixed. It rests uncon-
neous δ13Ccarb of –4.5‰, similar to the average value found by de formably on a ca. 3.0 Ga gneissic terrane, the Marmion Com-
Ronde et al. (1992). The sulfides, mainly pyrrhotite on the outside plex (Davis and Jackson, 1985; Wilks and Nisbet, 1988). The
edges of this zone, and pyrites in the center in bleb-like shapes, upper unit of the belt includes metavolcanic rocks with an age of
give a relatively narrow range of δ34S averaging +1.8 ± 0.4‰. 2.93 Ga (Davis, 1993 cited in Kusky and Hudleston, 1999) but
However, a variation of 3.64‰ over 1.5 cm in another sample it is partly allochthonous. It is thus likely but not certain that the
suggests that re-homogenization of δ34S, if it took place, was not Steep Rock Lake stromatolites are ~2.9–3.0 Ga old. The well-
complete, especially considering the low-grade greenschist facies preserved stromatolites clearly show biogenic structures on both
metamorphism that the rocks have experienced (Siebert, 2003). small and large-scales (Wilks and Nisbet, 1985; 1988). Crypto-
Kakegawa and Ohmoto (1999), who obtained similar δ34S zoon structures, branching walled and un-walled columnar forms
variations in pyrite grains at micro-scale within the Fig Tree up to 20 cm, are succeeded at the top of the unit by spectacular
Group, suggested both organic and inorganic origins for the domal structures up to 3 m in diameter. The presence of these
pyrite crystals. They pointed out that coarse-grained pyrites asso- very well preserved primary structures, ooliths and oncolites,
ciated with quartz veins, with δ34S from +1.1‰ to +3.6‰, have indicates that strain was very low.
an inorganic origin and are more likely precipitated from high- The samples studied are three 1-m-long hand drill cores
temperature hydrothermal fluids. Although the values of pyrites taken at different sites from near the larger domal structures. The
in the present study are in a similar range, none are associated carbonate is bluish-white, a mixture of calcite and dolomite, with
with quartz veins and they are agglomerates of fine rounded undulating bands of black material. These laminae are described
grains, which parallel the sedimentary bedding, suggesting that as organic kerogen (Hayes et al., 1983). Few analyses have been
they are a sedimentary or diagenetic feature. It is impossible to carried out on the formation; Hayes et al. (1983) obtained three
rule out a partial re-homogenization of the sulfides, but the δ34S δ13Cred values with a range from –26.4‰ to –22.1‰, and Schid-
values of +1.5‰ to +5.6‰ found locally seem to indicate at least lowski et al. (1983), three carbonates with δ13Ccarb from +1.1‰ to
some microbial reduction of the seawater sulfate. This was sug- +2.0‰. The authors inferred that these values were the result of
gested by Kakegawa and Ohmoto (1999), who concluded that the biological activity.
Barberton sea at 3.4–3.2 Ga was bearing an appreciable amount
of sulfate. Results from Steep Rock Lake Stromatolites
The isotopically homogeneous carbonate (δ13Ccarb ~ –4.5‰)
does not crosscut the structure of the shale. It occurs with pyrite The δ13Cred results (Table 3) obtained for the stromatolites
adjacent to a redox interface, outside which is pyrrhotite and are from –30.6‰ to –21.6‰ (n = 32), and δ13Ccarb from +0.1‰
reduced carbon with δ13C around –31‰. It is likely that this is a to +2.9‰ (n = 56), in agreement with the previously obtained
result of syngenetic or diagenetic hydrothermal processes. Shales values. The ranges for both reduced and carbonate carbon
enclosed by the carbonate-rich zones show some bleaching are narrow, especially the carbonate, with an average of +2.0
(Fig. 4), and significant exchange between the primary organic ± 0.6‰ (Fig. 5). The reduced carbon isotopic values have a
carbon and carbon from the hydrothermal fluid. The δ13Cred varies wider spread of 9‰. This could be original variation but might
from –27.6‰ to –5.7‰ close to the carbonate. This suggests that also indicate minor effects of metamorphism. Only one pyrite
only the δ13C values around –31‰ represent a primary organic has been found. It has a δ34S value of +5.0‰.
carbon signature. An alternative explanation for the relatively The δ13Cred average of –25.4‰ and the Δ13Cred-carb fraction-
heavy δ13Cred values is that they are a record of organisms using ation of 26‰–31‰ are interpreted as clear evidence of bio-
waste carbon dioxide or bicarbonate from methanotrophs. Either logical activity in these stromatolites, more particularly of the
way they are evidence for biological activity. carbon fractionation by Rubisco I. Consequently, considering
The view that light δ13Cred values and δ34S values of sul- that the biogenic structures were developed in shallow water,
fides are biogenic is in agreement with Kakegawa and Ohmoto on the margin of the Marmion Complex (Kusky and Hudleston,
(1999). The light δ13C of –31‰ suggest that the dark shales 1999), it is possible to propose that oxygenic photosynthesis
may be of deep-water facies, deposited in anoxic zone below was already established in the Steep Rock Group at 3.0 Ga, and
the photic zone. that cyanobacteria were fully active. This and the Mushandike
organic carbon- and pyrite-rich black shales (Jimmy Member)

of which the base is massive pyrite, capped by a thick mafic lava
sequence, the Reliance Formation. The study is mainly based
on the carbon- and sulfide-rich sedimentary units. The Spring
Valley and Shavi Members represent a shallow-water environ-
ment (Hunter et al., 1998). At the same stratigraphic level, the
discontinuous Rupemba Formation contains well-preserved
stromatolites that were deposited on slopes (Martin et al., 1980;
Abell et al., 1985a), and that have undergone low-grade meta-
morphism with a maximum temperature of 200 °C (Abell et
al., 1985a). Deposition of the Jimmy Member occurred below
wave base but still in the photic zone. It is complexly folded at
the contact with the Reliance Formation (Hunter et al., 1998;
Grassineau et al., 2002), but with well-preserved sub-rounded
and elongate sulfide structures below this.
The Cheshire Formation contains mostly shallow-water
sediments, mainly dark and ferruginous shales, ironstones,
conglomerates, important limestone reefs, with alternations of
lava and tuff. Two units were studied, the first one being the
1500-m-thick dark shales, above the basal volcanic sequence.
Their deposition was in shallow water, indicated by ripple mark
structures (Martin et al., 1980). The shales are well preserved
Figure 5. Distributions of δ13Cred and δ13Ccarb obtained for Steep Rock
Lake stromatolites. δ13Ccarb values have a narrow range of +2.0‰
and show evidence of only low-grade metamorphism. The sec-
± 0.6‰, and δ13Cred an average of approximately –26‰. ond unit is the 24-m-thick Macgregor stromatolites, one of the
largest reefs of the carbonate formation. They were deposited
on a volcanic slope and comprise 22 cycles of environmental
change (Martin et al., 1980) that occurred in the basin. Abell et
al. (1985a) observed little textural evidence for metamorphism
Formation in Zimbabwe are perhaps the oldest rock units where and from δ18Ocarb estimated the re-equilibration temperature for
an unambiguous “modern” carbonate isotopic ratio can be iden- these rocks to be 80 °C.
tified (Abell et al., 1985b).
Previous Stable Isotope Investigations
LATE ARCHEAN: THE BELINGWE GREENSTONE
BELT Until recently there were few δ13C data for the Manjeri
black shale units, with ranges of –32.2‰ to –8.8‰ for organic
The Belingwe Greenstone Belt in Zimbabwe is one of the carbon, and –7.2‰ to –5.7‰ for carbonate (from Strauss and
best preserved of all Archean successions. The two sequences, Moore, 1992). In contrast, Abell et al. (1985a) made an exten-
the Manjeri and Cheshire Formations, respectively the low- sive study on the Cheshire stromatolites (Table 4) with ranges
est and highest parts of the Ngezi Group (upper greenstones) of –0.6‰ to +1.0‰ for the carbonate, and –35.4‰ to –16.9‰
contain carbon- and sulfide-rich sediments and stromatolitic for the reduced carbon. They also obtained a few δ13Ccarb val-
limestones that have been affected only by low-grade meta- ues for the Rupemba stromatolites. Some Cheshire black shales
morphism (Martin et al., 1980; Abell et al., 1985a; Bickle and analyzed by Yong (1991) show a δ13Cred range of –40.8‰ to
Nisbet, 1993). Bolhar et al. (2002) obtained Pb-Pb ages of 2706 –32.0‰, and δ13Ccarb values from –13.3‰ to –8.0‰ for the
± 49 Ma for the Manjeri Formation, and 2601 ± 49 Ma for the carbonates in veins (Table 4). Some δ13Cred measurements from
Cheshire Formation, which is in agreement with Pb-Pb and Sm- the Manjeri carbon- and sulfide-rich sediments, with a range
Nd ages for the Reliance Formation of 2692 ± 9 Ma from imme- of –38‰ to −17‰, have been presented by Grassineau et al.
diately above the Manjeri Formation (Chauvel et al., 1993). (2001a, 2002). Grassineau et al. (2002) also obtained val-
The Manjeri sediments were deposited in a continental ues of –35.1‰ to –7.3‰ for the Rupemba stromatolites and
basin unconformably upon a 3.5 Ga tonalitic gneiss (Bickle and δ13Cred values of –35.1‰ and 36.6 wt% C for a soft black “ball”
Nisbet, 1993; Hunter et al., 1998). Four units of this formation ~0.5 mm in diameter found in the carbonate. This sample is
in the Nercmar drill core were investigated: at the base, 40 m bitumen and likely to be close to a pure kerogen composition.
of organic carbon-rich sediments with intercalated thin sulfide Finally a detailed study of δ34S presented by Grassineau et al.
layers (Spring Valley and Shavi Members), 56 m of volcano- (2001a, 2002) gives a range of –17.6‰ to +16.7‰ in the three
clastic rocks (Rubweruchena Member), and at the top, 10 m of sulfide-rich Manjeri units.
TABLE 4. BELINGWE GREENSTONE BELT RESULTS

Lithology G34S‰ n G13Cred‰ Carbon wt% N G13Ccarb‰ N
Manjeri Formation (2.7 Ga)
Jimmy Member –21.1 to +16.7‰ 206 –38.4 to –20.7‰ 0.04–20.0%† 27 –12.5 to –6.7‰ 19
S- and C-rich sediments (–3.4 ± 6.2‰) (–32.8 ± 4.2‰) (–9.4 ± 2.1‰)
Shavi Member –18.4 to +5.4‰ 53 –32.0 to –7.2‰ 0.13–3.18% 16 – –
S- and C-rich sediments (–3.5 ± 4.5‰) –25.0 ± 7.6‰
Spring Valley Member +1.2 to +2.9‰ 6 –31.9 to –23.3‰ 1.36–2.66% 3 – –
S- and C-rich sediments (+2.2 ± 0.7‰) –26.2 ± 4.0‰
Rupemba Member – – –35.1 to –7.3‰ 0.40–1.93% 19 –0.7 to +1.4‰ 44§
stromatolites (–22.3 ± 4.6‰) 36.65%* (0.1 ± 0.5‰)
Cheshire Formation (2.6 Ga)
Dark shales –6.0‰ 1 –43.8 to –32.0‰ 0.11–3.19% 39# –13.3 to –4.7‰ 50#
(–39.5 ± 3.0‰) (–8.7 ± 2.8‰)
McGregor Stromatolites – – –35.4 to –16.9‰ 0.02–1.99% 253§ –0.6 to +1.0‰ 463§
(–28.6 ± 3.3‰) (+0.2 ± 0.3‰)
Note: Averages are in parenthesis with 1 s.d.
†
High carbon concentrations from samples containing bitumen.
*Pure bitumen ball.
§
From Abell et al. (1985a), results obtained by McClory (1988).
#
Compilation of Yong (1991) and this study (McClory and Yong results are from conventional line techniques).
New Results Obtained and Discussion of the Belingwe

Greenstone Belt
Clastic Sediments
Further analyses have been made on the carbon- and sul-
fide-rich shallow-water and subtidal sediments from the Shavi
and Jimmy Members and the Cheshire dark shales (Table 4). The
results, obtained in pyrites, increase the extensive δ34S range in
the Manjeri units (Figs. 6A, 6B), from –21.1‰ to +16.7‰. This
range of nearly 38‰ is the widest range yet recorded in Archean
sediments.
In the Jimmy Member alone the δ34S range equals the total
spread found in the belt. There is a bimodal distribution of the
δ34S data (Fig. 6A). The sharp main peak is around 0‰. This
value occurs mainly at the top of the Jimmy Member, toward the
contact with Reliance volcanics. The proximity of the volcanism,
which is taken to be stratigraphically directly above the Manjeri
Formation (see discussion by Grassineau et al., 2002), suggests
that this value records hydrothermal processes that happened
during the deposition of the overlying volcanic unit or soon after,
producing local re-homogenization of the primary δ34S. The sec-
ond and smaller peak in the Jimmy Member histogram is around
–5‰, with a skewed distribution on both sides. This value is close
to the –6‰ recorded in the only pyrite found in the Cheshire
shales. The δ34S range in the Shavi Member, even if smaller, still
has a significant spread of nearly 24‰. The population is distrib-
uted around a less pronounced peak of –4‰ (Fig. 6B).
Such large δ34S ranges suggest microbial sulfate-reduction
processes, but locally values have been recorded where δ34S in
the pyrites are above +6‰. This occurs only in one black shale
sample of 10 cm length (TR51 at 67.05 m), where the pyrites are
in bleb-like shapes with internal concentric structure, which are
interpreted to be biogenic features. These 34S-enriched pyrites (up
to +17‰) are associated with adjacent blebs of dolomite, which Figure 6. δ34S distributions for sulfide minerals in three units of car-
bon- and sulfide-rich sediments of the Manjeri Formation (A, B) (BGB,
have light δ13Ccarb (−12‰ to –10‰). 2.7 Ga). The distribution is bimodal in the Jimmy Member (A), with a
The reduced carbon for the Manjeri shales has a δ13C spread main peak at 0‰, and a second skewed peak at –5‰. For information,
of 31‰. The distribution shows a main value around –29‰ the –6‰ value measured for the Cheshire Formation has been added.
(Fig. 7), and carbon concentrations from 0.04 to 20.0 wt‰. A ates (down to –13.3‰; Table 4). These results suggest methano-
few samples containing “soft” bitumen-like material have a high genic and methanotrophic activities at the time (e.g. Yong, 1991;
content of reduced carbon (from 14 to 20 wt%). In detail, the Grassineau et al., 2002), with methane oxidation by methano-
Shavi and Spring Valley Members show an average δ13Cred of trophs and sulfate reducers to generate the light δ13Ccarb values at
–25.0‰ ± 7.6‰ (Fig. 7A), whereas the Jimmy Member is much the oxic/anoxic interface (e.g., Peckmann and Thiel, 2004).
lighter, with a δ13Cred of –32.8‰ ± 4.2‰ (Fig. 7B).
The wide range found in δ13Cred indicates primary bacterial Detailed Study of BES49c
activity. The main peak value around –30‰ suggests fraction- Numerous samples have been analyzed in detail at millime-
ation by Rubisco I (Pierson, 1994). Shavi and Spring Valley units ter scale to search for evidence of microbial consortia, and large
appear to reflect mainly the enzyme signature for oxygenic pho- isotopic variations have been found for both δ34S and δ13Cred.
tosynthesis, but the lighter values in the Jimmy Member more BES49c, a 5-cm-long core section, roughly at the middle of the
likely record carbon cycling via methanogenesis, in addition to Jimmy Member is an example. It is a black shale with a gray
photosynthesis (Grassineau et al., 2001a, 2002). In comparison, chert layer at the bottom, and on the top, a 1.5-cm-thick sulfide
data for Cheshire shales are very light, averaging –39.5‰ ± 3.0‰ layer (Fig. 8). The sedimentary layering in BES49c suggests sig-
(Fig. 7C). These occur locally in association with light carbon- nificant diagenetic compaction.
The δ34S range is from –19.6‰ to –5.4‰, with large vari-
ations of more than 9‰ over distances of 2 or 3 mm. All sul-
fides are pyrite. In the upper part, they have slightly cataclazed
bleb-like shapes, rather rounded, consisting of microcrystalline
framboid-like agglomerations, with very small internal cavities,
resembling a “sponge” structure. The δ34S value of –18.6‰ inside
the top left large bleb seems to have been sealed from further
exchange since the time of its deposition (Fig. 8). The δ34S varia-
tions are evidence of very prolific biological activity. The heavi-
est value, at –5.4‰, is from a layer of disseminated small pyrite
crystals, interbedded in dark chert layers. It is difficult to say if
these sulfides are primary, but the –5‰ value is very common in
the Jimmy Member and partial homogenization by hydrothermal
fluids could have occurred. Most likely the value is a genuine
biological signature from isolated activity with a plentiful SO4
supply, as it is mainly found in scattered interbedded pyrites. The
Figure 7. Distribution of δ13Cred from sedimentary sequences in the Figure 8. Detailed isotopic study of core sample BES49c, from the
Belingwe Greenstone Belt. The samples are from various drill cores. Jimmy Member. BES49c is a carbon- and sulfide-rich black shale, with
The median value is –29‰ for the Shavi Member (A), –33‰ for the a layer of rounded bleb-like sulfides. The δ34S values vary between
Jimmy Member (B), and –40‰ for the Cheshire dark shales (C). –19.6‰ and –5.4‰, with the heavier values for the dispersed pyrites.
lighter values are typically in ~1–2-cm-size blebs of aggregated

microcrystallized sulfides and the isotopic compositions are
zoned in these structures. As the variations imply a biogenic ori-
gin, these rounded and granulated shapes might be an illustration
of what remains of the bacterial communities, with intense bio-
logical activity illustrated by large fractionations (Canfield and
Teske, 1996). In these multiple micro-ecosystems once hosted by
BES49c, the interactions between the bacteria created different
fractionation values, depending on the neighboring bacteria and
environment. Each value then could represent the identity of a
unique bacterial community.
The δ13Cred values in BES49c, −27.5‰ and –31.4‰, are also
evidence of biological activity (Fig. 8). They give an idea of the
organic processes that had most likely taken place. At the bot-
tom, the signature seems to reflect Rubisco I activity, and con-
sequently the existence of photosynthetic processes, whereas at
the top, near the sulfides, the lighter value more likely indicates
a combination of photosynthesis and probably methanogenesis.
Studying the isotopic variations at the scale of the sample gives
an instant snapshot of the interactions between different bacterial
activities present at 2.7 Ga.
Carbonate Reefs
Carbon isotopic results obtained for the stromatolites of the
Figure 9. Distributions of δ13Cred and δ13Ccarb for the Rupemba and
Rupemba Member at the base of Manjeri and of the Macgregor Macgregor (Cheshire Formation) stromatolitic sequences. The Mc-
Member in the Cheshire are shown in Figure 9. The samples are Gregor stromatolites have a Δ13Ccarb-red around 29‰.
1-m-long hand-drilled cores, and they show superbly preserved
biological textures. The values of the Macgregor Member are
from Abell et al. (1985a) and McClory (1988). Considering the
very narrow ranges for δ13Ccarb and δ13Cred (Table 4), we con- controlling early anoxygenic activity or may be a result of lower
cluded that the stromatolites represent nearly unaltered biogenic pCO2 in the water column. On the other hand, the single bitumen
sediments. value of –35.1‰ with 36.6 wt% C (Table 4) from the Rupemba
For the Macgregor stromatolites, the average δ13Ccarb is stromatolites indicates that methanogenic processes were occur-
+0.2‰ ± 0.3‰ and the δ13Cred values are mostly between –34‰ ring either at the base of stromatolitic layers, or post-deposition-
and –24‰, with a –28.6‰ ± 3.3‰ average (Fig. 9). With these ally if the limestone acted as a hydrocarbon reservoir.
results, a mean value of 28.8‰ for the Δ13Cred-carb fractionation can
be inferred. Given the stromatolitic textures, this fractionation is Summary
interpreted as the unambiguous signature of Rubisco I activity, The wide δ34S range of nearly 40‰ obtained for pyrite from
indicating that the Macgregor stromatolites were deposited by a the Manjeri Formation indicates that the biological sulfur cycle
cyanobacterial microbial ecology based on oxygenic photosyn- was mostly, if not completely, operational at 2.7 Ga (Grassineau
thesis (Abell et al., 1985a; Grassineau et al., 2002). et al., 2001a), well before the post-Archean increase of atmo-
Further analyses were made in the Rupemba stromatolites, spheric oxygen took place at 2.2 Ga (e.g., Habicht and Canfield,
which underwent more strain and are in a setting where metamor- 1996; Blank, 2004). Moreover, the isotopic variations at milli-
phic recrystallization has been more extensive. The δ13Ccarb values meter scale show that an important diversity of sulfur-dependent
range from –0.7‰ to +1.4‰ and average +0.1 ± 0.5‰, which organisms, mainly sulfate-reducing bacteria, were present. It is
are similar to Macgregor stromatolite values. The samples also also possible that sulfide-oxidizing bacteria were operational in
analyzed for δ13Cred give values mostly between –23.3‰ and – restricted localities during Jimmy Member deposition, as some
21.0‰. The <200 °C metamorphism mentioned earlier may have anoxygenic photosynthesizers can conduct sulfide oxidation
shifted δ13Cred slightly toward heavier values. Compared with the (Cohen et al., 1989). At the interface between anoxygenic and
Macgregor stromatolites, the evidence for Rubisco I is less obvi- oxygenic conditions, elemental sulfur is liberated and then re-
ous, but it cannot be ruled out because of the probable effect of reduced in part to sulfide, by sulfur disproportionation. At the
metamorphism, suggesting that the original values might have interface, 13C-depleted carbonate can also be formed, as seen in
been lighter. Alternatively, the 22‰ difference between δ13Ccarb TR51 sample (e.g., Wadham et al., 2004). More recent examples
and δ13Cred may be a primary fractionation record of Rubisco II of depleted carbonates associated with sulfide mats and globular
sulfides are found near methane seeps (Peckmann et al., 2004). and methanotrophy. The δ13Cred values in the Cheshire stromato-
The light carbon is a result of the anaerobic oxidation of the lites show the fractionation produced by the Rubisco I enzyme,
methane in association with sulfate reduction (e.g., Peckmann evidence of oxygenic photosynthesis by cyanobacteria, in a tidal
and Thiel, 2004). An alternative explanation of the positive sulfur to subtidal environment.
values could be Rayleigh fractionation of a finite sulfate reser-
voir. The data for sulfide below TR51 in the core are on aver- COMPARISON AMONG THE FOUR ARCHEAN
age 15‰ lighter, but they form a separate peak, which does not GREENSTONE BELTS
tail off to 34S-enrichment as expected by Rayleigh fractionation
(Fig. 6A). Therefore, a change to an environment typical of more Figure 10 (A, B) shows the narrow δ34S ranges obtained
modern methane- and sulfur-rich hydrothermal events seems for the Barberton and Isua Greenstone Belts. The metamorphic
to be the likely explanation. The relatively smaller range in the
Shavi Member (Fig. 6B) suggests that the rate of SO4 reduction
was higher than in the Jimmy Member (e.g., Schidlowski et al.,
1983). It is possible that in the Jimmy Member, the SO4 sup-
ply might have temporarily and locally changed, or been limited
at the deposition sites, as illustrated by the TR51 sample (δ34S
above +6‰).
The wide range of δ13Cred measured in the three Manjeri
carbon-rich units reveals the existence of a well-developed car-
bon cycle. The oxygenic photosynthesis operated by Rubisco I
existed in the Shavi and Spring Valley Members. This is consis-
tent with the interpretation of these units as shallow-water sedi-
ments (Hunter et al., 1998). At the same stratigraphic level, the
Rupemba stromatolites, despite some isotopic adjustment, were
most likely generated by the same activity, directed by cyanobac-
teria. The δ13Cred in the Jimmy Member may represent a complex
mixing of signatures from different bacterial processes, recording
composite bacterial communities. The Jimmy Member includes
finely laminated material probably deposited in quiet water
below wave base, deeper than the Shavi Member and with condi-
tions becoming dominantly anaerobic toward the top of the unit,
with δ13C red lighter than –30‰. Further evidence for the intro-
duction of methane into a previously oxic environment comes
from the δ13C of organic carbon. Ten centimeters below TR51,
the δ13Cred drops to –37‰ with an associated increase in carbon
content from <1 wt% to up to 18 wt% in the overlying 2.4 m.
At this stratigraphic level the sulfide isotopic compositions shift
from being strongly heterogeneous to having a restricted range
typical of hydrothermal sulfur (Fig. 6A). The apparent presence
of an “oxygenic photosynthesis” signature, at intervals mainly
observed in the lower part of this 2.4 m, may record the infall of
dead organisms from the water surface.
Highly evolved prokaryote mat communities operating
complex metabolic processes existed at 2.7 Ga (Grassineau et al.,
2001a, 2002). They are the result of various microbial ecosys-
tems due to a change of environments, from the subtidal or tidal
microbial mats of the Shavi Member to a deeper environment
below wave base with the appearance of anoxic processes in the
Jimmy Member. Figure 10. Comparison between δ34S values from this study for sulfide
Similar observations are noted in the Cheshire Formation. minerals from sedimentary sequences in Early, Middle, and Late Ar-
The dark shales give very light δ13Cred values, down to –44‰, chean. From small ranges, 7.2‰ in the IGB (A) and 4.1‰ in the BGB
(B), the range in the Belingwe belt reached 38‰ (C). Considering pre-
which argue in favor of anaerobic environments. The processes vious studies in the Barberton belt, it appears that the range slowly ex-
involved might have been occasionally anoxygenic photosynthe- panded through Early and Middle Archean but widened considerably
sis at the oxic/anoxic interface, but more likely methanogenesis at the dawn of the Late Archean.
events that affected the IGB, including the northeastern part,

might have reduced the original range of values, but the average
is probably representative of the primary signature. Even if there
is no additional evidence, a biological cause for this 4.9‰ range
cannot be ruled out. The fractionation is small, perhaps because
of a low supply of sulfate, but is still consistent with the presence
of sulfate reducers. The presence of high-temperature (>85 °C)
hyperthermophiles (Woese, 1987; Nisbet and Fowler, 1996; Nis-
bet and Sleep, 2001) could result in such a small isotopic frac-
tionation. Evidence for hydrothermal activity at Isua comes from
methane- and carbonate-rich inclusions in vesicles from the least
deformed pillow lavas (Appel et al., 2001), which are similar in
composition to some current off-ridge axis vent fields (Touret,
2003). In such settings inorganic processes may have been sig-
nificant, with elemental sulfur forming at vents (δ34S of 0‰ to
+1‰). Presumably little sulfate was present in the Isua oceans
and around these vents, and sulfur-oxidizing bacteria might have
existed (Jannasch, 1989). Under this condition, the fractionation
of sulfur due to biological activity will surely have been small,
whatever metabolic processes were occurring.
The range of δ34S values obtained for Barberton samples in
this study but also by previous authors (see list of references in the
BGB section) is slightly bigger that the one found for Isua. The
average is also enriched (+1.4‰) relative to the IGB (−0.2‰). In
the detailed study, the growth of replacement carbonate within
the formation and preferentially around the sulfide minerals may
have been responsible for some re-homogenization. However,
the recorded δ34S range is best explained as evidence of bacterial
sulfate reduction, in sediments or in the anoxic water column, at
3.24 Ga (Kakegawa and Ohmoto, 1999).
The impact of metamorphism and secondary processes on
the carbon isotopic compositions is not negligible, and perhaps
only locally has the δ13C partially preserved its original signa-
ture. This has to be taken into account in assessing claims for
biogenicity, especially in Isua, where Schidlowski et al. (1979),
Schidlowski (1988, 2001) and Rosing (1999) stressed that results
should be corrected for the metamorphism. Only the range of
–30‰ to –22‰ in the Isua sedimentary sequences can be dis-
cussed (though these values could have been lighter originally),
as it indicates the possibility of life. Rosing (1999) suggests the
presence of organisms similar to modern plankton in the Isua sea Figure 11. Comparison between δ13Cred values from this study for sedi-
(δ13Cred = −19‰). On the other hand, a process, probably anaero- mentary rocks in early (A), middle (B, C), and Late Archean (D–G),
bic, is proposed from the results of –29‰ to –28‰ in the sulfide- with a probable isotopic shift due to metamorphism in the IGB and
BGB. The δ13Cred values for Steep Rock Lake and Belingwe are light-
bearing BIF and the metaconglomerate. Given that the isotopic er (C, D). The two stromatolites, at 3.0 Ga and 2.6 Ga (C, G), show
range may have been reduced by metamorphism, values as light similar ranges for δ13Cred and δ13Ccarb. Stromatolites are represented in
as –29.6‰ in this BIF suggest that methanotrophic activity took dotted pattern, black shales in black. The thick black bars for each
place during deposition, although it is also possible that photo- diagram represent the ranges of δ13Ccarb values obtained, with their av-
synthetic processes had a wider range of fractionations at the erage value shown by the triangles.
time (Watanabe et al., 1997). Thus a complex community might
have been present as early as 3.8 Ga.
In the middle Archean, the δ13Cred range in the Barberton belt values may be linked to the formation of replacement carbonates.
is larger and lighter than in the Isua belt, with values down to However, the lower part of the range, –32‰ to –26‰, suggests
–32.4‰ (Figs. 11A, 11B). Again, allowance must be made for a that several bacterial activities are recorded, including photosyn-
small metamorphic shift, and for the possibility that the heavier thesis and methane-utilizing processes.
By the late Middle Archean, a significant change had The colonization of the continental margins by the cyanobacte-
occurred. The δ13Cred averaging –25.4‰ and the Δ13Cred-carb frac- rial mats was the beginning of a flourishing of activity that is well
tionation around 28‰ in Steep Rock Lake (Fig. 11C) are inter- illustrated at 2.7 Ga.
preted as clear evidence of biological activity and more particu- In the basal Manjeri Formation and the Cheshire Formation
larly of fractionation by Rubisco I. Therefore, as the biogenic in the 2.7 Ga Belingwe Greenstone Belt, the ranges for δ13Cred
structures were developed in shallow water associated with a and δ34S are wide, respectively ~37‰ and ~38‰. These values
carbonate platform, it is likely that oxygenic photosynthesis was indicate a large diversity of bacterial activities. The important
already established around 3.0 Ga, and that cyanobacteria were variations found at millimeter scale in samples indicate that these
fully active at that time. bacterial communities were interacting by various pathways.
By the Late Archean, the wide δ34S range obtained in the Oxygenic photosynthesizers and sulfate reducers were operating
Manjeri units strongly indicates an operational biological sulfur in microbial mats in shallow water reefs under subtidal and tidal
cycle at 2.7 Ga (Fig. 10C), with mainly sulfate-reducing bacteria, conditions. Organisms below wave base used mainly anoxygenic
but also possible sulfur oxidizers. The activities seem to depend photosynthesis and sulfate-reducing processes. Methanotrophs
on local supplies of sulfate, conditioned by the environment. A and sulfide oxidizers existed at the oxic/anoxic interface, with
fast SO4 supply will benefit a high rate of sulfate reduction in the methanogens below.
Shavi Member, but with a slower rate, additional but restricted
sulfur-oxidizing processes could have occurred within the Jimmy ACKNOWLEDGMENTS
Member at the oxic-anoxic interface. The wide δ13Cred range mea-
sured in the Manjeri Formation (Fig. 11D) reveals the existence Various funding bodies financed this study: Natural Envi-
of a well-developed carbon cycle with Rubisco I operational. The ronment Research Council (NERC) and the Leverhulme Trust
Shavi and Spring Valley Members and probably the Rupemba for the Belingwe work; the Geological Survey of Denmark and
stromatolites indicate oxygenic photosynthesis (Fig. 11F). The Greenland, the Isua Multidisciplinary Research Project, the Dan-
lighter δ13Cred values (below –30‰) recorded in the Jimmy Mem- ish Research Council, the Bureau of Minerals and Petroleum in
ber and Cheshire Formation suggest that processes in addition to Greenland, the Commission of Scientific Research in Greenland,
photosynthesis were occurring, most likely methanogenesis and the GEODE project, and the Royal Society for the Isua part. A
methanotrophy (Fig. 11E). Values for δ13Cred in the Macgregor particular thank you is addressed to C. Siebert and J. Kramers
stromatolites (Fig. 11G) indicate oxygenic photosynthesis with for the Fig Tree samples, courtesy of R. LeRoux and C. Rip-
cyanobacterial management. pon (AVGold, ETC division). The authors thank H. Ohmoto, T.
Over this one billion years of Earth’s history, the evolution Kakegawa and, Y. Shen for their constructive reviews, and S.
of life appears to have been marked by changes of environmen- Kesler for thorough editing of the manuscript.
tal conditions, as suggested by Nisbet and Sleep (2001). The
δ34S and δ13C obtained in the three Archean periods suggest that REFERENCES CITED
biological processes created wider isotopic fractionations from
3.8 Ga to 2.7 Ga. With better sulfate and oxygen supplies, organ- Abell, P.I., McClory, J., Martin, A., and Nisbet, E.G., 1985a, Archaean stromat-
olites from the Ngesi Group, Belingwe Greenstone Belt, Zimbabwe. Pres-
isms colonized a wide spectrum of environments on Earth, which ervation and stable isotopes—Preliminary results: Precambrian Research,
increased the biological diversity by the Late Archean. v. 27, p. 357–383, doi: 10.1016/0301-9268(85)90094-4.
Abell, P.I., McClory, J., Martin, A., Nisbet, E.G., and Kyser, T.K., 1985b, Petrog-
raphy and stable isotope ratios from Archaean stromatolites, Mushandike
CONCLUSIONS Formation, Zimbabwe: Precambrian Research, v. 27, p. 385–398, doi:
10.1016/0301-9268(85)90095-6.
In the Early Archean at 3.8 Ga, the δ34S spread in sedimen- Appel, P.W.U., 1979, Stratabound copper sulfides in a banded iron-formation
and in basaltic tuffs in the Early Precambrian Isua supracrustal belt: West
tary sequences of 7.2‰ around a 0‰ peak suggests that biologi- Greenland: Economic Geology, v. 74, p. 45–52.
cal activity was focused around high-temperature hydrothermal Appel, P.W.U., Fedo, C.M., Moorbath, S., and Myers, J.S., 1998, Recogniz-
vents. Sulfur processing could have existed, and possibly metha- able primary volcanic and sedimentary features in a low strain domain
of the highly deformed, oldest known (3.7–3.8 Ga) greenstone belt,
nogenesis (δ13Cred of –29‰ after metamorphism). Photosynthesis Isua: West Greenland: Terra Nova, v. 10, p. 57–62, doi: 10.1046/j.1365-
might also have occurred. 3121.1998.00162.x.
In the Middle Archean, at 3.24 Ga, microbial activity seems Appel, P.W.U., Rollinson, H.R., and Touret, J.L.R., 2001, Remnants of an early
Archaean (>3.74 Ga) sea-floor, hydrothermal system in the Isua Green-
to be more widespread, but mainly under anoxic conditions. Sul- stone Belt: Precambrian Research, v. 112, p. 27–49, doi: 10.1016/S0301-
fate-reducing processes have been recognized, using seawater 9268(01)00169-3.
sulfate. Methanogenic archaea were probably present, and pos- Blank, C.E., 2004, Evolutionary timing of the origins of mesophilic sulphate
reduction and oxygenic photosynthesis: A phylogenomic dating approach:
sibly sulfur oxidizers. For Steep Rock Lake at 3.0 Ga, there is Geobiology, v. 2, p. 1–20, doi: 10.1111/j.1472-4677.2004.00020.x.
unequivocal evidence of Rubisco I and therefore oxygenic pho- Bickle, M.J., and Nisbet, E.G., editors, 1993, The geology of the Belingwe
tosynthesis, with a fractionation between organic carbon and Greenstone Belt: A study of the evolution of Archaean continental crust:
Geological Society of Zimbabwe Special Publication 2: Rotterdam, A.A.
residual carbonate of –28‰. Therefore cyanobacteria were well Balkema, 239 p.
established by the end of the Middle Archean/early Late Archean. Bolhar, R., Hofmann, A., Woodhead, J.D., Hergt, J.M., and Dirks, P., 2002,
Pb- and Nd-isotope systematics of stromatolitic limestones from the 2.7 Ga Grassineau, N.V., Mattey, D.P., and Lowry, D., 2001b, Rapid sulphur isotopic
Ngezi Group of the Belingwe Greenstone Belt: Constraints on timing of analyses of sulphide and sulphate minerals by continuous flow- isotope
deposition and provenance: Precambrian Research, v. 114, p. 277–294. ratio mass spectrometry (CF-IRMS): Analytical Chemistry, v. 73, p. 220–
Cameron, E.M., 1982, Sulphate and sulphate reduction in early Precambrian 225, doi: 10.1021/ac000550f.
oceans: Nature, v. 296, p. 145–148, doi: 10.1038/296145a0. Grassineau, N.V., Nisbet, E.G., Fowler, C.M.R., Bickle, M.J., Lowry, D.,
Canfield, D.E., and Teske, A., 1996, Late Proterozoic rise in atmospheric oxy- Chapman, H.J., Mattey, D.P., Abell, P., Yong, J., and Martin, A., 2002,
gen concentration inferred from phylogenetic and sulphur-isotope stud- Stable isotopes in the Archaean Belingwe belt, Zimbabwe: Evidence for
ies: Nature, v. 382, p. 127–132, doi: 10.1038/382127a0. a diverse microbial mat ecology, in Fowler, C.M.R., Ebinger, C.J., and
Cavalier-Smith, T., 2002, The neomuran origin of archaebacteria, the negibac- Hawkesworth, C.J., eds., The early Earth: Physical, chemical and biologi-
terial root of the universal tree and bacterial megaclassification: Inter- cal development: Geological Society [London] Special Publication 199,
national Journal of Systematic and Evolutionary Microbiology, v. 52, p. 309–328.
p. 7–76. Habicht, K., and Canfield, D.E., 1996, Sulphur isotope fractionation in modern
Chauvel, C., Dupre, B., and Arndt, N.T., 1993, Pb and Nd isotopic correla- microbial mats and the evolution of the sulphur cycle: Nature, v. 382,
tion in Belingwe komatiites and basalts, in Bickle, M.J., and Nisbet, E.G., p. 342–343, doi: 10.1038/382342a0.
eds., The geology of the Belingwe Greenstone Belt, Zimbabwe: Harare, Hayes, J.M., Kaplan, I.R., and Wedeking, K.W., 1983, Precambrian organic
Geological Society of Zimbabwe, v. 2, p. 167–174. geochemistry, preservation of the record: Chapter 5, in Schopf, J.W., ed.,
Cohen, Y., Gorlenko, V.M., and Bonch-Osmolovskaya, E.A., 1989, Interaction Earth’s earliest biosphere: Its origin and evolution: Princeton, New Jersey,
of Sulphur and Carbon Cycles in Microbial Mats, in Brimblecombe, P., Princeton University Press, p. 93–134.
and Lein, A. Yu, eds., Evolution of the global biogeochemical sulphur Hayes, J.M., 1994, Global methanotrophy at the Archaean-Proterozoic transi-
cycle: SCOPE report no. 39: Chichester, UK, John Wiley and Sons, p. tion, in Bengtson, S., ed., Early life on Earth: Nobel Symposium 84: New
191–238. York, Columbia University Press, p. 220–236.
Coleman, D.D., Risatti, J.B., and Schoell, M., 1981, Fractionation of carbon and Hoefs, J., 1997, Stable Isotope Geochemistry (4th edition): Berlin, Springer-
hydrogen isotopes by methane-oxidizing bacteria: Geochimica et Cosmo- Verlag, 201 p.
chimica Acta, v. 45, p. 1033–1037, doi: 10.1016/0016-7037(81)90129-0. Hunter, M.A.H., Bickle, M.J., Nisbet, E.G., Martin, A., and Chapman, H.J.,
Davis, D.W., and Jackson, M.C., 1985, Preliminary U-Pb zircon ages from the 1998, A continental extension setting for the Archaean Belingwe Green-
Lumby Lake – Marmion Lake area, districts of Kerona and Rainy River, stone belt, Zimbabwe: Geology, v. 26, p. 883–886, doi: 10.1130/0091-
in Wood, J., White, O.L., Barlow, R.B., and Colvine, A.C., eds., Summary 7613(1998)026<0883:CESFTA>2.3.CO;2.
of fieldwork and other activities: Ontario Geological Survey Miscella- Jannasch, H.W., 1989, Sulphur emission and transformations at deep sea hydro-
neous Paper 126, p. 135–137. thermal vents, in Brimblecombe, P., and Lein, A. Yu., eds., Evolution of
de Ronde, C.E.J., and Ebbesen, T.W., 1996, 3.2 b.y. of organic compound the global biogeochemical sulphur cycle: SCOPE report no. 39: Chiches-
formation near sea-floor hot springs: Geology, v. 24, p. 791–794, doi: ter, UK, John Wiley and Sons, p. 181–190.
10.1130/0091-7613(1996)024<0791:BYOOCF>2.3.CO;2. Kakegawa, T., and Ohmoto, H., 1999, Sulfur isotope evidence for the origin
de Ronde, C.E.J., Spooner, E.C.T., de Wit, M.J., and Bray, C.J., 1992, Shear of 3.4 to 3.1 Ga pyrite at the Princeton gold mine, Barberton Green-
zone-related, Au quartz vein deposits in the Barberton greenstone belt, stone Belt, South Africa: Precambrian Research, v. 96, p. 209–224, doi:
South Africa: Field and petrographic characteristics, fluid properties, and 10.1016/S0301-9268(99)00006-6.
light stable isotope geochemistry: Economic Geology and the Bulletin of Komiya, T., Maruyama, S., Masuda, T., Nohda, S., Hayashi, M., and Okamoto,
the Society of Economic Geologists, v. 87, p. 366–402. K., 1999, Plate tectonics at 3.8–3.7 Ga: Field evidence from the Isua
Erez, J., Bouevitch, A., and Kaplan, A., 1998, Carbon isotope fractionation by accretionary complex, Southern West Greenland: Journal of Geology,
photosynthetic aquatic microorganisms: Experiments with Synechococ- v. 107, p. 515–554, doi: 10.1086/314371.
cus PCC7942, and a simple carbon flux model: Canadian Journal of Bot- Kramers, J.D., Meisel, T., Morel, P., Nägler, T., and Siebert, C., 2004, Clues
any, v. 76, p. 1109–1118, doi: 10.1139/cjb-76-6-1109. from PGE and Mo on the early redox history of the atmosphere [abs.]:
Faure, G., 1986, Principles of isotope geology (2nd edition): New York, John Geologiska Föreningen, 26th Nordic Geological Winter Meeting, Uppsala,
Wiley and Sons, 589 p. Sweden, v. 126, p. 11.
Fedo, C.M., 2000, Setting and origin for problematic rocks from the >3.7 Ga Kusky, T.M., and Hudleston, P.J., 1999, Growth and demise of an Archean car-
Isua Greenstone Belt, southern west Greenland: Earth’s oldest coarse bonate platform Steep Rock Lake, Ontario, Canada: Canadian Journal of
clastic sediments: Precambrian Research, v. 101, p. 69–78, doi: 10.1016/ Earth Sciences, v. 36, p. 565–584, doi: 10.1139/cjes-36-4-565.
S0301-9268(99)00100-X. Lepland, A., Arrhenius, G., and Cornell, D., 2002, Apatite in Early Archean
Fedo, C.M., Myers, J.S., and Appel, P.W.U., 2001, Depositional setting and Isua supracrustal rocks, southern West Greenland: Its origin, association
paleogeographic implications of earth’s oldest supracrustal rocks, the with graphite and potential as a biomarker: Precambrian Research, v. 118,
> 3.7 Ga Isua Greenstone belt: West Greenland: Sedimentary Geology, p. 221–241, doi: 10.1016/S0301-9268(02)00106-7.
v. 141, p. 61–77, doi: 10.1016/S0037-0738(01)00068-9. Londry, K.L., and Des Marais, D.J., 2003, Stable carbon isotope fractionation
Frei, R., and Rosing, M.T., 2001, The least radiogenic terrestrial leads; implica- by sulfate-reducing bacteria: Applied and Environmental Microbiology,
tions for the early Archean crustal evolution and hydrothermal-metaso- v. 69, p. 2942–2949, doi: 10.1128/AEM.69.5.2942-2949.2003.
matic processes in the Isua supracrustal belt (West Greenland): Chemical Lowe, D., and Byerly, G.R., 1999, Stratigraphy of the west-central part of
Geology, v. 181, p. 47–66, doi: 10.1016/S0009-2541(01)00263-7. the Barberton Greenstone Belt, South Africa, in Lowe, D., and Byerly,
Frei, R., Bridgwater, D., Rosing, M., and Stecher, O., 1999, Controversial Pb- G.R., eds., Geologic evolution of the Barberton Greenstone Belt, South
Pb and Sm-Nd isotope results in the early Archean Isua (West Greenland) Africa. Geological Society of America Special Paper 329, p. 1–36.
oxide iron formation: Preservation of primary signatures versus secondary Martin, A., Nisbet, E.G., and Bickle, M.J., 1980, Archaean Stromatolites of the
disturbances: Geochimica et Cosmochimica Acta, v. 63, 3/4, p. 473–488. Belingwe Greenstone Belt: Precambrian Research, v. 13, p. 337–362, doi:
Grassineau, N.V., and Appel, P.W.U., 2000, Variations in the stable isotope 10.1016/0301-9268(80)90049-2.
compositions for two gold deposits in the Isua Greenstone Belt, West Matthews, D.E., and Hayes, J.M., 1978, Isotope-ratio-monitoring gas chroma-
Greenland Sequences [abs.]: Eos (Transactions, American Geophysical tography-mass spectrometry: Analytical Chemistry, v. 50, p. 1465–1473,
Union), v. 81, no. 48, p. F1267–68. doi: 10.1021/ac50033a022.
Grassineau, N.V., Nisbet, E.G., and Fowler, C.M.R., 2000, A stable isotopic McClory, J.P., 1988, Carbon and Oxygen Isotopic Study of Archaean Stro-
search for biological signatures in the Archaean: A comparative study matolites from Zimbabwe [Ph.D. thesis]: Kingston, University of Rhode
of early (Isua, 3.8 Ga; Greenland) and late Archaean (Belingwe, 2.7 Ga; Island, 107 p.
Zimbabwe) sequences [abs.]: Eos (Transactions, American Geophysical Monster, J., Appel, P.W.U., Thode, H.G., Schidlowski, M., Carmichael, C.M.,
Union), v. 81, no. 48, p. F1257–58. and Bridgwater, D., 1979, Sulfur isotope studies in Early Archaean sedi-
Grassineau, N.V., Nisbet, E.G., Bickle, M.J., Fowler, C.M.R., Lowry, D., Mat- ments from Isua, West Greenland: Implications for the antiquity of bacte-
tey, D.P., Abell, P., and Martin, A., 2001a, Antiquity of the biological rial sulfate reduction: Geochimica et Cosmochimica Acta, v. 43, p. 405–
sulphur cycle: Evidence from S and C isotopes in 2.7 Ga rocks of the 413, doi: 10.1016/0016-7037(79)90205-9.
Belingwe Belt, Zimbabwe: Proceedings of the Royal Society of London, Moorbath, S., and Whitehouse, M.J., 1996, Age of the Isua supracrustal
ser. B, v. 268, p. 113–119, doi: 10.1098/rspb.2000.1338. sequence of West Greenland, in Chela-Flores, J., and Raulin, F., eds.,
Chemical evolution: Physics of the origin and evolution of life: Dordre- and rims, Isua greenstone belt, Southwest Greenland: Implications for
cht, Kluwer Academic Publishers, p. 87–95. postmagmatic alteration processes: Geochimica et Cosmochimica Acta,
Moorbath, S., and Kamber, B.S., 1998, Re-appraisal of the age of the oldest v. 67, p. 441–457, doi: 10.1016/S0016-7037(02)01094-3.
water-lain sediments, West Greenland, in Chela-Flores, J., and Raulin, F., Rollinson, H., 2002, The metamorphic history of the Isua Greenstone Belt,
eds., Exobiology: Matter, energy and information in the origin of evolu- West Greenland, in Fowler, C.M.R., Ebinger, C.J., and Hawkesworth,
tion of life in the universe. Dordrecht, Kluwer Academic Publishers, p. C.J., eds., The early Earth: Physical, chemical and biological develop-
81–86. ment: Geological Society [London] Special Publication 199, p. 329–350.
Mojzsis, S.J., Coath, C.D., Greenwood, J.P., McKeegan, K.D., and Harrison, Rollinson, H., 2003, Metamorphic history suggested by garnet-growth chro-
T.M., 2003, Mass-independent isotope effects in Archean (2.5 to 3.8 Ga) nologies in the Isua Greenstone Belt, West Greenland: Precambrian
sedimentary sulfides determined by ion microprobe analysis: Geochi- Research, v. 126, 3–4, p. 181–196.
mica et Cosmochimica Acta, v. 67, p. 1635–1658, doi: 10.1016/S0016- Robinson, J.J., Scott, K.M., Swanson, S.T., O’Leary, M.H., Horken, K., Tabita,
7037(03)00059-0. F.R., and Cavanaugh, C.M., 2003, Kinetic isotope effect and characteriza-
Myers, J.S., 2001, Protoliths of the 3.8–3.7 Ga Isua Greenstone Belt: West tion of form II RubisCO from the chemoautotrophic endosymbionts of the
Greenland: Precambrian Research, v. 105, p. 129–141, doi: 10.1016/ hydrothermal vent tubeworm Riftia pachyptila: Limnology and Oceanog-
S0301-9268(00)00108-X. raphy, v. 48, p. 48–54.
Naraoka, H., Ohtake, M., Maruyama, S., and Ohmoto, H., 1996, Non-bio- Rose, N.M., Rosing, M.T., and Bridgwater, D., 1996, The origin of metacarbon-
genic graphite in 3.8 Ga metamorphic rocks from the Isua district, ate rocks in the Archaean Isua supracrustal belt, West Greenland: Ameri-
Greenland: Chemical Geology, v. 133, p. 251–260, doi: 10.1016/S0009- can Journal of Science, v. 296, p. 1004–1044.
2541(96)00076-9. Rosing, M.T., 1999, 13C-depleted carbon microparticles in >3700-Ma sea-floor
Nisbet, E.G., and Fowler, C.M.R., 1996, Some like it hot: Nature, v. 382, sedimentary rocks from West Greenland: Science, v. 283, p. 674–676,
p. 404–406, doi: 10.1038/382404a0. doi: 10.1126/science.283.5402.674.
Nisbet, E.G., and Fowler, C.M.R., 1999, Archaean metabolic evolution of Rosing, M.T., and Frei, R., 2004, U-rich Archaean sea-floor sediments from
microbial mats: Proceedings of the Royal Society of London, ser. B, Greenland—Indications of >3700 Ma oxygenic photosynthesis: Earth
v. 266, p. 2375–2382, doi: 10.1098/rspb.1999.0934. and Planetary Science Letters, v. 217, p. 237–244, doi: 10.1016/S0012-
Nisbet, E.G., and Sleep, N.H., 2001, The habitat and nature of early life: Nature, 821X(03)00609-5.
v. 409, p. 1083–1091, doi: 10.1038/35059210. Rye, R.O., and Ohmoto, H., 1974, Sulphur and carbon isotopes and ore genesis:
Nutman, A.P., 1986, The early Archaean to Proterozoic history of the Isuka- A review: Economic Geology and the Bulletin of the Society of Economic
sia area, southern West Greenland: Copenhagen, Grønlands Geologiske Geologists, v. 69, p. 826–842.
Undersøgelse Bulletin, v.154, p. 80. Sakai, H., Des Marais, D.J., Ueda, A., and Moore, J.G., 1984, Concentrations of
Nutman, A.P., Bennett, V.C., Friend, C.R.L., and Rosing, M.T., 1997, ~3710 isotope ratios of carbon, nitrogen and sulfur in ocean-floor basalts: Geo-
and >3790 Ma volcanic sequences in the Isua (Greenland) supracrustal chimica et Cosmochimica Acta, v. 48, p. 2433–2441, doi: 10.1016/0016-
belt; structural and Nd isotope implications: Chemical Geology, v. 141, 7037(84)90295-3.
p. 271–287, doi: 10.1016/S0009-2541(97)00084-3. Schidlowski, M., 1988, A 3,800-million-year isotopic record of life
Ohmoto, H., 1986, Stable isotope geochemistry of ore deposits, in Valley, J.W., from carbon in sedimentary rocks: Nature, v. 333, p. 313–318, doi:
Taylor, H.P., and O’Neil, J.R., eds., Stable isotopes in high temperature 10.1038/333313a0.
geological processes: Washington, D.C., Mineralogical Society of Amer- Schidlowski, M., 2001, Carbon isotopes as biogeochemical recorders of
ica, Reviews in Mineralogy, v. 16, p. 491–559. life over 3.8 Ga of Earth history: Evolution of a concept: Precambrian
Ohmoto, H., 1992, Biochemistry of sulfur and the mechanisms of sulfide- Research, v. 106, p. 117–134, doi: 10.1016/S0301-9268(00)00128-5.
sulfate mineralization in Archean oceans, in Schidlowski, M., Golubic, Schidlowski, M., Eichmann, R., and Junge, C.E., 1975, Precambrian sedimen-
S., Kimberley, M.M., McKirdy, D.M., and Trundinger, P.A., eds., Early tary carbonates: Carbon and oxygen isotope geochemistry and implica-
organic evolution: Implications for mineral and energy resources. Berlin, tions for the terrestrial oxygen budget: Precambrian Research, v. 2, p. 1–
Springer-Verlag, p. 378–397. 69, doi: 10.1016/0301-9268(75)90018-2.
Ohmoto, H., and Felder, R.P., 1987, Bacterial activity in the warmer, sul- Schidlowski, M., Appel, P.W.U., Eichmann, R., and Junge, C.E., 1979, Carbon
phate-bearing, Archaean oceans: Nature, v. 328, p. 244–246, doi: isotope geochemistry of the 3.7×109 yr old Isua sediments, West Green-
10.1038/328244a0. land: Implications for the Archaean carbon and oxygen cycles: Geo-
Ohmoto, H., and Godhaber, M., 1997, Sulfur and carbon isotopes, in Barnes, chimica et Cosmochimica Acta, v. 43, p. 189–199, doi: 10.1016/0016-
H.L., ed., Geochemistry of hydrothermal ore deposits, Volume 3: New 7037(79)90238-2.
York, John Wiley and Sons, p. 517–612. Schidlowski, M., Hayes, J.M., and Kaplan, I.R., 1983, Isotopic interferences of
Ohmoto, H., Kakegawa, T., and Lowe, D.R., 1993, 3.4-billion-year-old bio- ancient biochemistries: Carbon, sulfur, hydrogen, and nitrogen, in Schopf,
genic pyrites from Barberton, South Africa: Sulfur isotope evidence: Sci- J.W., ed., Earth’s earliest biosphere: Its origin and evolution: Princeton,
ence, v. 262, p. 555–557. New Jersey, Princeton University Press, Princeton, v. 7, p. 149–186.
Pace, N.R., 1997, A molecular view of microbial diversity and the biosphere: Sirevåg, R., Buchanan, B.B., Berry, J.A., and Troughton, J.H., 1977, Mecha-
Science, v. 276, p. 734–740, doi: 10.1126/science.276.5313.734. nisms of CO2 fixation in bacterial photosynthesis studied by the carbon
Paris, I.A., 1987, The Barberton Greenstone succession, South Africa: Implica- isotope fractionation technique: Archives of Microbiology, v. 112, p. 35–
tions for modelling the evolution of the Archaean crust: Chichester, UK, 38, doi: 10.1007/BF00446651.
John Wiley and Sons Geological Journal (winter thematic issue), v. 22, Shen, Y., and Buick, R., 2004, The antiquity of microbial sulfate reduction:
p. 5–24. Earth-Science Reviews, v. 64, p. 243–272.
Peckmann, J., and Thiel, V., 2004, Carbon cycling at ancient meth- Shen, Y., Buick, R., and Canfield, D.E., 2001, Isotopic evidence for microbial
ane-seeps: Chemical Geology, v. 205, p. 443–467, doi: 10.1016/ sulphate reduction in the early Archaean era: Nature, v. 410, p. 77–81,
j.chemgeo.2003.12.025. doi: 10.1038/35065071.
Peckmann, J., Thiel, V., Reitner, J., Taviani, M., Aharon, P., and Michalis, W., Siebert, C., 2003, Molybdenum isotope fractionation and its application to stud-
2004, A microbial mat of a large sulfur bacterium preserved in a Miocene ies of marine environments and surface processes through geological time
methane-Seep limestone: Geomicrobiology Journal, v. 21, p. 247–255, [Ph.D. thesis]: Bern, Switzerland, University of Bern, 82 p.
doi: 10.1080/01490450490438757. Strauss, H., 2003, Sulphur isotope and the early Archaean sulphur cycle: Pre-
Perry, E.C., Jr., and Ahmad, S.N., 1977, Carbon isotope composition of graph- cambrian Research, v. 126, 3–4, p. 349–361.
ite and carbonate minerals from 3.8-AE metamorphosed sediments, Isu- Strauss, H., and Moore, T.B., 1992, Abundances and isotopic compositions of
kasia, Greenland: Earth and Planetary Science Letters, v. 36, p. 280–284, carbon and sulfur species in whole rock and kerogen samples, in Schopf,
doi: 10.1016/0012-821X(77)90210-2. J.W., and Klein, C., eds., The Proterozoic biosphere, Cambridge, UK,
Pierson, B.K., 1994, The emergence, diversification and role of photosynthetic Cambridge University Press, Ch. 17, p. 711–798.
eubacteria, in Bengtson, S., ed., Early life on Earth: Nobel Symposium Stevens, C.M., 1988, Atmospheric methane: Chemical Geology, v. 71, p. 11–
84: New York, Columbia University Press, p. 161–180. 21, doi: 10.1016/0009-2541(88)90102-7.
Polat, A., Hofmann, A.W., Münker, C., Regelous, M., and Appel, P.W.U., 2003, Taylor, B.E., 1986, Magmatic volatiles: isotopic variation of C, H, and S, in
Contrasting geochemical patterns in the 3.7–3.8 Ga pillow basalt cores Valley, J.W., Taylor, H.P., and O’Neil, J.R., eds., Stable isotopes in high
temperature geological processes: Washington, D.C., Mineralogical Soci- Viljoen, M.J., and Viljoen, R.P., 1969, An introduction to the geology of the
ety of America, Reviews in Mineralogy, v. 16, p. 185–226. Barberton granite-greenstone terrain, in Haughton, S.H., ed., Upper
Touret, J.L.R., 2003, Remnants of early Archaean hydrothermal methane and Mamie Project: Geological Society of South Africa Special Publication 2,
brines in pillow-breccia from the Isua Greenstone Belt, West Greenland: p. 9–28.
Precambrian Research, v. 126, 3–4, p. 219–233. Wadham, J.L., Bottrell, S., Tranter, M., and Raiswell, R., 2004, Stable isotope
Ueno, Y., Yurimoto, H., Yoshioka, H., Komiya, T., and Maruyama, S., 2002, evidence for microbial sulphate reduction at the bed of a polythermal high
Ion microprobe analysis of graphite from ca. 3.8 Ga metasediments, Isua Arctic glacier: Earth and Planetary Science Letters, v. 219, p. 341–355,
supracrustal belt, West Greenland: Relationship between metamorphism doi: 10.1016/S0012-821X(03)00683-6.
and carbon isotopic Composition: Geochimica et Cosmochimica Acta, Watanabe, Y., Naraoka, H., Wronkiewicz, D.J., Condie, K.C., and Ohmoto,
v. 66, p. 1257–1268, doi: 10.1016/S0016-7037(01)00840-7. H., 1997, Carbon, nitrogen, and sulfur geochemistry of Archean and
van der Meer, M.T.J., Schouten, S., van Dongen, B.E., Rijpstra, W.I.C., Fuchs, Proterozoic shales from the Kaapvaal Craton, South Africa: Geochi-
G., Sinninghe Damsté, J.S., de Leeuw, J.W., and Ward, D.M., 2001, mica et Cosmochimica Acta, v. 61, p. 3441–3459, doi: 10.1016/S0016-
Biosynthetic controls on the 13C contents of organic components in the 7037(97)00164-6.
photoautotrophic bacterium Chloroflexus aurantiacus: Journal of Bio- Wilks, M.E., and Nisbet, E.G., 1985, Archaean stromatolites from the Steep
logical Chemistry, v. 276, no. 14, p. 10971–10976, doi: 10.1074/jbc. Rock Group, northwestern Ontario, Canada: Canadian Journal of Earth
M009701200. Sciences, v. 22, p. 792–799.
van Zuilen, M.A., Lepland, A., and Arrhenius, G., 2002, Reassessing the Wilks, M.E., and Nisbet, E.G., 1988, Stratigraphy of the Steep Rock Group,
evidence for the earliest traces of life: Nature, v. 418, p. 627–630, doi: northwestern Ontario: A major Archaean unconformity and Archaean
10.1038/nature00934. stromatolites: Canadian Journal of Earth Sciences, v. 25, p. 370–391.
van Zuilen, M.A., and Lepland, A., Teranes, J., Fenarelli, J., Wahlen, M., and Woese, C.R., 1987, Bacterial evolution: Microbiological Reviews, v. 51,
Arrhenius, G., 2003, Graphite and carbonates in the 3.8 Ga old Isua p. 221–271.
Supracrustal Belt, southern West Greenland: Precambrian Research, v. Yong, J.N., 1991, Stable isotope ratios of shales from the Archaean Cheshire
126, p. 331–348. Formation, Zimbabwe [Ph.D. thesis]: Kingston, University of Rhode
Veizer, J., Hoefs, J., Ridler, R.H., Jensen, L.S., and Lowe, D.R., 1989, Geo- Island, 127 p.
chemistry of Precambrian carbonates: I. Archean hydrothermal systems:
Geochimica et Cosmochimica Acta, v. 53, p. 845–857, doi: 10.1016/0016-
7037(89)90030-6. MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA


Memoir 198
2006
Fingerprinting the metal endowment of early continental crust

to test for secular changes in global mineralization
Christien Thiart*
Africa Earth Observatory Network (AEON) and Department of Statistical Sciences,
University of Cape Town, Rondebosch 7701, South Africa
Maarten J. de Wit*
Africa Earth Observatory Network (AEON) and Department of Geological Sciences,
University of Cape Town, Rondebosch 7701, South Africa
ABSTRACT
Archean cratons are fragments of old continents that are believed to be more
richly endowed with mineral deposits than younger terrains. The mineral deposits of
different cratons are also diversely enriched with useful (to humankind) chemical ele-
ments. Cratons are therefore mineral diversity hotspots that represent regional geo-
chemical heterogeneities of early Earth, the evidence for which remains encoded on
each craton as unique metallogenic “fingerprints.” Using six selected elements groups
from our extensive in-house GIS database of Gondwana mineral deposits, we derive
the metallogenic fingerprints of 11 Archean cratons of the Southern Hemisphere,
and compare these against metallogenic fingerprints of the same selected elements in
younger crust of three of their host continents (Africa, Australia, and South America).
After adjusting the mineral inventory of each craton to account for underexploration
of regions lacking infrastructure and other political and economic conditions for min-
eral investment, we show that mineral deposit density and diversity of Earth’s conti-
nental lithosphere has decreased with time. We conclude that metallogenic elements
were transferred more efficiently from the mantle to the continental lithosphere in
the Archean and/or that subsequent (younger than 2.5 Ga) recycling of these elements
(mineral deposits) back into the mantle has become more effective. How most of these
fragments of old continents inherited their rich and diverse metallogenic character-
istics is unresolved, because different cratons are likely to represent only small rem-
nants of once much larger and possibly more varied Archean continents, and part of
the total metal inventory of Archean continents must have been recycled back into in
the mantle. The latter has implications for understanding the secular change in the
redox state of the Archean mantle and fluid envelope.
Keywords: cratons, Gondwana, metallogenesis, mineralization, exploration index,

geodynamics.
INTRODUCTION are underlain by relatively thin crust (~30–40 km) and thick man-
tle lithosphere (up to ~250–300 km; de Wit, 1998; James et al.,
The lithosphere of Archean cratons (older than 2.5 Ga) is 2001; Stankiewicz et al., 2002; Fouch et al., 2004). The origin of
distinct from that of younger continental lithosphere in that they these cratons is still a matter of intense debate. Part of that debate
*[email protected], [email protected]
Thiart, C., and de Wit, M.J., 2006, Fingerprinting the metal endowment of early continental crust to test for secular changes in global mineralization, in Kesler, S.E.,
and Ohmoto, H., eds., Evolution of Early Earth’s Atmosphere, Hydrosphere, and Biosphere—Constraints from Ore Deposits: Geological Society of America Memoir
198, p. 53–66, doi: 10.1130/2006.1198(03). For permission to copy, contact [email protected]. ©2006 Geological Society of America. All rights reserved.
53
54 C. Thiart and M.J. de Wit
centers on the origin of the cratons’ mineral deposits and, in turn, (Herrington et al., 1997). Although not all of these greenstone belt
how knowledge of these deposits may shed light on the forma- deposits are firmly dated, we assume they reflect inheritance of
tion of Earth’s early continents. However, surprisingly few quan- Archean gold. Thus, unless there is clear evidence to the contrary,
titative data are available in a collective format with which to some geodynamic process(es) served to concentrate precious ele-
embark on a comparative study of the riches found in continental ments together into the continental lithosphere of the Kaapvaal
crust of different ages. Near surface, the crust of most cratons is Craton during its formation in the Archean (Groves et al., 1987;
well endowed in concentrations of metallic elements useful and Tredoux et al., 1989; McDonald et al., 1995).
economic to humankind (hereafter referred to as mineral depos- The mantle lithosphere of most cratons is also invariably
its), but it is not known with any degree of certainty if the crust of enriched in diamonds that range in age from Mesoarchean to
the old continents was mineralized to a greater degree than that Phanerozoic (Hart et al., 1997; Shirey et al., 2002; Jelsma et al.,
of younger continents, as is often assumed intuitively. As a test 2004). Because the preservation potential of these minerals in the
of such a temporal change, here we compare the mineral endow- mantle lithosphere relates mostly to the relatively low heat flow
ment of old cratonic crust with that of younger crust, using our recorded across most cratons, we do not here focus further on
in-house mineral deposit database. diamonds. They do, however, serve to illustrate the resilience of
Cratons are metallogenically distinct in that their mineral cratons as Archean geologic archives.
deposits contain different mixtures of metallic elements from We first set out to verify that different Archean cratons have
craton to craton (Wilsher et al., 1993; Wilsher, 1995; de Wit et distinct metallogenic patterns, which we refer to as their metal-
al., 1999). For example, the Kaapvaal Craton of South Africa is logenic “fingerprints.” These fingerprints may reveal something
known to be relatively enriched in gold and platinum group ele- fundamental about the formation of cratons and Earth’s earliest
ments (PGE), the Zimbabwe Craton and the Yilgarn Craton of continents. Next we address the question of whether a unit of
Australia in gold and tungsten, the São Francisco Craton in gold Archean continental crust is more enriched in mineral deposits
and base metals (Cu/Pb/Zn), and the Amazonian Craton in gold than that of younger crust. We compare our results from Archean
and tin (Groves et al., 1987; Wilsher, 1995; Barley et al., 1998; cratons with those from younger crust (younger than 2.5 Ga) at
de Wit et al., 1999; de Wit and Thiart, 2005). We assume that different scales: first at a continental scale (e.g., South America,
mineral deposits of Archean cratons are contemporaneous with Australia, Africa), and then at a supercontinental scale using all
their host rocks and reflect processes of geochemical concentra- the known mineral deposits of Gondwana. This reveals important
tion that operated during the formation of the cratons. In most information about the evolution of continental crust.
cases this can be verified if the age of the mineral deposits is con- We are limited in our analyses to the Southern Hemisphere,
strained by geological relations or isotopic ages. In some cases because our mineral database is confined only to continental
it is not clear whether these mineral enrichments were inherited fragments of the former supercontinent Gondwana (see below).
from their host Archean craton or were added to the craton at a A shortcoming of our previous work (de Wit and Thiart, 2005)
later stage. For example, the platinum deposits of the Kaapvaal relates to the extent that the distribution of known mineral depos-
Craton are associated mostly with the igneous rocks of the its (as in our database) is skewed because some regions are
Bushveld Complex that intruded the center of this craton in the explored better than others. Less-developed countries with poor
Paleoproterozoic (2.05–2.06 Ga; Cawthorn and Walraven, 1998; infrastructure and political instability are less likely to have had
Eglington and Armstrong, 2004) and could have been added their mineral inventory tested to the same degree as better-devel-
from the asthenospheric mantle at that time. However, the PGE oped nations. Here, we attempt to adjust for underexploration by
geochemistry (and that of other siderophile elements like Ni, Cr, including a socioeconomic measure in our calculations.
Au) of lithospheric mantle xenoliths found in kimberlites across Establishing metallogenic fingerprints of different cratons by
the Kaapvaal Craton (McDonald et al., 1995), as well as that of using selected elements from their mineral deposits allows us to
Archean mid to lower crust (Hart et al., 2004) and of Archean address several controversial tectonic models that contrast early
mineral deposits in the greenstone belts of this craton (Tredoux Earth processes with those of the present. For example, modern
et al., 1989), suggests that both the lithospheric mantle and crust plate tectonic processes yield specific mineral deposit types in dis-
of this craton were already rich in PGE in Archean times. Indeed tinct plate tectonic environments (Sawkins, 1990; Windley, 1995).
it has been suggested that the PGE in the parental magmas of the Similar associations, if found in Archean cratons, would argue for
Bushveld Complex may have been inherited from contamination plate tectonic processes in the Archean. Conversely, on an early
with its underlying PGE-enriched Archean mantle lithosphere Earth dominated by vertical tectonics (e.g., driven by plume and
(McDonald et al., 1995). Similarly, most of the extraordinary diapir dynamics, as postulated by some workers [Hamilton, 1998;
enrichment of the Kaapvaal Craton in gold is hosted by Archean Zegers and van Keken, 2001; van Kranendonk et al., 2004]), metal-
sedimentary rocks of the Witwatersrand Basin and is thought to logenic provinces would be expected to display distinct metal asso-
have been derived largely by erosion of earlier crust, although ciations not found in the younger crust of the present continents
some gold may have been introduced and/or remobilized at a (e.g., Hutchinson, 1981, 1992). In addition, variations in the total
later time (Frimmel and Minter, 2002; Phillips and Evans, 2004). concentration of these elements in continental crust of different
Gold deposits are also common in older Archean greenstone belts ages may be used to test for changes in the rates of related tectonic
Fingerprinting the metal endowment of early continental crust 55
processes over time. This is one aim of this paper, but there are center, AEON. This database is restricted to the major continental
others: First, can metallogenic elements serve as chemical tracers fragments of Gondwana; its geological component is based on the
to establish that the earliest continents were assembled by tectonic geological map of Gondwana (de Wit et al., 1988), whereas the
processes as diverse as those at present? Second, can the mineral mineral layer of the database was constructed from open-access
diversity patterns of cratons help us to decipher the recycling his- literature sources (Wilsher, 1995). Originally the mineral layer
tory of Earth’s continental materials? Third, can mineral diversity comprised roughly 10,000 deposits across Gondwana, cover-
patterns (or mineral hotspots) of cratons and younger lithosphere ing 15 metallogenic elements (commodities). The mineral layer
be used to test reconstructions of past supercontinents like Gond- consists of shape files tied to 28 attribute tables, with attributes
wana and Rodinia? Finally, because mineralization in the crust ranging from commodities, size, type, and age of the mineral
requires concentration processes that often involve large fluid fluxes deposits, as well as information on the host rocks of the deposits
between the mantle and/or crust and the hydrosphere, it is of interest and the references from which the data were obtained. In 1998,
to ask whether secular change in the total mass of mineral deposits in collaboration with the BRGM (France), we embarked on an
in the crust represents a potential proxy for concomitant changes in “added-value” project to create a metallogenic-potential GIS of
chemical fluxes and redox states of Earth’s major reservoirs such as Gondwana. During this process the original database was updated
its mantle and fluid envelope. and revalidated. Roughly 6,000 new deposits were added to the
database. In the final stage of the project, this “updated” min-
DERIVING METALLOGENIC FINGERPRINTS OF eral database was integrated with geostatistical software and a
ARCHEAN CRATONS AND YOUNGER CRUST USING browser (GEORAMA). The end product yielded a metallogenic-
SELECTED ELEMENTS FROM THEIR MINERAL potential GIS of Gondwana (available on CD-ROM; for details
DEPOSITS see: http://gondwana.brgm.fr/index_eng.htm). Thus, currently
there are roughly 16,000 deposits in the mineral layer, ranging
from active mines to undeveloped occurrences. Figure 1 shows
The Database the typical density of these mineral deposits across Gondwana,
and from which our data are extracted. The origin and evolution
The geological and mineral deposit data used for this study of the database has been described in detail elsewhere (Wilsher
are incorporated in a GIS relational database, called GO-GEOID et al., 1993; de Wit et al., 1999; Thiart and de Wit, 2000; de Wit
(GOndwana – GEOscientific Indexing Database), housed at our et al., 2004).
Figure 1. Mineral deposit map of Gondwana. Data from the Gondwana mineral deposit GIS database housed at the Africa Earth Observatory
Network (AEON) (de Wit et al., 1999; Thiart and de Wit, 2000; http://www.uct.ac.za/depts/cigces/gondmin.htm).
Figure 2 shows the global distribution of cratons that we Methods

selected for this study. There are 11 Gondwana cratons (seven in
Africa, two in South America, and two in Australia) for which we Previously we defined metallogenic fingerprints of frag-
have sufficient mineral deposit data in our Gondwana database. ments of continental crust through their spatial association with
For general comparison, we also selected one Canadian craton, a combination of six element groups, and applied this at three
the Superior Province. The mineral deposit data for the latter are scales—cratons, continents, and supercontinents (de Wit and
from the Geological Survey of Canada (Kirkham et al., 1994, Thiart, 2005). The measure of spatial association (normalized to
2002; Jenkins et al., 1997; Eckstrand and Good, 2000; Kirkham area) we termed the spatial coefficient, rij, where
and Dunne, 2000, 2002; Eckstrand et al., 2002). We emphasize
that this Canadian-based data structure is significantly different N ( Ci ∩ D j ) / N ( D j )
from our Gondwana database, and because the subset of their rij = , (1)
A(Ci ) / A(C• )
data provided to us may not include all known mineral deposits
in the Superior Province, the results in the subsequent sections
should be evaluated in this light. in which A(Ci) is the area of the ith craton (Ci), and A(C•) is the
Eleven metallogenic elements were selected for our analy- total area of all the cratons in the study (A(C•) = ∑ A(Ci)). N(Dj) is
ses, and these were divided into six element groups according to the total number of deposits in the jth element group (Dj). N(Ci ∩
their geochemical affinities (e.g., lithophile, chalcophile, sidero- Dj) represents the number of deposits of group Dj in craton Ci .
phile), as well as their relative abundances in the database. The The spatial coefficient (equation 1) represents the propor-
six element groups, and the total number of deposits in which tion of deposits (say gold, j) of all the jth deposits (in cratons) that
these groups occur on each craton, are tabulated in Table 1A. In occur in the specified craton (i) per unit area of all 12 cratons. Our
total there are just over 6000 deposits spread over 12 identified spatial coefficient is similar to the measure of spatial association
cratons (Table 1A). used by Mihalasky and Bonham-Carter (2001) to measure the
Figure 2. Cratons examined in this study. 1—Kaapvaal; 2—Limpopo; 3—Zimbabwe; 4—Congo; 5—Tanzania; 6—Leo-Man; 7—Requibath;
8—Superior; 9—Amazonia; 10—São Francisco; 11—Pilbara; 12—Yilgarn.
TABLE 1A. NUMBER OF MINERAL DEPOSITS WITH SELECTED ELEMENTS

ON GONDWANA’S CRATONS AND THE SUPERIOR PROVINCE OF CANADA
Au CrNiPgeTi CuZnPbBa SnSb W UThREE Craton total† Area
(km2)
Superior 162 134 52 3 0 0 351 2.73E + 06
Pilbara 7 6 11 1 0 4 29 1.51E + 05
Yilgarn 97 50 17 1 0 11 176 6.25E + 05
Amazonia 434 38 109 645 8 13 1247 3.48E + 06
São Fra. 435 127 462 149 6 66 1245 9.64E + 05
Kaapvaal 894 241 152 96 5 77 1465 6.60E + 05
Limpopo 1 40 33 0 8 9 91 1.59E + 05
Zimbabwe 625 114 106 81 305 15 1246 3.52E + 05
Congo 53 2 24 1 0 17 97 2.31E + 06
Leo-Man 47 22 11 10 2 9 101 3.20E + 05
Tanzania 7 0 3 1 0 5 16 3.33E + 05
Requibath 0 5 21 0 3 4 33 1.68E + 05
Deposits total 2762 779 1001 988 337 230 6097 1.23E + 07
Note: Area is given in km2 rounded to 2 decimals. The total area of all the cratons is 1.2248155 × 107 km2.
†
Craton Total reflects only the total of those deposits in which the elements selected for this study groups occur as
its major component(s).
TABLE 1B. NATURAL LOG OF THE SPATIAL COEFFICIENT (rij,)

BETWEEN ELEMENT GROUPS AND CRATONS
Au CrNiPgeTi CuZnPbBa SnSb W UThREE Cra tot
Superior –1.33 –0.26 –1.46 –4.30 –1.35
Pilbara –1.58 –0.47 –0.11 –2.50 0.35 –0.95
Yilgarn –0.37 0.23 –1.10 –3.92 –0.06 –0.57
Amazonia –0.59 –1.76 –0.96 0.83 –2.48 –1.61 –0.33
São Fra. 0.69 0.73 1.77 0.65 –1.49 1.29 0.95
Kaapvaal 1.79 1.75 1.04 0.59 –1.29 1.83 1.50
Limpopo –3.58 1.37 0.93 0.60 1.10 0.14
Zimbabwe 2.06 1.63 1.30 1.05 3.45 0.82 1.96
Congo –2.29 –4.30 –2.06 –5.23 –0.94 –2.47
Leo-Man –0.43 0.08 –0.87 –0.95 –1.48 0.40 –0.46
Tanzania –2.37 –2.20 –3.29 –0.22 –2.34
Requibath –0.76 0.42 –0.43 0.24 –0.93
Note: These values are presented in Figures 4 and 5 as solid bars. No entries—no data.
TABLE 1C. WEIGHTED NATURAL LOG OF THE SPATIAL COEFFICIENT( r w , EQUATION 4), BETWEEN
ij
ELEMENT GROUPS AND CRATONS
Au CrNiPgeTi CuZnPbBa SnSb W UThREE Cra tot
Superior –2.10 –0.93 –2.21 –5.02 –2.12
Pilbara –2.31 –1.10 –0.82 –3.18 –0.44 –1.67
Yilgarn –1.10 –0.40 –1.81 –4.60 –0.86 –1.29
Amazonia –0.68 –1.75 –1.03 0.79 –2.79 –1.77 –0.41
São Fra. 0.61 0.73 1.70 0.61 –1.80 1.14 0.87
Kaapvaal 1.75 1.80 1.01 0.59 –1.55 1.72 1.46
Limpopo –3.59 1.46 0.94 0.37 1.03 0.13
Zimbabwe 2.30 1.95 1.55 1.32 3.46 0.99 2.20
Congo –1.72 –3.64 –1.48 –4.62 –0.44 –1.90
Leo-Man 0.19 0.79 –0.23 –0.28 –1.08 0.96 0.17
Tanzania –1.71 –1.52 –2.58 0.38 –1.67
Requibath 0.01 1.12 0.02 0.85 –0.25
Note: These values are presented in Figures 4 and 5 as striped bars.
spatial association between lithodiversity and minerals deposits mineral j and craton i; rij < 1 (e.g., where there are fewer depos-
in Nevada. The value of the spatial coefficient ranges from 0 to its than expected by chance) indicates a negative association.
infinity; it is equal to 1 if there is no spatial association between a Because all negative associations are compressed in the range
craton and an element group (e.g., if the proportion of jth mineral from 0 to 1, and all positive associations fall in the range of 1 to
is the same as the proportion of area occupied by the ith craton). infinity, we use the natural log of rij to eliminate this skewness.
For values of rij > 1 (e.g., where the expected number of deposits Thus, ln(rij) is a symmetric value around 0: positive associations
is greater than by chance), there is a positive association between are greater than 0, and negative associations are less than 0.
The above approach ignores the fact that each study region age of area that each country contributes to the total area of that
(e.g., defined cratons) has a different exploration history, a craton. The craton GNI-C is then calculated as a weighted aver-
bias that might influence the analyses. Intuitively, we expect, age (by area percentage) of all the GNI-C values of each of the
for example, that greater accessibility (e.g., infrastructure) and countries involved. For example the Kaapvaal Craton constitutes
political stability increases exploration and discovery rates and 14% of Botswana, 4% of Lesotho, 80% of South Africa, and 2%
mining activity. Therefore mineralization in some cratonic areas of Swaziland. The GNI-C for the entire Kaapvaal Craton then is
of less-developed nations may not be adequately represented in as follows: 14%*GNI-C for Botswana + 4%*GNI-C for Lesotho
our database and analyses. We address this problem by weighting + 80%*GNI-C for South Africa + 0.2%*GNI-C for Swaziland.
the spatial coefficient (equation 1), with an “exploration index,” To calculate an exploration index, we use the United States
using socioeconomic data from the World Bank development (USA) as a benchmark (e.g., the United States has an exploration
indicators database (World Bank, 2004). This “weighted” spa- index of 1). The exploration index for Cratoni is the ratio of the
tial coefficient (defined below in equation 4) can be used with GNI-C for USA to the GNI-C for Cratoni
increased confidence to evaluate the metal endowments of differ-
k
ent continental fragments. ⎡ GNI − C forUSA ⎤
Rich countries are better explored than poor countries ⎢ ⎥ (2)
⎣ GNI − C for Cratoni ⎦
because of their better-developed infrastructure and investment
regimes. We tested various world development indicators pub-
lished by the World Bank (2004) designed to quantify this dif- where k is a power function. For k = 1 our calculated exploration
ferential development. Here we use the measure of gross national indices range from 1 (for the United States) to 126 (for the poor-
income per capita (GNI-C, for 2002; Table 2) because it contains est craton, the Requibath Craton of northwest Africa). To contain
the most obtainable data for all the countries of our interest. To the index to a more manageable range, we apply various power
derive the GNI-C for a specific craton, we calculate the percent- functions (k from ½ to ¼), or use the natural log (ln (equation
TABLE 2. GNI PER CAPITA (GNI-C) FOR COUNTRIES CONTAINING THE SELECTED CRATONS
(OR PARTS THEREOFF) ACROSS GONDWANA AND CANADA
Country name Craton name Area of craton GNI-C, Atlas Craton Continent GNI-C
(%) method GNI-C
(current US$)
United States 35400 35400
34120
Canada Superior 100 22390 22390
Australia Pilbara 100 19530 19530
19530
Australia Yilgarn 100 19530 19530
Brazil Amazonia 100 2830 2830
3280
Brazil SãoFra. 100 2830 2830
Angola Congo 9 710
Cameroon Congo 4 550
Central African Republic Congo 14 250
Congo Congo 10 100
Equatorial Guinea Congo 1 930
Gabon Congo 6 3060
Sudan Congo 0 370
Uganda Congo 0 240
Zaire Congo 56 100 389
Botswana Kaapvaal 14 3010
Lesotho Kaapvaal 4 550
South Africa Kaapvaal 80 2500
Swaziland Kaapvaal 2 1240 2470
Guinea Leo–Man 40 410
Ivory Coast Leo–Man 14 620
Liberia Leo–Man 25 140
Africa 650
Mali Leo–Man 4 240
Senegal Leo–Man 3 470
Sierra Leone Leo–Man 15 140 330
Botswana Limpopo 58 3010
Mozambique Limpopo 1 200
South Africa Limpopo 13 2500
Zimbabwe Limpopo 28 480 2221
Mauritania Requibath 53 280
Saharawi, ADR Requibath 47 –* 280
Kenya Tanzania 8 360
Tanzania, United Republic of Tanzania 81 290
Uganda Tanzania 11 240 290
Botswana Zimbabwe 23 3010
Mozambique Zimbabwe 1 200
Zambia Zimbabwe 1 340
Zimbabwe Zimbabwe 75 480 1069
Note: Craton GNI-C is the weighted average (by % area). No data is available for the Arab Democratic
Republic of Saharawi; GNI-C for Requibath is therefore based on Mauritania.
2), with k = 1) of the GNI-C ratios. These exploration indices This “weighted” proportion (equation 3) is then used in the cal-
are given in Figure 3A. For this investigation, we chose to work culation of the “weighted” spatial coefficient rijw :
with the power function k = 1/3, as it closely follows the natural
log function and has the advantage that it ranges from 1 (United pijw / p•wj
rijw =
States) to roughly 5 (least explored cratons). A(Ci ) / A(C• ) (4)
For our analyses at a continental scale we derive the GNI-C
ratio for each continent in a similar manner, and use the com-
bined GNI-C of the United States and Canada (North America, where pijw is the weighted proportion of the ith craton and jth ele-
top of the last column of Table 2) as our benchmark. The GNI-C ment group; p•wj is the weighted proportion of the jth element
for Africa is from the African Development Indicators (2004). group (total of the jth element group); A(Ci) is the area of the ith
The GNI-C for North America and Australia are from the World craton; and A(C• ) is the total area of all the cratons. This weighted
Bank development indicators database (World Bank, 2004). The spatial coefficient is treated as before (in this text), in the sense
GNI-C for South America is from the country data of the World that we again take the natural log of rijw . The interpretation will
Bank development indicators database (World Bank, 2004). be the same as in that of the unweighted (natural log) spatial
The derived exploration indices for the continents are given in coefficient: positive values of natural log ( rijw ) is a measure of
Figure 3B. the positive spatial association between the ith craton and the jth
From the original mineral deposit data (Table 1A) we can element group, and negative values of ln(rij) is a measure of the
now calculate a proportion (pij) as the number of deposits in Cra- negative association between the ith craton and jth element group.
toni and Element Groupj divided by the total number of deposits
in the database (6097; Table 1A). This calculated proportion is RESULTS
then multiplied by the exploration index, wi,, where wi is the ratio
defined in equation 2 (with k = 1/3) of the ith craton, and divided Metallogenic Fingerprints of Selected Archean Cratons
by the sum of the product of wi, and pij. Thus
The number of mineral deposits of the six selected element
wi pij groups within each of the 12 cratons is given in Table 1A, and
pijw = . (3) the natural log of the spatial coefficients between the element
Σ Σ wi pij
i j group and each craton (ln(rij), rij defined in equation 1) is given
in Table 1B. The natural log of the “weighted” spatial coefficient
12
A
10
Figure 3. (A) Exploration indices (as a func-
8 tion of the state of development of countries
1/ 2 normalized to that of the USA) determined
Weight
6 1/ 3 for cratons using data from the World Bank

(2004). The Pilbara and Yilgarn Cratons are
7/ 24
4
shown combined under Australia (Aus), and
1/ 4 the Amazonia and São Francisco Craton are
ln shown combined as Brazil (Bra). The Su-
2
perior Province is listed as Canada (Can).
The indices are shown at various power
0 functions and as a (natural) log function
Usa Can Au s Bra Kp v L im Zim Con Lm n Tan Rqb (see text for explanation). Note the similar-
ity between the natural log function and k =
8 1/3. Con—Congo; Kpv—Kaapvaal; Lim—
Limpopo; Lmn—Leo-Man; Rqb—Requi-
B bath; Tan—Tanzania; Zim—Zimbabwe.
6
(B) Exploration index determined for the
1/2
three continents studied, normalized to that
Weight
4 1/ 3 of North America. The index is shown at

ln two power functions (k = ½ and k = 1/3), and
2 as a (natural) log function.
0
North-America Australia South-America Africa
(ln ( rijw ), equation 4) for the 12 selected cratons and the six the Tanzania and Congo Cratons represent the lowest values, and
selected element groups is given in Table 1C. Figures 4 and 5 these values may be unreliable because of the low number of dis-
summarize these data in graphical format to illustrate the mineral covered mineral deposits on these cratons (and thus in our data-
diversity of each craton. For comparison both the unweighted base). In turn this corroborates the need to incorporate into the
(ln(rij), solid bars) and the weighted spatial coefficient (ln( rijw ), analyses an exploration-index value that reflects the past history
striped bars) are given in the same figure. Figure 4 expresses the of (under)exploration of areas under consideration. Of additional
diversity among cratons for each set of elements of the six ele- interest here is that the relative differences between the weighted
ment groups, and provides 12 cratonic (spatial) coefficients for and unweighted results express a degree of underexploration,
each set of elements. Figure 5 represents the total spatial coeffi- and therefore provide a signal for potential near-surface mineral
cient combining all sets of elements (e.g., all six element groups) deposits still to be found in areas where infrastructure develop-
for each craton; these are the “metallogenic fingerprints” of the ment and exploration investment has lagged behind that of North
12 cratons. Each individual craton has a unique fingerprint even America. The greatest positive values are those of the Zimbabwe
though some may appear very similar (e.g., Congo and Tanzania Craton. This reflects a high count of mineral deposits covering all
Cratons). Where values are high (e.g., Zimbabwe) the imprint of six element groups in Zimbabwe and the relatively small size of
total mineralization is high relative to the other cratons: metaphor- this craton. It is the only craton that has no negative spatial coef-
ically a “strong fingerprint.” Values between –1 and 1 represent a ficient in any of the element groups. The similarity between its
random association between the total element set and the specific weighted and unweighted spatial coefficients indicates a mature
craton; values below 0 (Fig. 4) indicate a negative association for degree of exploration throughout this craton, in concert with a
the given element group and the specific craton. The values for long history of intense mineral exploration.
4 4
Gold CrNiPgeTi
2 2
0 0
-2 -2
-4 -4
-6 -6
4 4
CuZnPbB a SnSb
2 2
0 0
-2 -2
-4 -4
-6 -6
4 4
W UThREE
2 2
0 0
-2 -2
-4 -4
-6 -6
Ama
Ama
SaF
SaF
Kp v
Kp v
Sup
Con
Lm n
Rqu
Sup
Con
Lm n
Rqu
Lim
Zim
Lim
Zim
Pil
Yil
Pil
Yil
Tan
Tan
Figure 4. Mineral diversity of Archean cratons. Solid bars represent the “raw” spatial coefficient (ln(rij)); striped bars represent the weighted
spatial coefficient (ln( rijw )) between specific element(s) and the corresponding crust “weighted” with their exploration indices Note how each
craton has a unique diversity of elements that characterize its “metallogenic fingerprint” (see also Fig. 5). The difference between the weighted
and unweighted bars represents an exploration potential. See text for further information.
2
Figure 5. Mineral diversity measure be-
1 tween all combined elements of mineral
deposits of the 12 cratons investigated
0 in this study. The solid bars represent
the unweighted metallogenic finger-
-1 print (spatial coefficient (ln(rij)) and the
striped bars represent the “weighted”
-2 metallogenic fingerprint (ln( rijw )) of
each craton.
-3
Ama
SaF
Kpv
Sup
Con
Tan
Rqu
Lmn
Pil
Lim
Zim
Yil
Some negative spatial coefficients might be intuitively sur- 2.5 Ga) on each individual continent. The mineral deposit data for
prising (e.g., those for gold in the Superior Province). The spatial the old cratons and younger crust are summarized in Table 3A.
coefficient (equation 1) is defined as a ratio between two propor- The natural log of the spatial coefficients is given in Table 3B,
tions, that is, the number of deposits (say Au) in a specified cra- and the natural log of the “weighted” spatial coefficients is given
ton relative to the number of deposits in all the cratons, divided in Table 3C. Figure 6 graphically displays the unweighted natu-
by the proportion of area that the specific craton occupies relative ral log of the spatial coefficients (ln(rij), solid bars) against that of
to that of the area of all the cratons. If the ratio is greater than the natural log of the weighted coefficient (ln( rijw ), striped bars).
1, there is a positive association (more deposits expected than The total mineral inventory for the continental crust older and
by chance), whereas a value between 0 and 1 results in a nega- younger than 2.5 Ga (O and Y, respectively) can now be directly
tive association (fewer deposits expected than by chance). The compared also (Fig. 7). These results clearly depict the mineral
natural log of the spatial coefficient for the exceptionally large deposit diversity between cratons and younger crust; they also
Superior Province is probably negative because it expresses its depict a real difference (per unit area) in enrichment of specific
mineralization relative to other cratons: in the Superior Province elements between old and young crust on all three continents.
there are only 162 Au deposits in the Canadian database com- In general the older fragments are more enriched in all element
pared to a total of 2762 in all cratons combined; and the Supe- groups, except for tungsten in South America, and for the Sn, Sb,
rior province is the largest of all the cratons (2.73 × 106 km2). and UThREE groups in Australia. This result is probably inde-
We suspect that there are probably a lot more small deposits in pendent of the “exposure” problem mentioned above and there-
the Superior Province that are not incorporated into the Cana- fore a strong conclusion.
dian database, because they are in the GO-GEOID database. The The strongest spatial coefficient observed is between tin
results would also change if one could factor in the actual area and antimony (Sn and Sb) and the old crust (cratons) of South
of outcrop (most of the Superior Province is covered by glacial America. A strong spatial coefficient for Sn and Sb is also
drift (till, transported overburden). Geophysical and geochemi- observed for Africa, and may be controlled mostly by the West
cal exploration methods cannot “see” gold deposits through this African cratons that were part of the Amazonian Craton until
cover as easily as they can “see” base metal (VMS) deposits. Gondwana breakup. This supports a long (3.0–0.5 Ga) contermi-
Other cratons with residual overburden would not be affected by nous history for this old crust (Trompette, 1994). We have pre-
this complication. It is probably unwise therefore to compare the viously explored this Gondwana-tin association between West
two data sets directly, because they were constructed differently Africa and South America to refute a frequently advocated fit
and with different specifications. This supports a call to standard- between them and eastern North America (where there are no
ize global databases. For the further analysis below we therefore Archean–Mesoproterozoic tin deposits or occurrences), a rela-
eliminate the Superior Province from our input. tion that is important to models of the proposed Mesoproterozoic
supercontinent Rodinia (de Wit et al., 1999).
Mineral Inventory on a Continental Scale In summary, it seems that the younger crust of the continents
in the Southern Hemisphere is relatively enriched in the litho-
The analyses described above were repeated for three large phile element group (UThREE), whereas the older crust of all
continental fragments (Africa, South America, and Australia) of three continents is affiliated to a greater degree with concentra-
the former supercontinent Gondwana, to enable direct compari- tions of siderophile element group (Cr Ni PGE Ti). In contrast,
son of the metallogenic inventory of cratons (older than 2.5 Ga; there is no discernable degree of difference in concentrations of
old crust) with mineralization in younger crust (younger than chalcophile element group (base metals) between old and young
TABLE 3A. NUMBER OF MINERAL DEPOSITS OF CRATONS (OLD) AND YOUNGER CRUST (YOUNG) IN
THREE CONTINENTS: AFRICA, SOUTH AMERICA, AND AUSTRALIA
Au CrNiPgeTi CuZnPbBa SnSb W UThREE Total Area
(km2)
Afr old 1627 424 350 189 323 136 3049 4.31E + 06
Afr young 499 91 768 142 119 491 2110 2.06E + 07
Sam old 869 165 571 794 14 79 2492 4.44E + 06
Sam young 797 119 999 236 828 671 3650 1.07E + 07
Aus old 104 56 28 2 0 15 205 7.75E + 05
Aus young 70 2 15 22 5 157 271 6.01E + 06
Deposit total 3966 857 2731 1385 1289 1549 11777 4.68E + 07
Note: Abbreviations: Afr—Africa, Sam—South America, and Aus—Australia. The total area of younger crust =
7 6 2
3.73 × 10 , and that of the cratons = 9.52 × 10 km .
TABLE 3B. NATURAL LOG OF THE SPATIAL COEFFICIENT (rij) OF ELEMENT GROUPS OF OLD CRATONS
AND YOUNGER CRUST OF THREE ONTINENTS: AFRICA, SOUTH AMERICA, AND AUSTRALIA
Au CrNiPgeTi CuZnPbBa SnSb W UThREE Total
Afr old 1.50 1.68 0.33 0.40 1.00 –0.05 1.04
Afr young –1.25 –1.42 –0.45 –1.46 –1.56 –0.33 –0.90
Sam old 0.84 0.71 0.79 1.80 –2.17 –0.62 0.80
Sam young –0.13 –0.50 0.47 –0.29 1.03 0.64 0.30
Aus old 0.46 1.37 –0.48 –2.44 –0.54 0.05
Aus young –1.98 –4.01 –3.15 –2.09 –3.50 –0.24 –1.72
Note: Abbreviations: Afr—Africa, Sam—South America, and Aus—Australia. These values are presented in
Figures 6 and 7 as solid bars.
TABLE 3C. WEIGHTED NATURAL LOG OF THE SPATIAL COEFFICIENT ( rijw ) OF ELEMENT GROUPS AND
CRATONS (OLD) AS WELL AS YOUNGER CRUST (YOUNG) OF THREE CONTINENTS
Afr old 1.73 1.89 0.62 0.78 1.32 0.28 1.32
†
Afr young –1.02 –1.22 –0.16 –1.07 –1.24 0.00 –0.62
Sam old 0.53 0.37 0.54 1.65 –2.38 –0.83 0.54
Sam young –0.44 –0.83 0.22 –0.45 0.82 0.42 0.05
Aus old –0.44 0.44 –1.32 –3.18 –1.35 –0.80
Aus young –2.89 –4.94 –4.00 –2.84 –4.31 –1.05 –2.57
Note: Abbreviations: Afr—Africa, Sam—South America, and Aus—Australia. These values are presented in
Figures 6 and 7 as striped bars.
†
Exact value is –0.0026845.
crust. Our integrated results also support general statements that The data are summarized in Tables 4A to 4B and plotted in
old crust is more richly endowed with mineral deposits than Figure 8. From this it is clear that Archean crust is indeed miner-
young crust (Fig. 7). alized to a significantly greater degree than younger crust (except
for the strongly lithophile element group UThREE), thus hint-
Metallogenesis on a Gondwana Scale ing at the possibility that young crust may have inherited at least
some metal enrichment during remobilization of its embedded
The above analyses were repeated using all the data of the cratons (e.g., tin in South America, gold and PGE in Africa).
cratons and the three continents combined. This allows us to
compare and contrast the mineral inventory of all old (Archean) DISCUSSION AND CONCLUSIONS
crust with that of younger crust of the three continents combined
as one, as they would have been in Gondwana times (ca. 200– By applying an “exploration index” to our existing mineral
500 Ma). To some degree this corrects a bias in the analyses by deposit data, we have attempted to compare the mineral inventory
concentrations of recent mineral deposits related to one specific of different cratons across continents with variable exploration
present-day plate tectonic environment (such as subduction below histories. Although our exploration index needs more rigorous
South America) that may skew the results and interpretations testing, incorporating this type of data allows a more informed
when using a single continent only. On this subglobal continen- and robust comparative analysis between mineral riches of dif-
tal scale we combine all old crust (cratons) across all combined ferent cratons. The results confirm our earlier work (de Wit and
continents, but we apply no “weighting” to allow for differen- Thiart, 2005) that, per unit area of crust, there is a greater concen-
tial development and exploration; both the relatively “rich” and tration of mineral deposits in Archean cratons relative to younger
“poor” of the Southern Hemisphere are dealt with collectively. crust. Although the mineral inventory is greater in cratonic crust
2 2
Gold CrNiPGE
0 0
-2 -2
-4 -4
-6 -6
2 2
CuZnPbBa SnSb
0 0
-2 -2
-4 -4
-6 -6
2 2
W UThREE
0 0
-2 -2
-4 -4
-6 -6
Af r O Afr Y Sam O Sam Y Au s O Au s Y Af r O Afr Y Sam O Sam Y Au s O Au s Y
Figure 6. Mineral inventory of three separate continents (Aus—Australia; Afr—Africa; Sam—South America) showing the spatial coefficients
(ln(rij)) between their groups of selected elements and two ages of continental crust (older than 2.5 Ga [O] and younger than 2.5 Ga [Y]), The
solid bars represent their unweighted spatial coefficient (ln(rij)), and the striped bars represent their “weighted” spatial coefficients (ln( rijw )).
-2
-4
Afr O Af r Y Sam O Sam Y Au s O Aus Y
Figure 7. Total mineral diversity between old cratons (O) and younger crust (Y) for three separate continents (Afr—Africa, Sam—South Amer-
ica, and Aus—Australia). The solid bars represent their unweighted metallogenic fingerprint (spatial coefficient (ln(rij)), and the striped bars
represent their “weighted” metallogenic fingerprint (ln( rijw )). Note that, on each continent, the mineral inventory of its cratons is in general
greater than that of its younger surrounding crust. This suggests a decrease in mineral diversity (compare with Fig. 6) and mass per unit area in
Earth’s crust with time. Exceptions to this are tungsten and the lithophile elements U/Th/REE in South America, and possibly tin and antimony
in Australia, but data from cratonic Australia are insufficient for more robust analysis.
TABLE 4A. NUMBER OF MINERAL DEPOSITS OF SELECTED ELEMENT GROUPS

ON ALL CRATONS (OLD) AND YOUNGER CRUST (YOUNG)
OF THREE CONTINENTS COMBINED AS A SUPER-CONTINENT (GONDWANA)
Au CrNiPgeTi CuZnPbBa SnSb W UThREE Total Area
(km2)
Old 2600 645 949 985 337 230 5746 9.52E + 06
Young 1366 212 1782 400 952 1319 6031 3.73E + 07
Total 3966 857 2731 1385 1289 1549 11777 4.68E + 07
TABLE 4B. NATURAL LOG OF THE SPATIAL COEFFICIENT ( r ) BETWEEN ELEMENT GROUPS
ij
AND ALL CRATONS (OLD) AS WELL AS YOUNGER CRUST (YOUNG)
OF THREE CONTINENTS COMBINED AS A SUPER-CONTINENT (GONDWANA)
Old 1.17 1.31 0.54 1.25 0.25 –0.31 0.88
Young –0.84 –1.17 –0.20 –1.01 –0.08 0.07 –0.44
Note: These values are presented in Figure 8.
Crustal Age for some of these cratons on the basis of geologic and geophysi-
cal evidence (e.g., Herrington et al., 1997; Barley et al., 1998).
> 2.5 Ga < 2.5 Ga Cumulative evidence indicates that cratonic crust is more
1.5 enriched in mineral deposits than younger crust (younger than
2.5 Ga). The greater concentration of this Archean mineralization
1.0 may represent more efficient mineralization processes, perhaps
Au
related to higher heat and/or volatile loss from the early Earth
CrNiPgeTi
0.5 compared to today (Abbott et al., 1994; de Wit and Hart, 1993;
CuZnPbBa
de Wit and Hynes, 1995; Pollack, 1997; de Wit, 1998), in which
SnSb
0.0 case our results may be interpreted to reflect greater “partition
W
coefficients” of selected elements between cratonic crust and
-0.5 UThREE (now depleted) mantle during the formation of Archean litho-
Total sphere (Fig. 9).
-1.0 However, we cannot rule out the possibility that the higher
concentration of Archean mineralization represents a greater pres-
-1.5 ervation potential of cratons (and their mineral deposits) relative
to younger continents, as the presence of (Archean) diamonds in
Figure 8. Relative concentration in Archean cratons and younger
cratons might imply. In this case, the greater mineral wealth of
Gondwana crust of elements in all six element groups analyzed in this
study. In all cases (except U/Th/REE) total concentrations are greater cratons may be merely a consequence of greater rates of recy-
in the Archean cratons than in the younger crust of Gondwana. This cling of young continental crust relative to that of old Archean
suggests a decrease in the mass of new mineral deposits per unit area crust preserved in cratons (Fig. 9). Because the majority of the
of crust over time. cratonic crust is preserved at low grades of metamorphism (and
is thus representative of the upper crust) this seems an equally
valid interpretation.
than younger crust, we also confirm that a significant mineral If this interpretation has merit, it implies significant changes
diversity exists among cratons, and that each craton has a unique in the efficiency of recycling of young continental crust since
metallogenic fingerprint. These differences resemble variations the Archean. Because subduction is the principal mechanism
of Phanerozoic mineralization on continents at scales that clearly by which the hydrosphere recycles into the mantle to generate
link mineral deposits to different plate tectonic environments continental crust enriched in mineral deposits, secular change in
(e.g., oceanic arcs, continental subduction zones; Sawkins, 1990; crustal mineralization could be used to track secular changes in
Windley, 1995). If we assume a similar origin for the processes the chemical composition of Earth’s fluid envelope and the redox
of mineralization in Archean times, this might provide a strong state of its upper mantle.
basis of support for models that advocate that plate tectonics Either way, Earth’s crust appears to signal a decrease in new
operated on the Archean Earth. For example, many of the Neo- mineral deposit mass and diversity through time. But, as a word
archean cratons have strong Au and base-metal signatures that fit of caution, our results are based on only one mineral database tied
with mineralization of subduction-accretion models as proposed to only one specific exploration index. In addition only two rela-
Barley, M.E., Krapez, B., Groves, D.I., and Kerrich, R., 1998, The Archean
bonanza: metallogenic and environmental consequences of the interaction
between mantle plumes, lithospheric tectonics and global cyclicity: Precam-
brian Research, v. 91, p. 65–90, doi: 10.1016/S0301-9268(98)00039-4.
Cawthorn, R.G., and Walraven, F., 1998, Emplacement and crystallization time
for the Bushveld Complex: Journal of Petrology, v. 39, p. 1669–1687,
doi: 10.1093/petrology/39.9.1669.
de Wit, M.J., 1998, On Archean granites, greenstones, cratons and tectonics:
Does the evidence demand a verdict?: Precambrian Research, v. 91,
p. 181–226, doi: 10.1016/S0301-9268(98)00043-6.
de Wit, M.J., and Hart, R.A., 1993, Earth’s earliest continental lithosphere,
hydrothermal flux, and crustal recycling: Lithos, v. 30, p. 309–336, doi:
10.1016/0024-4937(93)90043-C.
de Wit, M.J., and Hynes, A., 1995, The onset of interaction between the hydro-
sphere and oceanic crust, and the origin of the first continental lithosphere,
in Coward, M., and Ries, A., eds., Early Precambrian processes: Geologi-
cal Society [London] Special Publication 95, p. 1–9.
de Wit, M.J., and Thiart, C., 2005, Metallogenic fingerprints of Archean Cra-
tons, in MacDonald, I., Boyce, A.J., Butler, I.B., and Herrington, R.J.,
eds., Mineral Deposits and Earth Evolution: Geological Society [London]
Special Publication 248, p. 59–70.
de Wit, M.J., Jeffery, M., Bergh, H., and Nicolaysen, L., 1988, Geological
map of sectors of Gondwana reconstructed to their disposition ~150 Ma,
scale 1:10.000.000: Tulsa, Oklahoma, American Association Petroleum
Geologists.
Figure 9. Summary of the main findings of this study as based on the de Wit, M.J., Thiart, C., Doucoure, C.M., and Wilsher, W., 1999, Scent of a
mineral deposit data analyzed in this study. The figures on top sche- supercontinent: Gondwana’s ores as chemical tracers—Tin, tungsten and
matically illustrate the present Earth with its average-size plates (left), the Neoproterozoic Laurentia- Gondwana connection: Journal of African
and the African plate with its Archean cratons (black; right); the lower Earth Sciences, v. 28, p. 35–51, doi: 10.1016/S0899-5362(98)00085-2.
figure shows the possible average size of Archean plates with embedded de Wit, M.J., Thiart, C., Doucouré, C.M., Milesi, J.P., Billa, M., Braux, C., and
cratons (black), as based on heat flow arguments (modified from Pol- Nicol, N., 2004, The “Gondwana Metal-Potential” GIS, a geological and
lack, 1997, and de Wit and Hart, 1993). Relative to young crust, Archean metallogenic synthesis of the Gondwana supercontinent at 1:10 million
cratons are mineral diversity hotspots, and these mineral hotspots are scale: Developed and published by CIGCES (South Africa) and BRGM
(France). (http://gondwana.brgm.fr/index_eng.htm)
diverse, so that each craton retains a unique metallogenic fingerprint. It Eckstrand, O.R., and Good, D.J., compilers, 2000, World distribution of nickel
is not possible at this stage to distinguish with certainty if the mineral deposits: Geological Survey of Canada, Open-File Report 3791a, 3 dis-
wealth and diversity of cratons reflects a decrease in transfer of the se- kettes.
lected elements from the mantle to the continental lithosphere from the Eckstrand, O.R., Good, D.J. and Gall, Q., 2002, Ni-PGE-Cr deposits: World
Archean onward (input), or that the return (output; recycling through Minerals Geoscience Database Project, Mineral Resources Division,
erosion and subduction) of continental mineral deposits back into the Geological Survey of Canada, unpublished database under revision.
mantle has become more efficient since Archean times. Eglington, B., and Armstrong, R.A., 2004, The Kaapvaal craton and adjacent
orogens, southern Africa: A geochronology database and overview of the
development of the craton: South African Journal of Geology, v. 107,
p. 13–32, doi: 10.2113/107.1-2.13.
Frimmel, H.E., and Minter, W.E.L., 2002, Recent developments concerning the
tively long time spans are compared (e.g., Archean and younger). geological history and genesis of the Witwatersrand gold deposits, South
Africa: Society of Economic Geologists Special Publication 9, p. 17–45.
It would be wise to explore further, using a different databases Fouch M.J., James, D.E., VanDecar, J.C., van der Lee, S., and the Kaapvaal
and development indices, as well as with greater time resolution, seismic group, 2004, Mantle seismic structure beneath the Kaapvaal and
before our general conclusions can be established beyond reason- Zimbabwe Cratons: South African: Journal of Geology, v. 107, p. 33–44.
Groves, D.I., Ho, S.E., Rock, N.M.S., Barley, M.E., and Muggeridge, M.T.,
able doubt. 1987, Archean cratons, diamonds and platinum: Evidence for long lived
crust mantle systems: Geology, v. 15, p. 801–805, doi: 10.1130/0091-
ACKNOWLEDGMENTS 7613(1987)15<801:ACDAPE>2.0.CO;2.
Hamilton, W.B., 1998, Archean magmatism and deformation were not products
of plate tectonics: Precambrian Research, v. 91, p. 143–181, doi: 10.1016/
We are grateful to Graham Bonham-Carter and Fritz Agter- S0301-9268(98)00042-4.
berg for their interest and encouragement in our work. Construc- Hart, J.H., de Wit, M.J., and Tredoux, M., 1997, Refractory trace elements
in diamonds: Further clues to the origin of the ancient cratons: Geol-
tive discourse on an early draft by Keith Long, Stephen Kes- ogy, v. 25, p. 1143–1146, doi: 10.1130/0091-7613(1997)025<1143:
ler, and an anonymous reviewer helped to clarify some of our RTEIDI>2.3.CO;2.
thoughts and improve this contribution significantly. This work is Hart, R.J., McDonald, I., Tredoux, M., de Wit, M.J., Carlson, R.W., Andreoli,
M., Moser, D.E., and Ashwal, L.D., 2004, New PGE and Re/Os isotope
supported through funds of the South African National Research data from lower crustal sections of the Vredefort dome and a reinterpre-
Foundation (NRF). This is AEON contribution number 008. tation of its “crust on edge” profile: South African Journal of Geology,
v. 107, p. 173–184, doi: 10.2113/107.1-2.173.
Herrington, R.J., Evans, D.M., and Buchanan, D.L., 1997, Metallogenic
REFERENCES CITED aspects, in de Wit, M.J., and Ashwal L.D., eds., Greenstone belts: Oxford,
UK, Oxford University Press, p. 176–220.
Abbott, D.H., Burgess, L., Linghi, J., and Smith, W.H.F., 1994, An empiri- Hutchinson, R.W., 1981, Mineral deposits as guides to supracrustal evolution, in
cal thermal history of the Earth’s upper mantle: Journal of Geophysical O’Connell, R.J., and Fyfe, W.S., eds., Evolution of the earth: Washington
Research, v. 99, p. 13835–13850, doi: 10.1029/94JB00112. D.C, American Geophysical Union, Geodynamics Series, v. 5, p. 120–140.
Hutchinson, R.W., 1992, Mineral deposits and metallogeny: Indicators of Earth Shirey, S.B., Harris, J.W., Richardson, S.H., Fouch, M.J., James, D.E., Car-
evolution, in Schidlowsky, M., Golubic, S., and Kimberley, M.M., eds., tigny, P., Deines, P., and Viljoen, F., 2002, Diamond genesis, seismic
Early organic evolution: Implications for mineral and energy resources: structure, and the evolution of the Kaapvaal-Zimbabwe Craton: Science,
Berlin, Springer-Verlag, p. 551–545. v. 297, p. 1683–1686.
James, D.E., Fouch, M., VanDecar, J., and van der Laan, S., 2001, Tectospheric Stankiewicz, J., Chevrot, S., van der Hilst, R.D., and de Wit, M.J., 2002,
structure beneath southern Africa: Geophysical Research Letters, v. 28, Crustal thickness, discontinuity depth and upper mantle structure beneath
p. 2485–2488, doi: 10.1029/2000GL012578. southern Africa: Constraints from body wave conversions: Physics of the
Jelsma, H.A., de Wit, M.J., Thiart, C., Skinner, E.M., Dirks, P.H., Viola, G., Earth and Planetary Interiors, v. 130, p. 235–251, doi: 10.1016/S0031-
Basson, I.J., and Anckar, E., 2004, Preferential distribution along trans- 9201(02)00012-2.
continental corridors of kimberlites and related rocks of southern Africa: Thiart, C., and de Wit, M.J., 2000, Linking spatial statistics to GIS: Exploring
South African Journal of Geology, v. 107, p. 301–304, doi: 10.2113/107.1- potential gold and tin models of Africa: South African Journal of Geology,
2.301. v. 103, p. 215–230, doi: 10.2113/1030215.
Jenkins, C.L., Vincent, R., Robert, F., Poulsen, K.H., Garson, D.F., and Blondé, Tredoux, M., de Wit, M.J., Hart, R.J., Armstrong, R.A., Lindsay, N., and
J.A., 1997, Index-level database for lode gold deposits of the world: Geo- Sellschop, J.P.F., 1989, Platinum group elements in a 3.5 Ga nickel-iron
logical Survey of Canada, Open-File Report 3490, diskette. occurrence: Possible evidence of a deep mantle origin: Journal of Geo-
Kirkham, R.V., and Dunne, K.P.E., 2000, World distribution of porphyry, por- physical Research, v. 94, B1, p. 795–813.
phyry-associated skarn, and bulk-tonnage epithermal deposits and occur- Trompette, R., 1994, Geology of western Gondwana (2000–500 Ma): Rotter-
rences: Geological Survey of Canada, Open-File Report 3792a, diskette. dam, Balkema, 350 p.
Kirkham, R.V., and Dunne, K.P.E., 2002, Sediment-hosted copper deposits: van Kranendonk, M.J., Collins, W.J., Hickman, A., and Pawley, M.J., 2004,
World Minerals Geoscience Database Project, Mineral Resources Divi- Critical tests of vertical vs. horizontal tectonic models for the Archean
sion, Geological Survey of Canada, unpublished database under revision. East Pilbara granite-greenstone terrane, Pilbara craton, Western
Kirkham, R.V., Carrière, J.J., Laramée, R.M., and Garson, D.F., compilers, Australia: Precambrian Research, v. 131, p. 173–211, doi: 10.1016/
1994, Global distribution of sediment-hosted stratiform copper deposits j.precamres.2003.12.015.
and occurrences: Geological Survey of Canada, Open-File Report 2915b, Wilsher, W., 1995, The distribution of selected mineral deposits across Gond-
map, report, and diskette. wana with geodynamic implications [Ph.D. thesis]: Rondebosch, Univer-
Kirkham, R.V., Carrière, J.J., Rafer, A., and Born, P., 2002, Sediment-hosted sity of Cape Town, 258 p.
copper deposits: World Minerals Geoscience Database Project, Mineral Wilsher, W., Herbert, R., Wullschleger, N., Naicker, I., Vitali, E., and de Wit,
Resources Division, Geological Survey of Canada, unpublished database M.J., 1993, Towards intelligent spatial computing for the earth sciences in
under revision. South Africa: South African Journal of Science, v. 89, p. 315–323.
McDonald, I., de Wit, M.J., Smith, C.B., Bizzi, L., and Viljoen, K.S., 1995, The Windley, B.F., 1995, The evolving continents (3rd edition): New York, John
geochemistry of the platinum-group elements in Brazilian and Southern Wiley and Sons, 385 p.
African kimberlites: Geochimica et Cosmochimica Acta, v. 59, p. 2883– World Bank, 2004a, World development indicators—Online database: (http://
2903, doi: 10.1016/0016-7037(95)00183-2. www.worldbank.org/data/onlinedatabases/onlinedatabases.html and
Mihalasky, M.J., and Bonham-Carter, G.F., 2001, Lithodiversity and its spatial http://www.worldbank.org/data/countrydata/countrydata.html).
association with metallic mineral sites, Great Basin of Nevada: Natural World Bank, 2004b, African development indicators [drawn from the World
Resources Research, v. 10, p. 209–226, doi: 10.1023/A:1012569225111. Bank Africa Database, published March 2004]: ISBN: 0-8213-5720-4
Phillips, G.N., and Evans, K.A., 2004, Role of CO2 in the formation of gold SKU: 15720, 424 p. (http://www4.worldbank.org/afr/stats/adi2004/
deposits: Nature, v. 429, p. 860–862, doi: 10.1038/nature02644. default.cfm).
Pollack, H.N., 1997, Thermal characteristics of the Archean, in de Wit, M.J., Zegers, T.E., and van Keken, P.E., 2001, Middle Archean continent formation
and Ashwal L.D., eds., Greenstone belts: Oxford, UK, Oxford University by crustal delamination: Geology, v. 29, p. 1083–1086, doi: 10.1130/0091-
Press, p. 223–232. 7613(2001)029<1083:MACFBC>2.0.CO;2.
Sawkins, F.J., 1990, Mineral deposits in relation to plate tectonics (2nd and
enlarged edition): Berlin, Springer-Verlag, 461 p. MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA


Memoir 198
2006
Discovery of the oldest oxic granitoids in the Kaapvaal Craton and its
implications for the redox evolution of early Earth
Shunso Ishihara
Geological Survey of Japan, Tsukuba, 305-8567, Japan
Hiroshi Ohmoto
Penn State Astrobiology Research Center of the NASA Astrobiology Institute and the Department of Geosciences,
Pennsylvania State University, University Park, Pennsylvania 16802, USA
Carl R. Anhaeusser
Economic Geology Research Institute, University of the Witwatersrand, Johannesburg, South Africa
Akira Imai
Kyushu University, Fukuoka, 812-8581, Japan
Laurence J. Robb
Economic Geology Research Institute, University of the Witwatersrand, Johannesburg, South Africa
ABSTRACT
Phanerozoic granitoids have been classified into magnetite and ilmenite series
based on the abundance of magnetite, which is related to the Fe2O3/FeO ratio of the
rock and the oxygen fugacity ( fO ) of its parent magma. We have examined the tempo-
2
ral and spatial distributions of both series in Archean granitoids from the Barberton
region and the Johannesburg Dome of the Kaapvaal Craton, South Africa. The old-
est syntectonic TTG (tonalite-trondhjemite-granodiorite) granitoids (ca. 3450 Ma in
age) were found to be ilmenite series, whereas some intermediate-series granitoids
occurred locally. Younger and larger syntectonic TTGs (e.g., the 3230 Ma Kaap Val-
ley plutons) comprise nearly equal quantities of magnetite and ilmenite series. The
major 3105 Ma calc-alkaline batholiths (e.g., Nelspruit batholith), emplaced during
the late-tectonic stage, comprise mostly magnetite-series granitoids, suggesting that
an oxidized continental crust already existed by this time.
The rare earth element ratios and δ18O values, as well as the Fe2O3/FeO ratios,
of the Archean magnetite-series granitoids suggest that their magmas were gener-
ated from the partial melting of subducted oceanic basalts that had been oxidized by
interaction with seawater on mid-oceanic ridges; the processes of magma generation
were much like those for Phanerozoic magnetite-series granitoids. This further sug-
gests that the concentrations of oxidants (O2 and/or SO42–) in the Archean oceans were
similar to those in Phanerozoic oceans.
Low concentrations of chlorine in the magmas, as well as deep levels of granite
erosion, appear to explain the absence of major mineral deposits associated with the
Kaapvaal granitoids.
Keywords: Archean, ilmenite series, magnetite series, granitoids, Kaapvaal.

Ishihara, S., Ohmoto, H., Anhaeusser, C.R., Imai, A., and Robb, L.J., 2006, Discovery of the oldest oxic granitoids in the Kaapvaal Craton and its implications for
the redox evolution of early Earth, in Kesler, S.E., and Ohmoto, H., eds., Evolution of Early Earth’s Atmosphere, Hydrosphere, and Biosphere—Constraints from
Ore Deposits: Geological Society of America Memoir 198, p. 67–80, doi: 10.1130/2006.1198(04). For permission to copy, contact [email protected]. ©2006
Geological Society of America. All rights reserved.
67
68 S. Ishihara et al.
INTRODUCTION 1/T x 103 (K)

1.5 1.0
-5
Granitoids represent an integrated component of the conti-
Po S
+
M
nental lithosphere, hydrosphere, oceanic crust, and upper mantle,
H
M
which were amalgamated by dynamic plate motion. Ishihara Magnetite series
(1977) has suggested that Phanerozoic granitoids can be classi- -10
QM
fied into two types, magnetite series and ilmenite series, based
CH O2 F
on the presence or absence of magnetite, which is determined by Porphyry
C
Cu + Au deposits
4
microscopic observation and modal analyses. However, when the
magnetite content of a rock is very low, microscopic observation -15
or modal analysis is difficult; in these situations, magnetite can be

easily detected by a magnetic susceptibility measurement, even
in the field (Ishihara et al., 2000). Therefore, Ishihara (1977) has its
Log fO2 (atm)

+
S
pos
-20
proposed a magnetic susceptibility value of 100 × 10−6 emu/g (or H y de Ilmenite series
2
P o
SO
3 × 10−3 SI unit) at SiO2 ≈70% as the boundary between ilmenite M
s
and magnetite series. Magnetic susceptibility of ilmenite-series S sit
H2
epo
granitoids can be as low as ~0.01 × 10−3 SI, and that of magnetite- W
d
-25
series granitoids can be as high as ~100 × 10−3 SI. Ishihara (1981)
-
4
SO
has also found that the Fe2O3/FeO ratios of granitoids generally
H
increase with increasing values of magnetic susceptibility, from
+M
less than 0.5 in weight ratio (or 0.22 in mole ratio) for ilmenite
Po
Py
-30
series to greater than 0.5 in magnetite series.
The magnetite- versus ilmenite-series granitoid classifica- Sn deposits
tion is primarily based on the oxidation state of magma: mag-

netite series indicate higher fO2 values and ilmenite series denote -35
300 400 500 600 700 800 900 1100
lower fO2 values at a given temperature and pressure condition.
Temperature (˚C)
Ishihara (1977) and Czamanske et al. (1981) suggest that the
fO2 boundary for the two series probably lies near the NNO Figure 1. Genetic conditions for magnetite- and ilmenite-series
(nickel+nickel oxide) buffer line (Fig. 1). granitoid magmas and depositional conditions for porphyry Cu-Au,
Ishihara (1977) has also recognized that Cu-Au and Cu-Mo porphyry Mo, porphyry Sn, porphyry W, and hydrothermal U de-
porphyry deposits are typically associated with magnetite-series posits (modified after Ohmoto and Goldhaber, 1997). The solid lines
granitoids, whereas W and Sn deposits are found with ilmenite- show fO -T conditions for the following well-known mineral buffers:
2
series granitoids. Such associations of granitoid and ore deposit quartz+fayalite+magnetite (QFM) and magnetite+hematite (MH); the
types occur because the redox state of magma strongly influences broken lines indicate those for mSO /mH S = 1 and mCO /mCH = 1. Total
2 2 2 4
the sulfur chemistry (e.g., SO2, H2S, and SO42–), as well as metal fluid pressure = 1 kbar. (The nickel+nickel oxide buffer line [NNO]
chemistry (e.g., Fe, Cu, Mo, Au, W, and Sn), of magmatic fluids lies approximately halfway between the QFM and HM lines).
(Burnham and Ohmoto, 1980).
Distinct differences also exist between magnetite- and
ilmenite-series granitoids in various geochemical parameters Compared with Phanerozoic granitoids, very little is known
(e.g., O and S isotope ratios; Ishihara et al., 2000; Ishihara and about the petrochemistry and genesis of Archean granitoids. No
Matsuhisa, 2002; Sasaki and Ishihara, 1979; Ishihara and Sasaki, study has evaluated the redox state of Archean granitoids. The main
2002), as well as in their Fe2O3/FeO ratio (Ishihara, 1977, 2004). objective of this study is, therefore, to determine whether magnetite-
Such data have been used to suggest that magnetite-series granit- and/or ilmenite-series granitoids formed during the Archean. The
oids were generated in subduction zones from the partial melting answer will provide important constraints on the chemical evolu-
of hydrated oceanic crust with an Fe2O3/FeO ratio that had been tion of the mantle, crust, oceans, and atmosphere. We have pursued
increased by reactions with O2- and SO42–-rich seawater at mid- this objective through magnetic, mineralogical, and geochemical
oceanic ridges, whereas ilmenite-series granitoids were gener- investigations of major granite batholiths and plutons (older than
ated from the partial melting of normal mantle amphibolites with 3150 Ma) in the Kaapvaal Craton, South Africa.
some contributions of metasediments from the continental crust
(e.g., shales, graywackes) (Burnham and Ohmoto, 1980; Ishihara FERRIC/FERROUS RATIOS AND OXIDATION STATE
and Matsuhisa, 1999). Note that by seawater-rock interactions, OF GRANITOID MAGMAS (THEORETICAL)
the Fe2O3/FeO (weight) ratios of oceanic basalts have typically
increased from ~0.08 (normal mantle value) to ~0.1 – ~1.0 (aver- Because our study attempts to relate the magnetic suscep-
age = 0.31) (e.g., Lécuyer and Ricard, 1999). tibility and Fe2O3/FeO ratios of granitoids to the oxidation state
Discovery of the oldest oxic granitoids in the Kaapvaal Craton 69
of magmas, it is necessary first to evaluate the relationships 125

among the magnetic susceptibility, relative abundances of Fe- (mt + hm)
(ferrous min + mt) hm
% Fe-bearing minerals
bearing minerals, Fe2O3/FeO ratio of granitoids, and the fugac- 100 no
mt only
ity of oxygen (fO2) in the magmas. ferrous magnetite hematite
75 minerals
Magnetic Susceptibility and Fe2O3/FeO Ratio of Granitoids
50
The magnetic susceptibility of granitoids, ranges from 0.01
× 10−3 to 100 × 10−3 SI and is primarily a measure of the amount
of magnetite present in the rock, because other magmatic mag- 25
netic minerals (e.g., pyrrhotite, ilmenite) are typically much less
abundant than magnetite. Magnetic susceptibility of magnetite is 0
-3 -2 -1 0 1 2 3
5 × 10−4m3kg1 (Thompson and Oldfield, 1986), and that of pyr-
Log(Fe2O3/FeO)mol
rhotite and ilmenite is known to be lower by two orders of mag-
nitude. Mafic silicates, common in granitoids such as hornblende
Figure 2. Relationship between the Fe2O3/FeO ratios and the relative
and biotite, are lower by three orders of magnitude in magnetic abundances of magnetite and ferrous-rich silicates in granitoids.
susceptibility, but higher by one order of magnitude in modal
abundance, than magnetite. Therefore, the measured magnetic
susceptibility represents a combination of these mafic minerals,
M = k Xmt ∑Fe (3)
but heavily depending on the magnetite contents.
Measurement of the magnetic susceptibility indicates
that the fO2 conditions of all (or most) granitoid magmas where k is a coefficient that relates magnetic susceptibility to the
should fall between the quartz+fayalite+magnetite (QFM) and amount of magnetite in a rock. Because the ∑Fe content gener-
magnetite+hematite (MH) buffer lines (Fig. 1). Oxygen fugac- ally decreases with increasing SiO2 content, the magnetic suscep-
ity conditions above the MH buffer line are unlikely because tibility of rocks with a given Fe2O3/FeO ratio generally decreases
hematite+ferrous silicates are a non-equilibrium assemblage. We with increasing SiO2 content.
ma y further assume that in granitoids (i) the ferrous compo-
nent (Fe2+ or FeO) mostly resides in ferrous-rich silicates (mostly Relationships between the Fe2O3/FeO Ratio and fO2 of
hornblende and biotite, rarely pyroxene) and magnetite, although Magma
some may be found in pyrrhotite and ilmenite; and (ii) the ferric
component (Fe3+ or Fe2O3) resides largely in magnetite, although For silicate melts that are undersaturated with [Fe2O3] (i.e.,
some may exist in biotite (Czamanske et al., 1981). Certain gran- aFe2O3 < 1), the redox equilibria between ferrous and ferric
itoids contain appreciable amounts of hematite that formed by oxides may be written as
subsolidus reactions with meteoric water (e.g., Taylor, 1968).
Thus, the mole fractions of Fe in silicates and magnetite (Fe2+)melt + H+ + 1/4O2 = (Fe3+)melt + 1/2H2O. (4)
(Xsil and Xmt, respectively) with respect to the total number of
moles of Fe in a rock can be expressed as For reaction (4), the equilibrium fO2 value is expressed as
Xsil + Xmt = 1. (1) logfO = 4log(XFe3+/XFe2+)melt + 4log(γFe3+/γFe2+)melt – 4logK

2
– 4logaΗ+ + 2logfH O (5)
2
Because magnetite (Fe3O4) has one FeO and one Fe2O3 com-
ponent, the Fe2O3/FeO mole ratio of a rock can be related to Xmt as where fi, ai, Xi, and γI are, respectively, the fugacity, activity, mole
fraction, and activity coefficient of component i, and K is the
(Fe2O3/FeO)mol = Xmt/(3 – 2Xmt). (2) equilibrium constant for reaction (4). Equation (5) may be sim-
plified to
This positive relationship between the Fe2O3/FeO ratio and
Xmt in rocks is shown in Figure 2. It illustrates that the maximum logfO = 4log(XFe3+/XFe2+)rock + C (5′)
2
Fe2O3/FeO ratio is 1 in mole ratio (or 2.2 in weight ratio) when
all the Fe atoms in rocks are in magnetite. If the Fe2O3/FeO ratio where
exceeds this value, it suggests that a significant amount of sec-
ondary hematite crystals occur in the granite. C = 4log(γFe3+/γFe2+)rock – 4logK – 4logaΗ+ + 2logfH O. (6)
2
Because the amount of magnetite in a rock is related to both
the Xmt and ∑Fe content of the rock, the magnetic susceptibility The activity coefficient ratio, γFe3+/γFe2+, most likely depends
(M) of a rock can be expressed as on the bulk chemistry of the rock or melt (e.g., SiO2 content), but
here we will assume unity. K depends on temperature (T) and (Fe2O3/FeO)wt.

total pressure (Ptotal). Therefore, C has a unique value at a given 0.09 0.22 -0.5 2.2
set of T, pH, fH O, and melt (or rock) composition. -5
2 M H
The relationship between the (Fe3+/Fe2+)melt and mole ratio of F QM
(Fe2O3/FeO)rock can be expressed as C
-10 900 °
LogfO2 (atm)
(XFe3+/XFe2+)melt = k’× 2(XFe O /XFeO)rock (7)
2 3
°C
where k’ is a distribution coefficient. -15 700
That is,
log(XFe3+/XFe2+)melt = log k’ + log2 + log(XFe O /XFeO)rock. (7′) -20 Magnetite

2 3
Ilmenite series
series
Substituting (7′) into (5′), we obtain
-25
-2 -1.5 -1 -0.5 0
logfO = 4log(XFe O /XFeO)rock + 4 logk’ + 4log2 + C. (8)
2 2 3 Log(Fe2O3/FeO)mol
Equation (8) indicates that, at a given set of T, Ptotal, PH O, and
2
bulk rock chemistry, the log fO2 value proportionally increases to Figure 3. Relationship between the Fe2O3/FeO ratios and fO2 values of mag-
four times the value of log(Fe2O3/FeO)mol (Fig. 3). For example, mas at Ptotal = 1 kbar and T = 700 °C and 900 °C. Thermodynamic data used
when Ptotal = 1 kb and T = 900 °C, the log fO2 - log(Fe2O3/FeO)mol in computations are summarized in Ohmoto and Kerrick (1977).
Figure 4. Geologic map of the Kaapvaal Craton, South Africa, showing the locations of the main Archean greenstone belts (Anhaeusser, 1976).
line extends from log fO2 = –7.25 and log(Fe2O3/FeO)mol = 0 (i.e.,

MH buffer) to log fO2 = –12.86 and log(Fe2O3/FeO)mol = −1.40
(i.e., QFM buffer). When Ptotal = 1 kb and T = 700 °C, the line
extends from log fO2 = –11.70 and log(Fe2O3/FeO)mol = 0 (i.e.,
MH buffer) to log fO2 = –17.35 and log(Fe2O3/FeO)mol = −1.41
(i.e., QFM).
The above examples of the fO2 versus Fe2O3/FeO ratio rela-
tionship illustrate an important point when comparing the redox
state of igneous rocks that crystallized at different T-P condi-
tions; the comparison should be made in terms of the Fe2O3/FeO
(or Fe3+/Fe2+) ratios of igneous rocks, rather than the fO2 values,
because the fO2 value varies depending on temperature and pres-
sure conditions.
GRANITOIDS IN THE KAAPVAAL CRATON
The Kaapvaal Craton (Fig. 4) is one of the oldest and best-

preserved Archean continental fragments on Earth. Its assem-
bly during the Archean eon is attributed to a complex combina-
tion of processes analogous to modern-day plate tectonics (Pou-
jol et al., 2003). Such processes took place episodically over
a 1000 million-year period (ca. 3500–2500 Ma) and involved
magmatic arc formation and accretion as well as the tectonic
amalgamation of numerous, discrete terranes or blocks (de Wit
et al., 1992; Lowe, 1994; Poujol and Robb, 1999).
The oldest rocks so far recognized are located in the Swa-
ziland-Barberton regions on the eastern side of the Kaapvaal
Craton (Fig. 4), where ages more than 3600 Ma have been
recorded. The early stages of shield development are also best
exposed in the Barberton Mountains, where it is now appar-
ent that continent formation took place by magmatic accretion
and tectonic amalgamation of small protocontinental blocks.
At Barberton, several diachronous blocks that formed between
3600 and 3200 Ma have been identified (Kamo and Davis,
1994), each of which represents a cycle of arc-related magma-
tism and sedimentation (Lowe, 1999; Poujol et al., 2003).
The Barberton Region

Figure 5. Geologic map of the Barberton region (modified after Ishi-
The Archean granitoids in the Barberton region occupy an hara et al., 2002b).
area of ~20,000 km2 (Fig. 5). They are divided into three groups
based on their intrusive ages in tectonic history: syntectonic, late- substage granitoids are mostly hornblende-biotite tonalites
tectonic, and post-tectonic (Anhaeusser and Robb, 1980a; Meyer and are free of magnetite and ilmenite, but contain titanite
et al., 1994). that occurs as euhedral wedges and mostly subhedral crys-
(1) Syntectonic tonalite-trondhjemite-granodiorite (TTG). S1 tals associated with mafic silicates, and which could have
substage: small intrusive bodies with the oldest age been converted from ilmenite. The color of the Z-axis of
(ca. 3450 Ma), e.g., Rooihoogte, Steynsdorp, Stolzburg, biotite is generally brown and similar to ilmenite-series
Theespruit, and other bodies; and 2 substage: larger bodies biotites of Phanerozoic age. The S2 substage granitoids
with younger ages (ca. 3230 Ma), e.g., Kaap Valley (KV in are also mostly magnetite-free, but the Kaap Valley plu-
Fig. 5) and Nelshoogte plutons. ton contains magnetite (up to 1.7% by volume), locally.
The syntectonic TTGs are closely associated with ultra- The magnetite forms polygonal to granular crystals and
mafic to mafic volcanic rocks of the greenstone belt and contains small inclusions of ilmenite and chalcopyrite.
intrude them concordantly. The largest Kaap Valley pluton In the most mafic intrusive phase, hematite blades occur
is circular in form, and intrudes them discordantly. The S1 along the 111 cleavage. However, the hematitization along
the cleavage and margin of the crystals is not seen in the plagioclase and microcline. Euhedral titanite and polyg-
magnetite of most magnetite-bearing phases, implying onal to rounded magnetite are common, but ilmenite
lack of oxidation at the latest stage of their crystallization. appears to be absent. No hematitization of the magnetite
(2) Late-tectonic calc-alkaline granitoids (ca. 3105 Ma); e.g., is observed.
Nelspruit, Mpuluzi and Heerenveen (HV in Fig. 5). The
late-tectonic calc-alkaline granites are batholithic in Johannesburg Dome
dimension, and biotite granite in composition, consisting In the central part of the Kaapvaal Craton, Paleoarchean
largely of unzoned plagioclase, microcline, and/or ortho- granitoids similar to those found in the Barberton region form
clase and quartz. Biotite is the dominant mafic silicate the Johannesburg Dome (~50 km in diameter). Evidence has
mineral and its Z-axis color is generally greenish brown been presented elsewhere showing that two types of granitoid
but rarely green, as is typical for fresh magnetite-series gneisses exist in this region. The oldest variety (ca. 3340 Ma;
granitoids of Phanerozoic age (Ishihara, 1977). The green Poujol and Anhaeusser, 2001) consists of leuco-biotite trondhje-
biotite, which is found locally, could be an alteration prod- mitic gneisses and associated migmatites that developed on the
uct because it is associated with chloritized biotite. Flaky northern half of the dome (TTG in Fig. 6) (Anhaeusser, 1973,
muscovite that is found in a few samples could also be an 1999). The younger variety, which is on the southern edge of
alteration product. Magnetite is composed of cubic to the dome, consists of hornblende-biotite tonalitic gneisses that
polygonal crystals associated with rare hematite blades or yielded a multiple zircon age of ca. 3170 Ma (Anhaeusser and
stringers. Ilmenite is absent, but titianite is abundant and Burger, 1982), and a more recent single zircon emplacement age
forms both euhedral wedge-shaped crystals and anhedral of ca. 3201 Ma (Poujol and Anhaeusser, 2001).
aggregates associated with biotite. Following the emplacement of trondhjemite-tonalite gneiss,
(3) Post-tectonic low-Ca (ca. 3070 Ma) and high-Ca another episode of magmatism took place on the Johannesburg
(ca. 2700 Ma) granitoids. These rocks include an older, Dome. This produced intrusions of Mesoarchean potassic grano-
low-Ca biotite granite, and younger, high-Ca alkaline diorites that occupy an area of batholithic dimension, extending
granitoids. The older granites are magnetite free, but across most of the southern portion of the dome (CA [calc-alka-
the younger granitoids contain magnetite. The younger line], Fig. 6). Two granodiorite phases have been distinguished:
granitoids of the Boesmanskop and Salisburgkop one on the southern and southeastern parts of the dome consists
stocks consist of minor quartz, alkaline amphiboles, mainly of medium-grained, homogeneous, gray granodiorites,
and greenish-brown biotite, with more abundant zoned whereas a second variety, found mainly on the southwestern part of
Figure 6. Distribution of the rock facies

(TTG, CA, and UD) and magnetic sus-
ceptibility values of the Johannesburg
Dome.
the dome, consists of porphyritic granodiorites (Anhaeusser, 1973). Temporal and Spatial Distributions of Granitoid Types in
Zircons extracted from the two granodiorite types yielded ages of the Barberton Region
ca. 3121 Ma for the homogeneous variety and ca. 3114 Ma for the
porphyritic variety (Poujol and Anhaeusser, 2001). In the Barberton region, all of the earliest S1, as well as S2,
Numerous pegmatite dykes and veins crosscutting the substage TTGs of the Nelshoogte pluton possess magnetic sus-
granodiorites are younger than 3114 Ma and may represent the ceptibility values less than 3 × 10−3 SI, indicating that they belong
final stages of magmatism associated with a batholith emplace- to very reduced ilmenite-series granitoids (Fig. 7). Anhaeusser
ment ca. 3000 Ma. and Robb (1980b) have shown that many of the TTGs interacted
with and assimilated greenstone wall rocks that have Fe2O3/
ANALYTICAL METHODS FeO (weight) ratios much less than 0.3 (Hunter, 1974; Hawkes-
worth and O’nions, 1977), suggesting that the reduced nature
Magnetic susceptibilities of 546 hand specimens from the of the TTGs may have been acquired (and/or enhanced) during
Barberton region and 346 specimens from the Johannesburg Dome emplacement, rather than at the magma sources.
were measured using a Kappameter KT-3. Most of the samples The S2 substage TTGs of the Kaap Valley tonalite pluton
used in this study were collected for earlier projects (Robb and (Robb et al., 1986) likewise possess generally low magnetic sus-
Anhaeusser, 1983; Robb et al., 1983; Robb et al., 1986), although ceptibility values, which suggests they are ilmenite series, but
~30 samples were collected especially for this study. contain sporadic values higher than 3 × 10−3 SI (Ishihara et al.,
Concentrations of major, trace, and rare earth elements 2002b). The distribution of magnetite-bearing rocks is so erratic
(REE) were determined on seven samples by XRF methods for that its genetic significance is difficult to evaluate. The post-tec-
most elements, titrimetry for FeO, and ICP-MS method for REE. tonic, low-Ca biotite granites also appear to be ilmenite series,
Concentrations of Cl and F, as well as the major element compo- although the number of samples examined was small.
sitions of minerals, were determined using an electron microbe The late-tectonic calc-alkaline granitoids, which form large
(EPMA) on 452 grains of biotite from eight representative gran- batholiths in the Barberton region, intrude the greenstones as
ite samples. Concentrations of Cl, F, and S in apatite crystals (258 well as the older TTGs (Anhaeusser and Robb, 1983). These
spot analyses) from two representative granite samples were also granitoids, along with the post-tectonic high-Ca granitoids, gen-
determined using an EPMA. erally possess magnetic susceptibilities higher than 3.0 × 10−3 SI,
indicating they are mostly magnetite series.
DISCUSSION OF THE ANALYSES In the ca. 3105 Ma Nelspruit batholith, 86% of the 243 sam-
ples exhibit magnetic susceptibility values greater than 3 × 10−3
The magnetic susceptibility of the ~900 samples range from SI (Fig. 7), which are characteristic of magnetite-series granit-
0.03 × 10−3 to 53 × 10−3 SI, indicating that both ilmenite series oids. The Nelspruit batholith is the largest granitic body in the
(<3 × 10−3 SI) and magnetite series (>3 × 10−3 SI) are present. region, and consists dominantly of homogeneous K-rich porphy-
Approximately 10% of the granitoid samples from the Barberton ritic granitic phases, associated potassic gneisses and migmatites
syntectonic plutons to ~80% in the late-tectonic plutons belong that occur together with remnants of older Na-rich gneisses and
to magnetite series (Fig. 7). Approximately 30% of the samples greenstones, and a homogeneous, medium-grained, granodioritic
from the Johannesburg Dome are magnetite series. phase (Robb et al., 1983).
The Mpuluzi batholith (Fig. 5) is the second largest granit-
oid massif in the region and consists of 53% magnetite series (n =
Magnetite series Ilmenite series ALL THE GRANITOIDS (N=12)
POST- 64). In contrast, the smaller Boesmanskop syenite pluton is 100%
TECTONIC High Ca granitoids (n=8)
GRANITOIDS magnetite series (n = 13).
Low Ca granitoids (n=4)
The post-tectonic high-Ca granitoids, found mainly in Swa-
ALL THE GRANITOIDS (N=402) ziland, occur as episodic, discrete intrusions crosscutting all the
LATE- Boesmanskop (n=13)
other Archean rock types. They are hornblende-biotite alkaline
TECTONIC
Nelspruit batholith (n=284) granitoids and have high magnetic susceptibilities similar to
GRANITOIDS
Mpuluzi body (n=64) alkaline granitoids.
Heerenveen batholith (n=41)
Temporal and Spatial Distributions of Granitoid Types in
ALL THE GRANITOIDS (N=132)
SYN-
the Johannesburg Dome
TECTONIC Kaap Valley pluton (n=34)
Theespruit body (n=40)
GRANITOIDS
Within the Johannesburg Dome area (Fig. 6), the Archean
Other TTG bodies (n=58)
0% 50 100
granitoids comprise older TTGs (ca. 3300–3200 Ma) and younger
calc-alkaline granitoids (ca. 3100–2900 Ma). Magnetic suscep-
Figure 7. Relative abundances of magnetite- and ilmenite-series gran- tibility measurements indicate that the TTGs consist mainly of
itoids in major batholiths and plutons of the Barberton region ilmenite series (78% of the measurements) and partly magnetite
series (22%); the calc-alkaline granitoids are also largely com- Chemical compositions of the representative magnetite-
posed of ilmenite series (83%) with some local (17%) magne- series granitoids are presented in Table 1. A sample of magne-
tite series (Ishihara et al., 2002a). Fe2O3/FeO ratios of bulk rock tite-series Kaap Valley tonalite (LKV7) reveals a much higher
samples range from 0.05 to 0.72 (Ishihara et al., 2002a) and are
generally below 0.5, implying a generally reduced nature for the
TABLE 1. REPRESENTAIVE CHEMICAL COMPOSITIONS OF TWO ILMENITE-
Johannesburg Dome pluton. SERIES AND FIVE MAGNETITE-SERIES GRANITOIDS, BARBERTON REGION
The weakly magnetic granitoids tend to occur in the central Theespruit-TTG Kaap Vly Nelspruit Batholith Mpuluzi Batholith
part of the Johannesburg Dome around an undifferentiated phase 23-4 23-5 LKV7 C31 D34 22-1 22-2
SiO2 70.05 75.34 65.47 69.78 71.05 66.49 70.02
(UD) at the interface between the TTG and calc-alkaline granit- TiO2 0.26 0.09 0.45 0.44 0.35 0.59 0.27
oids (Fig. 6). Al2O3 15.53 13.73 15.96 14.55 14.47 14.84 14.91
Fe2O3 0.49 0.39 2.26 1.88 0.97 2.46 1.24
FeO 1.61 0.80 1.68 1.26 1.57 1.94 1.15
Chemical Characteristics of Representative Granitoids MnO 0.03 0.04 0.05 0.05 0.04 0.08 0.05
Mg0 1.22 0.22 2.00 0.69 0.69 0.84 0.35
CaO 2.67 1.49 4.31 1.94 2.05 2.23 1.34
Partial chemical analyses were available for ~60 of the sam- Na2O 5.54 4.61 5.19 4.67 4.73 5.25 4.53
K2O 1.78 2.79 0.93 3.56 2.93 3.23 4.79
ples. Plots of magnetic susceptibility versus SiO2 content (Fig. 8) P2O5 0.08 0.03 0.16 0.24 0.13 0.35 0.12
show that the Nelspruit granitoids fall within the magnetite-series S <0.01 <0.01 <0.01 <0.01 <0.01 0.05 0.01
H2O+ 0.43 0.43 1.07 0.61 0.71 0.76 0.43
field (12 of 15 samples). The sporadic occurrence of magnetite- H2O– 0.17 0.09 0.42 0.22 0.17 0.19 0.19
bearing rocks in the Kaap Valley TTGs are also illustrated; one CO2 0.01 0.08 0.10 0.04 0.05 0.56 0.33
and three samples of the 16 specimens fall within the magnetite- SUM 99.87 100.13
Trace elements (ppm)
100.05 99.93 99.91 99.86 99.73
and intermediate-series fields, respectively. Thus, the Kaap Val- Rb 57 176 21 134 100 153 182
ley TTGs belong essentially to the ilmenite-series granitoids. All Cs
Sr
2.7
556
9.4
126
<1.5
576
3.2
563
1.3
467
4.0
812
5.7
785
other TTGs also have very low magnetic susceptibility values, in Ba 389 227 219 976 470 1276 1993
contrast to the high values of sodic and/or adakitic granitoids of Zr
Hf
110
3.1
77
3.4
102
2.4
262
6.5
169
5.6
412
10.3
293
7.1
Phanerozoic ages (e.g., Tanzawa pluton) (Fig. 8). Nb 4.2 13.5 3.0 23.4 14.9 23.2 16.0
Ta 2.3 2.6 2.3 3.2 4.1 2.6 2.2
Y 5 20 7 29 24 34 33
V 20 5 71 15 23 19 <4
Cr 58 41 109 30 77 52 37
Co 11 6 16 7 8 10 6
Ni 27 2 24 4 3 4 1
Cu 3 5 5 13 0 14 3
Zn 50 35 50 76 65 114 53
Pb 10 18 5 28 20 23 27
Ga 17.6 14.4 18.2 19.9 19.5 19.7 16.7
Ge 0.9 1.1 0.9 0.9 0.8 1.0 0.6
Se 0.2 0.2 0.2 0.5 0.4 0.3 0.3
Mo <0.2 0.2 0.5 1.4 0.7 1.7 1.7
W 1.1 1.9 <1.2 1.9 2.5 <1.4 1.9
Sn 1.5 6.3 1.4 3.3 3.5 6.5 5.0
Cd 0.7 <0.2 0.3 0.3 <0.2 0.3 0.2
Tl 0.6 1.5 0.6 2.8 1.6 1.6 1.9
Bi 0.4 0.3 <0.3 0.9 0.3 <0.3 0.4
Th 4.5 5.8 1.5 15.1 9.9 15.3 20.6
U 0.7 3.0 0.5 1.7 2.7 2.4 1.8
Rare earth elements (ppm)
La 21.3 8.93 13.7 133 125 109 42.2
Ce 37.4 16.2 29.1 285 244 216 86.2
Pr 3.60 1.65 3.35 29.4 23.5 21.6 9.35
Nd 12.5 5.85 13.7 108 79.3 75.4 34.3
Sm 2.07 1.71 2.71 18.0 13.2 12.3 6.16
Eu 0.673 0.383 0.920 4.15 2.89 2.54 1.28
Gd 1.56 2.21 2.37 12.8 10.0 8.62 5.30
Tb 0.21 0.47 0.32 1.60 1.38 1.17 0.82
Dy 1.10 2.91 1.63 7.43 6.83 5.84 4.41
Ho 0.20 0.60 0.30 1.19 1.19 1.02 0.91
Er 0.58 1.78 0.86 3.09 3.20 2.76 2.65
Tm 0.081 0.295 0.123 0.399 0.427 0.396 0.392
Yb 0.50 1.85 0.75 2.36 2.49 2.44 2.40
Lu 0.076 0.278 0.114 0.331 0.337 0.341 0.359
Y 6.2 19.7 8.8 36.3 36.2 31.7 26.4
-3
Kai (×10 ) 0.1 0.1 18.9 19.1 13.9 36.0 14.2
Figure 8. Magnetic susceptibility versus silica content comparison of Fe2O3/FeO 0.3 0.5 1.4 1.5 0.6 1.3 1.1
the TTGs and Nelspruit batholith; SiO2 data are from Anhaeusser and Rb/Sr 0.1 1.4 0.0 0.2 0.2 0.2 0.2
Note: Rock type and plutons: 23-4—very fine quartz diorite, foliated, Theespruit; 23-
Robb (1983). The fields for Japanese magnetite- and ilmenite-series 5—Medium, trondjemite, Theespruit; LKV7—Fine, quartz diorite, Kaap Valley; C31—
granitoids are shown for comparison (separated by dotted lines). Open Medium, pink K-feldspar porphyritic biotite granite, Nelspruit; D34—Coarse, pink K-
stars represent data on sodic Miocene gabbroid and granitoids of the feldspar porphyritic biotite granite, Nelspruit; 22-1—Fine, quartz diorite, Mpuluzi; 22-
Tanzawa-Niijima (Ishihara, unpublished data). 2—fine, biotite granite, Mpuluzi. Kai, magnetic susceptibility, × 10-3 SI unit.
Fe2O3/FeO ratio (1.35) than the magnetite-free TTGs of the Magnetite crystals in the calc-alkaline magnetite-series
Theespruit pluton (Fe2O3/FeO = 0.30–0.49), indicating a good granitoids tend to occur together with mafic silicates (mostly bio-
correlation between magnetic susceptibility and Fe2O3/FeO ratio. tite). The Mg/Mg+Fe atomic ratios of biotites from the Nelspruit
Trace amounts of V and Cr, which may be substitutes for Fe2+ and batholith (Table 2) are similar to those of typical magnetite-series
Fe3+ in magnetite, respectively, are higher in the magnetite-bear- biotite granite in the Sanin District, Japan (Fig. 10). The magne-
ing rocks than in the magnetite-free rocks. tite crystals contain rare hematite blades, which may have formed
Other TTG characteristics include high contents of Na2O during the subsolidus stage, corresponding to a high degree of
and Sr, but low contents of K2O, Rb, and Ba (Robb and Anhae- oxidation. Hematitization is only locally and weakly observed in
usser, 1983) (Table 1). The LKV7 tonalite and 23-4 trondhjemite the batholiths.
samples are high in Sr and low in Y, thus exhibiting adakitic char-
acteristics (Drummond et al., 1996). Compared with the late-tec- SIGNIFICANCE OF THE RESULTS
tonic granitoids, the TTGs are rich in MgO, Al2O3, CaO, Ni, and
Co (compare 23-4 and C31 in Table 1). Genesis of Archean Granitoids in the Kaapvaal Craton
Magnetite-rich rocks of the Nelspruit and Mpuluzi batholiths
have similarly high Fe2O3/FeO ratios (0.62–1.49). These batho- The Kaap Valley pluton contains local magnetite-bearing
liths, however, have differing feldspar chemistry; the Nelspruit rocks (Figs. 7 and 8), and has, as a whole, higher Fe2O3/FeO
granites are calcic, whereas the Mpuluzi granites are potassic. ratios than the other TTGs. The whole-rock δ18O values of the
The Mpuluzi granites also have higher Rb, Sr, and Ba contents Kaap Valley pluton are on average 2‰ higher than those of Qua-
than the Nelspruit granites. ternary low-K tholeiite (Faure and Harris, 1991). We interpret
The granitoids generally display enrichments of light rare these data to indicate that the magmas for the Kaap Valley pluton
earth elements (LREE) and depletions of heavy rare earth ele- were generated from partial melting of subducted seafloor basalt,
ments (HREE) with respect to chondrite (Fig. 9). There are which had increased its δ18O value by low-temperature interac-
almost no Eu anomalies in the studied samples, except for the tion with seawater.
23-5 Theespruit-TTG specimen, which, with a higher SiO2 The Fe2O3/FeO ratios of oceanic basalts probably increased
content and a higher Rb/Sr ratio compared to the other TTGs, from ~0.1 to ~1 by interaction with O2- and/or SO42–-bearing
represents a highly fractionated phase of the TTGs. The general seawater, much like the processes of seawater-rock interaction
REE patterns of the studied samples are similar to those of other in modern oceanic crust (e.g., Shanks et al., 1981; Lécuyer and
TTGs (Robb et al., 1986) and garnet-bearing metamorphic pro- Ricard, 1999). The possibility of SO42–-rich Archean seawater is
toliths; the HREE-depleted patterns are attributed to garnet and/ also suggested from the abundance of barite beds associated with
or hornblende fractionation. The magnetite-series granitoids, felsic volcanic rocks in older greenstones (ca. 3500 Ma Onver-
particularly of the Mpuluzi potassic granite, are most enriched wacht Group) and later ferruginous sediments (ca. 3230 Ma Fig
in REEs, whereas the ilmenite-series TTGs are generally most Tree Group). The geochemistry of banded iron-formations and
depleted in REE. associated rocks also led Ohmoto et al. (this volume) to conclude
Figure 9. REE patterns of selected magne-

tite-series (solid symbols) and ilmenite-series
(open symbol) granitoids. Sample numbers
correspond to those in Table 1.
TABLE 2. CHEMICAL COMPOSITION OF BIOTITES FROM REPRESENTATIVE GRANITOIDS OF THE BARBERTON REGION
Syn-tectonic granitoids Late- and post-tectonic granitoids
Ilmenite-series Magnetite-series Magnetite-series Magnetite-series
Theespruit Kaap Valley Boesmankop Salisburgkop Nelspruit Nelspruit
Magnetic susceptibility 0.12 × 10–3 0.05 × 10–3 53.0 × 10–3 19.9 × 10–3 24.0 × 10–3 6.5 × 10–3
Sample number 23–4 23–5 24–6 LKV–19 22–3 24–7B C31 D34
TTG–5 TTG–6 TTG–1 TTG–3
Number of analysis 85 22 123 16 26 22 94 64
ave (±1V) ave (±1V) ave (±1V) ave (±1V) ave (±1V) ave (±1V) ave (±1V) ave (±1V)
SiO2 (wt%) 36.99 (0.36) 35.73 (0.31) 37.16 (0.23) 37.08 (0.54) 39.53 (0.35) 37.66 (0.42) 38.35 (0.26) 37.54 (0.29)
Al2O3 14.70 (0.28) 17.26 (0.26) 15.04 (0.23) 14.91 (0.19) 11.25 (0.17) 14.31 (0.41) 14.32 (0.22) 14.47 (0.39)
TiO2 1.65 (0.29) 1.70 (0.13) 1.74 (0.22) 2.26 (0.60) 0.87 (0.06) 1.41 (0.17) 2.35 (0.29) 2.73 (0.29)
FeO* 19.01 (0.47) 24.07 (0.42) 17.34 (0.44) 17.11 (0.34) 17.79 (0.82) 16.59 (0.37) 16.63 (0.44) 17.12 (0.38)
MnO 0.24 (0.04) 0.44 (0.04) 0.79 (0.06) 0.28 (0.04) 0.55 (0.04) 0.68 (0.04) 0.59 (0.06) 0.59 (0.06)
MgO 11.79 (0.32) 5.92 (0.16) 12.15 (0.37) 12.99 (0.50) 13.91 (0.64) 13.26 (0.41) 11.98 (0.37) 12.04 (0.33)
CaO 0.08 (0.05) 0.04 (0.03) 0.04 (0.03) 0.14 (0.32) 0.01 (0.01) 0.02 (0.01) 0.02 (0.03) 0.02 (0.04)
Na2O 0.06 (0.03) 0.05 (0.02) 0.08 (0.23) 0.06 (0.03) 0.03 (0.02) 0.06 (0.02) 0.07 (0.02) 0.10 (0.02)
K2O 9.16 (0.42) 9.28 (0.21) 9.53 (0.27) 9.37 (0.50) 9.76 (0.09) 9.82 (0.11) 9.54 (0.13) 9.48 (0.14)
Cl 0.02 (0.01) 0.01 (0.01) 0.08 (0.01) 0.05 (0.01) 0.22 (0.02) 0.02 (0.01) 0.02 (0.01) 0.04 (0.01)
=O 0.00 0.00 –0.02 –0.01 –0.05 0.00 0.00 –0.01
F 0.25 (0.10) 0.25 (0.13) 0.45 (0.13) 0.13 (0.09) 1.78 (0.21) 1.06 (0.16) 1.06 (0.16) 0.98 (0.13)
=O –0.06 –0.06 –0.10 –0.03 –0.40 –0.24 –0.24 –0.22
Total 93.88 94.70 94.29 94.33 95.27 94.63 94.68 94.87
Si (atom, O = 22) 5.736 (0.035) 5.630 (0.023) 5.740 (0.029) 5.667 (0.050) 6.222 (0.027) 5.844 (0.062) 5.923 (0.025) 5.798 (0.027)
Al 2.688 (0.045) 3.206 (0.039) 2.739 (0.036) 2.686 (0.030) 2.087 (0.032) 2.617 (0.075) 2.607 (0.034) 2.634 (0.065)
Ti 0.192 (0.034) 0.201 (0.015) 0.203 (0.026) 0.260 (0.070) 0.104 (0.007) 0.164 (0.020) 0.273 (0.034) 0.317 (0.034)
Fe(2+)* 2.465 (0.060) 3.172 (0.068) 2.240 (0.058) 2.187 (0.038) 2.343 (0.118) 2.152 (0.046) 2.148 (0.058) 2.212 (0.051)
Mn 0.032 (0.006) 0.059 (0.006) 0.104 (0.008) 0.036 (0.006) 0.074 (0.006) 0.089 (0.005) 0.077 (0.008) 0.077 (0.009)
Mg 2.725 (0.077) 1.390 (0.036) 2.797 (0.082) 2.960 (0.115) 3.264 (0.139) 3.066 (0.095) 2.758 (0.084) 2.772 (0.077)
Ca 0.013 (0.009) 0.007 (0.004) 0.007 (0.005) 0.023 (0.053) 0.002 (0.002) 0.003 (0.002) 0.003 (0.007) 0.003 (0.007)
Na 0.017 (0.008) 0.016 (0.007) 0.023 (0.007) 0.018 (0.008) 0.010 (0.005) 0.017 (0.007) 0.021 (0.007) 0.030 (0.005)
K 1.812 (0.080) 1.866 (0.036) 1.878 (0.048) 1.826 (0.089) 1.961 (0.020) 1.943 (0.018) 1.879 (0.027) 1.868 (0.028)
Cl 0.006 (0.003) 0.002 (0.002) 0.021 (0.003) 0.012 (0.002) 0.060 (0.006) 0.004 (0.002) 0.005 (0.003) 0.010 (0.003)
F 0.123 (0.051) 0.126 (0.064) 0.222 (0.062) 0.063 (0.045) 0.885 (0.107) 0.520 (0.077) 0.519 (0.081) 0.479 (0.066)
Al (IV) 0.264 0.371 0.260 0.333 –0.222 0.156 0.077 0.202
XMg 0.525 (0.010) 0.305 (0.008) 0.555 (0.013) 0.575 (0.007) 0.582 (0.022) 0.588 (0.012) 0.562 (0.013) 0.556 (0.010)
*Indicates that the total Fe contents are expressed as FeO or Fe(2+).
that the Archean oceans were sulfate-rich and generally oxygen- melting of earlier TTGs (Anhaeusser and Robb, 1983; Robb et
ated, except in local basins. al., 1983). The Fe2O3/FeO weight ratios of earlier TTG mag-
The late-tectonic calc-alkaline granites of the Nelspruit and mas may have increased to >0.5 by the same processes as the
Mpuluzi batholiths are also composed of typical magnetite-series Kaapvaal plutons (i.e., the partial melting of altered and oxidized
rocks. They are thought to have been generated by the partial subducted mid-oceanic-ridge basalt). The magnetite-series calc-
alkaline granites (e.g., Nelspruit batholith) may have been gen-
erated from such a TTG lower crust by the heat from the upper
0.8 mantle and water from dehydration of the altered subducting
oceanic crust.
ula Mineralization Associated with Archean Granitoids

0.7 ins
en
aP
n aw
Fe/Fe+Mg (atomic)
a
Tak Highly oxidized magnetite-series magmas, rather than
it
ds,
u
r
G ilmenite-series magmas, are more favorable sources for the gen-
spr
0.6 s
se rie
ite- eration of ore-forming fluids for Cu, Mo, Pb, Zn, Ag, and Au
ee
n
lme
Th
I
it deposits (e.g., Burnham and Ohmoto, 1980). Many such exam-
0.5 pru ples are found in the circum-Pacific Rim (Ishihara, 1998). In
ls
Mag
netit Ne the Barberton region, orogenic gold deposits that occur in shear
e-s eries zones in the greenstones are the most prominent metallic miner-
0.4 Grds
, Dait
o-Yok alization type present, but these gold deposits are not considered
ota Are
a genetically related to the adjacent TTG and calc-alkaline gra-
nitic occurrences.
0.3 Along the northern flank of the Barberton greenstone belt,
50 60 70 80
SiO2 (%) there is a broad temporal overlap between a mineralized, fel-
Figure 10. Fe/Mg ratios of the magnetite-series Nelspruit batholith. sic porphyry (ca. 3126 Ma Fairview Mine porphyry; de Ronde
Japanese data are from Czamanske et al. (1981). et al., 1991) and the ca. 3105 Ma magnetite-series Nelspruit
TABLE 3. CHEMICAL COMPOSITION OF APATITES FROM REPRESENTATIVE GRANITOIDS OF THE BARBERTON REGION
Syn-tectonic granitoids Late- and post-tectonic granitoids
Ilmenite-series Magnetite-series Magnetite-series Magnetite-series
Theespruit Kaap Valley Boesmankop Salisburgkop Nelspruit Nelspruit
Magnetic susceptibility 0.12 × 10–3 0.05 × 10–3 53.0 × 10–3 19.9 × 10–3 24.0 × 10–3 6.5 × 10–3
Sample number 23–4 23–5 24–6 LKV–19 22–3 24–7B C31 D34
TTG–5 TTG–6 TTG–1 TTG–3
Number of analysis 85 27 104 23 130 42 145 113
ave (±1V) ave (±1V) ave (±1V) ave (±1V) ave (±1V) ave (±1V) ave (±1V) ave (±1V)
SiO2 (wt%) 0.11 (0.14) 0.15 (0.06) 0.26 (0.08) 0.16 (0.04) 0.60 (0.17) 0.32 (0.08) 0.27 (0.16) 0.28 (0.18)
Al2O3 0.01 (0.05) 0.01 (0.02) 0.01 (0.02) 0.01 (0.01) 0.01 (0.01) 0.01 (0.01) 0.01 (0.01) 0.01 (0.04)
TiO2 0.01 (0.02) 0.03 (0.03) 0.01 (0.02) 0.01 (0.02) 0.01 (0.02) 0.01 (0.02) 0.01 (0.01) 0.01 (0.02)
FeO* 0.13 (0.10) 0.36 (0.14) 0.13 (0.11) 0.22 (0.08) 0.23 (0.17) 0.16 (0.06) 0.19 (0.09) 0.09 (0.09)
MnO 0.03 (0.03) 0.53 (0.13) 0.10 (0.04) 0.04 (0.02) 0.04 (0.03) 0.10 (0.04) 0.07 (0.03) 0.09 (0.03)
MgO 0.01 (0.03) 0.01 (0.01) 0.01 (0.01) 0.01 (0.01) 0.01 (0.01) 0.01 (0.01) 0.01 (0.01) 0.01 (0.01)
CaO 55.62 (0.38) 54.65 (0.03) 55.44 (0.34) 55.51 (0.26) 54.06 (0.45) 54.77 (0.43) 55.01 (0.56) 55.02 (0.61)
Na2O 0.00 (0.01) 0.10 (0.02) 0.09 (0.06) 0.03 (0.02) 0.04 (0.04) 0.11 (0.08) 0.06 (0.04) 0.02 (0.02)
K2O 0.01 (0.02) 0.03 (0.02) 0.02 (0.02) 0.01 (0.01) 0.01 (0.02) 0.02 (0.01) 0.01 (0.01) 0.02 (0.01)
P2O5 39.88 (0.53) 39.61 (0.31) 39.04 (0.54) 39.16 (0.34) 38.41 (0.50) 38.80 (0.57) 38.99 (0.65) 38.00 (0.67)
Cl 0.02 (0.01) 0.01 (0.01) 0.03 (0.01) 0.11 (0.05) 0.01 (0.01) 0.01 (0.01) 0.01 (0.01) 0.01 (0.01)
=O 0.00 0.00 –0.01 –0.02 0.00 0.00 0.00 0.00
F 4.21 (0.31) 4.72 (0.40) 4.47 (0.35) 3.63 (0.33) 4.69 (0.37) 4.86 (0.34) 4.39 (0.45) 4.54 (0.44)
=O –0.95 –1.06 –1.01 –0.82 –1.06 –1.10 –0.99 –1.02
SO3 0.02 (0.02) 0.01 (0.01) 0.40 (0.19) 0.13 (0.04) 0.04 (0.11) 0.36 (0.24) 0.20 (0.07) 0.15 (0.04)
Total 99.11 99.15 99.00 98.19 97.10 98.44 98.23 97.22
*Indicates that the total Fe contents are expressed as FeO or Fe(2+).
granite batholith (Kamo and Davis, 1994). But the late-tectonic 0.6
Nelspruit granites may be too large in exposed dimensions, thus
exposing deeper parts of the batholiths, to be genetically related 0.5
Philippines: Barren Intrusives
to the porphyry that formed at a shallow crustal depth.
0.4
Several small cassiterite-bearing pegmatite dikes and veins
SO3 (wt %)
occur in the Mpuluzi batholith close to the Swaziland border. 0.3

Although tin mineralization is typically associated with reduced
(ilmenite-series) and/or highly fractionated granites (Ishihara, 0.2
1981), the redox state of granite related to cassiterite mineraliza- Philippines: Porphyry-Cu related Intrusives
tion in the Mpuluzi batholith is unknown. 0.1
The TTGs of the Barberton region are adakitic and simi-
lar to those in the Au-Cu mineralized regions of the Philippines 0.0
0 1 2 3 4 5
(Sajona and Maury, 1998), northern Chile (Oyarzun et al., 2001),
Cl (wt %)
and Kitakami Mountains (Tsuchiya and Kanisawa, 1994; Ishi-
hara and Murakami, 2004). However, an important difference is Figure 11. SO3 and Cl contents of apatite crystals in the magnetite-
their low redox state, which may have limited the concentration series Nelspruit and Mpuluzi granitoids (dots). Original data are from
of sulfur, and hence S-combined ore metals, to the magmatic Imai (2002), Ishihara et al. (2002b), and this study.
fluids (Burnham and Ohmoto, 1980; Ishihara et al., 1988).
Oxidizing conditions of porphyry Cu-related magmas are
represented by high SO3 contents of apatites, such as those in Porphyry Mineralization in Earth’s History
the Philippines (Imai, 2001, 2002, 2004). Chlorine is an impor-
tant metal carrier in magmatic fluids, and the Cl content of Porphyry Cu–Mo and Cu-Au deposits are typically associ-
rock-forming apatite is a good indicator of the Cl contents of ated with oxidized magnetite-series magmatism, whereas Sn-W
granitic magmas (Nedachi, 1980). EPMA analyses indicate that deposits generally occur with reduced ilmenite-series granit-
the SO3 contents of apatite crystals in magnetite-series granitoids (Fig. 1). Because mineral exploration has been carried out
oids from the Barberton region approach 0.36 wt.% (Table 3), much more extensively than regional granitoid-series studies, ore
which is similar to the value of porphyry Cu-related intrusions deposit data may provide a clue to the distributions of the two
from the Philippines (Imai, 2001, 2002). However, Cl con- types of granitoids in Earth’s history.
tents of apatites from the Barberton granitoids are extremely A histogram of the total Cu tonnages in all the discovered
low (below 0.1%) (Fig. 11). These low Cl concentrations in the porphyry-type deposits, plotted against their geologic age, is
magnetite-series granitoids were probably an important reason shown in Figure 12. It may suggest that the abundance of mag-
for the scarcity of S-combined metallic mineralizations in the netite-series granitoids, as approximated by porphyry Cu depos-
Barberton region. its, increased with younger geologic age. The Cenozoic peak in
Discovered ores, ×108 tons Cu Grasberg

0 2 4 6 8 10 Andina
0 Ma El Teniente
Cenozoic Chuquicamata
65 Bingham
Cretaceous
146
Jurassic Valley Copper
208
Triassic
Erdenet
245
Permian
290
Carboniferous Almalyk
363
Devonian
409
Silurian
439
Ordvician
510
Cambrian
570
Late Proterozoic
800
Early Proterozoic
2450
Late Archean
3500
Figure 12. Tonnages of porphyry copper ores versus geologic time. Original data are from Singer et al. (2002).
Figure 12 comes mostly from young and super-large deposits at Sn deposits, which are typically associated with ilmen-
Grasberg-Ertzberg (3 Ma, 28 million tons [MT] Cu), El Teniente ite-series granitoids, are found mostly in the Phanerozoic rock
(4 Ma, 56.8 MT Cu), Andina (4 Ma, 37.2 MT Cu), Los Pelambres (Meyer, 1985). They are concentrated in the late Paleozoic
(10 Ma, 20.5 MT Cu), Chuquicamata (31–34 Ma, 55.2 MT Cu), (southwestern Europe) and Triassic–Jurassic (Southeast Asia),
Escondida (37–38 Ma, 34.3 MT Cu), and Bingham (39–40 Ma, reflecting the predominance of ilmenite-series granitoids in these
23.2 MT Cu) (the age and metal tonnage statistics are from Wata- regions. Cenozoic-age Sn-bearing granitoids are limited to the
nabe, 2003). Quaternary volcanoes of the circum-Pacific belt are Bolivian Miocene granitic belt and Miocene Outer Zone gran-
composed predominantly of magnetite series, and porphyry-type itoids of southwestern Japan. Scarcity of the late Cenozoic Sn
deposits are thought to be forming below these volcanoes (e.g., deposits is likely due to paucity of ilmenite-series granitoids dur-
Mt. Pinatubo; Imai et al., 1993, 1996; Hattori and Keith, 2001). ing this period in the circum-Pacific region.
Most porphyry-related hydrothermal deposits are formed at Sizable Sn deposits are known in Precambrian terrains. The
shallow depths, typically less than ~5 km (Burnham and Ohmoto, late Proterozoic granites (1.0 Ga), of Rondonia, Brazil (Bet-
1980). Therefore, the observed trends of decreasing amounts of tencourt and Dall’angol, 1995), and early Proterozoic granites
Cu deposits with older geologic age are likely due to preserva- (2.1 Ga) of the Bushveld complex host many cassiterite-pegma-
tion effects. Pillow lavas and high-level granitoids are still locally tite-greisen-veins. Archean granites also host cassiterite associ-
preserved in the Precambrian terrains. Copper deposits at Malan- ated with Li and Ta pegmatites. At Greenbush, Western Australia,
jkhand, central India (2400 Ma, 6.5 MT Cu), may or may not Sn-Ta-Li pegmatite sheet (2.5 Ga) produced in 1888–1987 the tin
be porphyry-type (Sarkar et al., 1996; Panigrahi and Mookher- concentrate of 26,000 tons (~72% Sn), Ta2O5 concentrate of 2300
jee, 1997). However, the oldest (3300 Ma) Mo-Cu deposit at the tons, low iron spodumene concentrate (72% Li2O) of 39,700 tons
Coppin Gap, Pilbara Craton (102 MT; 0.152% Cu, 0.105% Mo; and crude kaolin ore of 9500 tons (Hatcher and Clynick, 1990).
Jones, 1990), is a large porphyry-type deposit. A reconnaissance Ta-pegmatite of the Mount Cassiterite orebody (2.9 Ga) in the
magnetic survey of the porphyry indicates that it is intermedi- Pilbara Craton contains 8785 tons Sn metal (Huston et al., 2001).
ate between ilmenite and magnetite series. These data, therefore, Archean pegmatite in Zimbabwe is said to contain cassiterite
suggest that the formation of porphyry Cu-Mo and Cu-Au depos- with 114,000 tons Sn metal. Cassiterite mineralization therefore
its, as well as magnetite-series granitoids, have occurred since at appears to be concentrated in Precambrian time, yet the examples
least ~3.3 billion years ago. are still too small to draw a definite conclusion.
Anhaeusser, C.R., 1999, Archaean crustal evolution of the central Kaapvaal

CONCLUSIONS Craton, South Africa: Evidence from the Johannesburg Dome: South
African Journal of Geology, v. 102, p. 303–322.
Amitsoq-gray tonalite gneisses from West Greenland Anhaeusser, C.R., and Burger, A.J., 1982, An interpretation of U-Pb zircon
ages for Archaean tonalitic gneisses from the Johannesburg-Pretoria gran-
(ca. 3700 Ma) have Fe2O3/FeO ratios that average 0.33 (n = ite dome: Transactions of the Geological Society of South Africa, v. 85,
23; Nutman and Bridgwater, 1986), thus being ilmenite series. p. 111–116.
However, because these tonalites were highly metamorphosed, Anhaeusser, C.R., and Robb, L.J., 1980a, Magmatic cycles and the evolution of
the Archaean granitic crust in the eastern Transvaal and Swaziland: Geo-
and metamorphism generally lowers the Fe2O3/FeO ratios of logical Society of Australia Special Publication 7, p. 457–467.
rocks by reactions with graphite (Ohmoto and Kerrick, 1977), Anhaeusser, C.R., and Robb, L.J., 1980b, Regional and detailed field and
the original tonalites may have been intermediate or magnetite geochemical studies of Archaean trondhjemitic gneisses, migmatites
and greenstone xenoliths in the southern part of the Barberton Moun-
series. Archean granitoids from the Kaapvaal Craton that are tain Land, South Africa: Precambrian Research, v. 11, p. 373–397, doi:
older than ca. 3230 Ma are mostly reduced-type (ilmenite series) 10.1016/0301-9268(80)90073-X.
with a Fe2O3/FeO weight ratio of 0.3–0.5. Because their Fe2O3/ Anhaeusser, C.R., and Robb, L.J., 1983, Geological and geochemical charac-
teristics of the Heerenveen and Mpuluzi batholiths south of the Barberton
FeO ratios appear to have been lowered by the assimilation of greenstone belt and preliminary thoughts on their petrogenesis: Geologi-
greenstones, however, their original magmas might also have cal Society of South Africa Special Publication 9, p. 131–151.
been more oxidized than normal ilmenite-series magmas. The Bettencourt, J.S., and Dall’angol, R., eds., 1995, The rapakivi granites of the
Rondonia tin province and associated mineralization: Para, Brazil, Fed-
3230 Ma Kaap Valley pluton and younger granitoids are mostly eral University, Excursion Guide, 48 p.
magnetite series, suggesting that by ca. 3200 Ma, oxidized proto- Burnham, C.W., and Ohmoto, H., 1980, Late-stage processes of felsic magma-
liths existed in the Kaapvaal continental crust. tism: Mining Geology Special Issue 8, p. 1–11.
Czamanske, G.K., Ishihara, S., and Atkin, S.A., 1981, Chemistry of rock-form-
Source rocks for the Archean TTGs, as suggested by their ing minerals of the Cretaceous-Paleogene batholith in southwestern Japan
REE patterns and geochemical characteristics, appear to have and implications for magma genesis: Journal of Geophysical Research,
been subducted oceanic crust that was subjected to a garnet-bear- v. 86, p. 10431–10469.
de Ronde, C.E.J., Kamo, S., Davis, D.W., de Wit, M.J., and Spooner, E.T.C.,
ing metamorphic grade. The late-tectonic calc-alkaline granitoids 1991, Field geochemical and U-Pb isotopic constraints from hypabys-
may have formed from the partial melting of such older Archean sal felsic intrusions within the Barberton greenstone belt, South Africa:
TTGs. Although these rocks form part of the magnetite series, Implications for tectonics and the timing of gold mineralization: Precam-
brian Research, v. 49, p. 261–280, doi: 10.1016/0301-9268(91)90037-B.
they have low Cl contents. The deep erosion level of the granit- de Wit, M.J., Roering, C., Hart, R.J., Armstrong, R.A., de Ronde, C.E.J., Green,
oids and low Cl contents of the magmas were probably the pri- R.W.E., Tredoux, M., Peberdy, E., and Hart, R.A., 1992, Formation of an
mary reasons why no metallic deposits are associated with mag- Archean continent: Nature, v. 357, p. 553–562, doi: 10.1038/357553a0.
Drummond, M.S., Defant, M.J., and Kepezhinskas, P.K., 1996, Petrogenesis of
netite-series granitoids in the Barberton region. slab-derived trondhjemite-tonalite-dacite-adakite magmas: Transactions
Porphyry-type Cu (-Au or -Mo) deposits are the most rep- of the Royal Society of Edinburgh, Earth Sciences, v. 87, p. 205–215.
resentative mineralization associated with oxidized, magnetite- Faure, K., and Harris, C., 1991, Oxygen and carbon isotope geochemistry
of the 3.2 Ga Kaap Valley tonalite, Barberton greenstone belt, South
series magmatism. The oldest example may be the 3300 Ma Cop- Africa: Precambrian Research, v. 52, p. 301–319, doi: 10.1016/0301-
pin Gap in the Pilbara Craton. The presence of such Cu deposits 9268(91)90085-O.
and the abundance of magnetite-series granitoids in the Kaapvaal Hatcher, M.I., and Clynick, G., 1990, Greenbushes tin-tantalum-lithium
deposit, in Hughes, F.E., ed., Geology of the mineral deposits of Australia
Craton suggest that the processes of seawater-rock interaction at and Papua New Guinea: Melbourne, The Australian Institute of Mining
mid-oceanic ridges and granite magma generation in subduction and Metallurgy, p. 599–603.
zones, as well as the redox chemistry of ocean water, have been Hattori, K.H., and Keith, J.D., 2001, Contribution of mafic melt to porphyry
copper mineralization: evidence from Mount Pinatubo, Philippines, and
basically the same since at least ca. 3200 Ma. Bingham Canyon, Utah, USA: Mineralium Deposita, v. 36, p. 799–806,
doi: 10.1007/s001260100209.
ACKNOWLEDGMENTS Hawkesworth, C.J., and O’nions, R.K., 1977, The petrogenesis of some
Archean volcanic rocks from southern Africa: Journal of Petrology, v. 18,
p. 487–520.
We are grateful to D.A. Singer for providing the world cop- Hunter, D.R., 1974, Crustal development in the Kaapvaal Craton, I, The
per tonnage statistics and to S.E. Kesler, B.R. Frost, and E.J. Archean: Precambrian Research, v. 1, p. 259–294, doi: 10.1016/0301-
9268(74)90002-3.
Essene for helpful comments and suggestions on earlier manu- Huston, D.L., Blewett, R.S., Sweetapple, M., Brauhart, C., Cornelius, H., and
scripts. Ohmoto acknowledges support from the National Sci- Collins, P.L.F., 2001, Metallogenesis of the north Pilbara granite-green-
ence Foundation (EAR-9706279, EAR-0229556), the NASA stones, Western Australia—A field guide: Western Australia Geological
Survey, Record 2001/11, 87 p.
Astrobiology Institute (NCC2-1057, CA#NNA04CC06A), and Imai, A., 2001, Generation and evolution of ore fluids for porphyry Cu-Au min-
the NASA Exobiology Program (CA#NNG04GK00G). eralization of the Santo Tomas II (Philex) deposit, Philippines: Resource
Geology, v. 51, p. 71–96.
Imai, A., 2002, Metallogenesis of porphyry Cu deposits of the western Luzon
REFERENCES CITED arc, Philippines: K-Ar ages, SO3 contents of microphenocrystic apatite
and significance of intrusive rocks: Resource Geology, v. 52, p. 147–161.
Anhaeusser, C.R., 1973, The geology and geochemistry of the Archaean gran- Imai, A., 2004, Variation of Cl and SO3 contents of microphenocrystic apatite
ites and gneisses of the Johannesburg-Pretoria Dome: Geological Society in intermediate to silicic igneous rocks of Cenozoic Japanese island arcs:
of South Africa Special Publication 3, p. 361–385. Implications for porphyry Cu metallogenesis in the Western Pacific Island
Anhaeusser, C.R., 1976, Archean metallogeny in southern Africa: Economic Arcs: Resource Geology, v. 54, p. 357–372.
Geology, v. 71, p. 16–45. Imai, A., Listanco, E.L., and Fujii, T., 1993, Petrologic and sulfur isotopic
significance of highly oxidized and sulfur-rich magma of Mt. Pinatubo, the oldest-known sial: Contributions to Mineralogy and Petrology, v. 94,
Philippines: Geology, v. 21, p. 699–702. p. 137–148, doi: 10.1007/BF00592931.
Imai, A., Listanco, E.L., and Fujii, T., 1996, Highly oxidized and sulfur-rich Ohmoto, H., and Goldhaber, M.B., 1997, Sulfur and carbon isotopes, in Barnes,
magma of Mount Pinatubo: Implication for metallogenesis of porphyry H.L., ed., Geochemistry of hydrothermal ore deposits (3rd edition): New
copper mineralization in the western Luzon arc, in Newhall, C.G., and York, John Wiley and Sons, p. 517–612.
Punongbayan, R.S., eds., Fire and mud: Eruptions and lahars of Mount Ohmoto, H., and Kerrick, D.M., 1977, Devolatilization equilibria in graphitic
Pinatubo, Philippines: Quezon City, Philippine Institute of Volcanology systems: American Journal of Science, v. 277, p. 1013–1044.
and Seismology: Seattle, University of Washington Press, p. 865–874. Oyarzun, R., Marquez, A., Lillo, J., Lopez, I., and Rivera, S., 2001, Giant ver-
Ishihara, S., 1977, The magnetite-series and ilmenite series granitic rocks: Min- sus small porphyry copper deposits of Cenozoic age in northern Chile:
ing Geology, v. 27, p. 293–305. Adakitic versus normal calc-alkaline magmatism: Mineralium Deposita,
Ishihara, S., 1981, Granitoid series and mineralization: Economic Geology, v. 36, p. 794–798, doi: 10.1007/s001260100205.
75th anniversary volume, p. 458–484. Panigrahi, M.K., and Mookherjee, A., 1997, The Malanjkhand copper (+molyb-
Ishihara, S., 1998, Granitoid series and mineralization in the Circum-Pacific denum) deposit, India: Mineralization from a low-temperature ore-fluid of
Phanerozoic granitic belts: Resource Geology, v. 48, p. 219–224. granitoid affinity: Mineralium Deposita, v. 32, p. 133–149, doi: 10.1007/
Ishihara, S., 2004, Redox state of granitoids relative to tectonic setting and s001260050080.
Earth history: The magnetite/ilmenite series 30 years later: Transactions Poujol, M., and Anhaeusser, C.R., 2001, The Johannesburg Dome, South Africa:
of the Royal Society of Edinburgh, Earth Sciences, v. 95, p. 23–33. New single zircon U-Pb isotopic evidence for early Archaean granite-
Ishihara, S., and Matsuhisa, Y., 1999, Oxygen isotopic constraints on the gen- greenstone development within the central Kaapvaal Craton: Precambrian
esis of the Miocene Outer Zone granitoids in Japan: Lithos, v. 46, p. 523– Research, v. 108, p. 139–157, doi: 10.1016/S0301-9268(00)00161-3.
534. Poujol, M., and Robb, L.J., 1999, New U-Pb zircon ages on gneisses and peg-
Ishihara, S., and Matsuhisa, Y., 2002, Oxygen isotopic constraints on the gene- matite from south of the Maurchison greenstone belts, South Africa:
sis of the Cretaceous-Paleogene granitoids in the Inner Zone of Southwest South African Journal of Geology, v. 102, p. 93–97.
Japan: Bulletin of the Geological Survey of Japan, v. 53, p. 421–438. Poujol, M., Robb, L.J., Anhaeusser, C.R., and Gericke, B., 2003, A review of
Ishihara, S., and Murakami, H., 2004, Granitoid types related to Cretaceous the geochronological constraints on the evolution of the Kaapvaal Craton,
plutonic Au-quartz vein and Cu-Fe skarn deposits, Kitakami Mountains, South Africa: Precambrian Research, v. 127, p. 181–213, doi: 10.1016/
Japan: Resource Geology, v. 54, p. 281–298. S0301-9268(03)00187-6.
Ishihara, S., and Sasaki, A., 2002, Paired sulfur isotopic belts: Late Cretaceous- Robb, L.J., and Anhaeusser, C.R., 1983, Chemical and petrogenetic character-
Paleogene ore deposits in the Inner Zone of Southwest Japan: Bulletin of istics of Archaean tonalite-trondhjemite gneiss plutons in the Barberton
the Geological Survey of Japan, v. 53, p. 461–477. Mountain Land: Geological Society of South Africa Special Publica-
Ishihara, S., Hashimoto, M., and Machida, M., 2000, Magnetite/ilmenite-series tion 9, p. 103–116.
classification and magnetic susceptibility of the Mesozoic-Cenozoic Robb, L.J., Anhaeusser, C.R., and van Nierop, D.A., 1983, The recognition of
batholiths in Peru: Resource Geology, v. 50, p. 123–129. the Nelspruit batholith north of the Barberton greenstone belt and its sig-
Ishihara, S., Sasaki, A., and Terashima, S., 1988, Sulfur in granitoids and its nificance in terms of Archaean crustal evolution: Geological Society of
role for the mineralization, in Proceedings of the 7th Quadrennial IAGOD South Africa Special Publication 9, p. 117–130.
(International Association of Genesis of Ore Deposits) Symposium: Stutt- Robb, L.J., Barton, J.M., Kable, E.J.D., and Wallace, R.C., 1986, Geology, geo-
gart, Schweizerbart’sche Verlag, p. 573–581. chemistry and isotopic characteristics of the Archean Kaap Valley pluton,
Ishihara, S., Anhaeusser, C.R., and Robb, L.J., 2002a, Granitoid-series evalu- Barberton Mountain Land, South Africa: Precambrian Research, v. 31,
ation of the Archean Johannesburg Dome granitoids, South Africa: Bul- p. 1–36, doi: 10.1016/0301-9268(86)90063-X.
letin of the Geological Survey of Japan, v. 53, p. 1–9. Sajona, F.G., and Maury, R.C., 1998, Association of adakites with gold and cop-
Ishihara, S., Anhaeusser, C.R., Robb, L.J., and Imai, A., 2002b, Granitoid- per mineralization in the Philippines: Comptes Rendus de l’Académie des
series in terms of magnetic susceptibility: A case study from the Bar- Sciences de Paris, Sciences de la Terre et des Planètes, v. 326, p. 27–34.
berton region, South Africa: Gondwana Research, v. 5, p. 581–589, doi: Sarkar, S.C., Kabiraj, S., Bhattacharya, S., and Pal, A.B., 1996, Nature, origin
10.1016/S1342-937X(05)70630-4. and evolution of the granitoid-hosted early Proterozoic copper-molybde-
Jones, C.B., 1990, Coppin Gap copper-molybdenum prospect, in Hughes, F.E., num mineralization at Malanjkhand: Mineralium Deposita, v. 31, p. 419–
ed., Geology of the mineral deposits of Australia and Papua New Guinea, 431, doi: 10.1007/s001260050049.
Volume I: Melbourne, Australasian Institute of Mining and Metallurgy, Sasaki, A., and Ishihara, S., 1979, Sulfur isotopic composition of the magnetite-
Monograph 14, p. 141–144. series and ilmenite-series granitoids in Japan: Contributions to Mineral-
Kamo, S.L., and Davis, D.W., 1994, Reassessment of Archean crustal develop- ogy and Petrology, v. 68, p. 107–115, doi: 10.1007/BF00371893.
ment in the Barberton Mountain Land, South Africa, based on U-Pb dat- Shanks, W.C., Bischoff, J.L., and Rosenbauer, R.J., 1981, Seawater sulfate
ing: Tectonics, v. 13, p. 167–192, doi: 10.1029/93TC02254. reduction and sulfur isotope fractionation in basaltic systems: Inter-
Lécuyer, C., and Ricard, Y., 1999, Long-term fluxes and budget of ferric iron: action of seawater with fayalite and magnetite at 200–350°C: Geochi-
Implication for the redox states of the Earth’s mantle and atmosphere: mica et Cosmochimica Acta, v. 45, p. 1977–1995, doi: 10.1016/0016-
Earth and Planetary Science Letters, v. 165, p. 197–211, doi: 10.1016/ 7037(81)90054-5.
S0012-821X(98)00267-2. Singer, D.A., Berger, V.I., Moring, B.C., 2002, Porphyry copper deposits of the
Lowe, D.R., 1994, Accretionary history of the Archean Barberton Greenstone world: Database, maps, and preliminary analysis: U.S. Geological Survey
Belt (3.55–3.22 Ga), southern Africa: Geology, v. 22, p. 1099–1102, doi: Open-File Report 02-268, 61 p.
10.1130/0091-7613(1994)022<1099:AHOTAB>2.3.CO;2. Taylor, H.P., Jr., 1968, The oxygen isotope geochemistry of igneous rocks:
Lowe, D.R., 1999, Geologic evolution of the Barberton greenstone belt and Contributions to Mineralogy and Petrology, v. 19, p. 1–71, doi: 10.1007/
vicinity, in Lowe, D.R., and Byerly, G.R., eds., Geologic evolution of the BF00371729.
Barberton Greenstone Belt, South Africa: Geological Society of America Thompson, R., and Oldfield, F., 1986, Environmental magnetism: London,
Special Paper 329, p. 287–312. Allen and Unwin, 227 p.
Meyer, C., 1985, Ore metals through geologic history: Science, v. 227, p. 1421– Tsuchiya, N., and Kanisawa, S., 1994, Early Cretaceous Sr-rich silicic magma-
1427. tism by slab melting in the Kitakami Mountains, northeast Japan: Journal of
Meyer, F.M., Robb, L.J., Reimold, W.U., and de Bruiyn, H., 1994, Contrasting Geophysical Research, v. 99, p. 22205–22220, doi: 10.1029/94JB00458.
low- and high-Ca granites in the Archean Barberton Mountain Land, south- Watanabe, Y., 2003, Porphyry copper deposits, in, Shikazono, N., Nakano, T.,
ern Africa: Lithos, v. 32, p. 63–76, doi: 10.1016/0024-4937(94)90021-3. and Hayashi, K., eds., Resource and environmental geology: Tokyo, Soci-
Nedachi, M., 1980, Chlorine and fluorine contents of rock-forming minerals of ety of Resource Geology, p. 35–44 [in Japanese].
the Neogene granitic rocks in Kyushu, Japan: Mining Geology, Special
Issue 8, p. 39–48.
Nutman, A.P., and Bridgwater, D., 1986, Early Archaean Amitsoq tonalites and
granites of the Isukasia area, southern West Greenland: Development of MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA

Memoir 198
2006
Secular variations of N-isotopes in terrestrial reservoirs

and ore deposits
R. Kerrich
Department of Geological Sciences, University of Saskatchewan, 114 Science Place, Saskatoon S7N 5E2, Canada
Y. Jia*
CSIRO Exploration and Mining, School of Geosciences, Monash University, P.O. Box 28, Victoria 3800, Australia
C. Manikyamba
S.M. Naqvi
National Geophysical Research Institute, Hyderabad 500 007, India
ABSTRACT
New δ15N analyses combined with a literature compilation reveal that shale
kerogen, VMS-micas, and late-metamorphic vein micas show a secular trend from
enriched values in the Archean, through intermediate values in Proterozoic terranes,
to the Phanerozoic mode of 3‰–4‰. Kerogen in metashales from the 2.7 Ga Sandur
Greenstone Belt, eastern Dharwar Craton, India, is characterized by δ15N 13.1‰ ±
1.3‰, and C/N 303 ± 93. A second population has δ15N 3.5‰ ± 0.9‰, and C/N 8 ±
0.4, close to the Redfield ratio of modern microorganisms, and is interpreted as pre-
cipitates of Proterozoic or Phanerozoic oilfield brines that penetrated the Archean
basement. Kerogen from 1.7 Ga carbonaceous shales of the Cuddapah Basin average
5.0‰ ± 1.2‰, close to the mode at 3‰–4‰ for kerogen and bulk rock of Phanerozoic
sediments. Biotites from late-metamorphic quartz-vein systems of the 2.6 Ga Kolar
gold province, E. Dharwar Craton, that proxy for average crust, are also enriched at
14‰–21‰ for three samples, confirming that the N–budget of the hydrothermal flu-
ids is dominated by sedimentary rocks. Muscovites from altered volcanic rocks in 2.7
Ga Abitibi belt VMS deposits have δ15N 12‰–20‰, in keeping with published data
for shale kerogen from the same terrane, whereas equivalents in the 1.8 Ga Jerome
VMS span 11.7‰–14.1‰.
15
N-enriched values in Precambrian rocks cannot be caused by N-isotopic shifts
due to metamorphism or Rayleigh fractionation because (1) pre-, and post-metamor-
phic samples from the same terrane are both enriched in 15N; (2) there is no covaria-
tion of δ15N with N, C/N ratios, or metamorphic grade; and (3) the magnitude of
fractionations of 1‰ (greenschist) to 3‰ (amphibolite facies) during progressive
metamorphism of sedimentary rocks, as constrained from empirical observations
and experimental studies, is very small. Nor can 15N-enriched values stem from long-
term preferential diffusional loss of 14N, as samples were selected from terranes where
40
Ar/39Ar ages are within a few million years of concordant U-Pb ages; nitrogen is
structurally bound in micas, whereas Ar is not.
It is possible that the 15N-enriched values stem from a different N-cycle in the
Archean, with large biologically mediated fractionations, yet the magnitude of the
*Corresponding author, present address: Silvercorp Metal Inc., 1588-609 Granville Street, Vancouver, B.C., V7Y 1G5 Canada; [email protected].
Kerrich, R., Jia, Y., Manikyamba, C., and Naqvi, S.M., 2006, Secular variations of N-isotopes in terrestrial reservoirs and ore deposits in Kesler, S.E., and Ohmoto,
H., eds., Evolution of Early Earth’s Atmosphere, Hydrosphere, and Biosphere—Constraints from Ore Deposits: Geological Society of America Memoir 198, p.
81–104, doi: 10.1130/2006.1198(05). For permission to copy, contact [email protected]. ©2006 Geological Society of America. All rights reserved.
81
82 R. Kerrich et al.
fractionations between atmospheric N2 and organic nitrogen observed exceeds any

presently known, and chemoautotrophic communities tend to depleted values. Earlier
results on Archean cherts show a range of δ15N from –6‰ to 30‰. Given the temporal
association of chert–banded iron formation (BIF) with mantle plumes, the range is
consistent with mixing between mantle N2 of −5‰ and the 15N-enriched marine reser-
voir identified in this study. The 15N-enriched Archean atmosphere-hydrosphere res-
ervoir does not robustly constrain Archean redox-state. We attribute the 15N-enriched
reservoir to a secondary atmosphere derived from CI-chondrite-like material and
comets with δ15N of +30‰ to +42‰. Shifts of δ15N to its present atmospheric value
of 0‰ can be accounted for by a combination of early growth of the continents with
sequestration of atmospheric N2 into crustal rocks, and degassing of mantle N ~−5‰.
If Earth’s surface environment became oxygenated ca. 2 Ga, then there were no asso-
ciated large N-isotope excursions.
Keywords: nitrogen, isotopes, kerogen, hydrothermal-systems, gold, VMS, Precam-

brian, Paleozoic, secular change.
INTRODUCTION AND SCOPE 104
Phs1
The present nitrogen cycle is quite well documented, includ-
ing N-isotope fractionations accompanying organic and inor- 103
Phs2
Phs3
ganic transfers of N between terrestrial reservoirs (Delwiche and Phs 4
Prs1 Ars1
Steyn, 1970; Macko et al., 1987; Rau et al., 1987; Williams et al.,
N (ppm)
Phs 5
1995; Kao and Liu, 2000). There are abundant data on modern 102 Phg1
Prs 2
organic compounds in the biosphere and sediments (Peters et al., Ars4 Phg2 Ars2
1978; Sweeney et al., 1978; Mazuka et al., 1991; Sadofsky and Ars5 Ars3
Arg1
Bebout, 2004, and references therein), and Phanerozoic rocks
101
(Haendel et al., 1986; Bebout and Fogel, 1992; Williams et al.,
1995; Mingram and Bräuer, 2001, and references therein). Hol-
loway and Dahlgren (2002) give a recent review. MORB
100 Upper Mantle
The isotopic compositions of N in the mantle (−5‰), organic -10 -5 0 5 10 15 20
compounds in sediments (0‰ to −6‰ with a mean of −4‰), δ15N (‰)
atmosphere (0‰), Phanerozoic black shales (kerogen, 3‰–4‰),
Figure 1. Nitrogen isotope compositions and concentrations in vari-
and granites (range 0 to10‰) are well constrained (Fig. 1). Isoto- ous geological reservoirs. Data represent mean value plus one standard
pic differences in terrestrial reservoirs stem from (1) nitrification, deviation from the following sources: Archean (Ars): shale kerogen,
denitrification, N-limitation, and other metabolic or inorganic Ars1 and Ars2 (Jia and Kerrich, 2004b) and Ars3 (Jia and Kerrich,
reactions of the near-surface N-cycle (Delwiche and Steyn, 1970; 2000); Meso- and Neoarchean chert kerogen, Ars4 and Ars5 (Beau-
Wada et al., 1975; Sweeney et al., 1978; Saino and Hattori, 1980; mont and Robert, 1999). Proterozoic shale kerogen (Prs): Prs1 (Jia and
Kerrich, 2004b) and Prs2 (Boyd and Philippot, 1998; Haendel et al.,
Macko et al., 1987; Rau et al., 1987; Hoch et al., 1994; Pinti 1986). Phanerozoic sediments and sedimentary rocks (Phs): Phs1 (Se-
and Hashizume, 2001; Lehmann et al., 2002); (2) equilibrium phton et al., 2002), Phs2 (Williams et al., 1995; Kao and Liu, 2000),
fractionation in the geosphere, for example fluid-rock interac- Phs3 (Bebout and Fogel, 1992; Busigny et al., 2003), Phs4 (Haendel et
tion (Hanschmann, 1981); and (3) kinetic effects associated with al., 1986; Mingram and Bräuer, 2001), and Phs5 (Peters et al., 1978).
mantle degassing (Marty and Zimmermann, 1999; Cartigny and Archean granitoids (Arg): Arg1 (Jia and Kerrich, 1999, 2000); Pha-
nerozoic granite (Phg): Phg1 (Boyd et al., 1993), Phg2 (Bebout et al.,
Ader, 2003, and references therein), or Rayleigh fractionation 1999). Mid-oceanic ridge basalt (MORB) source N (1–2 ppm and −5
during metamorphic devolatilization (Bebout and Fogel, 1992; ± 2‰) and upper mantle N (0.27 ± 0.16 ppm and −5 ± 2‰; Marty and
Mingram and Bräuer, 2001) (Fig. 1). Dauphas, 2003, and references therein).
However, little is known about the early evolution of the
N-cycle because records in Archean and Proterozoic rocks are
sparse and the oxidation state of the Archean atmosphere-hydro- tion (Hayes et al., 1983; Sano and Pillinger, 1990; Beaumont and
sphere system is uncertain (Ohmoto, 1997, 2004; Holland, 1999; Robert, 1999; Pinti et al., 2001a). N-isotope data on hydrother-
Phillips et al., 2001). Only a few data for N contents and N-iso- mal ore deposits, Precambrian or Phanerozoic, are limited (Jia
topes have been published on Archean shales (Zhang, 1988; Jia and Kerrich, 1999, 2000; Jia et al., 2001, 2003a).
and Kerrich, 2000, 2004a, 2004b), cherts and stromatolite-bear- Nitrogen-isotopes have been used to address a variety of
ing sediments (Gibson et al., 1985, 1986), and chert–iron forma- questions: (1) the isotopic composition of the Archean atmo-
Secular variations of N-isotopes in terrestrial reservoirs and ore deposits 83
sphere (Sano and Pillinger, 1990; Beaumont and Robert, 1999); have been described from carbonaceous cherts at the former local-
(2) secular variation of redox state of the atmosphere and oceans ity (Naqvi et al., 1987; Venkatachala et al., 1990). The eastern arc
(Beaumont and Robert, 1999); (3) bacterial metabolic pathways and western plateau terranes accreted post 2.6 Ga (Manikyamba
(Pinti and Hashizume, 2001; Pinti et al., 2001a); (4) origin of et al., 1997; Naqvi et al., 2002). Metamorphic grade varies from
Earth’s atmosphere-hydrosphere (Javoy, 1998; Tolstikhin and greenschist facies in the west to mid-amphibolite facies in the
Marty, 1998; Jia and Kerrich, 2004a, 2004b); (5) the N-isotope east (Manikyamba et al., 1997). Deformation is of low intensity
characteristics of Archean sedimentary rocks (Hayes et al., 1983; except proximal to faults or shear zones.
Zhang, 1988; Jia and Kerrich, 2000, 2004a, 2004b); (6) sources The intracratonic Cuddapah Basin developed over 1.9–1.7
of hydrothermal fluids involved in orogenic, or shear zone-hosted Ga (Fig. 2). Siliciclastic rocks are prevalent in this Proterozoic
mesothermal, gold deposits (Jia and Kerrich, 1999, 2000; Jia et sequence, including conglomerates, current-bedded and rippled
al., 2001, 2003a); (7) nitrogen budgets in convergent margins arenites, and diverse shales (Nagaraja-Rao et al., 1987). Well-
(Bebout and Fogel, 1992; Sadofsky and Bebout, 2004); and (8) preserved stromatolitic units and microfossils in cherts associ-
recycling of sedimentary rocks into the mantle (Marty and Dau- ated with stromatolites are present (Schopf and Prasad, 1978;
phas, 2003). Nagaraja-Rao et al., 1987). Samples of carbonaceous shales were
Several observations emerge from the limited database: (1) obtained from near Mangampeta and Marcapur. Metashales are
kerogen in Mesoarchean (3.4–2.9 Ga) cherts are 15N-depleted at prehnite-pumpellyite facies in the former locality, and at green-
relative to Neoarchean (2.9–2.5 Ga) counterparts; (2) kerogen in schist facies in the latter. The basin’s western margin lies uncon-
Archean black shales is 15N-enriched compared to Mesoarchean formably on Archean craton, whereas the eastern margin tectoni-
cherts and Phanerozoic equivalents; (3) the δ15N of hydrothermal cally underlies the Eastern Ghat Mobile Belt (EGMB), which was
K-micas in orogenic gold deposits of Phanerozoic accretionary thrust over the basin. From dating of anorthosite, alkali plutons,
terranes is comparable to that of contemporaneous shales; and and granitoids emplaced along the accretionary zone, accretion
(4) K-micas in Archean gold deposits are as enriched as contem- of the EGMB, which was part of Antarctica before the breakup
poraneous shale kerogen in 15N. Consequently, Archean cherts of Gondwanaland, to the Dharwar Craton is estimated to have
and shales may sample different N-reservoirs, and shales record occurred ca. 1600 Ma (Dasgupta and Sengupta, 2003). Carbona-
a secular variation of δ15N (Fig. 1). ceous shales were sampled from fresh rock in road or rail cuttings
In this paper, new N-isotope data are reported for 2.7 Ga and for both the Sandur belt and Cuddapah Group.
1.8 Ga carbonaceous shales from India to further test for secular
variations. The first data for 2.7 and 1.8 Ga volcanic hosted mas- Kolar Gold Province
sive base metal (Cu-Zn-Pb) sulfide (VMS) deposits are also pre-
sented, together with new data for hydrothermal K-micas from The Kolar terrane is one of several supracrustal greenstone
the ca. 2.6 Ga Kolar gold province, India. We compile existing sequences in the eastern Dharwar Craton. Tholeiitic basalts and
data together with the new results to evaluate the δ15N of vari- komatiites dated at 2.7 Ga are prevalent, at amphibolite facies.
ous lithologies through time. From this database we consider Gold mineralization was coeval with brittle-ductile deforma-
the origin of N in orogenic gold, rare metal pegmatite, and VMS tion (Hamilton and Hodgson, 1986) associated with accretionary
deposits, examine the implications of secular variations of δ15N, tectonics (Balakrishnan et al., 1999). Biotite-rich alteration sel-
and address the question of whether N-isotopes record “the great vedges bounding gold-bearing quartz-calcite veins were obtained
oxygenation event” ca. 2.3 Ga. from the Oriental reef (see Siddaiah and Rajamani, 1989, and ref-
erences therein). Nitrogen, as NH4+, may substitute for K given
GEOLOGICAL SETTING similar valence and ionic radius (Honma and Itihara, 1981; Bos
et al., 1988). Accordingly, K-silicates are preferred minerals for
Carbonaceous Shales N-isotope studies of hydrothermal ore systems.
Carbonaceous metashales were sampled from the 2.7 Ga VMS Deposits

Sandur Greenstone Belt (SGB), and the Paleoproterozoic Cud-
dapah Basin, India (Fig. 2). The Sandur belt is one of a series of VMS deposits reflect the conjunction of magmatic, hydro-
composite tectonostratigraphic supracrustal terranes, separated thermal, and sedimentary processes proximal to the seafloor. The
by granitoids, in the eastern Dharwar Craton (Manikyamba et al., Abitibi greenstone, composite, supracrustal terrane, dated to ca.
1997). The eastern SGB is dominated by arc-related tholeiitic to 2.7 Ga, is the largest and best-preserved greenstone belt, with
calc-alkaline flows. Minor sedimentary units include polymictic several VMS districts (for a review see Jackson and Fyon, 1991).
conglomerate, graywacke, and carbonaceous shales. Rhyolites Metamorphic grade is prehnite-pumpellyite to greenschist facies,
have zircon U-Pb ages of 2658 ± 14 Ma using the SHRIMP tech- and deformation intensity is low except proximal to regional
nique (Nutman et al., 1996). The western SGB is prevalently 2.7 shear zones. The VMS deposits formed on the seafloor during
Ga high-Mg basalts and komatiites. Shales were sampled from a hiatus of bimodal volcanism in several areas, notably Kidd
Vibhutigudda and Bhimangundi (Fig. 2). Fossil cyanobacteria Creek, Ontario, and the Noranda and Matagami districts of Que-
Figure 2. Simplified geological map of the Dharwar Craton showing the distribution of greenstone belts and shear zone complexes from the
western and eastern Dharwar Craton. (AJ) Ajjanahalli, (BA) Bababudan, (C) Chitradurga, (DH) Dharwar, (G) Gadag, (GD) Gadwal, (H) Hun-
gund, (HO) Holenarsipur, (HU) Hutti, (KA) Kadri, (KO) Kolar, (KU) Kudremukh, (MN) Mangalore, (N) Nellore, (NA) Narayanpet, (P) Pena-
kacherla, (R) Ramagiri, (RC) Raichur, (S) Sandur, (SH) Shimoga. Inset shows the location of the main map (modified after Sreeramachandra
Rao, 2001).
bec. Alteration of footwall volcanic rocks, by modified seawater- Arizona. Seafloor massive sulfides and mineralized breccias are
derived hydrothermal fluids, generated domains of muscovite and present, and “black smoker” chimneys are preserved. Muscovite-
chlorite bearing alteration (Franklin et al., 1981). Muscovite-rich rich, hydrothermally altered felsic volcanic rocks were sampled
samples were obtained from mafic volcanic flows in the footwall from the footwall. Metamorphic grade is greenschist facies
of the 2.7 Ga Kidd Creek VMS deposit, Ontario, and from fel- (Anderson et al., 1971; Sangster and Scott, 1976; Lindberg and
sic pyroclastic units subjacent to the Amulet deposit, Noranda, Gustin, 1987).
and Mattagami Lake deposit, Matagami, Quebec. Geological
relationships of Abitibi VMS deposits have been reviewed by SAMPLE DESIGN
Franklin et al. (1981), Bleeker et al. (1999), and Hannington et
al. (1999). Seawater-rock ratios in the footwall are known to be Studies of progressively metamorphosed sedimentary
large compared to possible mantle or magmatic contributions rocks show shifts of ~1‰ in δ15N from protoliths to greenschist
(Beaty and Taylor, 1982; Costa et al., 1983). facies counterparts, and ≤3‰ to amphibolite facies (Haendel et
The Jerome VMS deposit is located in the upper of two al., 1986; Bebout and Fogel, 1992; Mingram and Bräuer, 2001;
cycles of bimodal magmatism of the 1.8 Ga Ash Creek Group, Busigny et al., 2003; Jia, 2004). Alternatively, some authors
attributed variations of 27‰–36‰ in their data sets to shifts due rich (2000). Analytical precision (reproducibility, 1σ; n ≥ 3) is
to Rayleigh fractionation accompanying metamorphism (Beau- typically ~0.3‰ for kerogen δ15N, ~0.5‰ for muscovite δ15N,
mont and Robert, 1999; Pinti et al., 2001a). The former studies and ~0.3‰ for kerogen δ13C. The long-term reproducibilities
are all of siliciclastic sedimentary rocks whereas the latter are of for international nitrogen isotope standard materials are as fol-
kerogen from chert–banded iron formation (BIF). Consequently, lows: IAEA-N1 = 0.54 ± 0.07‰ (n = 15, accepted value 0.53‰);
differences between the studies could be sample or environment IAEA-N2 = 20.35 ± 0.08‰ (n = 10, accepted value 20.41‰);
dependent, rather than due to metamorphism, or alternatively and for the internal laboratory standard material BLN.SOIL, 5.20
they could reflect assumptions in Rayleigh modeling. ± 0.21‰ (n = 20, accepted value 5.15‰). Ten replicate analyses
Given the time-series and spatial association of chert–iron of “in house” muscovite separates of samples KAII-1 and CD11-
formation with volcanic sequences erupted from mantle plumes 21-1 yielded mean δ15N values of 19.4‰ ± 0.09‰ and 3.4‰
(Isley and Abbott, 1999; Condie et al., 2001), mantle N may be ± 0.06‰, respectively (Jia et al., 2003b). Nitrogen and carbon
incorporated into kerogen in chert–iron formation, but probably concentrations were obtained from each sample based on system
not in distal carbonaceous shales. Accordingly, we selected car- calibration using known standards. Blanks for this technique are
bonaceous shales distal from chert-BIF to obtain a marine bio- <0.075 μg N2 for routine runs of the types of samples analyzed in
genic signature. The magnitude of metamorphic shifts of δ15N this study. Isotope data are reported in standard δ-notation rela-
values was estimated by analysis of pre-metamorphic and late- tive to atmospheric N2 for nitrogen and to the Peedee Belemnite
metamorphic materials from the same terrane. Analyses of fine- limestone (PDB) standard for carbon.
grained hydrothermal K-micas from VMS deposits allow com-
parison with published data for kerogen in carbonaceous shales Isobaric Interferences and Modern Organic
from the same 2.7 Ga terrane. In turn, data from these two types Contamination
of pre-metamorphic samples are compared with late-metamor-
phic hydrothermal micas from the Kolar gold deposit, and with Studies show that isobaric interference of carbon monoxide
published data for micas from 2.7 Ga late-metamorphic gold (CO) may produce erroneous nitrogen isotope ratios. One percent
deposits. Alteration associated with gold deposits overprints CO in an analyte would cause ~7‰ errors in δ15N (Beaumont et
regional metamorphic fabrics and the deposits retain primary al., 1994). The analytical approaches used in this study have care-
fluid inclusions, equilibrium quartz-muscovite oxygen isotope fully eliminated any such isobaric interferences.
fractionations, and upper plateau 40Ar/39Ar ages. Accordingly the Contamination by organic nitrogen compounds during the
deposits have not been overprinted by a later metamorphic event preparation of the kerogen residue could change the initial iso-
(Kerrich and Cassidy, 1994; McCuaig and Kerrich, 1998). topic composition. However, samples were treated with dichlo-
romethane-ethanol, as in other studies of Precambrian rocks
ANALYTICAL METHODS (Beaumont and Robert, 1999). Consequently, the very positive
δ15N values of Archean samples cannot be attributed to contami-
Separation and Analysis nation because modern organic nitrogen compounds have δ15N
of close to 0‰, or negative down to −6‰ (Nadelhoffer and Fry,
Previous studies of Precambrian carbonaceous sedimentary 1988; Sachs and Repeta, 1999; Kao and Liu, 2000), and modern
rocks revealed low bulk rock N contents. Accordingly, kerogen kerogen ranges from 1‰ to 6‰ (Williams et al., 1995; Ader et
was separated for analysis of N and δ15N, revealing that kerogen al., 1998; Kao and Liu, 2000; Sephton et al., 2002). Also, sev-
N dominated the whole rock N-budget. Minor N is likely to be in eral samples have the extremely 13C-depleted values of –33‰ to
K-silicates (Hayes et al., 1983; Beaumont and Robert, 1999; Jia –48‰, characteristic of some Archean kerogen (Wellmer et al.,
and Kerrich, 2004b). Kerogen was separated using the following 1999); accordingly, significant contamination by modern hydro-
procedure. All selected sedimentary rock samples were washed carbons can be ruled out.
in dichloromethane-ethanol to remove modern organic contami-
nation, ground in a steel mortar to <50 μm, and then treated with Total Inorganic Carbon (TIC)/Total Organic Carbon (TOC)
HCl and HF to remove carbonates and silicates using the tech-
nique of Durand and Nicaise (1980). All kerogens in this study Total inorganic carbon, and total organic carbon were deter-
have been screened by X-ray diffraction (XRD); graphite peaks mined on carbonaceous shales using a Leco CR-12 carbon ana-
were at or below detection, in keeping with greenschist facies lyzer, following the procedure of Wang and Anderson (1994).
shale and the results of Landis (1971). Pure muscovite separates,
where N as NH4+ substitutes for K, from gold and VMS deposits RESULTS
were obtained by standard mineral separation procedures.
The analytical techniques used in this study involved a Data are reported in Table 1 and Figure 3. Also shown in
high-precision continuous flow-isotope ratio mass spectrometer Tables 2 and 3, for comparison, are compilations of recent data
(CF-IRMS) at the Soil Science Laboratory, University of Sas- on nitrogen contents and δ15N values of Archean to Phanerozoic
katchewan, and followed the techniques used by Jia and Ker- siliciclastic sediments, chert-BIF, granitoids, and ore deposits.
TABLE 1. NITROGEN AND CARBON ISOTOPIC COMPOSITIONS AND C/N ATOMIC RATIOS OF KEROGEN
FROM INDIAN PRECAMBRIAN CARBONACEOUS SHALES, AND FROM HYDROTHERMAL
K-MICAS VMS AND GOLD DEPOSITS
Era Sample no. N G15N C G13C C/N TOC/TIC
(ppm) (‰) (ppm) (‰) (atomic ratio)
†
Carbonaceous shales
Proterozoic
Mangambeta Locality (P3)
HCLCB1 454 5.6 15868 –30.3 41
HCLCB2 506 5.9 15322 –30.7 35
HCLCB3 419 5.5 14809 –29.9 41
HCLCB4 464 6.4 8756 –30.6 22
HCLCB6 370 6.8 11173 –31.0 35
Mean ± 1standard 442 ± 51 6.0 ± 0.6 13,185 ± 3085 –30.5 ± 0.4 35 ± 8
derivation
Marcapur locality (P3)
HCLCB8 289 3.8 26297 –28.6 106
HCLCB9 334 4.4 28111 –28.4 98
HCLCB10 415 4.1 46828 –28.5 132
HCLCB11 385 3.1 37070 –28.6 112
HCLCB12 313 4.0 8857 –27.5 33
Mean ± 1standard 347 ± 52 3.9 ± 0.5 297,430 ± 14,105 –28.3 ± 0.5 96 ± 37
derivation
Archean
Bhimangundi Locality (P2)
HCLC1 838 3.4 5690 –32.0 8 0.1505
HCLC4 847 3.1 5810 –32.0 8 0.1405
HCLC9 811 2.6 4980 –31.9 7 0.1598
HCLC10 867 4.7 5522 –32.0 7 0.1462
Mean ± 1standard 841 ± 24 3.5 ± 0.9 5500 ± 365 –32.0 ± 0.1 8 ± 0.4 0.149 ± 0.008
derivation
Vibutigudda Locality (P1)
HCLC11 88 14.6 19233 –28.3 256 0.0241
HCLC12 73 14.1 11197 –26.0 179 0.0309
HCLC14 80 13.0 17729 –26.3 259 0.0575
HCLC17 57 11.8 21165 –26.7 435 0.0591
HCLC19 64 13.4 20807 –28.1 380 0.0516
HCLC22 76 11.4 19952 –28.3 305 0.0451
Mean ± 1standard 73 ± 11 13.1 ± 1.3 18,345 ± 3710 –27.3 ± 1.1 303 ± 93 0.045 ± 0.014
derivation
VMS deposit
Archean
Kidd Creek deposit
K1441 15 18.6
K9943 10 16.3
Ansil deposit
91GIA 25 16.4
30EA 16 12.4
Matagami
M80-7 14 20.1
M80-8 35 18.4
M80-9 28 12.9
Proterozoic
Jerome
J92-A 47 12.7
J92-B 33 14.1
J92-C 28 11.7
J92-D 39 12.4
Gold deposits
Archean
Kolar
K432 18 19.4
K577 32 21.0
K616 47 13.7
†
N and C contents referenced to whole rock powders.
103 pret P2 data in terms of infiltration of Proterozoic or Phanerozoic

hydrocarbon-bearing formation brines locally into Archean base-
ment, forming a secondary hydrocarbon signature. Penetration of
102
Sedimentary kerogen Proterozoic and Phanerozoic formation brines, and Quaternary
N (ppm)
Archean
groundwaters, into Archean crust has been documented for sev-
Proterozoic
VMS deposits
eral cratons (for a review see Kerrich and Ludden, 2000).
101 Archean
Proterozoic Hydrothermal K-micas from Volcanogenic Massive Sulfide
Gold deposits (VMS) and Gold Deposits
Archean
100
0 5 10 15 20
The δ15N values of hydrothermal biotite from 2.6 Ga quartz
vein systems at Kolar are 14‰–21‰ and overlap the range for
δ15N ( )
P1 kerogen in the eastern Dharwar Craton; both sets of samples
Figure 3. Variations of N contents and nitrogen isotopic compositions of are enriched in 15N compared to Proterozoic or Phanerozoic N-
Indian Precambrian carbonaceous shales and the Kolar gold deposit, and reservoirs (Tables 1, 3; Figs. 1, 3). K-micas generated by modi-
Precambrian VMS deposits. Uncertainties for δ15N values and N content fied seawater-derived hydrothermal fluids in 2.7 Ga Abitibi
are all smaller than the plot symbols. Data sources are in Table 1.
VMS deposits are enriched in 15N at Kidd Creek (16‰ to19‰),
Amulet (12‰–16‰), and Matagami (13‰–20‰), overlapping
the range of δ15N values of kerogen in metashales from this ter-
rane. All micas from the Archean and Proterozoic VMS deposits
Precambrian Kerogen have low N contents of 10–25 ppm (Tables 1, 3; Fig. 3). In the
1.8 Ga Jerome VMS deposit, four δ15N values of muscovite are
Carbonaceous shales from the Archean Dharwar Craton have 12‰–14‰, intermediate between Phanerozoic kerogen or bulk
two compositional-isotopic populations. The six kerogens from sedimentary rocks, and Archean VMS counterparts or kerogen
Vibutigudda (P1) are characterized by relatively uniform N (73 ± (Tables 2, 3).
11 ppm) and C (18,345 ± 3710 ppm) contents, with C/N ratios of
179–435. The δ15N values (12‰ to 15‰) are much higher than COMPARISON WITH THE LITERATURE
those of most Phanerozoic sedimentary rocks (1‰–6‰; Table
3). Four kerogens from Bhimangundi define a second population Mantle and Diamonds
(P2); they feature greater N (841 ± 24 ppm) but lower C (5500 ±
365 ppm) contents, giving uniform C/N ratios of 7–8. Both the From worldwide sampling of diamonds, Boyd et al. (1987,
δ15N values (3.5‰ ± 0.9‰) and δ13C (−32.0‰ ± 0.1‰) are lower 1992) reported negative δ15N values (−8.7‰ to −1.7‰) with
than counterparts from Vibutigudda (Table 1). Proterozoic carbo- a mode of −6‰ to −5‰. Independently, Cartigny et al. (1997,
naceous shales constitute a third population (P3). Their average 1998) showed that mantle δ15N values of diamonds were between
N (395 ppm) and C (21,309 ppm) contents are greater than in −8‰ and −5‰ (see also Marty and Humbert, 1997; Marty and
P1, but average C/N ratios (73) and δ15N are lower (Table 1, Fig. Zimmermann, 1999; Javoy and Pineau, 1991). The diamonds are
3). Shale kerogens at the Mangampeta locality are enriched by mostly Archean, as old as 3.2 Ga (Richardson et al., 1984, 2001).
~2‰ relative to greenschist facies equivalents at Macapur. Given Accordingly, upper mantle δ15N may have been uniform at about
shifts of ~1‰ from sedimentary protoliths to greenschist facies −6 ± 1‰ since ~3.2 Ga (Fig. 1). However, lower δ15N values of
in several studies (see Discussion section), the 2‰ difference down to −20‰ were found in some rare diamonds, signifying
may reflect primary variations of δ15N. another N reservoir (Cartigny et al., 1997).
P2 “Archean” samples are characterized by C/N values Marty (1995) estimated the N content of undegassed mid-
close to the Redfield ratio, averaging 6.6 (C/N = 106:16) of mod- oceanic ridge basalt (MORB) to be 1–2 ppm; these values are
ern organic compounds (Chen et al., 1996; Fraga et al., 1998). In inferred from (1) observed covariations of N2/36Ar, 40Ar/39Ar, and
contrast, P1 and most Archean kerogens are characterized by C/ 3
He/4He in MORB; (2) an assumption that nitrogen, like He and
N spanning 35–600 (Hayes et al., 1983; Sano and Pillinger, 1990; CO2, behaves as an incompatible element during partial melting
Beaumont and Robert, 1999; Jia and Kerrich, 2004b). P2 sam- of rocks; and (3) the mantle carbon content, which is ~400 ppm.
ples also feature the conjunction of less-dispersed compositional The upper mantle would then have an N content of ~0.16 ppm
and isotopic values than the other two Precambrian populations (Porcelli and Turekian, 2004). According to Cartigny and Ader
(P1 and P3), with total organic carbon (TOC) to total inorganic (2003), uncertainties in these estimates for MORB may arise
(TIC) ratios greater than P1 (Table 1). The data plot to the lower from isotopic fractionation during partial degassing of basaltic
end member of the colinear array, in δ15N vs. δ13C coordinates, rocks. Alternatively, the mantle N content is estimated at ~40
of organic compounds in modern marine sediments of Peters et ppm from δ13C-N systematics for diamond, assuming that N is
al. (1978). Considering these lines of evidence together, we inter- not highly incompatible (Cartigny et al., 2001).
TABLE 2. SUMMARY OF N-ISOTOPIC COMPOSITIONS OF SEDIMENTARY ROCKS
Era Age N (ppm)† G15Nair (‰)† Corg (ppm) C/N ratio Refs.
Sample location (Ma) range mean ± 1V range mean ± 1V range mean ± 1V range mean ± 1V
Archean
S. Africa, Canada and W. Australia‡ 3500–2600 3–33 27 ± 27 (22) –4.8–29.9 3.6 ± 6.9 (22) 23–1438 457 ± 493 (22) 53–454 185 ± 117 (22) 1
Cherts or dolomitic cherts, S. Africa and W. Australia‡ 3500–2500 1.04–1.89 (Pmole) 1.31 ± 0.33 (5) 0.8–3.2 2.3 ± 1.0 (5) 114–260 (Pmole) 187 ± 54 (5) 90–250 150 ± 64 (4) 2
Shales, Western Australia‡ 2500 1.18 (Pmole) 5.7 125 (Pmole) 148 2
West Greenland and Western Australia 3800–2800 0.7–13.4 2.7 ± 3.7 (21) –3.8–12.2 6.7 ± 4.4 (12) 3
South Africa and Western Australia 3300–2600 0.7–13.4 –0.1–1.5 4
Vibutigudda locality, India‡ 2700 57–88 75 ± 11 (6) 11.4–14.6 13.1 ± 1.3 (6) 11,200–21,165 18,345 ± 3,710 (6) 179–435 305 ± 95 (6) 5
Abitibi greenstone belt, Canada‡ 2700 15–51 28 ± 13 (8) 11.8–17.3 15.3 ± 1.8 (8) 14,610–242,500 119,920 ± 81,550 (6) 51–928 390 ± 385 (6) 6
Abitibi greenstone belt, Canada‡ 2700 97–310 245 ± 71 (10) 14.8–17.9 16.0 ± 1.7 (10) 7
Penhalonga Formation, Botswana‡ 2700 26–60 39 ± 14 (5) 15.7–19.9 17.3 ± 1.9 (5) 311–1,811 807 ± 671 (5) 134–685 285 ± 230 (5) 7
Proterozoic
Ashanti belt, Ghana, West Africa‡ 2200–2100 140–575 300 ± 147 (11) 9.3–12.6 10.8 ± 1.1 (11) 76,790–107,330 90,245 ± 11,470 (8) 208–698 450 ± 230 (8) 7
Cherts or dolomitic cherts, S. Africa and W. Australia‡ 2000–800 0.90–2.85 (Pmole) 1.80 ± 0.66 (11) 2.8–9.9 5.7 ± 2.1 (11) 57–203 (Pmole) 133 ± 45 (11) 40–200 83 ± 45 (11) 2
Shales, Australia‡ 2000–800 1.86–3.79 (Pmole) 2.62 ± 0.77 (6) 3.0–8.5 4.8 ± 2.4 (6) 90–155 (Pmole) 134 ± 24 (6) 40–83 54 ± 17 (6) 2
Canada and Australia‡ 2100–700 3–106 13 ± 11 (11) 2.1–7.7 3.8 ± 1.8 (11) 7–348 91 ± 95 (11) 31–148 93 ± 38 (11) 1
Mangambeta and Marcapur localities, India‡ 1900–1700 289–506 395 ± 70 (10) 3.1–6.8 5.0 ± 1.2 (10) 8,755–46,830 21,310 ± 12,880 (10) 22–132 65 ± 40 (10) 5
Moine succession, Scotland 1500–1025 140–422 307 ± 99 (7) 8.4–16.6 13.7 ± 2.6 (7) 8
Tetsa Formation, British Columbia, Canada‡ 1600 570–590 580 ± 20 (5) 3.6–4.5 4.0 ± 0.4 (5) 16,000–18,200 17,220 ± 880 (5) 32–38 35 ± 3 (5) 7
Sachsisches Erzgebirge, Germany Neoproterozoic 35–115 55 ± 30 (9) 6.8–17.0 11.3 ± 4.0 (9) 9
Phanerozoic
Sachsisches Erzgebirge, Germany Ordovician 14–650 297 ± 207 (32) 3.2–10.8 3.7 ± 1.9 (32) 9
Erzgebirge schists, Germany Ordovician 76–896 282 ± 230 (30) 1.2–10.5 3.9 ± 2.2 (30) 10
Anthracites, Germany and Pennsylvania‡ Carboniferous 2.7–5.1 4.0 ± 0.8 (18) 11
Black shales, British Columbia, Canada‡ Jurassic 400–9300 –0.6–5.0 2.8 ± 1.9 (17) 1,600–40,700 16,170 ± 11,245 (17) 1–14 7 ± 4 (17) 12
Catalina Schist, California, U.S.A. Cretaceous 30–1075 390 ± 293 (46) 1.0–5.9 3.0 ± 1.2 (46) 13
Louisiana sandstone, U.S.A. Cretaceous 1600 ± 100 (64) 3.1 ± 1.4 (51) 14
Louisiana sandstone, U.S.A.‡ Cretaceous 100–3100 2.6–3.6 3.2 ± 0.3 (26) 14
Sediments, North America Tertiary 2.4–9.9 6.1 ± 1.8 (55) 15
Sediments, Izu-Bonin-Mariana margin Tertiary 18–661 277 ± 154 (36) –0.9–8.2 4.7 ± 1.7 (36) 16
Sediments, western Alps, Europe Tertiary 169–1721 586 ± 490 (16) 2.6–4.8 3.6 ± 0.7 (16) 17
Arctic Ocean sediments Tertiary 2,100–17,800 1,105 ± 1080 (37) 4.5–7.7 6.1 ± 0.8 (37) 2,600–18,000 9,705 ± 4,695 (37) 4–10 7 ± 2 (37) 18
Taiwan watershed, China ‡ Tertiary 700–800 800 ± 100 (5) 3.9–4.0‡ 3.9 ± 0.1 (5) 3,700–4,500 3,980 ± 340 (5) 5–7 6 ± 1 (5) 19
Note: Refs: 1—Beaumont and Robert (1999); 2—Hayes et al. (1983); 3—Sano and Pillinger (1990); 4—Pinti et al. (2001); 5—this study; 6—Jia and Kerrich (2000); 7— Jia and Kerrich (2004a,b); 8—Boyd and Philippot
(1998); 9—Haendel et al. (1986); 10—Mingram and Bräuer (2001); 11—Ader et al. (1998); 12—Sephton et al. (2002); 13—Bebout and Fogel (1992); 14—Williams et al. (1995); 15—Peters et al. (1978); 16—Sadofsky and
Bebout (2004); 17—Busigny et al. (2003); 18—Schubert and Calvert (2001); 19—Kao and Liu (2000).
†
Numbers in the parentheses are the numbers of analyzed samples.
‡
N-isotopic compositions were determined on kerogens isolated from sedimentary rocks.
TABLE 3. SUMMARY OF N-ISOTOPIC COMPOSITIONS OF HYDROTHERMAL MICAS

FROM OROGENIC GOLD DEPOSITS AND VMS DEPOSITS
Era Age N (ppm)† G15Nair (‰)† Refs.
Sample location (Ma) range mean ± 1V range mean ± 1V
Archean
Norseman terrane, Western Australia‡ 2700 20–70 32 ± 20 (12) 10.9 to 23.7 17.3 ± 4.3 (12) 1
Abitibi greenstone belt, Canada§ 2700 10–35 20 ± 9 (7) 12.4 to 20.1 16.5 ± 2.9 (7) 2
Abitibi greenstone belt, Canada‡ 2700 20–205 95 ± 72 (40) 10.1 to 21.0 16.3 ± 2.9 (40) 1, 3
Harare greenstone belt, Zimbabwe‡ 2700 22–25 23 ± 1.5 (4) 17.7 to 23.4 20.7 ± 3.0 (4) 4
Dharwar, South India‡ 2500 18–47 32 ± 15 (3) 13.7 to 21.0 18.0 ± 3.8 (3) 2
Proterozoic
Ashanti belt, Ghana, West Africa‡ 2200–2100 1160–1900 1655 ± 250 (9) 9.1 to 12.0 10.2 ± 1.5 (9) 4
Trans-Hudson orogen, Canada‡ 1800 105–120 112 ± 6 (6) 6.6 to 7.4 7.2 ± 0.4 (6) 4
Jerome, USA§ 1800 1028–47 37 ± 8 (4) 11.7 to 14.1 12.7 ± 1.0 (4) 2
Moine succession, Scotland* 1500–1025 435–1739 1048 ± 520 (21) 7.4 to 16.2 13.6 ± 2.3 (21) 5
Phanerozoic
‡
Lachlan fold belt, SE Australia Ordovician 861–895 733 ± 103 (20) 2.8 to 4.5 3.5 ± 0.4 (20) 6
Western Qilian orogen, North China‡ 210 to 225 650–2510 260 ± 760 (8) 1.7 to 6.6 4.0 ± 2.0 (8) 4
North American Cordillera‡ 190 to 50 130–3500 1535 ± 1080 (100) 1.6 to 6.1 3.0 ± 1.2 (100) 7
Western Alps, Italy ‡ Tertiary 280–290 288 ± 5 (4) 4.5 to 5.5 4.9 ± 0.6 (4) 4
Note: Refs: 1—Jia and Kerrich (1999); 2—this study; 3—Jia and Kerrich (2000, 2004b,c); 4—Jia and Kerrich (2002); 5—Boyd
and Philippot (1998); 6—Jia et al. (2001); 7—Jia et al. (2003a).
†
Numbers in the parentheses are the numbers of analyzed samples.
‡
Hydrothermal micas from orogenic gold deposits.
§
Hydrothermal micas from VMS deposits.
*Micas from amphibolite facies sedimentary rocks.
Siliciclastic Sediments low as −2.7‰. In contrast, intense denitrification, characterized

by a kinetic isotope effect, may generate δ15N values as high as
Phanerozoic and Modern Sediments 19‰ (Cline and Kaplan, 1975; Sweeney et al., 1978; Rau et al.,
Sedimentary rocks have similar bulk nitrogen isotopic com- 1987). For example, Mazuka et al. (1991) report δ15N values of
positions as indigenous kerogen, consistent with the bulk rock N 0.9‰–18.9‰, with a mode of 8‰, for Pliocene and younger
budget being dominated by kerogen N. The total range of δ15N marine sediments at ODP site 724.
in Phanerozoic bulk sedimentary rocks is 1.0‰–10.8‰ (Fig. 1, The principal process that controls changes in isotope com-
Table 2). However, the averages of most rock data sets cluster position of organic nitrogen is thermal decomposition of SOC
between 3‰ and 5‰, within one standard deviation, the global during diagenesis. Preferential release of 14N, given lower energy
average being 3.65‰ ± 0.55‰ (Table 2). Similarly, the total to break 14N-12C than 15N-12C bonds, shifts the residual kerogen
range of kerogen is –2.0‰ to 6.0‰, but averages of each data to ~3‰ (Delwiche and Steyn, 1970; Wada et al., 1975; Macko
set cluster between 1.5‰ and 4.6‰, and the global average is et al., 1987), implying a flux of 15N depleted nitrogen back to the
3.7‰ (Table 2). surface. Wada et al. (1975) and Sweeney et al. (1978) also consid-
Recent marine sediments have a δ15N range of −3‰–9‰, ered that denitrification may be responsible for nitrogen isotope
with an average of 3.9‰ ± 1.1‰ (Table 2; Peters et al., 1978; fractionation between atmospheric and organic nitrogen in kero-
Sweeney et al., 1978; Rau et al., 1987; Williams et al., 1995; gens, because denitrification, a bacterial process, involves pref-
Schubert and Calvert, 2001; Sadofsky and Bebout, 2004). Varia- erential loss of 14N to the atmosphere, leaving nitrate enriched in
tions in δ15N have been attributed to varying mixtures of terrestri- 15
N; nitrate is the source of marine organic matter, which retains
ally derived (13C and 15N depleted) and enriched marine-derived its isotopic signature in marine sedimentary rocks. Williams et
organic compounds, given a linear correlation between δ15N and al. (1995) report two stages of N production during progressive
δ13C (Peters et al., 1978; Sweeney et al., 1978; Minoura et al., burial: from microbiological activity during diagenesis, and then
1997). Alternatively, according to Sadofsky and Bebout (2004) release of N in the “oil window” when NH4+ becomes incorpo-
variations may stem from some combination of changes in bio- rated into illite during the smectite-illite transition. The remaining
productivity, complex diagenetic processes, and differing pro- organic matter as kerogen in sedimentary rocks increases in δ15N
portions of marine and terrestrial organic matter in continental relative to precursors. Freudenthal et al. (2001) report a decrease
margin versus intraoceanic arc settings. from 1000 to 250 ppm N during diagenesis of recent sediments
A larger range of N-isotope compositions has been recorded accompanied by a shift of 1.9‰ by Rayleigh fractionation.
from specific niches. δ15N values of –12‰ to 4‰ are docu- Fractionations between kerogen N (3.2‰ ± 0.3‰) and fixed
mented from chemoautotrophic bacteria in warm seeps on the N in mudstones (3.0‰ ± 1.4‰) are small (Williams et al., 1995).
seafloor (Conway et al., 1994). Reduced nitrate availability may Given that Phanerozoic upper crust is dominated by recycled
generate sedimentary organic compounds (SOC) with δ15N as sedimentary rocks (Taylor and McLennan, 1985), and the kero-
gen-sediment fractionation is insignificant, bulk upper crust has a and Kerrich, 2000). Two other ca. 2.7 Ga Archean data sets on
nitrogen isotope composition of 3‰–4‰, indistinguishable from unweathered rocks in drill core from the greenschist facies Pen-
kerogen (see Fig. 1 for references). halonga Formation, Botswana, and western Abitibi Greenstone
In summary, notwithstanding extreme δ15N values in spe- Belt also have overlapping enriched averages δ15N of 17.3‰ ±
cial niches, sedimentation, maturation, and diagenesis clearly 1.9‰, and 16.0‰ ± 1.7‰, respectively (Table 2; Figs. 1, and
generate kerogen with a restricted range of values (Figs. 1, 4E, 2A) (Jia and Kerrich, 2004b, 2004c). Contemporaneous samples
4G; Table 2). The most robust comparison for this study is of of this study from the Dharwar Craton endorse systematically
Precambrian carbonaceous shales with modern and Phanerozoic 15
N-enriched sedimentary kerogen in Neoarchean carbonaceous
counterparts (Tables 1 and 2). shales (Table 1, Fig. 1).
Proterozoic Precambrian Chert–Iron Formation

Carbonaceous shales sampled in the 2.2 Ga Paleoprotero-
zoic lower greenschist facies Birimian volcanic-sedimentary ter- Beaumont and Robert (1999) analyzed nitrogen concentra-
rane, Ghana, have δ15N values ranging from 9.3 to12.6‰ (Table tions and isotopic compositions of kerogen from Precambrian
2, Fig. 4C) (Jia and Kerrich, 2004b). Boyd and Philippot (1998) cherts, ranging in age from 3.5 to 0.7 Ga. They found a very large
reported bulk rock values of 8.4–16.6 ‰ for the mid-amphibolite range of δ15N values, from –4.7‰ to 30.0‰, with generally low
facies Moine succession, Scotland, deposited between 1.5 and 1.0 bulk rock N contents of 2–106 ppm. There is a general increase
Ga (Table 2). Micas separated from these samples yield similar of δ15N values from −4.7‰ to 5.9‰ in the Mesoarchean (3.5 to
isotopic compositions of 7.4‰–16.2‰. Assuming an oxidized 2.9 Ga), through –2.5‰ to 30.0‰ in the Neoarchean (2.8 to 2.5
atmosphere-hydrosphere system by this time in the Mesoprotero- Ga), to 2.1‰ to 7.7‰ in the Proterozoic (<2.5 to 0.6 Ga). They
zoic, and therefore contemporaneous organic matter similar to concluded that Archean atmospheric N2 had a similar isotopic
δ15N in modern marine sediments of −3‰ to 9‰ (Peters et al., composition to the present, and that the secular change of δ15N
1978; Sweeney et al., 1978; Rau et al., 1987; Williams et al., records the “great oxygenation event” (Fig. 4A).
1995; Schubert and Calvert, 2001; Sadofsky and Bebout, 2004), Using a step-heating procedure, Pinti et al. (2001a) docu-
they concluded that metamorphism to amphibolite facies had mented δ15N ranging from −7‰ to 20‰ in Archean chemical
induced a shift of ≥ 8‰. However, if shifts to amphibolite facies sedimentary rocks (3.8–2.8 Ga). Nitrogen released at low-tem-
are ≤3‰ (see Discussion, below), primary values may have been perature steps, in all samples, had negative or near present-
3‰–12‰. Haendel et al. (1986) documented 15N-enriched values day atmospheric δ15N, which was interpreted to reflect mod-
for the Proterozoic Saschsisches Erzebirge, Germany (11.3‰; ern atmospheric or/and organic compound contamination; on
Table 2). The 1.6 Ga Tetsa samples, and Cuddapah kerogen data the other hand, for all but one sample the nitrogen extracted
of this study have δ15N values that are indistinguishable from the from high-temperature steps showed positive δ15N, with most
Phanerozoic data set. In summary, three of the five Proterozoic between 2‰ and 20‰; enriched values were modeled as sec-
data sets are enriched in 15N compared to Phanerozoic counter- ondary shifts by Rayleigh volatilization during metamorphism
parts (Table 2; Figs. 1, 4C). (Pinti et al., 2001a).
Archean Granitoids
Hayes et al. (1983) reported an extensive database on
Precambrian sedimentary rocks spanning 3.8–0.8 Ga, includ- Phanerozoic
ing X-ray characteristics, H and C contents, and δD and δ13C In a study of the Cornubian batholith of southwest Eng-
values. They reported N contents and δ15N on a subset of ten land, Boyd et al. (1993) reported N concentrations of 8–187
samples 3.4 to 1.6 Ga in age, of which five are cherts mostly ppm and δ15N values in the range of 5.2‰–10.2‰. According
associated with BIF, two are shales, and three are stromatolitic to Bebout et al. (1999), the early Devonian Skiddaw peralumi-
dolomites. δ15N values range from 0.8‰–5.7‰. Chert kerogen nous granite in the English Lake District has nitrogen contents
has a similar range of δ15N values as in the larger database for of 49 ± 27 ppm and δ15N values of 3.5‰ ± 1.1‰ (n = 7). Given
Precambrian cherts of Beaumont and Robert (1999). Kerogen that Phanerozoic peraluminous granites have metasedimen-
N in chert could be mixtures of mantle N having δ15N of ~−5‰ tary precursors (Hawkesworth and Kemp, 2004), these results
with enriched oceanic N. are consistent with melting of sedimentary rocks having initial
Zhang (1988) documented δ15N ranging from 2‰ to 39‰ δ15N of 2‰–6‰, which is shifted by ~3‰ during metamor-
in Archean carbonaceous metashales, the most enriched values phism prior to partial melting (Table 2).
being in greenschist facies shales from the 2.6 Ga Ventersdorp
Group, South Africa (Hayes et al., 1983). Greenschist facies Archean
carbonaceous shales in the 2.7 Ga Archean Abitibi Greenstone Sparse data for 2.7 Ga tonalitic rocks in the Uchi subprov-
Belt are characterized by variably enriched δ15N values from ince, Superior Province, range from −5.3 to +5.2‰, averaging
12‰, with an average of 15.3‰ ± 1.8‰ (Table 2, Fig. 4A) (Jia −0.9‰ (Jia and Kerrich, 1999, 2000). The Archean tonalite-
Figure 4. Histograms showing variations in δ15N of sedimentary and/or metasedimentary rocks (left side: A, C, E, and G) and hydrothermal micas
(right side: B, D, F, and H) of Archean to Phanerozoic age. The figure displays, except for cherts, common secular evolution of both sedimentary
rocks and hydrothermal micas. Data sources are in Tables 1–4, and published data.
trondhjemite-granodiorite (TTG) suite is considered to have Hydrothermal micas in Phanerozoic gold quartz veins have
formed by partial melting of basaltic oceanic crust on a sub- δ15N values similar to those in contemporaneous sedimentary
ducting slab (Drummond and Defant, 1990; Martin, 1999; rocks (Tables 1–3, Fig. 4). This result is in keeping with the bulk
Smithies, 2000). Archean diamond data indicate that Archean mid to upper crustal N budget being dominated by sedimen-
oceanic crust may have had a δ15N similar to that of mod- tary N. Crustal sedimentary rocks contain (0.22 ± 0.10) × 1020
ern MORB (averaging −5‰). Secondary isotopic shifts could mol N2, mostly in shales where N abundances are hundreds to
arise from some combination of seawater alteration and addi- thousands ppm. The mass of N is comparable in lower crustal
tion of carbonaceous sediments, which are unconstrained, fol- igneous rocks (Zhang and Zindler, 1993, and references therein)
lowed by a shift of ≤3‰ during subduction metamorphism (Table 5)
prior to partial melting. Nitrogen contents of the tonalites at Averages of δ15N values for deposits in accretionary oro-
5–27 ppm are comparable with Phanerozoic metaluminous genic belts are as follows: 3.5‰ ± 0.4‰ for the Paleozoic Lach-
granitoids (Hall, 1999), implying similar concentrations of N lan fold belt in southeastern Australia; 3.0‰ ± 1.2‰ for the
in the source. western North American Cordillera, which hosts many Juras-
sic–early Tertiary gold-bearing quartz vein systems from the
Ore Deposits Mother Lode of southern California to Alaska; 4.0‰ ± 2.0‰
for Mesozoic quartz veins in the western Qilian orogen, North
Orogenic Gold China; and 4.9‰ ± 0.6‰ for the middle to late Tertiary Monte
Orogenic, or mesothermal, lode gold deposits constitute a Rosa quartz veins in the Alpine orogen (Table 2; Figs. 4F, 4H)
distinct class of Au- and Ag-rich structurally hosted vein sys- (Jia and Kerrich, 2004b).
tems. They developed post peak-metamorphism, late in the δ15N values for hydrothermal micas from gold-bearing
development of accretionary orogenic belts. This deposit type quartz veins in the 2.1 Ga Ashanti belt of Ghana, and the 1.8
is common in Neoarchean greenstone belts and Paleoprotero- Ga Trans-Hudson orogen, Canada, are 7‰–10‰, intermediate
zoic terranes, and formed continuously through the Phanerozoic, between Archean and Phanerozoic counterparts (Table 3, Fig.
including large metallogenic provinces in the lower Paleozoic 4D) (Jia and Kerrich, 2004b).
Lachlan orogen and Jurassic–Tertiary of the North American For the Archean Superior Province, Jia and Kerrich (1999,
Cordillera. The vein systems extend tens of kilometers laterally, 2000) reported data for hydrothermal micas from nine 2.7 Ga
and up to 6 km vertically (McCuaig and Kerrich, 1998; Goldfarb deposits (Abitibi belt: Kerr-Addison, Dome, Hollinger, Gold-
et al., 2001). hawk, Beaumont; Wawa belt: Hemlo; Geraldton-Beardmore
Muscovite, or more rarely biotite, is abundant as an altera- belt: Geraldton, Pickle Crow; Red Lake belt: Red Lake); the total
tion phase, and therefore a candidate for determining the N-iso- range of δ15N values is 11.8‰–21.0‰, with a mean of 16.3‰ ±
tope composition of the hydrothermal fluids and source reser- 3.0‰ (Table 3, Fig. 2B). Hydrothermal micas from the Norse-
voir. According to H, C, and O-isotope data, and K-Cs-Rb-Ba man terrane, Western Australia, and Harare Greenstone Belt,
systematics, the veins precipitated from metamorphic fluids. Zimbabwe, are also characterized by 15N-enriched values, aver-
Given a “deep later” P-T-t path where peak-metamorphism aging 17.3‰ ± 4.3‰ and 20.7‰ ± 3.0‰, respectively. Includ-
migrated down through the crust, metamorphic fluids were ing the 15N-enriched data for the Kolar deposit of this study, high
probably generated syn peak-metamorphism at mid-crustal lev- δ15N values of late-metamorphic hydrothermal micas have been
els, and advected to shallow crustal levels where vein minerals recorded on four Neoarchean cratons (Tables 1, 3; Fig. 4B).
precipitated post peak-metamorphism (McCuaig and Kerrich,
1998). In the 2.7 Ga Abitibi Greenstone Belt, the deposition of VMS Deposits
sedimentary-volcanic host rock sequences was temporally sepa-
rated from precipitation of the post-metamorphic hydrothermal δ15N values of micas in the 1.8 Ga Jerome deposit are within
micas by only 30 m.y. (Corfu et al., 1989; Kerrich and Cassidy, the range for Proterozoic siliciclastic rocks and kerogen. Hall
1994). (1989) documented an increase of N content from 1 ppm to 182
Stable (H, C, O, and S) and radiogenic (Sr, Nd, and Pb) ppm (average 53 ppm), correlated with secondary K addition,
isotope studies of these vein systems of Archean to Phanerozoic in basalts altered by seawater. High N contents of 144–238 ppm
age suggest that the fluids acquired a signature of bulk crust, as have been reported from submarine hydrothermal fluids in the
expected for fluids evolved during regional metamorphism (for Sea of Cortez (Von Damm et al., 1985). Nitrogen in zones of
reviews see McCuaig and Kerrich, 1998; Hagemann and Cas- potassic submarine hydrothermal alteration could be sourced in
sidy, 2000; Kerrich et al., 2000). The dilute, aqueous dominated, seawater, organic-rich sediments, hydrolysis of K-silicates in the
but carbonic- and N-bearing fluids, were generated by dehydra- host volcanic rocks, or some combination. Given bottom seawa-
tion of hydroxyl-silicates for H2O and NH4+, and by decarbon- ter N contents of 0.5 ppm (Létolle, 1980), we tentatively interpret
ation of carbonates and/or oxidation or hydrolysis of organic this variably enriched signature in the VMS alteration zone as
compounds for aqueous carbonic species and N2 (Jia and Ker- reflecting N-bearing organic compounds in ambient sediments.
rich, 2004b). Micas in the Archean deposits have the same high δ15N values as
contemporaneous shale kerogen, consistent with a sedimentary of post peak-metamorphic hydrothermal micas through geological
N source. time. Jia et al. (2001, 2003a) and Jia and Kerrich (2004c) showed
that hydrothermal micas in the Paleozoic Lachlan accretionary oro-
Rare Metal Pegmatites gen and the Mesozoic–Cenozoic western North American Cordil-
lera have δ15N of 3.5‰ ± 0.4‰ (n = 20) and 3.0‰ ± 1.2‰ (n =
Rare metal pegmatites containing Mo and W occur in per- 100) respectively, comparable to Phanerozoic bulk metasedimen-
aluminous, S-type domains of the post-metamorphic 2.6 Ga Pre- tary rocks and average crustal values (Haendel et al., 1986; Bebout
issac-Lacorne post-tectonic batholith, Abitibi Greenstone Belt, and Fogel, 1992; Mingram and Bräuer, 2001; Busigny et al., 2003),
Canada (Feng and Kerrich, 1992). Muscovites are characterized and Phanerozoic kerogens (Williams et al., 1995; Ader et al., 1998;
by δ15N values of 1.6‰–5.3‰, with an average +3.2‰ (Jia and Kao and Liu, 2000; Sephton et al., 2002). This result endorses the
Kerrich, 1999, 2000). Peraluminous granites are considered to use of hydrothermal micas to proxy for crust in the Precambrian.
result from melting of average mid-crust (Sylvester et al., 1997; Mean δ15N values of the 2.2–2.1 Ga pre-metamorphic carbo-
Hawkesworth and Kemp, 2004). Neoarchean greenstone belts naceous shales and post peak-metamorphic hydrothermal micas
are dominated by syn-tectonic TTG batholiths, with supracrustal hosted in the Birimian sediments overlap at 10.8‰ ± 1.1‰ and
volcanic-sedimentary sequences. Notwithstanding uncertainties 10.2‰ ± 1.5‰, respectively (Tables 2, 3; Figs. 4C, 4D). Kerogen
in melt-residue fractionations, these results are consistent with and hydrothermal biotites from the eastern Dharwar Craton are both
fusion of a mix of relatively 15N-depleted tonalites and 15N- enriched in 15N. Similarly, δ15N values of the 2.7 Ga Three Nations
enriched clastic sedimentary rocks (Table 1, Fig. 1). carbonaceous shales (16.0‰ ± 1.7‰) and VMS micas (16.5‰ ±
2.9‰) from the Abitibi belt are both enriched, and comparable to
Other Deposits the isotopic composition of hydrothermal micas (16.3‰ ± 2.9‰)
in the same terrane (Figs. 4A, 4B) (Jia and Kerrich, 1999, 2000,
Nitrate deposits formed in arid climates possess up to 163,000 2004c). The hydrothermal quartz-mica veins precipitated post
ppm N, and δ15N ~0‰ consistent with atmospheric deposition of peak-metamorphism (Kerrich and Cassidy, 1994), and accordingly
N (Böhlke et al., 1997). Sparse data on NH4+-bearing K-feldspars their δ15N values cannot have been influenced by peak-metamor-
from a variety of epithermal springs and deposits in the west- phic conditions.
ern United States range from 2700 to 19,000 ppm N, with δ15N These results demonstrate minimal isotope fractionation of
of −0.6‰–12.3‰ interpreted as N mobilized from sedimentary nitrogen between metasedimentary rocks, hydrothermal fluids, and
rocks (Krohn et al. 1993). minerals precipitated from the fluids. They also rule out any signifi-
cant shifts in δ15N during metamorphism to greenschist facies, or
DISCUSSION mid amphibolite for the Sandur shales, consistent with the results of
previous studies as indicated below.
Nitrogen isotopic compositions of Archean shale kerogen,
hydrothermal gold quartz vein systems that proxy for average Progressively Metamorphosed Sediments
crust, and seawater-altered micas in the VMS deposits all show
15
N-enriched values. The three types of samples show parallel Bebout and Fogel (1992) reported data for progressively
secular trends from 15‰–24‰ at 2.7 Ga, through intermedi- metamorphosed sedimentary rocks of the Catalina Schist com-
ate values in the Proterozoic, to 3‰–4‰ in the Phanerozoic plex, California (Tables 2, 4; Fig. 5A). They obtained δ15N of
(Tables 1–3, Fig. 4). We first constrain the magnitude of pos- 2.2‰ ± 0.6‰ in low-grade rocks (350 ºC) and 4.3‰ ± 0.8‰ in
sible isotope fractionation during metamorphism or Rayleigh amphibolite facies equivalents (600 ºC). They calculated fluid-
volatilization, then address possible long-term (chronic) diffu- rock (N2-NH4+) N-isotope fractionations of –1.5 ‰ ± 1‰ with
sional loss of N. the Rayleigh distillation equation at temperatures ranging from
350 to 600 ºC.
Effects of Metamorphism Mingram and Bräuer (2001) also found shifts of <2‰ in
δ15N from low-grade carbonaceous shales (300 °C) at 2.2‰ ±
The sample design of this study involves comparing pre-meta- 0.6‰, through greenschist facies equivalent (470 °C) at 3.5‰
morphic kerogens from carbonaceous shales with late-metamor- ± 0.9‰, to amphibolite facies mica schists (550 °C) at 3.9‰ ±
phic hydrothermal micas from the eastern Dharwar Craton. Also, 0.8‰ (Table 4, Fig. 5B). In a more recent report on N-isotope
published data for pre-, and late-metamorphic samples from other fractionation due to metamorphism, Busigny et al. (2003) found
Precambrian terranes are compared. These sample types permit that the δ15N values of metasedimentary rocks from the Schistes
testing for large shifts of δ15N during metamorphism, as proposed Lustrés complex (Western Alps), which were subducted from
by Boyd and Philippot (1998), Beaumont and Robert (1999), and shallow level to depth of 90 km, are between 3.1‰ and 4.8‰ and
Pinti et al. (2001a). show no systematic isotopic shifts with increasing metamorphic
From inspection of Tables 1–3 and Figure 4, it is clear that grade (Tables 2, 4; Fig. 5C). A more closed system behavior may
the δ15N values of sedimentary kerogen are consistent with those explain why shifts were smaller than in the Catalina.
TABLE 4. EMPIRICAL AND EXPERIMENTAL STUDIES ON NITROGEN ISOTOPE FRACTIONATIONS DURING METAMORPHISMS
Author/location Geological setting isotopic composition Isotope fractionations
Expirical observations
Bebout and Fogel (1992) The complex contains lawsonite to amphibolite facies 2.2–4.3‰ Calculated bulk fluid-rock
Catalina complex, metasedimentary rocks that represent fractionations = –1.5 ±
California, USA metamorphism at temperatures of 350°–750 °C and 1‰
pressures corresponding to 15–45 km depth.
Haendel et al. (1986) The metasedimentary rocks consist of Phanerozoic (a) 5.1‰ in phyllites, Range of G15N reflects
Sächsisches Erzgebirge, phyllites and schists to Proterozoic gneisses, with (b) 7.4‰ for schists, age, not metamorphic
Germany corresponding metamorphic grades from low (c) 11.8‰ for gneisses grade
greenschist to amphibolite facies.
Jia and Kerrich (2000) Veins are hosted in metamorphic terranes. 16.3–17.2‰ Do not show significant
Quartz veins, Canada Metamorphism ranges from low greenschist facies isotope fractionations
and western Australia (270°–300 °C/1–2 kb) to lower amphibolite facies
(420°–500 °C/2–4 kb).
Mingram and Bräuer The Paleozoic schists contain progressively (a) 2.2‰ in low-grade, Only 1.7‰ shift from low-
(2001) metamorphosed sedimentary rocks from low-grade (b) 3.5‰ for phyllites, grade to amphibolite
Erzgebirge Schists, unit (300 °C/2kb), garnet-phyllite(470 °C/9kb), to (c) 3.9‰ in schists facies rocks
Germany mica schist unit (550 °C/12kb).
Busigny et al. (2003) The complex is an homogeneous sequence of pelagic 2.6–4.8‰ No specific trend with
Schistes Lustres nappe sediments subducted to depths of 0–90 km. increasing
western Alps, Europe Metamorphic grade increases from lower metamorphic grade
greenschist (300 °C/0.8 GPa) to amphibolite (650
°C/3.0 GPa) facies.
Jia (2004) The Paleozoic complex contains low-grade (a) 2.4–3.2 ‰ in green- < 2‰ shift from
Cooma complex, greenschist (300 °C/2–4kb) to high-grade granulite schist facies rocks, greenschist to
Australia facies (730 °C/ 2–4kb) metasedimentary rocks. (b) 3.8–4.3‰ for amphibolite amphibolite facies
facies zone; rocks. Large
(c) 12.3–12.9‰ in granulite fractionations of 8–
facies rocks 10‰ at granulite facies
Experimental studies
Ader et al. (1998) The coals represent a wide range of rank from 4.40–4.45‰ No shift due to change of
Anthracites, USA and anthracite to metaanthracite. Experimental facies or hydrothermal
Germany simulation of the denitrogenation process was disturbance
conducted at 600 °C and 2 kb conditions.
Haendel et al. (1986) documented shifts of ~1‰ from sedi- Experimental and Theoretical Studies
ments to greenschist facies counterparts, and ≤3‰ to amphibolite
facies in progressively metamorphosed siliciclastic sequences of Ader et al. (1998) observed a decrease of nitrogen content
the Sachsisches Erzgebirge, Germany (Table 4, Fig. 5D). How- but uniformity of δ15N in anthracite at temperatures of up to 600
ever, Haendel et al. (1986) emphasized that the total range of °C, in keeping with the empirical studies of N in progressively
δ15N may not be attributable to progressive metamorphism alone metamorphosed sedimentary rocks as indicated above (Table 4).
because (1) the different sedimentary lithologies may have had Hanschmann (1981) calculated nitrogen isotope fractionations
intrinsically different primary N contents and δ15N values; (2) between NH4+ in solid and fluid (N2) phases; interpolation of the
there were multiple metamorphic events in some sequences, but data yields fractionation of ~–2.25‰ at temperatures of 350–600
not others; (3) there is a range of age of the lithologies from Ordo- °C, in accord with the results of Bebout and Fogel (1992) and
vician to Late Precambrian; and (4) the Precambrian gneisses Haendel et al. (1986). Collectively, studies of progressively meta-
alone are characterized by the most 15N-enriched values. morphosed terranes, theoretical and empirical studies, as well as
Hence, there is systematic 15N-enrichment in the three Pro- data for hydrothermal micas, show that the isotope composition
terozoic data sets (see Table 2), the Erzegebirge (δ15N = 11.3‰
± 4.0‰), Moine (13.7‰ ± 2.6‰), and Ashanti belt (10.8‰ ±
1.1‰), notwithstanding the fact that the former two are at amphib-
olite facies, but the Ashanti belt is prevalently lower greenschist Figure 5. δ15N versus N content for progressively metamorphosed
facies (Tables 2 and 4, Jia and Kerrich, 2004b). Accordingly, the sedimentary rocks. Values on the right-hand side are mean (and me-
dian) δ15N plus one standard deviation, and numbers in parentheses
“metamorphic trend” of Haendel et al. (1986) could be reinter- are the sample sizes. Data sources: A, Bebout and Fogel (1992); B,
preted in terms of a secular evolution of crustal N and δ15N, in Mingram and Bräuer (2001); C, Busigny et al. (2003); and D, Haendel
keeping with this study et al. (1986).
of N2 in metamorphic fluids is close to that of the source rock -25

reservoir (Hanschmann, 1981; Bebout and Fogel, 1992; Jia et al.,
-26
A
2001, 2003a).
-27
Rayleigh Processes
δ13C (‰)
-28
Rayleigh fractionation generates large isotopic shifts where -29

volatilization nears completion (Valley, 1986). Pinti et al. (2001a)
-31
model their range of data as a shift of ≥27‰ from primary values
Archean
of −7‰, stemming from Rayleigh devolatilization using frac- -30
tionation factors of Hanschmann (1981) and no constrained f (the Paleoproterozoic
fraction of residual nitrogen in the rocks) values. Modeling was -32
0 50 100 150 200 250 300 350 400 450 500
on a mix of data from 3.8 Ga amphibolite facies chert-BIF and 3.4 C/N (atomic ratio)
Ga greenschist facies cherts. Two 15N-enriched samples of Pinti 15
et al. (2001a) are amphibolite facies metasedimentary rocks hav- B
ing C/N ratios of ~2000 and 5000 that have experienced multiple
Archean metamorphic events and a Proterozoic disturbance (e.g.,
Rollinson. 2002). Of the remaining four samples, two metamor- 10
δ15N (‰)
phosed to high-grade have δ15N intermediate between two low-
grade samples. Accordingly, there is a data cluster, rather than a
correlation of δ15N values with either C/N or metamorphic grade,
5
and hence no progressive Rayleigh fractionation with increasing
grade (Fig. 6 of Pinti et al., 2001a).
Within the seven Precambrian data sets there are only small
trends of increasing δ15N to lower N contents as expected for 0
Rayleigh processes (Table 2; Figs. 3, 4). As a corollary, in the 0 50 100 150 200 250 300 350 400 450 500
Onverwacht data of Beaumont and Robert (1999) two samples C/N (atomic ratio)
with the most positive δ15N have greater N contents than two of -25
the more negative. -26

C
Rayleigh effects can also be addressed via the quartz-mica
vein systems, which are precipitated from fluids generated by -27
δ13C (‰)
loss of volatile species during progressive metamorphism. If -28

Rayleigh effects were significant during metamorphic devolatil-
ization, then H, C, O, N, and S isotope values would be depleted -29
in the veins relative to ambient crust, but this is not the case -31
(McCuaig and Kerrich, 1998).
-30
Constraints from C/N and δ C 13
-32
0 5 10 15
Precambrian kerogens collected in this study do not show δ15N (‰)
correlations between C/N ratios and δ13C or δ15N values (Fig. 6).
Both δ15N and δ13C values of samples, for each given age, shift Figure 6. δ13C versus C/N (A), δ15N versus C/N (B), and δ13C versus
≤3‰ over the span of C/N ratios, consistent with previous esti- δ15N (C) of kerogen in Precambrian carbonaceous shales of India.
mates (Haendel et al., 1986; Watanabe et al., 1997).
A range of δ13C values (−18‰ to −48‰) has been reported
from greenschist facies sedimentary units in 2.7 Ga Superior
Province supracrustal terranes (Schoell and Wellmer, 1981; Botswana, each from greenschist facies carbonaceous shales,
Wellmer et al., 1999). According to Wellmer et al. (1999), no possess distinct δ13C populations averaging −20‰ and −26‰,
correlations exist between δ13C and H/C or C/N. Rather, they respectively (Table 1 of Jia and Kerrich, 2004b), consistent with
interpreted variations in δ13C values of kerogens as resulting different processes during deposition, as suggested by Wellmer et
from organic matter synthesized by different metabolic path- al. (1999). These results complicate the evaluation of metamor-
ways. The two Paleoproterozoic sample sets from the Cuddapah phic effects on kerogen samples compiled in Table 2, including
Basin (−27‰ and –31‰) are interpreted in this way (Table 1). new data from India, but endorse preservation of primary signa-
Two Archean kerogen sample sets from the Abitibi belt and tures (Table 1, Fig. 4).
Long-Term Diffusional Effects of sedimentary structures; (2) large distance from VMS or gold
deposits; (3) large distance from faults; (4) coherent REE pat-
Possible long-term kinetically induced diffusional loss of terns and Eu anomalies; and (5) coherent LILE/REE/HFSE sys-
14
N from kerogen or hydrothermal micas has been evaluated tematics (see Jia and Kerrich, 2004b) (Table 2).
with respect to the K-Ar system. Whereas N, as NH4+, substi-
tutes for K in K-silicates, radiogenic Ar gas is not structurally IMPLICATIONS AND CONCLUSIONS
sited in silicates. Diffusion-dependent blocking temperatures are
known for amphibole (~500 ºC), biotite (300 ºC), and muscovite Precambrian Oxidation State
(350 ºC) (McDougall and Harrison, 1988). The sample design for
kerogen and N-isotope analysis in four terranes 2.7 Ga in age and Two conflicting models based on various lines of evidence
three Proterozoic terranes was predicated on U-Pb zircon and have been developed for the evolution of oxygen in the atmo-
40
Ar/39Ar ages on granitoids within a few tens of million years. sphere in the Precambrian: low pO2 in the Archean, with a rapid
Given undetectable Ar loss in these terranes, N-isotope values increase in oxygen ca. 2.3 Ga (Holland, 1999, 2002 and refer-
should also be primary (Jia and Kerrich, 2004b). ences therein), or alternatively, an oxygenated atmosphere from
Given that Ar-loss is grain-size and temperature depen- the early Archean (Ohmoto, 1997, 2004; and references therein).
dent (McDougall and Harrison, 1988), the compliance of 15N- Beaumont and Robert (1999) interpreted their secular isotopic
enriched data between fine-grained pre-metamorphic micas in “trend” as a record of changes in the redox potential of Earth.
VMS deposits and kerogen, with coarse-grained late-metamor- Negative δ15N values reflect an unspecified metabolic isotopic
phic micas from the same terrane provides further constraints fractionation under anoxic conditions, with microorganisms
against preferential diffusional loss of 14N (Table 3). using reduced forms of nitrogen, whereas positive δ15N values
The effects on preferential loss of 14N of (1) structural sitting reflect an increase in pO2 after the Paleoproterozoic, which pro-
of N between kerogen and silicates (Jia and Kerrich, 2004b), (2) moted the biologic production of nitrate species (Beaumont and
metamorphism, and (3) age can also be evaluated from data sets Robert, 1999).
for Phanerozoic rocks. There are no obvious differences in δ15N Beaumont and Robert reported data for 12 Mesoarchean
between these three classes of samples from the Tertiary to Cam- cherts (33 analyses), 9 Neoarchean cherts, of which one is split
brian (Table 5, Fig. 7). into bedded and homogeneous domains (19 analyses), and 12
Proterozoic cherts (21 analyses). Some samples were analyzed
Hydrothermal Alteration once, others had duplicate or triplicate analytes prepared, and
measurements of some analyses were made more than once. Sta-
Kerogen for this study, as well as other Precambrian kero- tistically, for a given population multiple analyses of a sample
gens analyzed by Jia and coworkers, was selected for minimal reveal “within sample” variance, whereas differences between
secondary disturbance on the following criteria: (1) preservation samples represent “between sample” variance (Searle, 1971).
TABLE 5. A COMPILATION OF DATA FOR N-ISOTOPES IN PHANEROZOIC SEDIMENTARY KEROGENS AND ROCKS,
AND POST PEAK-METAMORPHIC HYDROTHERMAL K-MICAS
Metamorphic facies Age N content (ppm) G15N (permil) Refs.
Sample location Range mean ± 1V Range mean ± 1V
Unmetamorphosed:
(1) Kerogen, Taiwan Tertiary 700–800 800 ± 100 (5) 3.9–4.0 3.9 ± 0.1 (5) 1
(1) Lavagna, western Alps, Europe Tertiary 226–762 3.1–4.7 2
(4) Kerogen, Louisiana Cretaceous 100–3100 2.6–3.6 3.2 ± 0.3 (26) 4
(5) Sandstone, Louisiana Cretaceous 400–3500 1601 ± 100 (64) 0.2–3.2 3.1 ± 1.4 (51) 4
(9) Kerogen, British Columbia Late Jurassic 400–9300 -0.6–5.0 2.8 ± 1.9 (17) 7
(13) Mudstone, English Lake District Ordovician 730–910 476 ± 147 (10) 3.1–4.2 3.7 ± 0.3 (10) 9
Subgreenschist - greenschist facies:
(6) Catalina complex, California Cretaceous 100–1100 530 ± 260 (27) 1.1–3.1 2.2 ± 0.6 (27) 5
(8) K-micas, North American Cordillera 190 to 50 Ma 130–3500 1535 ± 1080 (100) 1.6–6.2 3.0 ± 1.2 (100) 6
(10) K-micas, W. Qinlin orogen, N. China Mesozoic 650–2510 1260 ± 760 (8) 1.7–5.6 4.0 ± 2.0 (8) 3
(11) Kerogen, Pennsylvania Carboniferous 4.1–5.1 4.6 ± 0.4 (10) 8
(12) Kerogen, Germany Carboniferous 2.7–3.7 3.2 ± 0.4 (8) 8
(14) Slate, north Wales Cambrian 129–168 148 (2) 3.9–5.2 4.5 (2) 9
(15) K-micas, Lachlan fold belt, SE. Australia Paleozoic 650–895 733 ± 103 (20) 2.8–4.5 3.5 ± 0.4 (20) 10
Greenschist - amphibolite facies:
(2) Schistes Lustrés, western Alps, Europe Tertiary 169–1721 586 ± 490 (16) 2.6–4.8 3.6 ± 0.7 (16) 2
(3) K-micas, W. Alps Tertiary 280–290 288 ± 5 (4) 4.4–5.5 4.9 ± 0.6 (4) 3
(7) Catalina complex, California Cretaceous 35–810 190 ± 150 (14) 3.6–5.9 4.3 ± 0.8 (14) 5
Note: Refs.: 1—Kao and Liu (2000); 2—Busigny et al. (2003); 3—Jia and Kerrich (2004b,c); 4—Williams et al. (1995); 5—Bebout
and Fogel (1992); 6—Jia et al. (2003a); 7—Sephton et al. (2002); 8—Ader et al. (1998); 9—Bottrell et al. (1988); 10—Jia et al. (2001).
6
(5)
(16) (9)
(13) (4)
5 (11) (2)
(7)
(10)
(1) (15)
(3) (17)
4
δ15N (‰)
(6) (14)
3 (8)
2
(12)
1
Relative age increase Relative age increase Relative age increase
Unmetamorphosed Sub-, to greenschist facies Greenschist to amphibolite facies
0
Metamorphic grade
Figure 7. Variation in N-isotopic composition of Phanerozoic kerogen (square), sedimentary rocks (diamond), and hydrothermal K-micas (cir-
cle). Open symbols represent mean values of δ15N. Vertical bars represent ± 1 standard deviation of N-isotope data. See Table 5 for data sources
and locations (numbers in parentheses).
Consequently, samples with multiple analyses are over-rep- depleted values in the data set of Beaumont and Robert (1999), is
resented as plotted by Beaumont and Robert (1999), Holland for chemoautotrophic processes.
(2002), and Marty and Dauphas (2003).
When the data are replotted to take into account “between Chert-BIF
sample” variance alone, the major shift is from δ15N averages of All data sets from Precambrian chert-BIF, metamorphosed
–0.5‰ in Mesoarchean chert kerogen, through 7.2‰ in Neoar- from greenschist to amphibolite facies, are characterized by (1)
chean kerogen to Paleoproterozoic and Neoproterozoic averages some depleted δ15N values (−5‰); (2) a large range of δ15N val-
of 3.9 ‰ and 2.7‰ respectively (Figs. 8, 9). If the model of an ues, up to 30‰, in greenschist facies cherts; and (3) an absence
anoxic environment until the “great oxidation event” ca. 2.3 Ga of correlation of δ15N with either N content or C/N ratios. The
is correct, then a rapid rise of pO2 is not reflected in the Archean third observation rules out Rayleigh fractionation as a cause of
to Proterozoic N-isotope record, with the implication that the N enriched values.
cycle was not profoundly changed ca. 2.3 Ga. The redox state of Precambrian chert and BIF are often interbedded; their spa-
the early atmosphere, and the timing and mechanism(s) of redox tial and temporal association with volcanic sequences erupted
transitions, remain controversial (Kasting, 1993; Ohmoto, 1997; from mantle plumes has been used as evidence for a hydrother-
Holland, 1999; Phillips et al., 2001; Ohmoto, 2004) and beyond mal origin of these sediments (Barley et al., 1998, and references
the scope of this paper. New approaches are required as suggested therein). Isley and Abbott (1999) and Condie et al. (2001) estab-
by Anbar and Knoll (2002). lished a statistical correlation of the time-series for BIF and vol-
canic sequences erupted from mantle plumes, corroborating the
Chemoautotrophs empirical association (Fig. 9).
Consequently, the large spread of values reported by Beau-
Pinti et al. (2001a) reported N and C contents and δ15N from mont and Robert (1999) and Pinti et al. (2001a) may be primary.
a variety of Archean chert-BIF, where δ15N spans from –7‰– Pinti et al. (2001a) measured δ15 N of –7.4‰ in sample Pano D-
20‰. They identify the most depleted values of −7‰, obtained 136 from 3.5 Ga cherts of the Pilbara Craton, Western Australia,
in high-temperature heating steps, as primary and related to met- for which 40Ar/36Ar at 58,500 and N2/36Ar ratios are close to those
abolic isotopic fractionation of NH4+ by chemoautotrophic bacte- estimated for the upper mantle of δ15N = −5 ± 2‰, 40Ar/36Ar =
ria at Archean hydrothermal vents. In a related discussion paper, 42,000, and N2/36Ar = (5 ± 2) × 106 (Marty and Zimmermann,
Pinti and Hashizume (2001) also suggested that 15N-depleted 1999). Independently, Pinti et al. (2001b) reported Xe-isotope
compositions in Archean cherts could stem from input of mantle evidence for input of mantle volatiles to the Pilbara cherts.
N, but their preferred explanation for values of −6‰ obtained In light of these observations and constraints on the magni-
in low-temperature steps by Sano and Pillinger (1990), and 15N- tude of metamorphic δ15N shifts to ≤3‰, we reinterpret the range
A Neoproterozoic (0.7-0.8 Ga) E Neoproterozoic (0.7-0.8 Ga)

2.5 ± 1.4‰ (3) 2.75‰ (2)
B Paleoproterozoic (1.6-2.1 Ga) F Paleoproterozoic (1.6-2.1 Ga) Figure 8. N-isotope compositions of

Precambrian cherts and/or iron forma-
3.8 ± 2.2‰ (18) 3.9 ± 1.8‰ (10) tion (Data are from Beaumont and
Robert, 1999). (A–D) Histograms re-
produced from Beaumont and Rob-
ert (1999) who plot all their analyses.
(E–G) Histograms plotted in this study
C Neoarchean (2.6-2.7 Ga) G Neoarchean (2.6-2.7 Ga) based on averages for replicate analyses
8.2 ± 10.6‰ (19) 7.2 ± 9.2‰ (9) of the same sample. Average and one
standard deviation are listed above each
of the eight histograms, with numbers in
parentheses on left and right sides rep-
D Mesoarchean (3.4-3.5 Ga) H Mesoarchean (3.4-3.5 Ga) resenting numbers of analyses and of
actual samples, respectively.
-0.3 ± 4.8‰ (33) -0.5 ± 3.7‰ (12)
-8 -4 0 4 8 12 16 20 24 28 32 36 -4 0 4 8 12 16 20 24 28
δ15N (‰) δ15N (‰)
of δ15N values from depleted to enriched in all of the data sets for
0.5
Precambrian chert-BIF as mixing between a 15N-depleted mantle
A
0.4 Banded Iron Formation source emitted from plumes and a 15N-enriched marine sedimen-
HEIGHT
0.3 tary kerogen component, identified by Jia and Kerrich (2004a,

0.2 2004b) and this study (Fig. 3A). Given a mantle plume associa-
tion, BIFs are not a record of near-surface redox conditions.
0.1
0.0
Secular Trends
0.4
B Nitrogen contents of micas from orogenic gold deposits
0.3 Global Plumes
mirror the secular variation in N content of carbonaceous shales
HEIGHT
0.2 (Table 2, Fig. 4B). This result is consistent with the breakdown of
sedimentary kerogen as the primary source of N in the hydrother-
0.1
mal fluids from which the micas were deposited, with dehydra-
0.0 tion of K-micas as a secondary source of N. This interpretation is
also in keeping with the second-order observation that Archean
15 C micas are characterized by a larger range of δ15N than Phanero-
zoic equivalents (Tables 2, 3, 5). Jia et al. (2001) accounted for
10
this distribution in that Archean terranes are volcanic domi-
δ15N (‰)
7.2‰
5 2.75‰
3.9‰ nated, with subordinate sedimentary units, whereas Phanero-
zoic accretionary belts possess a higher proportion of turbidites.
-0.5‰
0 Consequently, metamorphism of Archean terranes would gener-
ate metamorphic fluids having a variable N content and a δ15N
-5
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 depending on the proportions of metaigneous (low N content,
AGE (Ga) depleted δ15N) and metasedimentary (high N content, enriched
δ15N) rocks. Metamorphism of Phanerozoic terranes would gen-
Figure 9. Time-series of occurrences of banded iron formation (A) and erate fluids in which the N-budget was dominated by metasedi-
global plumes (B) from Isley and Abbott (1999). N-isotopic composi- mentary rocks. This interpretation is also consistent with fluid
tions of chert–iron formation (C) from Beaumont and Robert (1999)
as plotted in Figure 8 D–G. These time-series are generated by sum- inclusions, characterized by high N2 content (0.6–99.0 mol) from
ming Gaussian distributions of unit area using mean ages and standard quartz-gold veins in organic-rich slate belts (Bottrell et al., 1988;
deviations. Ortega et al., 1991; de Ronde et al., 1992).
Hayes et al. (1983) documented a general trend of increas- Earth reservoirs (OER: atmosphere-hydrosphere, continental
ing δ13C with decreasing organic C content for Precambrian sedi- crust, MORB-source mantle) with average carbonaceous chon-
mentary rocks. They modeled the trend as a Rayleigh fraction- drites (CC) on an Al-normalized basis. Normalized abundances
ation where δ13C of residual organic C shifted by ~6‰ for each of I, Br, N, and C decrease relative to CC in a manner consistent
factor-of-ten loss of C. However, both δ13C and TOC overlap with mass-dependent hydrodynamic loss during energetic impacts
for their Archean and Proterozoic data sets (Figure 8 of Hayes and extreme UV (EUV) solar radiation. Kramers accounted for
et al., 1983). In addition, differences of δ13C and δ15N between overabundance of H and Cl by their residence dominantly in an
populations of samples within given Archean and Proterozoic ocean, and for Ne-isotope characteristics of mantle plumes by
terranes are preserved (Table 1) (Wellmer et al., 1999; Jia and incorporation of ~10−7 of Earth’s volatile budget from solar atmo-
Kerrich, 2004b). We interpret increasing bulk N contents of sedi- sphere. The results of this study are consistent with the model of
mentary rocks as follows: Most Archean siliciclastic sequences Kramers (2003). An implication is that if isotopic fractionation
are first-cycle volcanogenic turbidites shed off of bimodal arcs of N accompanied its hydrodynamic loss, then Earth’s secondary
into tectonically active basins. Stable passive margins developed atmosphere could have acquired an initial δ15N >CI chondrite.
in the Proterozoic, receiving cratonic detritus at relatively slow Several authors have proposed an E-chondrite model for
sedimentation rates (Taylor and McLennan, 1985). Accordingly, Earth based on rare 15N depleted diamonds (Javoy, 1998; Tols-
organic compounds were diluted by detritus in the former setting tikhin and Marty, 1998). Those diamonds at ~–25‰ (Cartigny et
relative to the latter (Jia et al., 2001). al., 1997, 1998) are close to the range of δ15N in E-chondrites of
The implication from the shale and mica record is that N- –15‰ to −43‰ (Kung and Clayton, 1978; Grady et al., 1986).
fixing microorganisms have progressively drawn down atmo- However, a largely pure E-chondritic mantle can be ruled out
spheric N2, to sequester it as NH4+ in crustal K-silicates, increas- from chemical and isotopic compositional data of Earth (e.g.,
ing the crustal N-inventory from an original basaltic value of 1–2 Allègre et al., 2001; Drake and Righter, 2002), Cr-isotope data of
ppm (Table 6). These processes concurrently shift atmospheric the mantle and various classes of meteorites (Shukolyukov and
and crustal δ15N values down. An analogous process is seques- Lugmair, 1998), and 15N-enriched Archean carbonaceous shales
tration of atmospheric CO2, with transfer to the crustal carbon- (this study).
ate and kerogen budgets. According to Delsemme (1998; 2001), The best material to fit the isotope characteristics for 15N-
the crustal carbonate budget translates into a Hadean atmosphere enriched atmosphere and carbonaceous shales is CI-chondrite-
with 20 times the present CO2. like material having δ15N of 30‰–42‰ (Kerridge, 1985) at the
end of Earth accretion ca. 4.5 Ga, and/or comets formed in the
Origin of the Archean Atmosphere-Hydrosphere vicinity of Jupiter, which likely had the same N-isotopic compo-
sition as CI-chondrites, because they originated in the same zone
Archean sedimentary kerogens and crustal hydrothermal sys- of the early solar system. Volatiles in CI-chondrites and comets
tems as recorded in this study, together with compilations from may be similar (Delsemme, 2001).
the literature, signify systematically 15N-enriched values of 15‰– After considering the δ18O and other isotopes, and the major
24‰ (Tables 2, 3; Fig. 4). If the isotopic fractionation between element abundances (e.g., Mg/Si vs. Al/Si) in Earth and chon-
kerogens and atmosphere in the Precambrian was approximately dritic meteorites, Drake and Righter (2002) ruled out either
the same as at present, then it implies a 15N-enriched atmosphere E- or CI-chondrites as source materials for Earth. Rather, they
in the Archean (13‰–21‰ ca. 2.7 Ga). It is possible that the 15N- considered Earth to have formed by accretion in a narrow feed-
enriched values stem from a different N-cycle in the Archean, ing annulus at ~1 astronomical unit (AU), close to its present
with large biologically mediated fractionations, yet the magni- orbital radius. They argued that Earth’s budget of water cannot
tude of the fractionations exceeds any presently known. have been acquired primarily from a late veneer of CI chondrites
Kramers (2003) conducted an evaluation of Earth’s volatile in the asteroid belt, and that comets originating in the Oort cloud
budget. He compared volatile element abundances in the outer at 2.6 AU can deliver no more than 50% of Earth’s minimum
water budget; consequently, Earth’s water is indigenous. How-
ever, Robert (2001) argued that the source of Earth’s water is
consistent with current understanding of the water content of the
TABLE 6. GLOBAL NITROGEN INVENTORIES
Mass × 1024 Abundance‡ Mass N %
asteroid belt, as inferred from the chondritic meteorite record and
(g)† (g) from the mean of D/H ratios of clay minerals in carbonaceous
-3 1 21
Atmosphere
Continental crust
5.1 × 10
2.2 × 101
78%
50 ppm2
3.9 × 10
1.1 × 1021
30.1
8.5
chondrites, which is close to standard mean ocean water. Accord-
Upper mantle 6.2 × 10 2
0.27 ppm 3
1.7 × 1020
1.3 ing to Morbidelli et al. (2000), from early stages of accretion to
3 3 21
Lower mantle
†
2.9 × 10 2.7 ppm 7.8 × 10 61.1 late stage gas-free sweep-up of planetesimals, water was deliv-
The mass of the atmosphere is from Jacobson et al. (2000). The
Continental crust, upper mantle, and lower mantle are calculated based on ered to Earth from a mix of objects in the asteroid belt, in the
their respective (percentage) 0.374%, 10.3%, and 49.2% of the earth's mass vicinity of the giant planets, and in the Kuiper Belt.
at 5.98 × 1027 (g).
‡
Abundances are from: 1—Kramers (2003); 2—Wedepohl (1995); 3— If the 15N-enriched values of Archean samples reflect a com-
Marty And Dauphas (2003). mensurately enriched atmosphere, then it is possible that N is
Barley, M.E., Krapez, B., Groves, D.I., and Kerrich, R., 1998, The Late
not dominantly indigenous, but was acquired from 15N- enriched Archaean bonanza: Metallogenic and environmental consequences of
solar system materials including CI-chondrite type material and the interaction between mantle plumes, lithosphere tectonics and global
comets. The secular trend of 15N in shale kerogen documented cyclicity: Precambrian Research, v. 91, p. 65–90, doi: 10.1016/S0301-
9268(98)00039-4.
in this study reflects a corresponding atmospheric trend from Beaty, D.W., and Taylor, H.P., Jr., 1982, Some petrologic and oxygen isotopic
>21‰ ca. 2.7 Ga to 0‰ now: the shift can be accounted for by a relationships in the Amulet Mine, Noranda, Quebec, and their bearing on
combination of (1) mantle degassing (continuous addition of N2 the origin of Archean massive sulfide deposits: Economic Geology and
the Bulletin of the Society of Economic Geologists, v. 77, p. 95–108.
with δ15N –5‰); (2) progressive sequestration of atmospheric N2 Beaumont, V., and Robert, F., 1999, Nitrogen isotope ratios of kerogens in
into crustal rocks by nitrogen fixing organisms, with a return flux Precambrian cherts: A record of the evolution of atmosphere chem-
of 15N-depleted N2 stemming from diagenetic fractionation; and istry: Precambrian Research, v. 96, p. 63–82, doi: 10.1016/S0301-
9268(99)00005-4.
(3) recycling of 15N-enriched Archean sediments into the mantle. Beaumont, V., Agrinier, P., Javoy, M., and Robert, F., 1994, Determination of
Such nitrogen recycling has been proposed by Zhang and Zindler the CO contribution to the 15N/14N ratio measured by mass-spectrometry:
(1993). Given the limited database for Archean rocks this inter- Analytical Chemistry, v. 66, p. 2187–2189, doi: 10.1021/ac00085a039.
Bebout, G.E., and Fogel, M.L., 1992, Nitrogen-isotope compositions of
pretation is of necessity speculative, and other explanations may metasedimentary rocks in the Catalina Schist, California: Implications
emerge. Isotopic mass balance for N is difficult given uncertain- for metamorphic devolatilization history: Geochimica et Cosmochimica
ties in mantle N-content, mass of subducted material, and the Acta, v. 56, p. 2839–2849, doi: 10.1016/0016-7037(92)90363-N.
Bebout, G.E., Ryan, J.G., Leeman, W.P., and Bebout, A.E., 1999, Fractionation
conservative or volatile behavior of N at convergent margins. of trace elements by subduction-zone metamorphism – effect of con-
Kerrich and Jia (2004) attempted a simplified model, which indi- vergent-margin thermal evolution: Earth and Planetary Science Letters,
cated that subduction has caused a shift of <0.1‰ in the mantle v. 171, p. 63–81, doi: 10.1016/S0012-821X(99)00135-1.
Bleeker, W., Parrish, R.R., and Sager, K.A., 1999, High-precision U-Pb geo-
δ15N using the “chert”-like kerogen recycling assumptions of chronology of the late Archean Kidd Creek Deposit and Kidd volcanic
Marty and Dauphas (2003). complex: Economic Geology Monograph 10, p. 43–69.
Böhlke, J.K., Eriksen, G.E., and Revesz, K., 1997, Stable isotope evidence
for an atmospheric origin of desert nitrate deposits in northern Chile and
ACKNOWLEDGMENTS southern California, USA: Chemical Geology, v. 136, p. 135–152, doi:
10.1016/S0009-2541(96)00124-6.
We thank Steve Kesler and Hirishi Ohmoto for the invitation Bos, A., Duit, W., Eerden, M.J., and Jansen, J.B.H., 1988, Nitrogen storage in
biotite: An experimental study of the ammonium and potassium partition-
to submit this manuscript. We are grateful to M. Stocki for assis- ing between IM-phlogopite and vapour at 2 kb: Geochimica et Cosmochi-
tance with the nitrogen analysis in the Department of Soil Sci- mica Acta, v. 52, p. 1275–1283, doi: 10.1016/0016-7037(88)90281-5.
ences, University of Saskatchewan; P. Lindgren for orchestrating Bottrell, S.H., Carr, L.P., and Dubessy, J., 1988, A nitrogen-rich metamorphic
fluid and coexisting minerals in slates from North Wales: Mineralogical
a field trip to Jerome; and K.M. Ansdell, T. Oberthür, B. Pratt, A. Magazine, v. 52, p. 451–457.
Still, and S. Vearncombe for providing some of the samples. Y. Boyd, S.R., and Philippot, P., 1998, Precambrian ammonium biogeochemis-
Jia acknowledges receipt of a CSIRO postdoctoral fellowship, try: A study of the Moine metasediments, Scotland: Chemical Geology,
v. 144, p. 257–268, doi: 10.1016/S0009-2541(97)00135-6.
an honorary position at Monash University, and a research grant Boyd, S.R., Mattey, D.P., Pillinger, C.T., Milledge, H.J., Mendelssohn, M., and
from SEG Foundation, Inc., USA. R. Kerrich acknowledges a Seal, M., 1987, Multiple growth events during diamond genesis: An inte-
Natural Sciences and Engineering Research Council (NSERC) grated study of carbon and nitrogen isotopes and nitrogen aggregation
state in coated stone: Earth and Planetary Science Letters, v. 86, p. 341–
Discovery Grant, an NSERC MFA grant, and the George McLeod 353, doi: 10.1016/0012-821X(87)90231-7.
endowment to the Department of Geological Sciences, University Boyd, S.R., Pillinger, C.T., Milledge, H.J., Mendelssohn, M., and Seal, M.,
of Saskatchewan. The critiques of H. Ohmoto, S. Kesler, D. L. 1992, C and N isotopic composition and the infrared absorption spectra of
coated diamonds: Evidence for the regional uniformity of CO2-H2O rich
Pinti and an anonymous journal reviewer significantly improved fluids in lithospheric mantle: Earth and Planetary Science Letters, v. 109,
an earlier version of the manuscript. p. 633–644, doi: 10.1016/0012-821X(92)90121-B.
Boyd, S.R., Hall, A., and Pillinger, C.T., 1993, The measurement of δ15N in
crustal rocks by static vacuum mass spectrometry: Application to the
REFERENCES CITED origin of the ammonium in the Cornubian batholith, southwest Eng-
land: Geochimica et Cosmochimica Acta, v. 57, p. 1339–1347, doi:
Ader, M., Boudou, J., Javoy, M., Goffe, B., and Daniels, E., 1998, Isotope study 10.1016/0016-7037(93)90070-D.
on organic nitrogen of Westphalian anthracites from the Western Middle Busigny, V., Cartigny, P., Philippot, P., Ader, M., and Javoy, M., 2003, Massive
field of Pennsylvanian (U.S.A) and from the Bramsche Massif (Ger- recycling of nitrogen and other fluid-mobile elements (K, Rb, Cs, H) in a
many): Organic Geochemistry, v. 29, p. 315–323, doi: 10.1016/S0146- cold slab environment: Evidence from HP to UHP oceanic metasediments
6380(98)00072-2. of the Schistes Lustrés nappe (western Alps, Europe): Earth and Planetary
Allègre, C., Manhès, G., and Lewin, È., 2001, Chemical composition of the Science Letters, v. 215, p. 27–42, doi: 10.1016/S0012-821X(03)00453-9.
Earth and the volatility control on planetary genetics: Earth and Planetary Cartigny, P., and Ader, M., 2003, A comment on “The nitrogen record of crust-
Science Letters, v. 185, p. 49–69, doi: 10.1016/S0012-821X(00)00359-9. mantle interaction and mantle convection from Archean to Present” by
Anbar, A.D., and Knoll, A.H., 2002, Proterozoic ocean chemistry and evolu- B. Marty and N. Dauphas: Earth and Planetary Science Letters, v. 216,
tion: A bioinorganic bridge?: Science, v. 297, p. 1137–1142, doi: 10.1126/ p. 425–432, doi: 10.1016/S0012-821X(03)00505-3.
science.1069651. Cartigny, P., Boyd, S.R., Harris, W.J., and Javoy, M., 1997, Nitrogen isotopes
Anderson, A., Jr., Clayton, R.N., and Mayeda, T.K., 1971, Oxygen isotope ther- in peridotitic diamonds from Fuxian, China: The mantle signature: Terra
mometry of mafic igneous rocks: Journal of Geology, v. 79, p. 715–729. Nova, v. 9, p. 175–179, doi: 10.1046/j.1365-3121.1997.d01-26.x.
Balakrishnan, S., Rajamani, V., and Hansen, G.N., 1999, U-Pb ages for zircon Cartigny, P., Harris, J.W., Phillips, D., Girard, M., and Javoy, M., 1998, Subduc-
and titanite from the Ramagiri area, southern India: Evidence for accretion-related diamonds?: The evidence for a mantle-derived origin from
tionary origin of the eastern Dharwar craton during the Late Archean: coupled δ13C–δ15N determinations: Chemical Geology, v. 147, p. 147–
Journal of Geology, v. 107, p. 69–86, doi: 10.1086/314331. 159, doi: 10.1016/S0009-2541(97)00178-2.
Cartigny, P., Jendrzejewski, J.N., Pineau, F., Petit, E., and Javoy, M., 2001, Vol- Gibson, E.K., Carr, L.P., and Pillinger, C.T., 1985, Nitrogen isotopic composi-
atiles (C, N, Ar) variability in MORB and the respective roles of mantle tion of Archean samples: Evidence of the earth’s early atmosphere? [abs.]:
source heterogeneity and degassing: The South West Indian Ridge’s case: Lunar and Planetary Science, v. XVI, p. 270–272.
Earth and Planetary Science Letters, v. 194, p. 241–247, doi: 10.1016/ Gibson, E.K., Carr, L.P., Gilmour, I., and Pillinger, C.T., 1986, Earth’s atmo-
S0012-821X(01)00540-4. sphere during the Archean as seen from carbon and nitrogen isotopic
Chen, C.T., Gong, A.G.C., Wang, S.L., and Bychkov, A.S., 1996, Redfield ratios analysis of sediments [abs.]: Lunar and Planetary Science, v. XVII,
and regeneration rates of particulate matter in the Sea of Japan as a model p. 258–259.
of closed system: Geophysical Research Letters, v. 23, p. 1785–1788, doi: Goldfarb, R.J., Groves, D.I., Gardoll, S., 2001, Orogenic gold and geologic
10.1029/96GL01676. time; a global synthesis: Ore Geology Reviews, v. 18, p. 1–75.
Cline, J.D., and Kaplan, I.R., 1975, Isotope fractionation of dissolved nitrate Haendel, D., Mühle, K., Nitzsche, H., Stiehl, G., and Wand, U., 1986, Isotopic
during denitrification in the Eastern Tropical North Pacific: Marine Chem- variations of the fixed nitrogen in metamorphic rocks: Geochimica et Cos-
istry, v. 3, p. 271–299, doi: 10.1016/0304-4203(75)90009-2. mochimica Acta, v. 50, p. 749–758, doi: 10.1016/0016-7037(86)90351-0.
Condie, K.C., Des Marais, D.J., and Abbott, D., 2001, Precambrian super- Hagemann, S.G., and Cassidy, K.F., 2000, Archean orogenic lode gold depos-
plumes and supercontinents: A record in carbonaceous shales, carbon its, in Hagemann, S.G., and Brown, P.E., eds., Gold in 2000: Reviews in
isotopes, and paleoclimates?: Precambrian Research, v. 106, p. 239–260, Economic Geology, v. 13, p. 9–68.
doi: 10.1016/S0301-9268(00)00097-8. Hamilton, J.V., and Hodgson, C.J., 1986, Mineralization and structure of the
Conway, N.M., Kenicutt, M.C., and Van Dover, C.L., 1994, Stable isotopes in Kolar gold field, India, in MacDonald, A.J., ed., Gold ’86: Willowdale,
the study of marine chemosynthetic-based ecosystems, in Lajtha, K., and Ontario, Konsult International, p. 270–283.
Michener, R.H., eds., Stable isotopes in ecology and environmental sci- Hannington, M.D., Barrie, C.T., and Bleeker, W., 1999, The giant Kidd Creek
ence: Oxford, UK, Blackwell Scientific, p. 158–186. volcanogenic massive sulfide deposit, western Abitibi Subprovince,
Corfu, F., Krogh, T.E., Kwok, Y.Y., and Jensen, L.S., 1989, U-Pb zircon geo- Canada; summary and synthesis: Economic Geology Monograph 10,
chronology in the southwestern Abitibi greenstone belt, Superior Prov- p. 661–672.
ince: Canadian Journal of Earth Sciences, v. 26, no. 9, p. 1747–1763. Hanschmann, G., 1981, Berechnung von isotopieeffekten auf quantenchemi-
Costa, U.R., Barnett, R.L., and Kerrich, R., 1983, The Mattagami Lake Mine scher grundlage am beispiel stickstoffhaltiger molecule: Zfl-Mitteilun-
Archean Zn-Cu sulfide deposit, Quebec: Hydrothermal coprecipitation of gen, v. 41, p. 19–31.
talc and sulfides in a sea-floor brine pool—Evidence from geochemistry, Hall, A., 1989, Ammonium in spilitized basalts of southwest England and its
18
O/16O, and mineral chemistry: Economic Geology and the Bulletin of implications for the recycling of nitrogen: Geochemical Journal, v. 23,
the Society of Economic Geologists, v. 78, p. 1144–1203. p. 19–23.
Dasgupta, S., and Sengupta, P., 2003, Indo-Antarctic correlation: A perspec- Hall, A., 1999, Ammonium in granites and its petrogenetic significance: Earth-
tive from the Eastern Ghats granulite belt, India, in Yosida, M., Windley, Science Review, v. 45, p. 145–165, doi: 10.1016/S0012-8252(99)00006-9.
F., and Dasgupta, S., eds., Proterozoic East Gondwana: Supercontinent Hayes, J.M., Kaplan, I.R., and Wedeking, K.W., 1983, Precambrian organic
assembly and breakup: Geological Society [London] Special Publication geochemistry, preservation of the record, in Schopf, W.J., ed., Earth’s Ear-
206, p.131–143. liest Biosphere: Cambridge University Press, p. 92–134.
Delsemme, A., 1998, Our Cosmic Origin from the Big Bang to the Emergence Hoch, M.P., Fogel, M.L., and Kirchman, D.L., 1994, Isotope fractionation dur-
of Life and Intelligence: Cambridge, UK, Cambridge University Press, ing ammonium uptake by marine microbial assemblages: Geomicrobio-
269 p. logical Journal, v. 12, p. 113–127.
Delsemme, A., 2001, An argument for the cometary origin of the biosphere: Holland, H.D., 1999, When did the Earth’s atmosphere become oxic? A Reply:
American Scientist, v. 89, p. 432–442, doi: 10.1511/2001.5.432. Geochemical News, v. 100, p. 20–23.
Delwiche, C., and Steyn, P.L., 1970, Nitrogen isotope fractionations in soils Holland, H.D., 2002, Volcanic gases, black smokers, and the great oxidation
and microbial reactions: Environmental Science and Technology, v. 4, event: Geochimica et Cosmochimica Acta, v. 66, p. 3811–3826, doi:
p. 929–935, doi: 10.1021/es60046a004. 10.1016/S0016-7037(02)00950-X.
de Ronde, C.E.J., Spooner, E.T.C., De Wit, M.J., and Bray, C.J., 1992, Shear Holloway, J.M., and Dahlgren, R.A., 2002, Nitrogen in rock: Occurrences
zone-related Au quartz vein deposits in the Barberton greenstone belt, and biogeochemical implications: Global Biogeochemical Cycles, v. 16,
South Africa: Field and petrological characteristics, fluid properties, and doi:10.1029/2002GB001862.
light stable isotope geochemistry: Economic Geology and the Bulletin of Honma, H., and Itihara, Y., 1981, Distribution of ammonium in minerals of
the Society of Economic Geologists, v. 88, p. 366–402. metamorphic and granite rocks: Geochimica et Cosmochimica Acta,
Drake, M.J., and Righter, K., 2002, Determining the composition of the Earth: v. 45, p. 983–988, doi: 10.1016/0016-7037(81)90122-8.
Nature, v. 416, p. 39–44, doi: 10.1038/416039a. Isley, A.E., and Abbott, D.H., 1999, Plume-related mafic volcanism and the
Drummond, M.S., and Defant, M.J., 1990, A model for trondhjemite-tonalite- deposition of banded iron formation: Journal of Geophysical Research,
dacite genesis and crustal growth via slab melting: Archean to modern v. 104, p. 15461–15477, doi: 10.1029/1999JB900066.
comparisons: Journal of Geophysical Research, v. B95, p. 21503–21521. Jackson, S.L., and Fyon, J.A., 1991, The western Abitibi subprovince in Ontario:
Durand, B., and Nicaise, G., 1980, Procedures for kerogen isolation, in Burand, Ontario Geological Survey, Special Volume 4, pt. 1, p. 405–482.
B., ed., Kerogen: Paris, Technip, p. 35–53. Jacobson, M.C., Charlson, R.J., Rodhe, H., and Orians, G.H., 2000, Earth sys-
Feng, R., and Kerrich, R., 1992, Geodynamic evolution of the Southern Abitibi tem science from biogeochemical cycles to global change: San Diego,
and Pontiac terranes: Evidence from the geochemistry of granitoid Academic Press, 527 p.
magma series (2700 Ma–2630 Ma): Canadian Journal of Earth Sciences, Javoy, M., 1998, The birth of the Earth’s atmosphere: The behavior and fate
v. 29, p. 2266–2286. of its major elements: Chemical Geology, v. 147, p. 11–25, doi: 10.1016/
Fraga, F., Ríos, A.F., Pérez, F.F., and Figueiras, F.G., 1998, Theoretical lim- S0009-2541(97)00169-1.
its of oxygen/carbon and oxygen/nitrogen ratios during photosynthesis Javoy, M., and Pineau, F., 1991, The volatiles record of a “popping” rock from
and mineralization of organic matter in the sea: Scientia Marina, v. 62, the Mid Atlantic Ridge at 14° N: Chemical and isotopic composition of
p. 161–168. gas trapped in the vesicles: Earth and Planetary Science Letters, v. 107,
Franklin, J.M., Lydon, J.W., and Sangster, D.F., 1981, Volcanic-associated p. 598–611, doi: 10.1016/0012-821X(91)90104-P.
massive sulfide deposits: Economic Geology, 75th Anniversary Volume, Jia, Y., 2004, Nitrogen isotope compositions of the Cooma metamorphic com-
p. 485–627. plex, Australia: Implications for 15N-enriched reservoirs: Geochimica et
Freudenthal, T.T., Wagner, T., Wezhöfer, F., Zabel, M., and Wefer, T., 2001, Cosmochimica Acta, v. 68, Suppl. 1, p. A42–A42.
Early diagenesis of organic matter from sediments of the eastern subtropi- Jia, Y., and Kerrich, R., 1999, Nitrogen isotope systematics of mesothermal lode
cal Atlantic: Evidence from stable nitrogen and carbon isotopes: Geochi- gold deposits: Metamorphic, granitic, meteoric water, or mantle origin?:
mica et Cosmochimica Acta, v. 65, p. 1795–1808, doi: 10.1016/S0016- Geology, v. 27, p. 1051–1054, doi: 10.1130/0091-7613(1999)027<1051:
7037(01)00554-3. NISOML>2.3.CO;2.
Grady, M.M., Wright, I.P., Carr, L.P., and Pillinger, C.T., 1986, Composi- Jia, Y., and Kerrich, R., 2000, Giant quartz vein systems in accretionary oro-
tional differences in enstatite chondrites based on carbon and nitrogen genic belts: The evidence for a metamorphic fluid origin from δ15N and
stable isotope measurements: Geochimica et Cosmochimica Acta, v. 50, δ13C studies: Earth and Planetary Science Letters, v. 184, p. 211–224, doi:
p. 2799–2813, doi: 10.1016/0016-7037(86)90228-0. 10.1016/S0012-821X(00)00320-4.
Jia, Y., and Kerrich, R., 2004a, A reinterpretation of the crustal N-isotope Lindberg, P., and Gustin, M.S., 1987, Field-trip guide to the geology, structure,
record: Evidence for a 15N-enriched Archean atmosphere?: Terra Nova, and alteration of the Jerome, Arizona ore deposits, in Davis, G.H., and
v. 16, p. 102–108, doi: 10.1111/j.1365-3121.2004.00535.x. VandenDolder, E.M., eds., Geologic diversity of Arizona and its margins:
Jia, Y., and Kerrich, R., 2004b, Nitrogen 15-enriched Precambrian kerogen and Excursions to choice areas: Geological Society of America, 100th Annual
hydrothermal systems: Geochemistry, Geophysics, and Geosystems, v. 5, Meeting, Field-trip Guidebook: Arizona Bureau of Mines and Mineral
doi:10.1029/2004GC000716. Technology Special Paper 5, p. 176–187.
Jia, Y., and Kerrich, R., 2004c, Secular evolution of δ15N in crustal fluids: Impli- Macko, E.S., Fogel, M.L., Hare, P.E., and Hoering, T.C., 1987, Isotopic frac-
cations for the origin of Earth’s early atmosphere and hydrosphere, in tionation of nitrogen and carbon in the synthesis of amino acids by
Wanty, R.B., and Seal, R.R., eds., Proceedings of the 11th International microorganisms: Chemical Geology, v. 65, p. 79–92, doi: 10.1016/0009-
Symposium on Water-Rock Interaction: New York, A.A. Balkema, 2541(87)90196-3.
p. 945–949. Manikyamba, C., Naqvi, S.M., Gnaneshwar-Rao, T., Balaram, V., Ramesh,
Jia, Y., Li, X., and Kerrich, R., 2001, Stable isotope (O, H, S, C, and N) system- S.L., and Reddy, G.L.N., 1997, Geochemical heterogeneities of metagray-
atics of quartz vein systems in the turbidite-hosted Central and North Deb- wackes from the Sandur schist belt; implications for active plate margin
orah gold deposits of the Bendigo gold field, central Victoria, Australia: processes: Precambrian Research, v. 84, p. 117–138, doi: 10.1016/S0301-
Constraints on the origin of ore-forming fluids: Economic Geology and 9268(97)00022-3.
the Bulletin of the Society of Economic Geologists, v. 96, p. 705–721. Martin, H., 1999, Adakitic magmas: Modern analogues of Archean granitoids:
Jia, Y., Kerrich, R., and Goldfarb, R., 2003a, Genetic constraints of orogenic Lithos, v. 46, p. 411–429, doi: 10.1016/S0024-4937(98)00076-0.
gold-bearing quartz vein systems in the western North American Cordil- Marty, B., 1995, Nitrogen content of the mantle inferred from N2-Ar correlation
lera: Constraints from a reconnaissance study of δ15N, δD, and δ18O: Eco- in oceanic basalts: Nature, v. 377, p. 326–329, doi: 10.1038/377326a0.
nomic Geology and the Bulletin of the Society of Economic Geologists, Marty, B., and Dauphas, N., 2003, The nitrogen record of crust-mantle interac-
v. 98, p. 109–213. tion and mantle convection from Archean to present: Earth and Planetary Sci-
Jia, Y., Kerrich, R., Gupta, A.K., and Fyfe, W.S., 2003b, 15N enriched Gond- ence Letters, v. 206, p. 397–410, doi: 10.1016/S0012-821X(02)01108-1.
wana lamproites, eastern India: Crustal N in the mantle source: Earth Marty, B., and Humbert, F., 1997, Nitrogen and argon isotopes in oceanic
and Planetary Science Letters, v. 215, p. 43–56, doi: 10.1016/S0012- basalts: Earth and Planetary Science Letters, v. 152, p. 101–112, doi:
821X(03)00426-6. 10.1016/S0012-821X(97)00153-2.
Kao, S.J., and Liu, K.K., 2000, Stable carbon and nitrogen isotope systematics Marty, B., and Zimmermann, L., 1999, Volatiles (He, C, N, Ar) in mid-ocean
in a human-disturbed watershed (Lanyang-His) in Taiwan and the estima- ridge basalts: Assessment of shallow-level fractionation and character-
tion of biogenic particular organic carbon and nitrogen fluxes: Global Bio- ization of source composition: Geochimica et Cosmochimica Acta, v. 63,
geochemical Cycles, v. 14, p. 189–198, doi: 10.1029/1999GB900079. p. 3619–3633, doi: 10.1016/S0016-7037(99)00169-6.
Kasting, J.F., 1993, Earth’s early atmosphere: Science, v. 259, p. 920–926. Mazuka, A.N.N., Macko, S.A., and Pedersen, T.F., 1991, Stable carbon and
Kemp, A.I.S., and Hawkesworth, C.J., 2004, 3.11. Granitic perspectives on the nitrogen isotope compositions of organic matter from sites 734 and 725,
generation and secular evolution of the continental crust, in Rudnick, R., Oman Margin: Proceedings of the Ocean Drilling Program, Scientific
ed., The crust: Amsterdam, Elsevier, Treatise on Geochemistry, v. 3,p. Results, v. 117, p. 561–586.
349–410. McCuaig, T.C., and Kerrich, R., 1998, P-T-t-deformation-fluid characteristics
Kerrich, R., and Cassidy, K.F., 1994, Temporal relationships of lode-gold min- of lode gold deposits: Evidence from alteration systematics: Ore Geology
eralization to accretion, magmatism, metamorphism and deformation, Reviews, v. 12, p. 381–453, doi: 10.1016/S0169-1368(98)00010-9.
Archean to present: A review: Ore Geology Reviews, v. 9, p. 263–310, McDougall, I., and Harrison, T.M., 1988, Geochronology and thermochronol-
doi: 10.1016/0169-1368(94)90001-9. ogy by the 40Ar/39Ar Method: New York, Oxford University Press, 212 p.
Kerrich, R., and Jia, Y., 2004, A comment on “The nitrogen record of crust- Mingram, B., and Bräuer, K., 2001, Ammonium concentration and nitrogen iso-
mantle interaction and mantle convection from Archean to Present” by tope composition in metasedimentary rocks from different tectonometa-
B. Marty and N. Dauphas: Earth and Planetary Science Letters, v. 225, morphic units of the European Variscan Belt: Geochimica et Cosmochi-
p. 435–440, doi: 10.1016/j.epsl.2004.07.004. mica Acta, v. 65, p. 273–287, doi: 10.1016/S0016-7037(00)00517-2.
Kerrich, R., and Ludden, J., 2000, The role of fluids during formation and evolution Mingram, B., Hoth, P., Lüders, V., and Harlov, D., 2005, The significance of
of the Southern Superior Province lithosphere: A review: Canadian Journal of fixed ammonium in Palaeozoic sediments for the generation of nitrogen-
Earth Sciences, v. 37, p. 135–165, doi: 10.1139/cjes-37-2-3-135. rich natural gases in the North German Basin: International Journal of
Kerrich, R., Goldfarb, R., Groves, D., Garvin, S., 2000, The geodynamics of Earth Sciences, v. 94, p. 1010–1022.
world-class gold deposits: Characteristics, space-time distribution, and Minoura, K., Hoshino, K., Nakamura, T., and Wada, E., 1997, Late Pleistocene-
origins: Reviews of Economic Geology, v. 13, p. 501.544. Holocene paleoproductivity circulation in Japan Sea: Sea-level control
Kerridge, J.F., 1985, Carbon, hydrogen and nitrogen in carbonaceous chon- on δ13C and δ15N records of sediment organic material: Paleogeography,
drites: Abundances and isotopic compositions in bulk samples: Geochi- Paleoclimatology, Paleoecology, v. 135, p. 41–50.
mica et Cosmochimica Acta, v. 49, p. 1707–1714, doi: 10.1016/0016- Morbidelli, A., Chambers, J., Lunine, J.I., Petit, J.M., Robert, F., Valsecchi,
7037(85)90141-3. G.B., and Cyr, K.E., 2000, Source regions and timescales for the delivery
Kramers, J.D., 2003, Volatile element abundance patterns and an early liquid of water to the Earth: Meteorite and Planetary Science, v. 35, p. 1309–
water ocean on Earth: Precambrian Research, v. 126, p. 379–394, doi: 1320.
10.1016/S0301-9268(03)00106-2. Nadelhoffer, K.J., and Fry, B., 1988, Controls on natural 15N and 13C abun-
Krohn, M.D., Kendall, C., Evans, R.J., and Fries, T.L., 1993, Relations of dances in forest soil organic matter: Soil Science Society of America Jour-
ammonium minerals at several hydrothermal systems in the western U.S.: nal, v. 52, p. 1633–1640.
Journal of Volcanology and Geothermal Research, v. 56, p. 401–413, doi: Nagaraja-Rao, B.K., Rajurkar, S.T., Ramalingaswamy, G., and Ravindra, B.B.,
10.1016/0377-0273(93)90005-C. 1987, Stratigraphy, structure, and evolution of the Cuddapah Basin: Geo-
Kung, C.C., and Clayton, R.N., 1978, Nitrogen abundances and isotopic com- logical Society of India Memoir 6, p. 33–86.
positions in stony meteorites: Earth and Planetary Science Letters, v. 38, Naqvi, S.M., Venkatachala, B.S., Shukla, M., Kumar, B., Natarajan, R., and
p. 421–435, doi: 10.1016/0012-821X(78)90117-6. Sharma, M., 1987, Silicified cyanobacteria from the cherts of Archaean
Landis, C.A., 1971, Graphitization of dispersed carbonaceous material in meta- Sandur schist belt; Karnataka, India: Journal of the Geological Society of
morphic rocks: Contributions to Mineralogy and Petrology, v. 30, p. 34– India, v. 29, p. 535–539.
45, doi: 10.1007/BF00373366. Naqvi, S.M., Uday, R.B., Subba, R.D.V., Manikyamba, C., Nirmal, C.S.,
Lehmann, M.F., Bernasconi, S.M., Barbieri, A., and McKenzie, J.A., 2002, Balarm, V., and Srinivasa, S.D., 2002, Geology and geochemistry of aren-
Preservation of organic matter and alteration of its carbon and nitrogen ite-quartzwacke from the late Archaean Sandur schist belt; implications
isotope composition during simulated and in situ early sedimentary dia- for provenance and accretion processes: Precambrian Research, v. 114,
genesis: Geochimica et Cosmochimica Acta, v. 66, p. 3573–3584, doi: p. 177–197, doi: 10.1016/S0301-9268(01)00227-3.
10.1016/S0016-7037(02)00968-7. Nutman, A.P., McGregor, V.R., Friend, C.R.L., Bennett, V.C., and Kinny, P.D.,
Létolle, R., 1980, Nitrogen-15 in the natural environment, in Fritz, P. and Fon- 1996, The Itsaq Gneiss Complex of southern West Greenland; the world’s
tes, J.C., eds., Handbook of environmental isotope geochemistry: Amster- most extensive record of early crustal evolution (3900–3600 Ma): Pre-
dam, Elsevier, p. 407–433. cambrian Research, v. 78, p. 1–39, doi: 10.1016/0301-9268(95)00066-6.
Ohmoto, H., 1997, When did the Earth’s atmosphere become oxic? Geochemi- Schubert, C.J., and Calvert, S.E., 2001, Nitrogen and carbon isotopic composi-
cal News, v. 93, p. 12–13, 26–27. tion of marine and terrestrial organic matter in Arctic Ocean sediments:
Ohmoto, H., 2004. Archean atmosphere, hydrosphere and biosphere, in Erik- Implications for nutrient utilization and organic matter composition: Deep-
kson, P.G., Alterman, W., Nelson, D.R., Mueller, W.U., and Catuneanu, Sea Research, v. 48, p. 789–810, doi: 10.1016/S0967-0637(00)00069-8.
O., eds., The Precambrian Earth: Tempos and events: Amsterdam, Else- Searle, S.R., 1971, Topics in variance component estimation: Biometrics, v. 27,
vier, Developments in Precambrian Geology, v. 12, p. 361–388. p. 1–76.
Ortega, L., Vindel, E.M., and Beny, C., 1991, C-O-H-N fluid inclusions associ- Sephton, M.A., Amor, K., Franchi, I.A., Wignall, P.B., Newton, R., and
ated with gold-stibnite mineralization in low-grade metamorphic rocks, Zonneveld, L.P., 2002, Carbon and nitrogen isotope disturbances and an
Mari Rosa mine, Cacers, Spain: Mineralogical Magazine, v. 55, p. 235– end-Norian (Late Triassic) extinction event: Geology, v. 30, p. 1119–1122,
247. doi: 10.1130/0091-7613(2002)030<1119:CANIDA>2.0.CO;2.
Peters, K.E., Sweeney, R.E., and Kaplan, I.R., 1978, Correlation of carbon and Shukolyukov, A., and Lugmair, G.W., 1998, Isotopic evidence for the Creta-
nitrogen stable isotope ratios in sedimentary organic matter: Limnology ceous-Tertiary impactor and its types: Science, v. 282, p. 927–929, doi:
and Oceanography, v. 23, p. 598–604. 10.1126/science.282.5390.927.
Phillips, G.N., Law, J.D., and Myers, R.E., 2001, Is the redox-state of the Siddaiah, N.S., and Rajamani, V., 1989, The geologic setting, mineralogy, geo-
Archean atmosphere constrained?: Society of Economic Geologists chemistry, and genesis of gold deposits of the Archean Kolar schist belt,
Newsletter, v. 47, p. 9–18. India: Economic Geology and the Bulletin of the Society of Economic
Pinti, D.L., and Hashizume, K., 2001, 15N-depleted nitrogen in early Archean Geologists, v. 84, p. 2155–2172.
kerogens: Clues on ancient marine chemosynthetic-based ecosystems? Smithies, R.H., 2000, The Archaean tonalite-trondhjemite-granodiorite (TTG)
A comment to Beaumont, V., and F. Robert, F. (1999): Precambrian series is not an analogue of Cenozoic adakite: Earth and Planetary Science
Research, v. 96, p. 62–82: Precambrian Research, v. 105, p. 85–88, doi: Letters, v. 182, p. 115–125, doi: 10.1016/S0012-821X(00)00236-3.
10.1016/S0301-9268(00)00100-5. Sreeramachandra Rao, K., 2001, Regional surveys and exploration for gold in
Pinti, D.L., Hashizume, K., and Matsuda, J., 2001a, Nitrogen and argon sig- the greenstone-granite terranes of Andhra Pradesh: Geological Survey of
natures in 3.8 to 2.8 Ga metasediments: Clues on the chemical state of India Special Publication 58, p. 11–27.
the Archean ocean and the deep biosphere: Geochimica et Cosmochimica Sweeney, R.E., Liu, K.K., and Kaplan, I.R., 1978, Oceanic nitrogen isotopes
Acta, v. 65, p. 2301–2315, doi: 10.1016/S0016-7037(01)00590-7. and their uses in determining the source of sedimentary nitrogen, in Rob-
Pinti, D.Z., Matsuda, J., and Maruyama, S., 2001b, Anomalous xenon in inson, B.W., ed., Stable isotopes in the earth sciences: DSIR Bulletin, v.
Archean cherts from Pilbara craton, Western Australia: Chemical Geol- 220, p. 9–26.
ogy, v. 175, p. 387–395, doi: 10.1016/S0009-2541(00)00331-4. Sylvester, P.J., Campbell, I.H., and Bowyer, D.A., 1997, Niobium/uranium
Porcelli, D., and Turekian, K.K., 2004, The history of planetary degassing as evidence for early formation of the continental crust: Science, v. 275,
recorded by noble gases, in Keeling, R, ed., The atmosphere. Amsterdam, p. 521–523, doi: 10.1126/science.275.5299.521.
Elsevier, Treatise on Geochemistry, v. 4, p. 281–318. Taylor, S.R., and McLennan, S.M., 1985, The continental crust: Its composition
Rau, G.H., Arthur, M.A., and Dean, W.E., 1987, 15N/14N variations in Creta- and evolution: Blackwell Scientific Publications, p.312.
ceous Atlantic sedimentary sequence: Implication for past changes in Tolstikhin, I.N., and Marty, B., 1998, The evolution of terrestrial volatiles: A
marine nitrogen biogeochemistry: Earth and Planetary Science Letters, review from helium, neon, argon, and nitrogen isotope modeling: Chemi-
v. 82, p. 269–279, doi: 10.1016/0012-821X(87)90201-9. cal Geology, v. 147, p. 27–52, doi: 10.1016/S0009-2541(97)00170-8.
Richardson, S.H., Gurney, J.J., Erlank, A.J., and Harris, J.W., 1984, Origin Valley, J.W., 1986, Stable isotope geochemistry of metamorphic rocks: Reviews
of diamonds in old enriched mantle: Nature, v. 310, p. 198–202, doi: in Mineralogy, v. 16, p. 445–489.
10.1038/310198a0. Venkatachala, B.S., Manoj, S., Mukund, S., and Naqvi, S.M., 1990, Archaean
Richardson, S.H., Shirey, S.B., Harris, J.W., and Carlson, R.W., 2001, Archean microbiota from the Donimalai Formation, Dharwar Supergroup,
subduction recorded by Re-Os isotopes in eclogitic sulfide inclusions in India: Precambrian Research, v. 47, p. 27–34, doi: 10.1016/0301-
Kimberley diamonds: Earth and Planetary Science Letters, v. 191, p. 257– 9268(90)90028-O.
266. Von Damm, K.L., Edmond, J.M., Measures, C.I., and Grant, B., 1985, Chem-
Robert, F., 2001, The origin of water on earth: Science, v. 293, p. 1056–1058, istry of submarine hydrothermal solutions at Guaymas Basin, Gulf of
doi: 10.1126/science.1064051. California: Geochimica et Cosmochimica Acta, v. 49, p. 2221–2237, doi:
Rollinson, H., 2002, The metamorphic history of the Isua Greenstone Belt, 10.1016/0016-7037(85)90223-6.
West Greenland, in Fowler, C.M.R., Ebinger, C.J., and Hawkesworth, Wada, E., Kadonaga, T., and Natsuo, S., 1975, 15N abundance in naturally
C.J., eds., The early Earth: Physical, chemical and biological develop- occurring substances and global assessment of denitrification from isoto-
ment: Geological Society [London] Special Publication 199, p. 328–350. pic viewpoint: Geochemical Journal, v. 9, p. 139–148.
Sachs, J.P., and Repeta, S., 1999, Oligotrophy and nitrogen fixation during east- Wang, D., and Anderson, D.W., 1994, Determination of soil pedogenic carbon-
ern Mediterranean sapropel events: Science, v. 286, p. 2485–2488, doi: ates by stable carbon isotope ratios [abs.]: Canadian Society of Soil Sci-
10.1126/science.286.5449.2485. ence Annual Conference, v. 74, p. 357.
Sadofsky, S.J., and Bebout, G.B., 2004, Nitrogen geochemistry of subducting Watanabe, Y., Naraoka, H., Wronkiewicz, D., Condie, K.C., and Ohmoto,
sediments: New results from the Izu-Bonin-Mariana margin and insights H., 1997, Carbon, nitrogen, and sulfur geochemistry of Archean and
regarding global nitrogen subduction: Geochemistry, Geophysics, and Proterozoic shales from the Kaapvaal Craton, South Africa: Geochi-
Geosystems, v. 5, Q03I15, doi: 10.1029/2003GC000543. mica et Cosmochimica Acta, v. 61, p. 3441–3459, doi: 10.1016/S0016-
Saino, T., and Hattori, A., 1980, 15N natural abundance in oceanic suspended 7037(97)00164-6.
particulate matter: Nature, v. 283, p. 752–754, doi: 10.1038/283752a0. Wellmer, F.W., Berner, U., Hufnagel, H., and Wehner, H., 1999, Carbon isotope
Sangster, D.F., and Scott, S.D., 1976, Precambrian, strata-bound, massive Cu- geochemistry of Archean carbonaceous horizons in the Timmins area:
Zn-Pb sulfide ores of North America, in Wolf, K.H., ed., Cu, Zn, Pb, and Economic Geology Monograph 10, p. 441–456.
Ag deposits: New York, Elsevier, Handbook of Strata-Bound and Strati- Williams, L.B., Ferrell, R.E., Jr., Hutcheon, I., Bakel, A.J., Walsh, M.M., and
form Ore Deposits, v. 6, p. 129–222. Krouse, H.R., 1995, Nitrogen isotope geochemistry of organic matter
Sano, Y., and Pillinger, C.T., 1990, Nitrogen isotopes and N2/Ar ratios in cherts: and minerals during diagenesis and hydrocarbon migration: Geochi-
An attempt to measure time evolution of atmospheric δ15N value: Geo- mica et Cosmochimica Acta, v. 59, p. 765–779, doi: 10.1016/0016-
chemical Journal, v. 24, p. 315–325. 7037(95)00005-K.
Schoell, M., and Wellmer, F.W., 1981, Anomalous δ13C depletion in early Zhang, Y., and Zindler, A., 1993, Distribution and evolution of argon and nitro-
Precambrian graphites from Superior Province, Canada: Nature, v. 290, gen in Earth: Earth and Planetary Science Letters, v. 117, p. 331–345, doi:
p. 696–699, doi: 10.1038/290696a0. 10.1016/0012-821X(93)90088-Q.
Schopf, J.W., and Prasad, K.N., 1978, Microfossils in Collenia-like stromato- Zhang, D., 1988, Nitrogen concentrations and isotopic compositions of some
lites from the Proterozoic Vempalli Formation of the Cuddapah basin, terrestrial rocks [Ph.D. thesis]: Chicago, University of Chicago, 157 pp.
India: Precambrian Research, v. 6, p. 347–366, doi: 10.1016/0301-
9268(78)90022-0. MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA

Memoir 198
2006
The sedimentary setting of Witwatersrand placer mineral deposits

in an Archean atmosphere
Wyatt Ernest Lawrence Minter
Department of Geosciences, University of Cape Town, Rondebosch 7700, South Africa
ABSTRACT
The 3.05 Ga U-Pb dating of uraninite grains in Dominion and Witwatersrand

conglomerates has established that they were older than the onset of Witwatersrand
sedimentation at 2.97 Ga. and therefore that they are detrital in origin. A precise Re-
Os isochron age of 2.99 Ga obtained for rounded pyrite grains associated with the
uraninite indicates that the pyrite is also detrital. Evidence of detrital forms in the
gold has confirmed its placer origin prior to modification during metamorphism. Fur-
thermore, rhenium-depletion ages ranging from 3.5 to 2.9 Ga for Witwatersrand gold
support numerous other lines of evidence that have been used in the past to interpret
the gold, uranium and pyrite concentrates as Archean placers.
The sedimentary and stratigraphic history of the Witwatersrand succession indi-
cates that uraninite, pyrite, and gold were part of a sub-aerial sediment load over a
period of 180 m.y. Net sedimentation rate during the accumulation of the Central
Rand Group is estimated at less than 14 m per million years, reflecting stratigraphic
losses due to repeated reworking. Individual paleoplacers would have been exposed
over areas of up to 400 km2 but collectively they covered a region exceeding 2000 km2
and contained more than 243 million tons of pyrite, 1.5 million tons of uraninite, and
80,000 tons of gold.
Because there is no record of either detrital uraninite or pyrite in Proterozoic red-
bed sediments on the Kaapvaal Craton, it is concluded that a change in the composi-
tion of the atmosphere took place there after 2.64 Ga.
Keywords: paleoplacer, Witwatersrand, Central Rand Group, Dominion Group, West

Rand Group, Klipriviersberg Group.
INTRODUCTION being deposited it had changed to an oxygen-rich atmosphere in

which detrital uraninite and pyrite could not survive.
Early concepts that the Witwatersrand ore deposits repre- Converse concepts that the Witwatersrand ore deposits
sented a giant composite of coalescing sub-aerial placers (Pre- were the result of hydrothermal infiltration into the pebbly sand-
torius, 1975) led to the opinion that for the detrital uraninite and stone host rocks have been raised repeatedly (Pretorius, 1975;
pyrite mineralization to have been preserved it must have been Barnicoat, et al., 1997; Phillips and Law, 2000) and acceptance
in equilibrium with the atmosphere. Therefore, because uraninite would of course negate using Witwatersrand mineralogy to indi-
and pyrite are unstable in Earth’s present atmosphere it has been cate a time of low atmospheric oxygen availability.
proposed that the atmosphere during the late Archean had a lower A recent review, however, of the status of the debate con-
oxygen content, and that by the time Proterozoic red beds were cerning these two genetic concepts (Frimmel and Minter, 2002a),
E-mail: [email protected]
Minter, W.E.L., 2006, The sedimentary setting of Witwatersrand placer mineral deposits in an Archean atmosphere, in Kesler, S.E., and Ohmoto, H., eds., Evolu-
tion of Early Earth’s Atmosphere, Hydrosphere, and Biosphere—Constraints from Ore Deposits: Geological Society of America Memoir 198, p. 105–119, doi:
10.1130/2006.1198(06). For permission to copy, contact [email protected]. ©2006 Geological Society of America. All rights reserved.
105
106 W.E.L. Minter
has highlighted a precise rhenium-osmiridium isochron age of inferred uranium resource in the world, were followed and buried
2.99 Ga for Witwatersrand pyrite, and rhenium-depletion ages by extensive bimodal volcanism dated at 3074 ± 6 Ma (Arm-
ranging from 3.5 to 2.9 Ga for Witwatersrand gold (Kirk et al., strong et al., 1991). The standard deviation of ± 50 Ma for the
2001). These findings confirm that the pyrite and gold are older Rundle and Snelling date for the detrital uraninite allows one to
than the onset of Witwatersrand sedimentation and support the place its age at between 3086 and 3074 Ma.
detrital origin of the mineralization previously indicated by a The second stage of basin development is recorded in the
3.05 Ga date for the uraninite (Rundle and Snelling, 1977). rocks of the West Rand Group. They comprise a 5150 m thick-
The purpose of this account then is to describe the sedimen- ness of sandstones and siltstones deposited after 2970 Ma (Bar-
tary environments, within which the placer minerals survived ton et al., 1989) in a shallow marine setting, which transgressed
transport and deposition, by referring to interpretations based the Dominion Group to the north and west. A small number of
on numerous systematic sedimentological studies conducted fluvial intervals containing minor paleoplacers (referred to as
since 1964. reefs) occur, particularly in the Government Subgroup (Fig. 2). A
laterally extensive lava formation near the top of the West Rand
STRUCTURAL AND STRATIGRAPHIC SETTING Group has been dated at 2914 ± 8 Ma (Armstrong et al., 1991),
providing a minimum age for the West Rand Group. The rate of
The “Witwatersrand basin” may be thought of as a set of sedimentation was at least 92 m/m.y.
cratonic successor basins overlapping each other and resting on The third stage of basin development is recorded in the Cen-
middle Archean granite-greenstone terrane (Fig. 1) that is dated tral Rand Group of rocks that unconformably overlie the West
at 3086 ± 3 Ma (Robb et al., 1992). The first stage of basin devel- Rand Group (Fig. 1). It comprises a thickness of 2880 m that
opment is recorded in the Dominion Group rocks deposited in a is dominated by fluvial cycles of coarse siliciclastic sedimentary
continental rift. The basal unit comprises a thin sequence of flu- rocks, which are separated by erosional unconformities. The
vial sediments including placers that contain uraninite, which has upper age limit of this sequence is set by the age of overlying
been dated at 3050 ± 50 Ma (Rundle and Snelling, 1977). These lavas of the Klipriviersberg Group at 2714 ± 8 Ma (Armstrong
placers, which are extensive and represent the largest unmined et al., 1991).
A Ma
B
o
27 E Stratigraphy
2023 Vredefort impact
4 2054- Bushveld Igneous Complex
Carleton- 3 2 2059 Rooiberg Group
ville Johannesburg
6 5
7 1 2250 Transvaal Supergroup
EVANDER Pretoria
GOLD FIELD
Evander Group
2350
Klerksdorp
8 2432 Chuniespoort
o
27 S 27oS Group
Vredefort
N 2642 Black Reef Quartzite Fm.
Pniel/Wolkberg Groups
0 50 100 km 2709 Ventersdorp Platberg Group
Supergroup
2714 Klipriviersberg Gr.
2764 upper Central
Goldfields: Paleoslope Rand Group
2837
1 - Evander Welkom Major fault
9
Witwatersrand
2 - East Rand lower Central

Supergroup
Witwatersrand
Goldfield Rand Group

Supergroup
3 - Central Rand <2894

4 - West Rand Central Rand Group
2914
5 - South Deep West Rand Group West Rand
6 - Western Areas Group
Dominion Group <2970
7 - Carletonville
Archean granitoid 3074-
8 - Klerksdorp 3086
Dominion Group
9 - Welkom o
Greenstone
27 E Middle Archean basement
Figure 1. (A) Simplified surface and subsurface geological map of the Witwatersrand Basin (Frimmel and Minter, 2002a), also showing the
distribution of Archean granitoid domes, the location of the principal goldfields, major faults, and paleocurrent directions of Witwatersrand reefs
(from Minter and Loen, 1991, and Frimmel and Minter, 2002b). (B) Main stratigraphic units and ages (for sources see text).
Lithology
Subgroup
Reefs
Group
Meters Formation
Venterspost Ventersdorp Contact Reef
South Deep
2000 Mondeor
Turffontein
Bastard, EA, Upper Elsburg
Central Rand Elsburg/

Eldorado
Denny’s, Beatrix, Composite
Kimberley Crystalkop, A, Kalkoenkrans, B, Kimberley
1000 Booysens
Johannesburg
Krugersdorp
Basal, Steyn, Vaal, Saaiplaas, Leader
Luipaardsvlei
Livingstone
Randfontein
Main Middelvlei
Blyvooruitzicht
North, Main, South, Carbon Leader, Commonage
0 Ada May, Beisa
Maraisburg
Jeppestown
Roodepoort
Crown
Babrosco
Veldskoen, Inner Basin reefs
1000 Rietkuil M
Koedoeslaagte Buffelsdoorn, Outer Basin reefs
Afrikander
Elandslaagte Government
Government
Palmietfontein
M
Tusschenin Rivas Reef

Coronation
Coronation M, D
2000
Promise
West Rand
D
Bonanza
Bonanza
M
Hospital Hill
3000
Brixton
LEGEND
M Magnetic shale
M
D - Diamictite
4000 M Reef
Conglomerate
Parktown Sandstone
Shale
M Lava
5000 Orange Grove
Figure 2. Generalized stratigraphic column for the Witwatersrand Supergroup, exemplified by a section through the Klerksdorp goldfield, as pro-
posed by the South African Committee for Stratigraphy (unpubl.); also listed are the stratigraphic positions of the main auriferous conglomerate
(reef) horizons in the various goldfields (Frimmel and Minter, 2002b).
108 W.E.L. Minter
The net rate of sedimentation of the Central Rand Group was directions into the basin (Fig. 1). The top surfaces of the placer
only 14 m/m.y., reflecting the cumulative amount of degradation sediment packages were leveled off either by wind deflation
associated with the numerous unconformities (Sadler, 1981). in temporarily inactive parts of the depositional environments,
This truncation and reworking of the stratigraphy was apparently or by wave action after inundation by a body of water (Fig. 4).
up to seven times more prevalent than in the underlying West Lag deposits, especially on basal unconformities, degradation
Rand Group succession and might account for the greater inci- scour surfaces, clast-supported conglomerates, and winnowed
dence in the Central Rand Group of detrital heavy mineral con- reworked top-surfaces are preferentially mineralized (Fig. 5).
centrations represented by the paleoplacer orebodies that have
been exploited during the past 118 yr of mining. Mineral Concentration
SEDIMENTOLOGY Although the clastic sediment loads that produced paleopla-

cers would have traveled in suspension during flood conditions,
Stratigraphic and paleocurrent data indicate that the Central and been initially sorted by settling processes as the flow veloc-
Rand Group was deposited in a shrinking continental basin. The ity decreased, it has become evident from detailed studies that
paleoplacers were deposited at locations peripheral to the basin in selective entrainment of the less dense particles during bedload
a variety of sedimentary environments that ranged from alluvial transport (Slingerland and Smith, 1986) was the most effective
fans to braid plains and braid deltas (Minter, 1978; Minter and concentrating mechanism of smaller, more dense particles. Fig-
Loen 1991). In general, the paleoplacers are hosted by gray-col- ure 6 illustrates a number of consecutive beds of heavy-mineral
ored, medium-grained to coarse-grained sandstones, composed lags, dominated by rounded detrital pyrite, which demonstrate
of quartz-arenite, that occupied fluvial channels in braided drain- the process.
age systems and contained a generally oligomictic gravel compo- Measurements of the areal size frequency distribution of
nent as lags, bars, and sheets (Fig. 3). Primary sedimentary struc- rounded pyrite in both the Basal and Steyn paleoplacers of the
tures within the host sediments indicate unidirectional paleoflow Welkom goldfield (Minter, 1978) demonstrate further how selec-
A B
C
Figure 3. (A) Emerging longitudinal gravel bar in the Kicking Horse
River, at Field, British Columbia, with lateral accretion of sand facies
draping the left-hand side, and sand dunes occupying the channel be-
low water level. Shovel for scale. (B) A remnant pillar of the Elsburg
No. 5 Reef at Western Reefs Gold Mine in the Klerksdorp Goldfield.
The face of the pillar is 2 m high and exposes a transverse section
through a mineralized longitudinal conglomerate bar with laterally ac-
creted planar cross-bedded sandstone to the left and overlying trough
cross-bedded sandstone filling the channel (Smith and Minter, 1980).
(C) A diagram illustrating sample sites that indicate the distribution of
gold in the paleoplacer lithofacies described above. The blocks num-
bered from Tr (Trace) to 78 represent gold content measured in grams
per ton and illustrate association between gold tenor and particular
lithofacies.
Sedimentary setting of Witwatersrand placer mineral deposits 109
Figure 4. An underground exposure of the Leader Reef at Free State

Saaiplaas Gold Mine in the Welkom Goldfield. The stope face repre-
sents a section across a channel filled with clean gray-colored paleo-
placer sediment. Note the higher bed relief to the left and full channel
depth to the right. The arrangement of inclined conglomerate facies
indicates that gravel shed off the high relief and lined the channel,
which was then filled by a lens-shaped sheet of trough cross-bedded
sand. The sequence was topped by a planar layer of gravel that re- Figure 6. In this sample of the Basal Reef from Free State Geduld Gold
flects reworking of the surface by either eolian deflation or wave ac- Mine in the Welkom Goldfield, imbricated pebbles lying on pyrite lay-
tion. Fine-grained rounded pyrite, gold, and an assemblage of other ers indicate that the current flowed from the left, winnowing the un-
heavy detrital minerals are located in the gravel facies, particularly in derlying layer of pyrite grains and removing them in the turbulent zone
the bottom and top units where gold concentrations of up to 100 g/t upstream of pebbles, which then tilted into the depression. The small
occur, and distributed as cross-bedded laminations in the sand facies. ellipsoidal pyrite pebbles have also been imbricated in the same sense
The sediment overlying and underlying the gray paleoplacer sediment as the quartz pebbles have been. The open framework of the dense
has a yellowish color due to its argillaceous content. Hammer handle pyrite bed has, in this instance, provided spaces in which finer-grained
(30 cm) for scale. gold and other small heavy mineral particles were trapped. This re-
sembles the bed of a Malaysian tin jig. Width of sample is 15 cm.
tive entrainment produces a regional measure of sorting insofar

as the size of rounded pyrite decreases from a mean diameter of
7 mm to 1 mm over a transport distance of 25 km.
Calculations of the threshold shear stress required to entrain
the detrital mineral grains concentrated together in a set of cross-
bedded laminations (Figs. 7, 8) indicated that at the velocities
associated with dunes (trough cross-bedding), which is the most
common bedform preserved in the quartz arenite lithofacies of
paleoplacers, the quartz, zircon, pyrite, chromite and gold par-
ticles could be entrained (Table 1). The size frequency distribu-
tion of the disc-shaped gold grains is lognormal, as for natural
sediments, with a mean nominal diameter of 136 μ and a standard
deviation of 0.48, which indicates good sorting.
Inclined degradation surfaces sampled in a section across
a distal Steyn Reef channel on Free State Saaiplaas Gold Mine
in the Welkom Goldfield (Minter, 1989) illustrates how selec-
Figure 5. A bed of conglomerate in the Elsburg No. 5 Reef at Western tive entrainment reworks placer sediment to leave a concentrated
Reefs Gold Mine in the Klerksdorp Goldfield, which illustrates selec- assemblage of denser finer-grained minerals. The correlation
tive accumulation of rounded pyrite and other heavy detrital minerals between zircon, uraninite, and gold demonstrates this (Figs.
in the upper, clast-supported, portion of the conglomerate. This is in-
terpreted to be the result of selective separation of fine-grained dense 9A–C). The greater the reworking becomes, the pebblier the lag
minerals from passing sand facies and their protection by the gravel concentrate becomes, and the greater the detrital mineral concen-
framework from removal. Width of image is 20 cm. tration is. The highest concentrations collecting in this way occur
110 W.E.L. Minter
Figure 7. A rare specimen of cross-bedded heavy minerals from the

Basal Reef at Free State Geduld Gold Mine, containing abundant gold B
particles that have retained their original detrital forms (Minter et al.,
1993b). Measurement of these grains has provided an opportunity to
compare the relative sizes of associated heavy minerals in the same
foreset laminations. The four tangential foreset planes visible converge
toward the right where the toeset meets the bottomset. Width of speci-
men is 8 cm.
on major distal disconformities or unconformities like those

associated with the Vaal Reef, the Basal and Steyn Reefs, the
Carbon Leader Reef, and the Crystalkop Reef. The duration of
these longer-term surfaces of degradation is also displayed by the
Chemical Index of Alteration of the strata immediately underly-
ing the unconformity (Frimmel and Minter, 2002a). Values of up
to 90% have been recorded (Fig. 10). C
The spatial and temporal association of paleoplacer miner-
alization with the surface of degradation is amply demonstrated
by contemporaneous loadcasting of channel sediments in the
Saaiplaas Reef, into the underlying thixotropic sediments of
the Harmony Formation (Fig. 11). This soft-sediment deforma-
tion of paleoplacer deposits in erosion channels above the Main
Reef, the Kimberley Reef, and the Black Reef deposits has been
objectively described by Papenfus (1957). Attempted mining of
such disrupted paleoplacer beds has been frustrated and generally
unsuccessful.
Sedimentary Settings
Most of the goldfields shown in Figure 1 represent multiple

composites of any one of these sub-environments or combina- Figure 8. (A) Left: preserved disc-shaped 300-μ-diameter toroidal gold
tions of a number of different sub-environments (Minter, 1991). grains, representative of thousands extracted (Minter et al., 1993) from
the specimen illustrated in Figure 7, which demonstrate their detrital
Each deposit might range from 100 to 400 km2 in area and com- origin. Right: Toroid from Baaga stream in Yakutia which is reworking
posites might exceed 600 km2. an ancient wind deflation surface (Minter, 1999). (B) Sections through
Alluvial fans containing paleoplacers are limited to the most toroids from the Basal Reef (left) and the Baaga stream (right) indicat-
proximal preserved remnants of the basin margin. Examples ing a unique form believed to have been formed by the impact (peen-
include the Livingstone Reefs on the West Rand, the Massives ing) of sand grains on gold flakes in an eolian environment (Minter,
1999). (C) A concentrate of sand-blasted platinoid and gold grains from
at Western Areas on the Far West Rand, and the EA Reefs in the a modern raised beach placer deposit in New Zealand, South Island. An
Eldorado Formation of the northern part of the Welkom gold- evolution of 200-μ-diameter toroidal gold shapes from primary rims to
field (Minter et al., 1986). The EA sediments, for example, form almost hollow spherical forms is evident (Mitchell, 1996).
a bajada of coalescing alluvial fan lobes that seem to be spaced

TABLE 1. RELATIONSHIP BETWEEN MEASURED at 1500 m intervals along the north-striking folded structure that
AND THEORETICAL PLACER GRAIN DIAMETERS defines the western border of the Welkom goldfield. These fans
Mineral Specific Observed Theoretical size Threshold shear
gravity size (mm) Quartz basis stress extend downdip for 750 m eastward into the depository (Gray, et
(mm) (dynes/cm2) al., 1994). They were shed off the upturned western margin of the
Quartz 2.65 0.384 2.41
Central Rand Group as a syn-sedimentary response to active tec-
Zircon 4.68 0.133 0.225 16.7
tonic compression during the final stages of Central Rand Group
Pyrite 5.00 0.158 0.213 11.2
deposition. At the Target mine, individual fans comprise up to 15
Gold 19.32 0.136 0.072 76.3
repeated sequences of oligomictic mineralized conglomerates,
Note: Bed shear stresses in dune regimen for median grain size, D50 =
0.4mm, range from 20 to 80 dynes/cm 2 (James and Minter, 1999).
quartz-arenites, and lithic arenites in a 200-m-thick succession.
They evidently derived placer concentrates by reworking sub-
cropping paleoplacers in the upturned footwall.
100
C
A Gravel
Sand
Gravel & Sand
Au (ppm)
10
1
r = 0.65
r = 0.45
300
Zr (ppm)
B 200
100
1 10 100 1000
U (ppm)
100
r = 0.23
Au (ppm)
Figure 9. (A) A transverse underground exposure of the inner bend of a channel 10

edge in the Steyn Reef at Free State Saaiplaas Gold Mine in the Welkom Gold-
field. Inclined degradation surfaces on which reworking has taken place have
permitted selective entrainment of the sand fraction to leave behind pebbles and
fine-grained dense detrital minerals like gold, uraninite, and zircon. Clinorule (1
m) for scale. Photo located at meter mark 7 shown in Section B. (B) Section il- 1
lustrating with color packages of sediment wedges separated by master bedding
plane surfaces of degradation. Vertical and horizontal scale in meters. (C) The
results of chemical analysis indicate correlations between these minerals that
support their collective hydraulic placer origin (Frimmel and Minter, 2002a). 10 100 1000
Zr (ppm)
112 W.E.L. Minter
of clast-supported conglomerates, and on eolian deflation sur-

0 faces associated with ventifacts. These were then buried beneath
a prograding Eldorado Formation fan (Minter et al., 1988). In the
Meters above reef
distal black mudstone lithofacies of the prograding Eldorado fan,

heavy minerals were entrained from the paleosurface by debris
-10 Crystalkop Reef flows and suspended in an unsegregated dispersion quite atypical
9 Shaft, 59 Level of the average paleoplacer. The Beatrix Reef environment cov-
ered an area of ~78 km2 in which mixed lithofacies are recorded,
indicating fluvial, eolian, and semi-arid mud-cracked playa lake
-20
sediments.
Many of the Witwatersrand paleoplacers were deposited in
70 80 90 100 gravelly braid plains resembling the outwash sheets in southern
CIA Iceland (Minter, 1982a, p. 133). Examples are found in the Leader
Reef in the Welkom goldfield (Smith and Minter, 1980; Minter et
Figure 10. Variation in Chemical Index of Alteration (CIA) with dis-
tance from the Crystalkop Reef as measured in a profile at 9 Shaft, 59 al., 1986) and in the Main Reef and Kimberley Reefs of the West
Level, Vaal Reefs Gold Mine, in the Klerksdorp Goldfield (Frimmel Rand and Central Rand goldfields. A classic example of a braid-
and Minter, 2002a). plain channel-fill paleoplacer is illustrated in an exposure of the
Kalkoenkrans paleoplacer at the Oryx Goldmine (Fig. 12).
The most extensive paleoplacers in the Witwatersrand were
braid deltas associated with shorelines (McPherson et al., 1987).
They include the Basal and Steyn Reefs (Minter, 1978), the Vaal
Reef (Minter, 1976, Antrobus et al., 1986), the Carbon Leader
Reef (Buck and Minter, 1985), and the Kimberley UK9 Reefs
(Armstrong, 1966). Active flowing fluvial drainage systems that
deposited the placers were on average 50 cm deep and covered
only a small part of the deposits at any one time. Therefore the
remaining subaerial regions were exposed to eolian deflation.
Ventifacts recovered from many of the paleoplacers attest to this
(Fig. 13), and it is likely that wind deflation played an impor-
tant role in condensing the stratigraphy and thereby concentrat-
ing mineral accumulations (Minter, 1999). These deflation sur-
faces are preserved as thin ventifact lags (Fig. 14) on planar to
gently undulating unconformity surfaces, possibly representing
groundwater-table control on deflation, and as planar winnowed
Figure 11. An exposure of the Saaiplaas Reef in Crosscut 100 on 74 surfaces on gravel-bar tops. Measurements of facet orientations
Level, 4 Shaft at President Steyn Gold Mine in the Welkom Goldfield. on in situ ventifacts in the Crystalkop Reef in the Klerksdorp
The paleoplacer channel, which originally comprised of light gray- goldfield have indicated that the paleowind direction was from
colored quartz arenite containing trough cross-bedded laminations 12° east of present-day north (Minter et al., 1993a).
of rounded pyrite, and overlay a basal pebbly lag concentrate, now The Basal and Steyn paleoplacers in the Welkom goldfield
displays contemporaneous loadcasting into the thixotropic underly-
ing (waxy yellow) argillaceous sandstone of the Harmony Formation were adjacent braid deltas that entered the depositional area from
in such a way that the pebble lag surrounds the oval-shaped loadcast sources to the west and southwest, respectively, and terminated
and the cross-bedded quartz arenite interior is deformed. Mineraliza- in a standing body of water (Fig. 15) (Minter, 1982b). The Steyn
tion is associated with the pebble lag surface. Geological pick head paleoplacer is slightly younger than the Basal paleoplacer in that
for scale. it has truncated and overlapped the southern edge of the Basal
paleoplacer fan. Both paleoplacers were eventually inundated
by water and buried beneath siltstone deposited from suspension
Another paleoplacer associated with alluvial fan develop- (khaki shale). They have been preserved over an area of ~600
ment is the Beatrix Reef, which occurs at the southern end of the km2. In the case of the Steyn paleoplacer, not only were the peb-
Welkom goldfield. It lies on the pediment of an early Eldorado bles sorted by size downstream, from 70 mm in diameter in the
Formation alluvial fan where Aandenk paleoplacers outcropped. truncated proximal zone to 11 mm in the most distal zone, but
The recycled detrital minerals eroded from the Aandenk were so was the pyrite, from a mean diameter of 0.33 mm to 0.13 mm
evidently spread out as a concentrated deflation surface downdip (Minter, 1978, 1991), thus supporting its detrital origin (Fig. 16).
from the outcrop strike. These were distributed by shallow chan- The detrital nature of accompanying uraninite is illustrated by its
nels and preserved in arenite lithofacies as laminae, in the matrix correlation with zircon and gold (Frimmel and Minter, 2002a)
Figure 12. This Kalkoenkrans Reef exposure at the Oryx Gold Mine in Figure 13. (A) Ventifact from the Crystalkop Reef at Vaal Reefs Gold
the Welkom Goldfield records the tapered edge of a heavily mineral- Mine. Scale in mm divisions; (B) Ventifact 20 mm in diameter from
ized longitudinal gravel bar located at the base of the channel. It has the Vaal Reef at Hartebeestfontein Gold Mine, in the Klerksdorp
been buried beneath white trough cross-bedded quartz arenite (dunes Goldfield.
constituting the face of the exposure), which filled the channel, and
then during abandonment was covered by black mud (on the roof of
the exposure) that dried and cracked in the subaerial environment. The
mud polygon boundaries have filled with light-colored sandstone.
DETRITAL MINERALOGY
and by a steadily increasing uranium/gold ratio down the paleo-
slope, which is a result of mineral sorting. Uraninite is therefore From the position of an observer, standing at the southern
generally enriched in the more distal fluvial facies of Witwa- edge of the Basal and Steyn Reef deposit (Fig. 15), the view to
tersrand paleoplacers (Minter, 1978). the north for 30 km and to the east for 25 km would have been
The Witwatersrand alluvial-fan and braid-plain models a gray, pebbly sand flat mineralized with bronze-colored pyrite,
can be extended to other ancient pyrite- and uraninite-bear- black ilmenite and chromite, pink zircon, and black uraninite, in
ing paleoplacers. These include the Ventersdorp Contact Reef that order of abundance, with lesser amounts of gold. The gold
and the Black Reef, which overlie the Witwatersrand (Minter, content across such surfaces as seen in the Carbon Leader Reef,
1991), paleoplacers in the Pongola Supergroup (Minter, 1991), the Vaal Reef, and the Steyn and Basal Reefs averages about a
the Bababudan paleoplacers in India (Srinivasan and Ojakangas, million ounces per square kilometer.
1986), paleoplacers on the Pilbara Craton in Australia (England The pyrite is heterogeneous in origin, up to six species occur-
et al., 2002), paleoplacers at Elliot Lake (Theis, 1976) on the ring, chemically distinguished by different sulfur isotopic signa-
Canadian Shield and elsewhere in North America (Roscoe and tures (England et al., 2002), and has been abraded to coarse- to
Minter, 1993), and the Moeda and Jacobina paleoplacers on the fine-grained rounded forms, the outer margins truncating internal
São Francisco Craton in Brazil (Minter, et al., 1990, Garayp et structures and textures (Fig. 17). Concentrations of pyrite range
al., 1991, and Minter, 1991). from 50,000–200,000 ppm.
114 W.E.L. Minter
Figure 15. An artist’s impression of the distribution of the Steyn

(left) and Basal (right) paleoplacers in the Welkom goldfield, facing
west. The drainage patterns, and distal depositional limits, are based
on underground data and borehole intersections. The upslope front,
which was truncated by unconformable onlapping formations (Minter,
1982b) extends for 30 km to the north. Black dots mark the sites of
discovery boreholes.
Figure 14. A polished slab of the Vaal Reef from Stilfontein Gold Mine,
which represents a thin pebbly sandstone lag containing ventifacts and
fine-grained rounded detrital pyrite. Note thin carbon seam on basal
contact. Diameter of large quartz pebble is 20 mm.
Figure 16. An exposure illustrating brassy-colored granular pyrite con-
centrated along the upper parts of cross-bedded foresets (inclined to
the right) in the Steyn Reef at President Steyn Gold Mine in the Wel-
The ilmenite is fine-grained and occurs throughout the kom Goldfield.
Witwatersrand sequence, concentrated on all scour surfaces.
In Witwatersrand ores it has been altered during syn-deposi-
tional weathering to leucoxene, a composite of rutile needles between all these minerals and uraninite, illustrating an hydrau-
and authigenic quartz. In the Composite Reef on the West lic control over detrital mineral concentration (Slingerland and
Rand, the TiO2 content ranges between 420 and 4000 ppm Smith, 1986).
(Tucker, 1980), which is similar to contents of 500–11,700 ppm Uraninite was first identified in Witwatersrand ores by Coo-
at Elliot Lake (Theis, 1976) and 1310 ppm in the Moeda in per (1923) and then quantified by Weston Bourret in 1944–45
Brazil (Minter et al., 1990). during the Manhattan Project (Bourret, 1975). In the original
Chromite is common throughout, and averages ~3000 ppm survey in 1944, uranium was also recorded from Black Reef on
in the Steyn Reef (Frimmel and Minter, 2002a). Zircon is also the East Rand at Vogelsstruisbult where a gold and pyrite rich
abundant throughout. In the Composite Reef the ZrO2 content specimen assayed 3350 ppm U3O8. This is an exceptionally high
ranges between 83 and 1455 ppm (Tucker, 1980) and in the value because average ore grades at Modderfontein East were
Steyn Reef ~410 ppm, which are similar to values between 10 between 26 and 37 ppm. Recent borehole intersections of Black
and 600 ppm at Elliot Lake (Theis, 1976), and 54 ppm in the Reef from the East Rand (Camden-Smith, 2004, personal com-
Moeda (Minter et al., 1990). There are significant correlations mun.) range from 2 to 37 ppm over the full width.
were exploited. However, Chernobyl and other disasters resulted

in lobbies against nuclear power and the price fell to the point
where closure of many of these operations took place.
DISCUSSION
Today there is general agreement that the majority of Wit-

watersrand gold particles, which appear late in the paragenetic
sequence, are indeed hydrothermal precipitates. Thus, the debate
has shifted to the question of whether the source of the gold in the
auriferous hydrothermal fluid(s) was proximal, fluvially depos-
ited, detrital gold within the conglomerate beds, subsequently
mobilized by post-depositional fluids over short distances (modi-
fied placer model), or whether the source of the gold was exter-
nal to their host rocks. The latter case would imply long-dis-
Figure 17. An enlargement of part of Figure 6 to illustrate how the mar-
gins of coarse-grained rounded pyrites display an abraded truncation tance chemical transport by hydrothermal fluids (hydrothermal
of internal layers and crystal structures in the pyrite. The large pyrite model).
grain in the center of the image is 5 mm in diameter. The modified paleoplacer model has achieved strong sup-
port from the discovery of two distinct morphological types
of gold that occur together on a millimeter scale; one type dis-
The first “yellow cake” was produced from a pilot plant at plays morphological features typical of detrital gold, whereas
Blyvooruitzicht in the Carletonville goldfield in 1951. Commer- the other occurs as irregularly shaped aggregates or well-crys-
cial production began in 1952 at West Rand Consolidated in the tallized, euhedral overgrowths (Minter et al., 1993b). In this
West Rand goldfield where the Monarch Reef, represented by a model, transport of detrital gold particles into the host sediments
small-pebble distal placer, was mined at a mean grade of 2860 is believed to have taken place by fluvial processes. Physical
ppm U3O8, ten times richer than the average Witwatersrand ore. deformation of these particles as a result of post-burial compac-
Uranium in the Witwatersrand ores also occurs in the form tion, which pressed sand or heavy-mineral grains into the sur-
of brannerite, an alteration product of uraninite, which is attached faces of discs, is evident (Minter et al. 1993b). Chemical mobi-
to leucoxene. It is interesting to note that the ratio of uraninite lization of the gold as a result of short-range (micrometer- to
to brannerite in Witwatersrand ores varies. In the Welkom gold- meter-scale) transport, took place largely within the host rock,
field this variation appears to be systematic insofar that it changes by infiltrating hydrothermal fluids and/or degradation of in situ
up the stratigraphic column from 8.7 in the Steyn Reef to zero hydrocarbon or hydrous phases.
in the Beatrix Reef (Minter et al., 1988). All the uranium in the In Brazil, comparable Paleoproterozoic sedimentary cover
Black Reef, which is the uppermost reef in the sequence, reports rocks above greenstone belts also contain auriferous quartz peb-
to brannerite (England, 2004, personal commun.). Could this ble conglomerates with a few large gold deposits at the base
be related to repeated reworking in an increasingly oxygenated of fluvial and beach sequences (e.g., Jacobina deposit in Bahia
evolving atmosphere? and Moeda deposit, Quadrilatero Ferrifero, in the northern and
Up to 15% of the uranium in the Vaal Reef is attached to southern São Francisco Craton, respectively). An estimated
leucoxene and requires a high-pressure leach to extract. Accord- 50 tons (t) of Au have been recovered from conglomerates in
ing to Camisani-Calzolari et al. (1985), the average metallurgical the Serra de Jacobina region (Teixeira et al., 2001). A foreland
recovery of uranium has been between 20% and 80% at an aver- setting for the Jacobina Basin has been proposed (Ledru et al.,
age grade of 271 ppm. The Elliot Lake paleoplacers produced 1997), with U-Pb data for detrital zircon and a Rb-Sr whole-rock
an average grade of 1000 ppm and the grade of a typical Moeda age for a post-collisional granitoid in the Jacobina-Contendas-
intersection is 157 ppm (F.E. Renger, 2004, personal commun.). Mirante belt providing maximum and minimum age constraints
It was estimated in 1974 (Armstrong, 1975) that 35% of for the Basin fill of 2086 and 1883 Ma, respectively. The bulk of
known world uranium resources were contained in quartz pebble the gold extracted from the Jacobina deposits came from pyritic
conglomerates in South Africa and Canada and that the slimes quartz pebble conglomerate and minor quartz arenite intercala-
dumps alone contained 500 million pounds of U3O8. The Arab oil tions of the lowest formation, interpreted as representing alluvial
embargo in 1973 apparently stimulated the installation of nuclear fan and braided stream deposits. The gold occurs as fibrous to
power reactors, increased the demand for uranium and therefore oval-shaped particles, typically associated with euhedral pyrite.
its market price, and generated worldwide exploration for further Detrital phases include zircon and chromite, whereas metamor-
resources. Processing the pyrite, uranium, and gold contents of phic/hydrothermal phases comprise pyrrhotite and pyrite, fuch-
the Witwatersrand slimes dumps became a lucrative exercise, as site, white mica, minor andalusite, rutile, and tourmaline (Ledru
demonstrated by the Ergo operation, and most of these deposits et al., 1997).
116 W.E.L. Minter
By analogy with the Witwatersrand, the likely hydrother- carry an O isotopic signature comparable with orogenic vein
mal gold source was detrital gold in the local conglomerate quartz (Vennemann et al., 1995). The lack of gold in these
beds. A possibly more intense metamorphic or hydrothermal deposits might therefore simply reflect a lack of suitable
overprint would have led to a complete obliteration of the source rocks in the eroded hinterland.
original detrital microtextures and morphological forms of Uraninite and pyrite occur throughout the Witwatersrand
the gold, as is the case in some Witwatersrand reefs (e.g., the succession from Dominion at 3086 Ma to the upper Central
Ventersdorp Contact Reef). The S isotopic composition of the Rand at 2837 Ma and then into the Ventersdorp Contact Reef
pyrite points to a magmatic origin (Teixeira et al., 2001), but at 2714 Ma. They occur again in the Black Reef dated at 2642
it remains unclear whether the hydrothermal pyrite is related Ma (Eriksson, et al., 1995). This is the last known placer
to sulfidation of originally detrital Fe- oxides or reflects mobi- deposit before the onset of red beds on the Kaapvaal Craton.
lized detrital pyrite as in the Witwatersrand. Although the oxidation state of the contemporaneous
In contrast to the Witwatersrand, the Moeda deposits atmosphere plays a crucial role in the mineralogy of the detri-
(Minter et al., 1990) are not in a thick siliciclastic succession but tal minerals associated with the gold, it does not control the
rather at the base of only a thin (120 m) siliciclastic, predomi- distribution of detrital gold particles as evidenced by the Fe-
nantly arenitic succession, whose age is loosely constrained oxide bearing auriferous conglomerates of the Tarkwa Basin.
between 2.8 and 2.2 Ga. The gold occurs together with heavy The second largest known paleoplacer gold deposits outside
mineral concentrates, predominantly coarse-grained rounded the Kaapvaal Craton are hosted by the Tarkwaian System of
pyrite, in the matrix of cobble conglomerates that form dis- Ghana, whose age is constrained between 2133 and 2097 Ma
crete bars and sheets above an angular unconformity. Sedi- (Oberthür et al., 1998; Pigois et al., 2003). Past production was
ment transport was from the northwest into individual basins ~310 t Au (Pigois et al., 2003). Reserves in 2002 totaled 131 Mt
that were separated by basement ridges of Archean chert and at 1.6 g/t for 215 t Au within a resource of 340 Mt at 1.6 g/t for
iron formation. Mineralogy, geochemistry, and gold concen- 534 t Au (Gold Fields Ltd., unpublished annual report, 2003).
trations are similar to those of the Witwatersrand deposits The known amount of gold in that depository is thus two
(Minter et al., 1990). orders of magnitude less than in the Witwatersrand. Similar
Apart from gold, associated mineral phases that are key to the Witwatersrand, the gold is hosted by quartz pebble
to any plausible interpretation of the Witwatersrand deposits conglomerate beds, interbedded with sandstone units, all of
are pyrite and uraninite. Between 1952 and 1975, up to 1.5 which experienced greenschist facies metamorphism that led
million tons (Mt) of U3O8 were produced from Witwatersrand to recrystallization and redistribution of many of the miner-
quartz pebble conglomerates at an average grade of 271 ppm als, including gold. In contrast to the Witwatersrand, however,
U3O8. Iron oxides are conspicuously lacking, with the prin- heavy minerals associated with the gold are hematite and
cipal Fe-bearing phase being pyrite. Both pyrite and urani- magnetite rather than pyrite, in addition to rutile and zircon.
nite occur predominantly in rounded form. In the modified Uranium-bearing minerals are rare to absent and there is no
placer model, these rounded pyrite and uraninite grains are significant bitumen. The individual gold grains are located
considered detrital, as is the gold (Frimmel, 1997). In contrast, in the matrix, concentrated along heavy-mineral foresets and
in the hydrothermal models, they are interpreted as pseudo- small fractures, on quartz-pebble boundaries, in pressure-
morphic replacements of detrital Fe-oxides, Fe-pisolite, fer- solved quartz between quartz pebbles, as inclusions in hema-
ricrete, banded iron formation, and Fe-rich shale (Phillips and tite derived from the metamorphic oxidation of detrital mag-
Law, 2000) and/or products of post-depositional dissolution netite, and associated with metamorphic chlorite and white
and re-precipitation mechanisms (Phillips and Myers, 1989; mica (Hirdes and Nunoo, 1994; Pigois et al., 2003).
Barnicoat et al., 1997). As both pyrite and uraninite are sensi- Another analogue to Witwatersrand-type deposits is sited
tive to redox conditions, the genetic interpretations of rounded in the Roraima Supergroup, in northern South America (San-
pyrite and uraninite grains in the Witwatersrand metasedi- tos et al., 2003). It consists of largely undeformed sandstones,
mentary rocks constrain models of gold genesis as well as minor conglomerates, shale, and felsic ash-fall tuffs, all of
understanding of the evolution of the Archean atmosphere which rest on a 2.25–2.00 Ga granitoid-greenstone terrane
(for contrasting views see Rasmussen and Buick, 1999, and (Trans-Amazon province) of the Guyana Shield in the north-
Ohmoto et al., 1999). west of the craton. That terrane contains lode-gold deposits
There are other uraninite and pyrite placer deposits that and probably represents a continuation of the lode-gold bear-
bear many similarities to those of the Witwatersrand, although ing, Eburnean granitoid-greenstone belts of West Africa. The
none of these contain significant amounts of gold. These Roraima sedimentary rocks represent braided fluvial sedi-
include the 2.9–2.6 Ga Bababudan Group in India (Sriniva- ments deposited in alluvial plain to subaerial braided delta,
san and Ojakangas, 1986) and the 2.45 Ga Elliot Lake Group, and possibly shallow marine environments, for which a fore-
Huronian Supergroup, in Canada (Krogh et al., 1984, Sutton land setting has been suggested (Santos et al., 2003). The flu-
and Maynard, 1993). Interestingly, in contrast to the Witwa- vial sediments appear enriched in Au with a background con-
tersrand, none of the quartz pebbles in the Elliot Lake Group centration in the arenaceous fraction of 10 ppb (i.e., more than
the average 6 ppb for Witwatersrand arenites) and maximum Armstrong, G.C., 1966, A sedimentological study of the U.K.9 Kimberley
gold grades in quartz pebble conglomerates on basal degrada- reefs in part of the East Rand [M.S. thesis]: Johannesburg, University
of the Witwatersrand, 65 p.
tion surfaces of as much as 26 g/t over 5 cm thickness. In con- Armstrong, R.A., Compston, W., Retief, E.A., William, L.S., and Welke, H.J.,
trast to the other examples above, no significant metamorphic 1991, Zircon ion microprobe studies bearing on the age and evolution of
or hydrothermal overprint is indicated for the Roraima host the Witwatersrand triad: Precambrian Research, v. 53, p. 243–266, doi:
10.1016/0301-9268(91)90074-K.
rocks. Similar to the Witwatersrand deposits there is an appar- Barnicoat, A.C., Henderson, I.H.C., Knipe, R.J., Yardley, B.W.D., Napier,
ent sedimentological control on Au grade and, combined with R.W., Fox, N.P.C., Kenyon, A.K., Muntingh, D.J., Strydom, D., Win-
the occurrence of gold as well rounded nuggets, the evidence kler, K.S., Lawrence, S.R., and Cornford, C., 1997, Hydrothermal gold
mineralization in the Witwatersrand basin: Nature, v. 386, p. 820–824,
supports a paleoplacer model. The age of the Roraima aurif- doi: 10.1038/386820a0.
erous conglomerates is constrained by a U-Pb single zircon Barton, E.S., Compston, W., Williams, I.S., Bristow, J.W., Hallbauer, D.K.,
age of 1901 ± 1 Ma for a tuff bed overlying a conglomerate and Smith, C., 1989, Provenance ages for the Witwatersrand Supergroup
and the Ventersdorp Contact Reef: Constraints from ion-microprobe U-
bed (H.E. Frimmel, 2001, personal commun.). The O isotopic Pb ages of detrital zircons: Economic Geology and the Bulletin of the
composition of the quartz pebbles indicates derivation of the Society of Economic Geologists, v. 84, p. 2012–2019.
vast majority of the pebbles from orogenic vein quartz (Minter Bourret, W., 1975, Investigation of Witwatersrand uranium-bearing quartz-
pebble conglomerates in 1944–45, in Armstrong, F., ed., Genesis of
et al., 2002). This, together with paleocurrent directions and uranium- and gold-bearing Precambrian quartz-pebble conglomerates:
detrital zircon ages (Santos et al., 2003), points to a source U.S. Geological Survey Professional Paper 1161-A, p. A1–A6
for the sediment, and by implication the gold, in the 2.0 Ga Buck, S.G., and Minter, W.E.L., 1985, Placer formation by fluvial degrada-
tion of an alluvial fan sequence: The Proterozoic Carbon Leader placer,
Trans-Amazon greenstone belt to the north and northeast of Witwatersrand Supergroup, South Africa: Journal of the Geological
the Roraima Basin. Society [London], v. 142, p. 757–764.
It is evident from this comparison that the Witwatersrand Camisani-Calzolari, F.A.G.M., de Klerk, W.J., and van der Merwe, P.J.,
1985, Assessment of South African uranium resources: Methods and
deposits are not unique in terms of style of mineralization. results: Transactions of the Geological Society of South Africa, v. 88,
Furthermore, it has to be recognized that Witwatersrand-style p. 83–97.
gold mineralization occurred not only in a number of cratons, Cooper, R.A., 1923, Mineral constituents of Rand concentrates: Journal of
the Chemical, Metallurgical and Mining Society of South Africa, v. 24,
but also at different times. It cannot be ascribed to a single, no. 4, p. 90–95.
outstanding, gold-forming event. England, G.L., Rasmussen, B., Krapez, B., and Groves, D., 2002, Palaeoen-
The distribution of these deposits is global and over vast vironmental significance of rounded pyrite in siliciclastic sequences of
the Late Archaean Witwatersrand Basin: Oxygen-deficient atmosphere
areas, 31,000 km 2 in the Witwatersrand, 15,000 km 2 in the or hydrothermal alteration?: Sedimentology, v. 49, p. 1133–1156, doi:
Elliot Lake, and 17,000 km 2 in the Moeda. Because there is 10.1046/j.1365-3091.2002.00479.x.
no evidence for any correlation between organic carbon in the Eriksson, P.G., Hatting, P.J., and Altermann, W., 1995, An overview of the
geology of the Transvaal Sequence and Bushveld Complex, South
reefs and the pyrite and uraninite content, it is unlikely that Africa: Mineralium Deposita, v. 30, p. 98–111.
localized redox conditions could have preserved these detrital Frimmel, H.E., and Minter, W.E.L., 2002a, Recent developments concerning
minerals. If the 2200 Ma age of the Moeda is correct (Kirk, the geological history and genesis of the Witwatersrand gold deposits,
South Africa: Society of Economic Geologists Special Publication, v. 9,
et al., 2002, and 2004, personal commun.), the rate of change p. 17–45.
in the oxygen content of the atmosphere might have been very Frimmel, H.E., and Minter, W.E.L., 2002b, An overview of geological pro-
sudden (Maarten de Wit, 2004, personal commun.). One may cesses that controlled distribution of gold in the Witwatersrand deposits,
in Cooke, D.R., and Pongratz, J., eds., Giant ore deposits, characteris-
well consider this time of change in the atmospheric composi- tics, genesis and exploration: Hobart, Tasmania, Centre for Ore Deposit
tion to be the boundary between the Archean and the Protero- and Exploration Studies: CODES Special Publication 4, p. 221–242.
zoic Eras. Frimmel, H.E., 1997, Detrital origin of hydrothermal Witwatersrand gold-
a review: Terra Nova, v. 9, p. 192–197, doi: 10.1046/j.1365-3121.1997.
d01-23.x.
ACKNOWLEDGMENTS Garayp, E., Minter, W.E.L., Renger, F.E., and Siegers, A., 1991, Moeda
placer gold deposits in the Ouro Fino Syncline, Quadrilatero Ferrifero,
Brazil, p. 601–608, in Ladeira, F.A., ed., Brazil Gold ’91: Rotterdam,
The invitation and subvention received from Hiroshi A.A. Balkema, 823 p.
Ohmoto and Steve Kesler to participate in the Pardee sympo- Gray, N.K., Tucker, R.F., and Kershaw, D.J., 1994, The Sun project. 1. Dis-
sium is gratefully acknowledged. covery of a major new Witwatersrand goldfield, in XVth Council of
Mining and Metallurgical Institutions Congress: Johannesburg, South
African Institute of Mining and Metallurgy, v. 3, p. 95–102.
REFERENCES CITED Hirdes, W., and Nunoo, B., 1994, The Proterozoic paleoplacers at Tarkwa
Gold Mine, SW Ghana: Sedimentology, mineralogy, and precise age
Antrobus, E.S.A., Brink, W.C.J., Brink, M.C., Caulkin, J., Hutchinson, R.I., dating of the Main Reef and West Reef, and bearing on the investiga-
Thomas, D.E., van Graan, J.A., and Viljoen, J.J., 1986, The Klerksdorp tions on the source area aspects, in Oberthür, T., ed., Metallogenesis
goldfield, in Anhaeusser, C.R., and Maske, S., eds, Mineral deposits of selected gold deposits in Africa: Geologisches Jahrbuch, v. D100,
of Southern Africa, Volume 1: Johannesburg, Geological Society of p. 247–311.
South Africa, p. 549–598. James, C.S., and Minter, W.E.L., 1999, Experimental flume study of the
Armstrong, F.C, 1975, Genesis of uranium- and gold-bearing Precambrian deposition of heavy minerals in simulated Witwatersrand sandstone
quartz-pebble conglomerates: U.S. Geological Survey Professional unconformity: Economic Geology and the Bulletin of the Society of
Paper 1161-A. Economic Geologists, v. 94, p. 671–688.
118 W.E.L. Minter
Kirk, J., Ruiz, J., Chesley, J., Titley, S., and Walshe, J., 2001, A detrital model Ashanti belt of southern Ghana: Precambrian Research, v. 89, p. 129–143,
for the origin of gold and sulfides in the Witwatersrand basin based on doi: 10.1016/S0301-9268(97)00075-2.
Re-Os isotopes: Geochimica et Cosmochimica Acta, v. 65, p. 2149–2159, Ohmoto, H., Rasmussen, B., Buick, R., and Holland, H.D., 1999, Redox
doi: 10.1016/S0016-7037(01)00588-9. state of the Archean atmosphere: Evidence from detrital heavy miner-
Kirk, J., Ruiz, J., Chesley, J., Walshe, J., and England, G., 2002, A major als in ca. 3250–2750 Ma sandstones from the Pilbara Craton, Australia:
Archean, gold- and crust-forming event in the Kaapvaal craton, South Comment and Reply: Geology, v. 27, p. 1151–1152, doi: 10.1130/0091-
Africa: Science, v. 297, p. 1856–1858, doi: 10.1126/science.1075270. 7613(1999)027<1151:RSOTAA>2.3.CO;2.
Krogh, T.E., Davis, D.W., and Corfu, F., 1984, Precise U-Pb zircon and bad- Papenfus, J.A., 1957, The geology of the Government gold mining areas and
deleyite ages for the Sudbury area, in Pye, E., ed., The geology and ore new state areas, with particular reference to channel sediments and the
deposits of the Sudbury Structure: Ontario Geological Survey Special origin of minerals of economic importance [Ph.D. thesis]: Pretoria, Uni-
Volume 1, p. 431–446. versity of South Africa,.
Ledru, P., Milesi, J.P., Johan, V., Sabate, P., and Maluski, H., 1997, Foreland Phillips, G.N., and Law, J.D.M., 2000, Witwatersrand gold fields: Geology,
basins and gold-bearing conglomerates: A new model for the Jacobina genesis and exploration: Society of Economic Geology Reviews, v. 13,
Basin (São Francisco Province, Brazil): Precambrian Research, v. 86, p. 439–500.
p. 155–176, doi: 10.1016/S0301-9268(97)00048-X. Phillips, G.N., and Myers, R.E., 1989, Witwatersrand gold fields, Part II: An
McPherson, J.G., Shanmugan, G., and Moiola, R.J., 1987, Fan-deltas and braid origin for Witwatersrand gold during metamorphism and associated
deltas: Varieties of coarse-grained deltas: Geological Society of Amer- alteration, in Keays, R.R., Ramsay, W.R.H., and Groves, D.I., eds., The
ica Bulletin, v. 99, p. 331–340, doi: 10.1130/0016-7606(1987)99<331: geology of gold deposits: The perspective in 1988: Economic Geology
FABDVO>2.0.CO;2. Monograph 6, p. 598–608.
Minter, W.E.L., 1976, Detrital gold, uranium and pyrite concentrations related Pigois, J.-P., Groves, D.L., Fletcher, I.R., McNaughton, N.J., and Snee, L.W.,
to sedimentology in the Precambrian Vaal Reef placer, Witwatersrand, 2003, Age constraints on Tarkwaian palaeoplacer and lode-gold forma-
South Africa: Economic Geology and the Bulletin of the Society of Eco- tion in the Tarkwa-Damang district: SW Ghana: Mineralium Deposita,
nomic Geologists, v. 71, p. 157–176. v. 38, p. 695–714, doi: 10.1007/s00126-003-0360-5.
Minter, W.E.L., 1978, A sedimentological synthesis of placer gold, uranium and Pretorius, D.A., 1975, The depositional environment of the Witwatersrand gold-
pyrite concentrations in Proterozoic Witwatersrand sediments, in Miall, fields: A chronological review of speculations and observations: Minerals,
A.D., ed, Fluvial sedimentology: Canadian Society of Petroleum Geolo- Science and Engineering, v. 7, p. 18–47.
gists Memoir 5, p. 801–829. Rasmussen, B., and Buick, R., 1999, Redox state of the Archean atmosphere:
Minter, W.E.L., 1982a, The golden Proterozoic, in Tankard, A.J., Jackson, Evidence from detrital heavy minerals in ca. 3250–2750 Ma sandstones
M.P.A., Eriksson, K.A., Hobday, D.K., Hunter, D.R., and Minter, W.E.L., from the Pilbara Craton, Australia: Geology, v. 27, p. 115–118, doi:
eds., Crustal evolution of southern Africa: 3.8 billion years of earth his- 10.1130/0091-7613(1999)027<0115:RSOTAA>2.3.CO;2.
tory: Berlin, Springer-Verlag, p. 115–150. Robb, L.J., Davis, D., Kamo, S.L., and Meyer, F.M., 1992, Ages of altered gran-
Minter, W.E.L., 1982b, The Witwatersrand goldfield: Optima, v. 30, no. 4. ites adjoining the Witwatersrand Basin with implications for the origin of
Minter, W.E.L., 1989, Witwatersrand placer concentrates associated with gold and uranium: Nature, v. 357, p. 677–680, doi: 10.1038/357677a0.
inclined surfaces [abs.], in Programme and Abstracts, 4th International Roscoe, S.M., and Minter, W.E.L., 1993, Pyritic paleoplacer gold and uranium
Conference on Fluvial Sedimentology, Sitges, Spain: International Asso- deposits, in Kirkham, R.V., Sinclair, W.D., Thorpe, R.I., and Duke, J.M.,
ciation of Sedimentologists, p. 185. eds., Mineral deposit modelling: Geological Association of Canada Spe-
Minter, W.E.L., 1991, Ancient placer gold deposits, in Foster, R.P., ed., Gold cial Paper 40, p. 103–124.
metallogeny and exploration: London, Blackie and Sons, p. 283–308. Rundle, C.C., and Snelling, N.J., 1977, The geochronology of uraniferous min-
Minter, W.E.L., 1999, Irrefutable detrital origin of some Witwatersrand gold erals in the Witwatersrand Triad: An interpretation of new and existing
and evidence of eolian signatures: Economic Geology and the Bulletin of U-Pb data on rocks and minerals from the Dominion Reef, Witwatersrand
the Society of Economic Geologists, v. 94, p. 665–670. and Ventersdorp Supergroups: Philosophical Transactions of the Royal
Minter, W.E.L., and Loen, J.S., 1991, Palaeocurrent dispersal patterns of Witwa- Society of South Africa, v. 286, p. 567–583.
tersrand gold placers: South African Journal of Geology, v. 94, p. 70–85. Sadler, P.M., 1981, Sediment accumulation rates and the completeness of strati-
Minter, W.E.L., Hill, W.C.N., Kidger, R.D., Kingsley, C.S., and Snowden, P.A., graphic sections: Journal of Geology, v. 89, p. 569–584.
1986, The Welkom goldfield, in Anhaeusser, C.R., and Maske, S., eds., Santos, J.O.S., Potter, P.E., Reis, N.J., Hartmann, L.A., Fletcher, I.R., and
Mineral deposits of southern Africa, Volume 1: Johannesburg, Geological McNaughton, N.J., 2003, Age, source, and regional stratigraphy of
Society of South Africa, p. 497–535. the Roraima Supergroup and Roraima-like outliers in northern South
Minter, W.E.L., Feather, C.E., and Glathaar, C.W., 1988, Sedimentological America based on U-Pb geochronology: Geological Society of America
and mineralogical aspects of the newly discovered Witwatersrand placer Bulletin, v. 115, p. 331–348, doi: 10.1130/0016-7606(2003)115<0331:
deposit that reflect Proterozoic weathering, Welkom gold field, South ASARSO>2.0.CO;2.
Africa: Economic Geology and the Bulletin of the Society of Economic Slingerland, R., and Smith, N.D., 1986, Occurrence and formation of water-laid
Geologists, v. 83, p. 481–491. placers: Annual Review of Earth and Planetary Sciences, v. 14, p. 113–
Minter, W.E.L., Renger, F.E., and Siegers, A., 1990, Early Proterozoic gold 147, doi: 10.1146/annurev.ea.14.050186.000553.
placers of the Moeda Formation within the Gandarela Syncline, Minas Smith, N.D., and Minter, W.E.L., 1980, Sedimentological controls of gold and
Gerais, Brazil: Economic Geology and the Bulletin of the Society of Eco- uranium in two Witwatersrand paleoplacers: Economic Geology and the
nomic Geologists, v. 85, p. 943–951. Bulletin of the Society of Economic Geologists, v. 75, p. 1–14.
Minter, W.E.L., Cheatle, A., and Hartnady, C.J.H., 1993a, Signatures left by Srinivasan, R., and Ojakangas, R.W., 1986, Sedimentology of quartz-pebble
Archean paleowinds in the Witwatersrand gold deposits and their geotec- conglomerates and quartzites of the Archean Bababudan Group, Dharwar
tonic significance [abs.], in Abstracts and Program, Stratigraphic record craton, South India—Evidence for early crustal stability: Journal of Geol-
of global change, Pennsylvania State University, University Park, August ogy, v. 94, p. 199–214.
8-12: Society of Economic Paleontologists and Mineralogists, p. 59. Sutton, S.J., and Maynard, J.B., 1993, Petrology, mineralogy, and geochemistry
Minter, W.E.L., Goedhart, M.L., Knight, J., and Frimmel, H.E., 1993b, Mor- of sandstones of the lower Huronian Matinenda Formation: Resemblance
phology of Witwatersrand gold grains from the Basal Reef: Evidence for to underlying basement rocks: Canadian Journal of Earth Sciences, v. 30,
their detrital origin: Economic Geology and the Bulletin of the Society of p. 1209–1223.
Economic Geologists, v. 88, p. 237–248. Teixeira, J.B.G., de Souza, J.A.B., da Silva, M.d.G., Leite, C.M.M., Barbosa,
Minter, W.E.L., Frimmel, H.E., Kirk, J., and Vennemann, T., 2002, Paleoplacer J.F., Coelho, C.E.S., Abram, M.B., Filho, V.M.C., and Iyer, S.S.S., 2001,
gold potential in the Early Proterozoic Roraima Group: Geological Soci- Gold mineralization in the Serra de Jacobina region, Bahia Brazil: Tec-
ety of America Abstracts with Programs, v. 34, no. 6, p. 517. tonic framework and metallogenesis: Mineralium Deposita, v. 36, p.
Mitchell, M., 1996, Alluvial platinum group minerals from southern New Zea- 332–344.
land [Ph.D. thesis]: Dunedin, New Zealand, University of Otago, 277 p. Theis, N.J., 1976, Uranium-bearing and associated minerals in their geochemi-
Oberthür, T., Vetter, U., Davis, D.W., and Amanor, J.A., 1998, Age constraints cal and sedimentological context, Elliot Lake, Ontario [Ph.D. thesis]:
on gold mineralization and Paleoproterozoic crustal evolution in the Kingston, Ontario, Queen’s University, 158 p.
Tucker, R.F., 1980, The sedimentology and mineralogy of the Composite nium-bearing Witwatersrand and Huronian Supergroup quartz-pebble
Reef on Cooke Section, Randfontein Estates Gold Mine, Witwa- conglomerates: Economic Geology and the Bulletin of the Society of
tersrand, South Africa [M.S. thesis], Johannesburg, University of the Economic Geologists, v. 91, p. 322–342.
Witwatersrand, 355 p.
Vennemann, T.W., Kesler, S.E., Frederickson, G.C., Minter, W.E.L., and
Heine, R.R., 1995, Oxygen isotope sedimentology of gold- and ura- M ANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA


Memoir 198
2006
Witwatersrand gold-pyrite-uraninite deposits do not support

a reducing Archean atmosphere
Jonathan Law
CSIRO c/- 6 Genoa Court, Mount Waverley, 3149, Victoria, Australia
Neil Phillips
School of Earth Sciences, University of Melbourne c/- P.O. Box 3, Central Park, 3145, Victoria, Australia
ABSTRACT
The first serious suggestion that the Archean atmosphere was reducing was based
on the interpretation of round uraninite and pyrite grains in the Witwatersrand Basin
in the early 1950s. It was then inferred that these minerals were detrital and that they
reflected equilibrium with a reducing Archean atmosphere.
Over the past 20 years the understanding of the Witwatersrand Basin has
changed dramatically with more integrated studies of the basin and the recognition
of widespread alteration in close spatial association with the mineralization in every
goldfield. Post-depositional mobility of gold, sulfur, and uranium during alteration is
widespread and supports hydrothermal ore genesis, or at least substantial modifica-
tion of the original mineral assemblage. Pseudomorphic replacements of pre-existing
detrital minerals (e.g., pyrite after titano-magnetite), and precipitation and/or chemi-
cal rounding to generate round mineral shapes (e.g., uraninite in carbon seams) have
all been documented.
The recognition that the carbon seams formed by the post-depositional introduc-
tion and maturation of migrated hydrocarbons is a dramatic departure from earlier
models of coalified algal material deposited with the sediments. The enrichment of
both gold and uraninite in carbon seams implies that these minerals are hydrother-
mal and that their shapes do not reflect detrital processes. Uranium mobility in basin-
al waters may in fact require a relatively oxidizing atmosphere.
None of the existing arguments for the Witwatersrand mineralization unambigu-
ously support a placer or modified placer model for the mineralization. Consequently,
round uraninite and pyrite of the Witwatersrand Basin do not provide support for a
reducing Archean atmosphere.
Keywords: atmosphere, Archean, Witwatersrand, pyrite, hydrothermal.
INTRODUCTION ambiguous and controversial (see reviews by Dimroth and Kim-

berley, 1976; Clemmey and Badham, 1982; Palmer et al., 1987;
Since the early 1950s, it has been suggested that the Archean Ohmoto, 1996; Phillips et al., 2001).
atmosphere was “reducing,” and that it may not have contained In fact it was the discovery of widespread uraninite in the
significant amounts of free oxygen until ca. 2 Ga. However, the Witwatersrand Basin in the 1950s that led to the first serious sug-
geological evidence for a reducing Archean atmosphere is both gestion of a reducing Archean atmosphere. This hypothesis arose
*E-mail: [email protected]; [email protected]
Law, J., and Phillips, N., 2006, Witwatersrand gold-pyrite-uraninite deposits do not support a reducing Archean atmosphere, in Kesler, S.E., and Ohmoto, H., eds.,
Evolution of Early Earth’s Atmosphere, Hydrosphere, and Biosphere—Constraints from Ore Deposits: Geological Society of America Memoir 198, p. 121–141,
doi: 10.1130/2006.1198(07). For permission to copy, contact [email protected]. ©2006 Geological Society of America. All rights reserved.
121
122 J. Law and N. Phillips
Holland et al., 1994; Kasting, 1993; 2001; Rye and Hol-

from combining the widely held beliefs that the gold grains in
land, 1998; Farquhar et al., 2000).
the Witwatersrand conglomerates were detrital in origin, and
that round uraninite and pyrite grains spatially associated with There are several different understandings of what constitute
the gold probably had a similar origin. As the uraninite grains “reducing” and “oxidizing” conditions in the early atmosphere.
could not have survived detrital transportation under oxidizing Holland (1984) has used the composition of inferred “paleosols”
atmospheric conditions, a reducing Archean atmosphere was to express O2 levels relative to present atmospheric compositions,
proposed. Subsequently, paleosols were inferred in and around whereas others have argued that such calculations are premature
the Witwatersrand Basin (and elsewhere), and their chemical before conclusive evidence is presented demonstrating that the
composition was inferred to support the reducing atmosphere paleosols are unaltered, or indeed are paleosols at all (Palmer et
hypothesis (Button and Tyler, 1981). al., 1987; 1989).
These models for a reducing Archean atmosphere were We base our working definition of atmospheric redox state
developed in an era of overwhelming dominance of the placer on the geochemistry of iron, and this definition is implicit in many
and modified placer models for the Witwatersrand gold miner- other studies. For example, supporters of the “reducing” model
alization. As a result, research focused on the sedimentology assume that the level of atmospheric O2 is too low to stabilize
of the host rocks, and there was a widespread belief that Wit- significant Fe3+ (e.g., when drawing conclusions on iron mobility
watersrand rocks were “unaltered” and “unmetamorphosed,” in inferred “paleosols”; Holland, 1984; 1994). Iron is one of the
effectively precluding the possibility of a hydrothermal origin most important components in Earth’s crust, and knowledge of
for the mineralization. Over the past 20 years this picture has its behavior and its oxidation states is possibly the most impor-
changed dramatically with more integrated studies of the basin tant factor related to the early atmosphere affecting our under-
and the recognition of widespread alteration in close spatial standing of early geological processes. The chemical behavior
association with the mineralization in every goldfield. In the of iron is strongly influenced by whether near-surface environ-
light of these developments, it is appropriate to critically review ments, including aqueous solutions, had abundant Fe2+ and Fe3+,
new and existing data for the Witwatersrand to assess their rel- or simply Fe2+. If the former (i.e., oxidizing atmosphere), then
evance as constraints on the Archean atmosphere. In particular many near-surface processes would be similar to today; if the
we review the arguments linking mineralization to the Archean latter (i.e., reducing atmosphere), then none of the near-surface
atmosphere, i.e., whether or not the gold, pyrite, and uraninite processes that involve and require abundant iron in two oxidation
are primary detrital grains. states (e.g., laterite formation and iron pisolith formation) would
The possible link between mineralization and the Archean be operative. An “oxidizing” environment in terms of this defi-
atmosphere has important implications for ore genesis beyond nition does not necessarily imply an atmospheric composition
the Witwatersrand. Although a reducing Archean atmosphere exactly the same as today: it means oxidizing enough to facilitate
is “built in” to many existing genetic models for a variety of geological processes that require the presence of two oxidation
orebodies, the alternative of an oxidizing atmosphere would states of iron. Calculations of the amount of free oxygen neces-
require critical re-evaluation of the origins of many Archean sary to stabilize ferric iron minerals suggest that free oxygen may
orebodies that assume reducing atmospheric conditions. Also, have been several orders of magnitude less than the present atmo-
new exploration opportunities for Witwatersrand-style explora- spheric level (e.g., Krupp et al., 1994; Holland, 1994). Any more
tion targets may exist in younger sedimentary basins. precise definition is premature given uncertainties in interpreting
basic geological relationships and then inferring the associated
DEFINITION OF ATMOSPHERIC REDOX STATE chemical environment. For example, absolute thermodynamic
calculations require many assumptions on the origin and evolu-
Most models for the early atmosphere assume that it was tion of the mineral assemblages, chemical boundary conditions,
reducing when Earth formed, and subsequently changed to be and kinetic controls on reaction progress.
oxidizing as it is today. A reducing atmosphere in the very earliest The stability of detrital pyrite and uraninite are directly
Archean is common to all models and, in the absence of geologi- linked to atmospheric redox conditions. In a reducing atmo-
cal constraints, is based exclusively on conceptual studies of the sphere, detrital uraninite and pyrite would be meta-stable along
early evolution of Earth, Sun, and the solar system. Two basic with a suite of other sulfide minerals, and iron might be leached
models have been proposed: from detrital iron oxides. In an oxidizing atmosphere, many iron
1. An oxidizing Archean atmosphere: an oxidizing atmo- oxide minerals would be stable and uranium would be highly
sphere evolved very early in the evolution of Earth and soluble, destabilizing detrital uraninite. Oxidation of U4+ miner-
was already well developed in the early Archean (i.e., by als such as uraninite and brannerite occurs readily under modern
3.5 Ga; Palmer et al., 1987, 1989; Ohmoto, 1996). atmospheric conditions, making their persistence during detrital
2. A reducing Archean atmosphere: Earth’s atmosphere transport highly unlikely in an oxidizing atmosphere (Davidson,
was reducing until ca. 2.4–2.0 Ga, when a proliferation of 1953, 1957; Davidson and Cosgrove, 1955; Holland, 1994),
photosynthetic organisms resulted in gradually increasing though remotely possible given localized modern occurrences
levels of oxygen in the atmosphere (Krupp et al., 1994; (Robinson and Spooner, 1984; Maynard et al., 1991).
Witwatersrand gold-pyrite-uraninite deposits 123
OVERVIEW OF PYRITE, GOLD AND URANIUM IN also contain one or more subordinate reef horizons that are eco-
CONGLOMERATES nomic over smaller areas, and in the case of the West Rand gold-
field there are at least ten such reefs (Phillips and Law, 2000).
Geological Overview and Distribution Both placer and hydrothermal models have been proposed
for Witwatersrand-type pyrite, uraninite, and gold mineral-
Gold, pyrite, and uraninite in fluvial sediments of continental ization. Each of these models acknowledges the relationship
basins have been variably described as Witwatersrand-type, con- between gold distribution and rock type but provides differing
glomerate-hosted, quartz pebble-associated, and quartz arenite- explanations. According to the unmodified placer model gold
associated mineralization. Each of these descriptors focuses on and uranium were introduced as placer concentrations during
the nature of the host rocks and each implicitly relies on a placer sedimentation. The model assumes that detrital gold, pyrite, and
origin of the mineralization to provide useful discrimination from uraninite grains have the same shape, composition, and location
other deposit styles. In fact, the genesis of conglomerate-hosted now as they had at their time of burial and hence can be used
pyrite and uraninite ore bodies was once thought to be so well to constrain source terranes and their compositions. The modi-
understood that the most commonly used name for the orebodies fied placer model is similar but assumes that although gold and
included the genetic descriptor “paleoplacer.” related minerals were detrital in origin they have been locally
Numerous occurrences of gold (with or without pyrite and remobilized and recrystallized after burial to account for second-
uraninite) in conglomerates are known from around the world ary grain shapes and compositions. A scale of movement in the
but only a few have been significant economic producers (Fig. millimeter to centimeter range is generally inferred. The hydro-
1). Of all these, the Witwatersrand Basin in South Africa is by thermal replacement model proposes that the gold and uranium
far the largest producer of both gold and uranium. All-time Wit- mineralization is not detrital and was introduced by the circula-
watersrand production now exceeds 50,000 t of gold from seven tion of hydrothermal fluids in the basin after burial (see Robb and
goldfields. In each goldfield there is typically one reef horizon Meyer, 1995 and Phillips and Law, 2000 for opposing views on
that has produced the majority of the gold (e.g., Kimberley Reef the merits of these models).
at Evander, Main Reef Leader in the Rand goldfields, Carbon The placer and modified placer models have dominated
Leader Reef at Carletonville, Vaal Reef at Klerksdorp, and Basal thinking throughout much of the twentieth century and have been
Reef at Welkom). With the exception of Evander, all goldfields used to support the reducing Archean atmosphere hypothesis.
Huronian
Tarkwaian
Fortescue
Jacobina
Black Reef
Witwatersrand
Dominion
Reduced
Oxidized
Figure 1. Location of major reduced gold-pyrite-uraninite conglomerates and the oxidized Tarkwaian gold-magnetite-hematite conglomerates.
Time Distribution of “Oxidized” and “Reduced” intruding granites) by Bossière et al. (1996). Uraninite and pyrite
Witwatersrand-type Deposits are absent from the Tarkwaian ores and iron oxides dominate the
heavy mineral suite. Available geochronological constraints for
All pyrite-, gold-, and uraninite-bearing conglomerates are oxidized mineralization at Tarkwa and reduced mineralization at
older than 1900 Ma and together span the transition from a reduc- Jacobina indicate that they may overlap in time. The transition
ing to oxidizing atmosphere proposed by the reducing Archean between reduced and oxidized styles of conglomerate-hosted
atmosphere hypothesis (i.e., between 2.4 and 2.0 Ga; Fig. 2). In gold mineralization has been used to support a transition from
South Africa, Witwatersrand-type gold, pyrite, and uranium min- a reduced to oxidizing atmosphere; however, the available age
eralization is sporadically represented over a considerable period data indicate that they are broadly coeval and cannot both reflect
of time from local occurrences in the Archean greenstone base- global atmospheric composition.
ment, through the Dominion Reef, Witwatersrand Supergroup,
and Black Reef Formation of the Transvaal Sequence (Fig. 3). Common Features of Reduced Deposits
The Jacobina deposits are the youngest economically sig-
nificant “reduced” conglomerate-hosted gold-pyrite-uranium In spite of the variety in terms of age and geological setting,
deposits. Sedimentation at Jacobina is bracketed between 2086 some aspects of the gold-pyrite conglomerates listed in Table 1
Ma (age of the youngest detrital zircon) and 1883 Ma (age of (from South Africa and around the world) are remarkably consis-
intrusive post-tectonic granite) (Milesi et al., 2002). In contrast, tent and imply that similar processes have operated in each:
the oxidized conglomerate-hosted gold ores such as those at • All deposits are located in clastic continental sediments
Tarkwa have been inferred to postdate the oxygenation of the closely associated with active uplift during sedimentation
atmosphere and have been dated between 2124 ± 9 Ma (the age and associated unconformity development.
of the youngest detrital zircon) and 1991 ± 12 Ma (the age of • All contain elevated gold and uranium, although the rela-
tive abundance of each varies dramatically both within
and among deposits.
• All contain round pyrite grains.
Mass independent sulfur fractionation • Detrital magnetite and ilmenite are virtually absent in all
deposits.
• All contain uraninite and a variety of U-Ti minerals,
Inferred paleosols including brannerite and leucoxene, that at least in part
reflect post-depositional modification of pre-existing ura-
Mt Roe 1&2
Denison & Pronto
niferous minerals.
Ville Marie • All deposits have an overprint of diagenetic and hydro-
Hekpoort thermal alteration that modifies the primary detrital
Hokkalampi mineralogy.
Drakenstein
Flin Flon
• All contain migrated hydrocarbons intimately associated
with uranium minerals and inferred to be syn- or post-
uranium mineralization in age.
Gold ±uranium deposits • All contain gold that is mostly paragenetically late in the
Witwatersrand mineralization sequence and is variably interpreted either
Dominion
Pongola
as hydrothermal in origin or as remobilized detrital gold.
Fortescue Although the Witwatersrand gold production is more than
100 times larger than any of the other reduced deposits known
Black Reef
Huronian
from elsewhere in the world, the genetic models proposed for
Jacobina each are similar.
Tarkwaian
THE WITWATERSRAND BASIN
2.8 2.6 2.4 2.2 2.0 1.8
Ga Structural and Sedimentological Setting
Figure 2. Age constraints on the deposition of the major global gold-
pyrite-uraninite occurrences relative to inferred atmospheric composi- The Witwatersrand orebodies are confined to a tectonically
tion based on altered zones interpreted as “paleosols” (from Holland preserved remnant of a more extensive sedimentary basin with
and Rye, 1997) and the mass independent fractionation of sulfur (from a complex syn- and post-depositional history (Phillips and Law,
Farquhar et al., 2000). The oxidation state inferred by these authors is
indicated by line thickness (narrow—reducing; bold—oxidizing). The 2000). The Witwatersrand sedimentary sequence (Witwatersrand
transition from mass dependent to mass independent styles of sulfur Supergroup) was deposited on the stable granite-greenstone crust
fractionation is indicated by a dotted line. of the Kaapvaal Craton. The preserved structural basin is elon-
Pretoria Group Regional unconformity
TRANSVAAL SEQUENCE
subsidence reactivation
Basin
Chuniespoort / Carbonates,
Ghaap Group banded iron stone,
minor shale
Thermal
& sandstone
2557 ± 49 Ma
Black Reef Quartzite
Allanridge Formation Tholeiitic flood basalts
VENTERSDORP SUPERGROUP
Bothaville Formation Sandstone & conglomerate

2709 ± 4 Ma
development
Graben
Platberg Group Bimodal volcanics, Schematic Witwatersrand Supergroup
immature clastics gold lithology
distribution
Klipriviersberg Group v v Klipriviersberg Group
Tholeiitic flood basalts
o o o
o o VCR
2714 ± 9 Ma o o o
Rand Group
Fluvial sands,
Central
ooooo
Central Rand Group conglomerates,
WITWATERSRAND SUPERGROUP
Foreland basin
minor shale Booysens shale 1000 m
Dominantly marine
2914 ± 5 Ma Main Reef
v Crown Lava 0
Sands, greywacke, 2914 Ma
West Rand Group
magnetic shales &
subsidence
Thermal
argillites
West Rand Group

o conglomerate
quartzite
shale
v metavolcanics
3074 ± 5 Ma
Basalts,
Rifting
Dominion Group volcaniclastics &

minor quartzite
v v Dominion Group
Basement Granite & greenstone
Figure 3. Generalized stratigraphy of the Kaapvaal Craton in the vicinity of the Witwatersrand Basin showing the positions of the main mineral-
ized horizons. Ages refer to volcanic horizons in the stratigraphy at the positions shown.
TABLE 1. SUMMARY OF SELECTED BASINS WITH CONGLOMERATE-HOSTED GOLD

AND/OR URANIUM MINERALIZATION AND DOMINANT MINERALOGICAL ASSOCIATIONS
Basin, location Redox state Mineralogical associations
Gold Pyrite Uraninite Hydrocarbon Iron oxides
Jacobina, Brazil Reduced XX XX XX XX -
Tarkwaian, Ghana Oxidized XX - - - XX
Huronian, Canada Reduced X XX XX XX -
Transvaal (Black Reef Reduced XX XX XX XX -
Formation only), South Africa
Witwatersrand, South Africa Reduced XX XX XX XX -
Fortescue, Australia Reduced X XX XX XX -
Dominion, South Africa Reduced X XX XX XX -
Note: XX—common component; X—minor component; -—absent to very rare. Note that
absolute abundances of minerals vary dramatically within and between ore bodies. “Redox state”
refers to the presence of either pyrite (reduced) or iron oxide (oxidized) in the ore assemblage
(Krupp et al., 1994).
gate to the NE and is ~350 km long, 150 km wide, and up to 8 graphic sequence were deposited in response to local syn-deposi-
km thick. In a general sense, the stratigraphic record preserves tional deformation near the currently preserved margin (McCar-
a progressively upward-coarsening depositional sequence domi- thy, 1994; Coward et al., 1995). Most of the gold mineralization
nated by marine sedimentary rocks toward the base and clastic is in the upper sequence of continental sedimentary rocks.
continental sedimentary rocks toward the top (Fig. 3). The lower The Witwatersrand Supergroup unconformably overlies the
marine sequence was considerably more extensive than the cur- Dominion Group and surrounding Archean granite-greenstone
rently preserved structural basin. The upper parts of the strati- basement and is itself overlain unconformably by the Ventersdorp
Supergroup. All three units are Archean in age. The Ventersdorp whole rock age from the Chuniespoort Group of the Transvaal
is overlain by variable thicknesses of the Archean to Proterozoic sequence of 2557+49 Ma suggests that the base of the Transvaal
Transvaal Sequence and generally flat-lying Mesozoic sedimen- may be much earlier and that the lower Transvaal rocks are of
tary rocks of the Karoo Sequence. Rapid lateral variations in the late Archean age (Jahn et al., 1990). The stratigraphically high-
thickness of the cover sequence around the Witwatersrand Basin est gold mineralization in the Witwatersrand goldfields has been
result from regional unconformities and local structural complex- attributed to the mineralized conglomerate at the base of the
ity. Dikes and sills of Ventersdorp, Bushveld, and Karoo age are Black Reef Formation.
common throughout the Basin.
The basement granite-greenstone terranes are mostly from Metamorphism and Alteration
3.3 Ga to 3.1 Ga in age (Robb et al., 1990a, 1990b), and vol-
canic rocks from the Dominion Group have been dated at 3.07 Metamorphic assemblages indicate that greenschist facies
Ga (Armstrong et al., 1991). By synthesizing available age con- conditions of 300–400 °C and 2–3 kbars were reached in each
straints, Robb et al. (1990a) inferred that that Witwatersrand sedi- goldfield (Phillips and Law, 1994). Metamorphism was accom-
mentation commenced ca. 3.0 Ga and was terminated by extru- panied by widespread alteration that affected virtually the entire
sion of the Klipriviersberg volcanic rocks ca. 2714 Ma. Within upper Witwatersrand succession in every goldfield and resulted
the Witwatersrand Supergroup, the only reliable age is 2914 Ma in the progressive conversion of pre-existing detrital and dia-
from the Crown metabasalt toward the top of the West Rand genetic assemblages to muscovite, chlorite, pyrophyllite, and
Group (Armstrong et al., 1991). chloritoid. Pyrite is very widespread in samples in the Central
The structural evolution of the basin has been reviewed by Rand Group. Fluid flow has been channeled along bedding-
Coward et al. (1995), who have linked basin formation to the subparallel brittle-ductile faults particularly in heterogeneous
progressive tectonic evolution of the surrounding region during sedimentary units overlying unconformity surfaces (Phillips,
the Archean: 1988; Barnicoat et al., 1997). Alteration crosscuts stratigraphy,
• Pre-Witwatersrand rifting and deposition of the Dominion is locally focused by thick, regionally persistent shales, and is
Group comprising dominantly basaltic and felsic volcanic spatially related to mineralization (Phillips and Law, 2000). Iso-
rocks and minor sedimentary rocks. topic ages from alteration assemblages typically reflect younger
• Post-Dominion thermal subsidence and deposition of resetting events, but a minimum age on alteration is inferred
the dominantly clastic marine lower Witwatersrand from large-scale extensional faults of Platberg age that displace
succession. both alteration and mineralization on all scales (i.e., alteration
• A progressive change to a compressional tectonic regime and mineralization predate 2709 ± 4 Ma; Phillips and Law,
and the development of an emergent fold-thrust belt along 2000). Mineralization in the Black Reef and specifically pyro-
the northern and western basin margins. The upper Wit- phyllite alteration and related deformation in mineralized Black
watersrand succession thus comprises a generally upward- Reef at South Deep Mine are inconsistent with a pre-Platberg
coarsening succession of dominantly fluvial sedimentary age. This evidence has been used to infer hydrothermal gold
rocks forming an asymmetrical foredeep in the north and introduction with at least some of this gold post–Black Reef in
west, thinning progressively toward the south and east. age (Wall et al., 2004).
• Extrusion of the Klipriviersberg flood basalts during the In addition to the regional-scale alteration, mesoscopic chlo-
waning stages of compression that effectively terminated rite veinlets, retrogression around shear zones, and late micas and
Witwatersrand sedimentation at 2714 Ma. calcite are all common small-scale features in the Witwatersrand
• Post-Klipriviersberg extension and deposition of the Plat- and not unlike what is found in other sedimentary basins. These
berg Group sedimentary and bimodal volcanic rocks in small-scale alteration systems typically overprint the regional
asymmetric grabens related to northwest-southeast exten- alteration and are not related to mineralization.
sion ca. 2709 Ma. Alteration has resulted in the loss of mobile cations, includ-
The overlying Transvaal Sequence covers an area much ing Na, Ca, and locally K in quartzites and conglomerates, to
greater than the Witwatersrand Basin and comprises extensive stabilize the highly aluminous alteration assemblages. Given
clastic and chemical sedimentary rocks and minor volcanic the similarity between cation loss during alteration and that dur-
rocks. At the base of the Transvaal, the Black Reef Formation ing weathering, several authors have argued that the unusual
unconformably overlies the Ventersdorp Supergroup and a range bulk-rock compositions reflect weathering at the source, during
of older stratigraphic units, and is developed over a wide area of transport, or after deposition of the sediment (Sutton et al., 1990;
the Kaapvaal Craton both within and around the Witwatersrand Reimer, 1985). However, recent studies have shown that altera-
Basin. It comprises a basal conglomerate overlain by quartzite tion zones cut stratigraphy on a local and regional scale and are
and black shale of fluvial and marine origin, respectively (Tan- spatially related to bedding-subparallel shear zones and brittle
kard et al., 1982; Els et al., 1995). Platform carbonates of the fractures that channeled fluid flow (Barnicoat et al., 1997; Phil-
Chuniespoort or Ghaap Groups and fine-grained clastic rocks of lips and Law, 1997; 2000). The large-scale addition of K is a
the Pretoria Group successively overlie the Black Reef. A Pb-Pb feature of alteration, but less likely during weathering.
Alteration of Witwatersrand sedimentary rocks has even The recognition that the carbon seams formed by the post-
greater implications for the interpretation of the Archean atmo- depositional introduction and maturation of migrated hydrocar-
sphere in that many of the early examples of reduced “paleosols” bons is a dramatic departure from earlier models that interpreted
from the Witwatersrand are now recognized as alteration zones the seams as coalified material derived from the burial of algal
(Palmer et al., 1987; 1989). Similarly, studies of arenite textures material interbedded with the sediments. The enrichment of both
and unaltered sedimentary rocks show extensive evidence of gold and uraninite in carbon seams implies that these minerals
hydrothermal alteration superimposed on preexisting weathering are hydrothermal in origin and that their round shapes do not
and/or diagenetic assemblages (Phillips et al., 1990; Law et al., reflect detrital rounding. U mobility in basinal waters may in fact
1990; Law, 1991; Phillips and Law, 2000) require a relatively oxidizing atmosphere.
Mineralization and Paragenesis THE DILEMMA OF WITWATERSRAND

MINERALOGY
Gold, uranium, pyrite, and hydrocarbons are important
components of all “Witwatersrand-style” deposits worldwide All of the economically mineralized Witwatersrand con-
although their relative abundance varies dramatically both within glomerates accumulated on long-lived regional unconformities
and among deposits (Table 1). In the Witwatersrand, paragenetic and some are moderately enriched in durable minerals such as
studies on the timing of gold grains and related minerals have zircon and chromite. On the basis of their grain shape, chemical
yielded diametrically opposing conclusions: that some (but not and mechanical character, and textural setting, these two miner-
all) of the gold is texturally early (i.e., detrital; Minter, 1999) or als almost certainly represent detrital placer accumulations. By
that the gold is late (i.e., hydrothermal or remobilized; Davidson, inference, many other heavy minerals should have been accu-
1960; Ramdohr, 1958). Nevertheless, there is now a broad con- mulated at the same time, including magnetite, garnet, ilmenite,
sensus in the literature that the bulk of the gold is texturally late and possibly hematite. Most of these minerals exist in significant
with (Frimmel et al., 1993; Minter, 1999) or without (Phillips quantities in potential source rocks in southern Africa (e.g., mag-
and Law, 1997; Barnicoat et al., 1997) a significant placer gold netite in banded iron formations) and they are well represented
contribution. Differences still exist as to the timing and parage- in the overlying and underlying stratigraphic units spatially
netic relationships among the most closely related ore compo- removed from mineralization (Fig. 4) (Fuller, 1958; Phillips and
nents, i.e., the gold, carbon, uraninite and pyrite; in spite of local Law, 2000; Meyer et al., 1990). However, any visual examina-
complexities, however, most studies indicate a similar temporal tion shows the reefs to be devoid of iron oxides (except as mag-
sequence of early uraninite followed by hydrocarbons followed netite preserved within BIF clasts; e.g., Hirdes and Saager, 1983)
by gold (see review in Phillips and Law, 2000; see also Barnicoat whereas U and Ti oxides including rutile, brannerite, leucoxene,
et al., 1997; England, 1999; England et al., 2001). and uraninite are common. Iron-bearing minerals are common,
Phillips and Law (2000) have suggested that this paragenetic but Fe-oxides are virtually absent.
sequence is compatible with the inferred evolution of the basin Another dilemma comes from the implications for the source
with uranium introduction by meteoric water during syn–Central area if the sulfides are inferred to be detrital. Some very common
Rand Group tectonic uplift, followed by diagenetic maturation sulfides (especially of the base metals) are under-represented,
and migration of hydrocarbons, and final introduction of gold by and some unusual ones are over-represented (e.g., gersdorffite
hydrothermal processes. In contrast, Barnicoat et al. (1997) and and cobaltite), despite no evidence to suggest this relationship in
Jolley et al. (1999) also support a hydrothermal model but sug- any reasonable Archean source area. This discrepancy between
gest that the gold, uranium, and hydrocarbon relate to the one the predicted placer assemblage and the observed Witwatersrand
fluid event. Reef assemblage must be examined in light of the post-deposi-
On a larger scale, detailed structural studies of alteration tional history of the sedimentary sequence but even with local
and deformation show a network of bedding-subparallel shears, remobilization, the base metal sulfides should be better repre-
faults, quartz veins, and fractures that control alteration (Law and sented given likely source areas.
Spencer, 1992; Jolley et al., 1999; 2004). Individual fractures Despite the emphasis on sedimentological aspects of the
within these deformation zones host carbon seams that contain reefs in the modern literature, there are relatively few published
a substantial proportion of the mineralization in many orebod- whole rock geochemical data for the orebodies and especially
ies. In the case of the Carbon Leader Reef, the mineralization is for the surrounding lithologies. One consequence is that marked
locally hosted entirely within the carbon seam. More commonly, chemical differences between reefs and their enclosing litholo-
mineralization is hosted by a variety of rock types but hydrocar- gies are not widely reported. A comparison of compositional dif-
bon nodules and seams are commonly important. A complete pro- ferences between the Leader Reef in the Welkom goldfield and
gression from discrete individual nodules to aggregated nodules the surrounding upper Witwatersrand arenites (Fig. 5) illustrates
and finally continuous seams along the margins of shear zones several key relationships with implications for ore genesis:
suggests that the hydrocarbons were emplaced during deforma- 1. There is a dramatic enrichment in the orebody of gold and
tion (Law and Spencer, 1992). uranium relative to all other elements.
SOURCE AREA DETRITAL ASSEMBLAGES 2. Major elements constituting the silicate assemblage are
similar in both units with the exception of Mg and Fe,
reflecting the increased abundance of chlorite and iron
sulfides in the ore assemblage.
3. Cr and Zr, which reflect detrital concentrations of chro-
Orebodies mite and zircon, respectively, are slightly enriched in the
and altered Leader Reef. However, the concentrations do not support
sediments substantial enrichment of the heavy mineral suite, espe-
cially not enough to explain the enormous enrichment of
gold and uranium. Fe and Ti also reflect a substantial detri-
tal contribution either as detrital pyrite or as the missing
“black sands” component of the reef, including magnetite
Unaltered and ilmenite (prior to sulfidation of the reefs).
surrounding
stratigraphy 4. Chalcophile metals including Ni, Co, As, and Sb are con-
siderably enriched in the Leader Reef. Although As and Sb
are common in many hydrothermal gold deposits, Ni and
Co are rare, although they are well represented in hydro-
thermal uranium deposits, notably the unconformity-asso-
ciated deposits in Australia and Canada. Cu and Zn are
SHAPE MINERAL
moderately enriched in the Leader Reef but their absolute
euhedral Pyrite abundances are relatively low in both the reef and enclos-
ing quartzites.
round Magnetite / ilmenite
Enrichments in iron sulfides and associated chalcophile
Base metal sulphides metals similar to those in the Leader Reef in Welkom have been
reported from many other Witwatersrand reefs (e.g., Tucker,
Figure 4. Schematic illustration of the “Witwatersrand dilemma” with
respect to the mineralogy of likely source rocks, the orebodies, and 1980; Pretorius, 1976a, 1976b; Minter 1978).
other Archean stratigraphic units. The inferred granite-greenstone ter-
rane is likely to contain iron oxides including magnetite, ilmenite, and Ore Mineralogy: Pyrite, Gold and Uraninite
a complex suite of sulfides including pyrite and base metals. Given a
reducing atmosphere, each of these components should be well rep- Pyrite
resented in placer mineralization. In contrast, the orebodies contain
round pyrite to the virtual exclusion of base metals (excluding Ni and One of the most striking features of the Witwatersrand
Co, which are moderately enriched) and the “black sand” or iron oxide conglomerates is the concentration of pyrite. The relationship
component typical of modern placers. Away from the orebodies and as- between pyrite and gold is clear in a stratigraphic sense; most
sociated alteration the overlying and underlying units (in the West Rand reef horizons have significantly more pyrite than average for
Group) contain iron oxides but commonly lack any round sulfides. This the upper Witwatersrand although not all pyrite rich zones are
distribution pattern favors post-depositional sulfidation as the most
likely control on the observed abundances of heavy minerals. auriferous. In a lateral sense, there is also a general correlation
1000
Figure 5. Ratios of element concentra-

tions (geometric means) for the Leader
Central Rand Group quartzites
100
Reef and surrounding unmineralized
Ore assemblage quartzite from the Central Rand Group
(quartzite data from Law, 1991; Leader
Leader Reef
Heavy minerals
10 Reef geochemistry from Callow, 1989,
personal commun.). Gold is enriched in
a suite of chalcophile metals including
Silicate assemblage
Ni, Co, As, and Sb. Cu and Zn are mod-
1
erately enriched but absolute abundanc-
es in all lithologies are low. Fe, Ti, Zr,
and Cr form part of the heavy mineral
suite reflecting chromite, zircon, and
0.1
iron/titanium oxides, respectively.
Si Al K Na Ca Mg Fe Ti Cr Zr Zn Cu Th Co Sb Ni As U Au
Element
2. Porous round or “mudball” pyrite comprising round aggre-

between modal pyrite abundance and gold grades within some
gates of numerous smaller pyrite grains that are 1 mm to 1
reefs but this is less clear when data from many reefs are plot-
cm in diameter and commonly display radial, concentric,
ted together (Phillips and Law, 2000). Pyrite takes many forms
or oolitic textures (Figs. 7A [lighter colored nodules to the
and has been extensively studied and morphologically classified
right of scale bar], 7C) (Ramdohr, 1958; Hallbauer, 1986).
(e.g., Saager, 1970; see summary in Hallbauer, 1986).
An interesting feature of these pyrite grains is that they are
An overwhelming proportion of round pyrite grains are spa-
considerably larger than other heavy minerals, such as zir-
tially related to scours and other sedimentological features (Fig.
con, chromite, and normal (compact round) pyrite found
6A) and to heavy minerals that almost certainly form part of the
in the conglomerates (e.g., Coetzee, 1965). Porous round
detrital suite (e.g., zircon, chromite). There is thus broad agree-
pyrites could represent syn-sedimentary growth of sulfides
ment that round pyrite grains are located in positions controlled
at or near the depositional site (e.g., Hallbauer, 1986) or
by the sedimentary dynamics during deposition of the sediments.
post-depositional sulfidation of iron oxide pisoliths (Phil-
The atmosphere debate has thus centered on whether the grains
lips and Law, 2000).
were deposited as pyrite or as detrital iron oxide minerals that
3. Laminated round pyrites (Fig. 6D) comprising banded lay-
were pseudomorphously replaced by pyrite.
ers of fine-grained pyrite giving the appearance of rounded
Three main types of round (“buckshot”) pyrite are common
fragments of pyrite. These may represent banded sulfide
although other textural varieties are also present (see summary in
clasts or replacement of other iron-bearing minerals in a
Hallbauer, 1986):
banded precursor (e.g., ferruginous chert).
1. Compact round pyrites (Fig. 6B) comprising single pyrite
grains with round shapes and diameters of less than 1 Euhedral pyrite grains (Fig. 6E) and euhedral overgrowths
mm that have been widely interpreted as detrital. Alterna- on round grains (Fig. 6F) are very common in samples from Wit-
tively they could be rounded magnetite and titano-magne- watersrand mines. There is also abundant evidence for sulfida-
tite grains that were sulfidized after burial. Round pyrite tion of a variety of other minerals throughout the Witwatersrand
grains commonly host gold in crosscutting, post-deposi- goldfields (Ramdohr, 1958; Phillips and Dong, 1993; Myers et
tional fractures (Fig. 6C). al., 1993) (Figs. 8A–C). In some round pyrite grains, remnants of
Figure 6. Photomicrographs. (A) Pyrite on subhorizontal sedimentary bedding surfaces, Black Reef, Consolidated Modderfontein mine, showing
the strong sedimentological control on the distribution of iron typical of most orebodies. (B) Compact round pyrite (P), detrital quartz (Q), and
phyllosilicate matrix (part reflected and part cross-polarized). Field of view is 1.3 mm. From Phillips and Dong, 1993. Pressure solution has result-
ed in pyrite indenting detrital quartz grains. (C) Brecciated, compact, round pyrite with gold filling fractures, Ventersdorp Contact Reef. Reflected
light. Field of view is 0.5 mm. (D) Banded, porous, round pyrite with secondary overgrowth. Reflected light. Field of view is 2 mm. (E) Cubic
pyrite overgrowths on spheroidal cores (etched by HNO3). Field of view is 1.3 mm. From Phillips and Dong, 1993. (F) Pyrite overgrowth on a
round pebble with a skeletal texture indicated by differences in relief (reflected light). Field of view is 0.66 mm. From Phillips and Dong, 1993.
Figure 7. (A) Photograph of pisolitic

iron nodules from a modern pediment
(left) and porous round pyrite from the
Basal Reef in the Welkom goldfield
(right) showing the similarity in size and
morphology. (B) Photograph of modern
concentrically structured iron-rich con-
cretion in which dark layers are Fe-rich.
Figure from Brewer, 1964, reproduced
with permission. Field of view is ~1 cm.
(C) Photograph of porous round pyrite
from the Ventersdorp Contact Reef,
Deelkraal gold mine of the Witwa-
B C tersrand Basin in which Fe-rich layers
are light colored. We infer mechanical
abrasion of this central grain prior to
burial and before sulfidation. Field of
view is ~1.5 cm. Note the wavy nature
of some layers in each sample. Sample
(B) collected by M. Mullins.
Figure 8. Photos of replacement textures in Witwatersrand orebodies. (A–C) Partially sulfidized chert pebbles with increased sulfidation around
the pebble margin, reflecting sulfidation after burial. Scale bars in millimeters. (D) Inferred former titano-magnetite grain with lamellae of rutile
(light gray) formed during Ti-exsolution. The grain matrix of magnetite is replaced by pyrite (light yellow) reflected light. Field of view is 1.3
mm. From Phillips and Dong, 1993. (E) SEM photomicrograph of rutile needles (R) with a texture indicating dissolution of titano-magnetite with
the preservation of ilmenite exsolution lamellae, Kimberley Reef, Evander Goldfield. Photo by Andy Barnicoat. (F) SEM photomicrograph of
nodular hydrocarbon (black) with inclusion of uraninite and galena (white). Nodules are located along a fracture zone in association with round
pyrite (light gray), quartz, and chlorite. Pyrite is locally corroded by hydrocarbon, and gold is indicated by the arrow. Photo by Andy Barnicoat.
unsulfidized ilmenite are preserved (Fig. 8D) (Phillips and Dong, occur in textural sites that postdate deposition of the sediments
1993). In other cases, exsolution lamellae of rutile are preserved (e.g., Frimmel, 1997; Frimmel and Gartz, 1997; Barnicoat et al.,
within round pyrite grains indicating their origin by sulfidation 1997; Phillips and Law, 2000). Some possible exceptions have
of titano-magnetite (Fig. 9D). Rare examples of leached titano- been noted by Minter (1999) but the interpretation of these grain
magnetite and ilmenite grains have also been described (Fig. 8E). shapes remains controversial (Barnicoat et al., 2001).
MacLean and Fleet (1989 have described some compact round Individual Witwatersrand gold grains are compositionally
pyrite grains with growth zones truncated by round grain bound- homogenous with respect to Ag and Hg (Utter, 1979; Hirdes
aries (Figs. 9A, 9B). and Saager, 1983; von Gehlen, 1983; Hallbauer, 1986; Oberthür
Sulfur isotope studies of pyrite grains from the Witwa- and Saager, 1986; Reid et al., 1986; Frimmel et al., 1993). On
tersrand provide conflicting evidence partly because of conflict- the scale of a single hand specimen, greater variability has been
ing data sets depending on the scale of observation (Fig. 10). described. Some samples show significant between-grain vari-
Conventional bulk grain analyses show little departure from ability whereas all grains in other samples are compositionally
0‰ (see review in Phillips and Law, 2000, p. 478). In contrast, homogeneous. In a study of the Basal Reef, Frimmel and Gartz
SHRIMP (sensitive high-resolution ion microprobe) ion micro- (1997) demonstrated substantial within-sample variability and
probe analysis of micron-sized spots on individual grains show inferred a detrital origin for the gold. However, there is a strong
variations of 6‰ within single pyrite grains, 9‰ between adja- mineralogical control on gold grain composition (Fig. 11) with
cent touching grains, and 11‰ in single samples (Eldridge et al., relatively restricted compositional variability in grains spatially
1993; Phillips and Law, 2000), and 21‰ in round porous pyrite associated with chlorite and pyrite and far greater variability
(England, 1999). within grains associated with quartz grains. This pattern implies
an in situ control on gold compositions.
Gold A recent study on within-sample gold homogeneity has
There is now widespread agreement that the majority of described grain shapes from a single sample of the Basal Reef in
Witwatersrand gold grains have secondary grain shapes and/or the Welkom goldfield (Minter et al., 1993; Frimmel et al., 1993).
A 0.05 mm
B
0.05 mm
C 0.05 mm
D 0.1 mm
Pyrite
overgrowth
Figure 9. Sketches illustrating important processes responsible for round pyrite grains in the Witwatersrand. (A, B) Compact round pyrite grains with
oscillatory polygonal to colloform growth banding defined by As-rich and As-poor bands, Basal Reef, Welkom Goldfield. KMnO4 stained. Redrawn
from MacLean and Fleet (1989). MacLean and Fleet have argued that the truncation of growth banding by round grain margins indicates that the
grain was pyrite at the time of rounding and that these textures preclude sulfidation of pre-existing round minerals. (C) Oscillatory-zoned pyrite from
the hydrothermal ores of the Agnico-Eagle gold mine showing sector zoned core, polyhedral growth banding and irregular margin. KMnO4 stained.
Redrawn from Fleet et al. (1989). Original zoned grains have been partly dissolved and overgrown by later low-As pyrite. Dissolution of this type
provides an alternative rounding mechanism for pyrite grains in the Witwatersrand such as (A) and (B). (D) Round pyrite grain from the Witwa-
tersrand showing characteristic titano-magnetite exsolution lamellae preserved as rutile. The original titano-magnetite grain has been sulfidized to
form pyrite (redrawn from Ramdohr, 1958). Sulfidation may have occurred pre- or post-burial.
These authors have inferred that 75% of the gold preserves pri-
SHRIMP single point
mary detrital gold shapes and that the remaining 25% has typical
Euhedral hydrothermal characteristics. Other workers studying exactly the
same sample inferred that all of the gold grains are hydrothermal
Round in origin (Barnicoat et al., 2001), that locations of the grains are
controlled by fractures, and that their shapes reflect intergrowths
Conventional bulk grain with other secondary minerals. Irrespective of the origin of the
gold, the compositions and compositional ranges of all grains are
Euhedral
similar, requiring either a remarkably homogeneous source ter-
Round rane (if detrital) or a post-depositional control on gold composi-
tions (Fig. 12).
-8 -4 0 4 8 12 16 20
δ 34S
20
Figure 10. Summary of traditional “bulk-grain” and SHRIMP micro- A
probe analyses of sulfur isotopes for Witwatersrand pyrites (data from
Hoefs et al., 1968; Palmer, 1986; Eldridge et al., 1993; England, 1999; 15
Ag (wt %)
Phillips and Law, 2000). VCR W Driefontein
VCR Driefontein
Vein
A
10
in VCR Basal Reef
chlorite
5 5
80 85 90 95
gold with: Au (wt %)
Hg (wt %)
4
Quartz
Chlorite 6
3 Pyrite
5
B
Galena
Vein in VCR
2 4
Hg (wt %)
pyrite 3
VCR Driefontein
1
2 Basal Reef
n = 17 VCR W Driefontein
0 1
6 7 8 9 10 11 12
0
Ag (wt %) 5 10 15 20
B Ag (wt %)
gold with: Figure 12. Compositional variability in Witwatersrand gold grains

Quartz from the Ventersdorp Contact Reef (VCR) and the Basal Reef. (A) Ag
5 chlorite
Chlorite
versus Au, and (B) Hg versus Ag. Each field represents the variability
y in gold grain compositions for one or more samples from the localities
Pyrite
indicated and discussed in the text. Compositional variability for any
Hg (wt %)
pyrite Pyrrhotite
4 likely source terrane is likely to span the entire range illustrated and
should be reflected in random detrital samples. Gold grain composi-
tional variability between sites is limited and implies post-depositional
3
controls on gold grain composition. Data compiled from Frimmel et
al., 1993 and Frimmel and Gartz, 1997.
2
1
Figure 11. Within-sample compositional variation in Hg and Ag for gold
particles in two samples from the Ventersdorp Contact Reef, Klerksdorp
0
6 7 8 9 10 11 12
goldfield (reproduced from Frimmel and Gartz, 1997). Compositional
variability is reduced in spatial association with pyrite and chlorite im-
Ag (wt %) plying a post-depositional control on gold grain compositions.
On a regional scale, there are substantial variations in the and its implications for the Archean atmosphere be carefully
compositions of gold grains among samples. For example, gold reassessed. Three questions are critical: (1) Are the round pyrite
grains in two samples from the Ventersdorp Contact Reef (VCR) grains detrital? (2) Do gold compositions reflect atmospheric
at East Driefontein mine are compositionally distinct from those composition? (3) Are round uraninite grains detrital?
at West Driefontein mine some 10 km away on the same strati-
graphic horizon (Fig. 12). The degree of variability at each of Are the Round Pyrite Grains Detrital?
these locations is similar to that reported by Frimmel et al. (1993)
and discussed previously. In addition to weathering and mechanical abrasion, post-depo-
sitional mechanisms that have modified Witwatersrand detrital
Uraninite assemblages include chemical rounding, grain dissolution/replace-
Round to “muffin” shaped uraninite grains are common in ment and sulfidation. The significance of these processes for inter-
the Witwatersrand reefs (see summaries in Liebenberg, 1955, pretation of round pyrite grains is discussed below.
Saager, 1968, and England et al., 2001). They typically occur
as rounded to sub-rounded grains although rare euhedral grains Chemical Rounding of Iron Sulfides
have been reported. Grains are typically less than 250 μm in Many of the textural varieties described above display com-
diameter and occur in two dominant associations: (1) as isolated plex internal structure that is locally truncated by round grain
grains often spatially associated with, and partially disaggregated margins (e.g., Ramdohr, 1958). Several recent studies of the Wit-
and replaced by, hydrocarbons (Fig. 13); or (2) as concentrated watersrand have highlighted the aggressive nature of hydrother-
clusters within bedding-subparallel fractures known as carbon mal alteration in the reefs, especially in association with hydro-
seams (e.g., Fig. 8F). carbons, suggesting the potential for widespread dissolution and
modification of original grain shapes (Gray et al., 1998; Jolley et
IMPLICATIONS OF WITWATERSRAND al., 1999; Phillips and Law, 2000; England et al., 2001). Some
MINERALOGY FOR THE ARCHEAN ATMOSPHERE compact round pyrite grains have euhedral growth zones trun-
cated by round grain boundaries (Figs. 9A, 9B; MacLean and
The placer model for Witwatersrand gold has been the Fleet, 1989). Similar grains, with textures similar to the Wit-
mainstay of the reducing Archean atmosphere hypothesis since watersrand, have been described by the same authors (Fleet et
the 1950s. Over the past 20 years the geological framework for al., 1989) from the greenstone hosted hydrothermal ores at the
the Witwatersrand has changed significantly and many of the Agnico-Eagle Mine of the Abitibi greenstone belt of Canada
assumptions underpinning the placer model are no longer valid. (Fig. 9C) where hydrothermal pyrite grains have been modi-
In particular, the recognition of post-depositional alteration in fied by later hydrothermal processes. The Witwatersrand pyrites
and around each of the orebodies requires that the placer model may thus have formed by similar hydrothermal processes, but
MacLean and Fleet have argued that these textures demonstrate
that the pyrites were rounded by detrital processes and depos-
ited as a sulfide phase. In the case of the hydrothermal Agnico-
Eagle mineralization, this conclusion is demonstrably wrong and
similar textures cannot be used to constrain the origin of Witwa-
tersrand sulfides.
Dissolution of Iron Oxides

Iron is highly mobile in reduced, low-sulfur fluids near
Earth’s surface. It is thus possible that iron could be leached from
detrital iron oxides during transport under reducing atmospheric
conditions. If such a process occurred, it would be difficult to
constrain the extent of the dissolution. However, in some cases
textural evidence suggests that some titano-magnetite and ilmen-
ite grains did survive transport to the basin and were subsequently
10mm 10mm
leached in situ to leave skeletal titanium oxides that are probably
Figure 13. Sketches of an SEM photomicrograph showing the progres- too delicate to have survived detrital transport (Fig. 8E).
sive fragmentation and dissolution of uraninite grains (white) by hy- The dissolution of heavy minerals by reduced basinal fluids
drocarbon (black). Drawn from photographs by England et al., 2001. is a common process in sedimentary basins (Morton and Halls-
The nodule in left (A) is coated with brannerite and florencite (too worth, 1999) and has been important in the Witwatersrand. How-
small to illustrate). The nodule in right (B) is highly fragmented and
rounding of individual fragments reflects the partial dissolution urani- ever, the relative importance of this process and the location of
nite by hydrocarbons. The nodule is surrounded by chlorite, pyrite and the dissolution process (i.e., in transport versus in situ) remain
gold (not illustrated). effectively unconstrained.
Sulfidation of Detrital Iron Oxides the basis of shape alone. This ambiguity currently prevents any
Reduced Witwatersrand-type pyrite and uraninite mineral- definite conclusions to be made about the original proportions
ization is invariably associated with round pyrite and the virtual of pyrite in the conglomerates and precludes the use of pyrite to
absence of the “black sand” components typical of modern plac- constrain the composition of the Archean atmosphere.
ers, including magnetite, ilmenite, and hematite, all of which are This does not necessarily preclude the presence of some
common components of any likely Witwatersrand source area. detrital pyrite grains—it is possible that uplift of older basin sedi-
Several authors have suggested that these minerals were sul- ments exposed local sources of detrital pyrite that was formed
fidized, dissolved, or replaced prior to deposition (e.g., Reimer during diagenesis, to erosion and redeposition. The extent of this
and Mossman, 1990). Others have argued that because detrital process and its implications for the composition of the Archean
magnetite and ilmenite are common detrital phases in stratigra- atmosphere cannot be determined in the absence of data on other
phy above and below the Witwatersrand orebodies, post-deposi- key variables such as the duration of transport and the physical
tional processes must be responsible for their removal (Dimroth and chemical nature of the depositional environment.
and Kimberley, 1976; Clemmey and Badham, 1982). Redox gradients in the near-surface are commonly extreme,
There is abundant supporting evidence of in situ sulfidation and many modern near-surface waters are not in equilibrium with
of a variety of other minerals throughout the Witwatersrand gold- the atmosphere. Even within modern weathering profiles, the
fields (Figs. 8A–C). Furthermore, euhedral pyrite grains indicate redox potential changes significantly with both time and position
the presence of a fluid capable of mobilizing gold and sulfur and in the profile, often in proximity to the water table. Inferences
the potential to dissolve and/or re-precipitate pyrite during altera- of atmospheric composition based on mineral assemblages and
tion. This process is widely accepted, but the location and timing paleosols thus require a clear demonstration of equilibrium with
of the sulfidation are unconstrained and could have occurred at the atmosphere, and that remains a challenge for most studies.
source, during transport into the basin, or after deposition. Diag-
nostic textural relationships such as replacement parallel to clast Do Gold Compositions Reflect Atmospheric Composition?
margins are the exception rather than the rule, and the extent of
the sulfidation is thus unconstrained. Detrital gold grains in equilibrium with the modern atmo-
Perhaps the most telling argument is the regional distri- sphere are commonly (but not universally) depleted in silver rela-
bution of the round sulfide assemblages. In each example of tive to their inferred source, reflecting the preferential leaching of
reduced Witwatersrand-type pyrite and uraninite mineralization silver in an oxygenated atmosphere (Morrison et al., 1991). In con-
worldwide (Table 1), round pyrites are common, black sand com- trast, Witwatersrand gold grains typically contain significant silver
ponents are virtually absent, and the evidence for sulfidation, in (around 10 wt%; range 0–30 wt%). It has thus been suggested that
the form of secondary euhedral pyrites and partially sulfidized compositional differences between gold in modern placers and in
mineral grains, is widespread. If sulfidation occurred outside the the Witwatersrand may reflect differences in atmospheric composi-
basin, the detrital assemblage (moving up the sequence) must tion. This hypothesis is based on two key assumptions: (1) Witwa-
have changed from oxide-dominated to sulfide-only, and then tersrand gold is detrital, and (2) there has been no secondary remo-
back again, to account for the stratigraphic distribution of assem- bilization and/or post-depositional alteration of gold chemistry (i.e.,
blages noted above. Furthermore, this process must have been assuming and unmodified placer model).
repeated in several localities throughout the world. In situ sul- There is now a broad consensus in the literature that composi-
fidation by hydrothermal processes appears to be more likely. tions of gold grains in the Witwatersrand reflect post-depositional
If the current isotopic distribution has not been reset during processes either by local remobilization and/or re-equilibration
retrogression, the wide range of sulfur isotopic values for pyrite of detrital gold (e.g., Frimmel et al., 1993; Frimmel and Gartz,
(Fig. 10) reflects either a source area that supplied heterogeneous 1997) or by a hydrothermal origin for the gold (e.g., Barnicoat et
detrital pyrite or sulfidation by solutions with variable sulfide- al., 1997; Phillips and Law, 2000). Gold grain compositions (e.g.,
sulfate ratios. The detrital pyrite model (with the variable δ34S Figs. 10, 11) show limited variability, with a range less than that
thus reflecting processes in the source terrane) has the conflict- observed in gold grains from auriferous quartz veins cutting the
ing requirement of a reducing atmosphere to stabilize pyrite dur- orebodies. As a result, within-sample variability cannot be used
ing transport, and a quite oxidizing atmosphere to provide the to suggest that the grains are derived from a variety of different
wide range of δ34S values. If such grains were not the result of source rocks and is compatible with a hydrothermal genesis.
an oxidizing atmosphere, a suitable mechanism to generate the In summary, textural studies indicate that the overwhelming
observed variability in the absence of an oxygenated atmosphere majority of gold grains are in structural sites that were not present
must be defined. at the time of deposition and cannot be detrital (unmodified) in
On the basis of the abundance of pyrite, the absence of iron origin. Local remobilization of any detrital grains may be pos-
and iron-titanium oxides, and the textural evidence for pseudo- sible but gold compositions now reflect post-depositional pro-
morphous replacement, the most reasonable conclusion is that the cesses and cannot be used to constrain atmospheric compositions
origin of any specific round pyrite grain in the Witwatersrand and at the time of sedimentation or to date the age of inferred source
other altered sedimentary sequences is likely to be ambiguous on rocks for detrital gold.
Are Round Uraninite Grains Detrital? pre-existing grains. In the latter case, rounding of pre-existing
grains may be an inevitable consequence of the disaggregation
The interpretation of round uraninite grain shapes is con- process. England et al. (2001) have argued that the hydrocar-
troversial. Davidson (1960, p. 155) drew attention to round bons are precipitated around detrital uraninite grains. Other
pitchblende grains in vein deposits from Freiberg, Germany. workers argue that uraninite and hydrocarbon are both localized
More recently Phillipe et al. (1993) showed that epigenetic ura- along structurally controlled fractures (e.g., Law and Spencer,
ninite grains in unconformity uranium deposits of the Atha- 1992; Barnicoat et al., 1997; Jolley et al., 1999; 2004; Phillips
basca Basin are round and have compositions similar to those and Law, 2000).
in the Witwatersrand. These grains demonstrate that round- In addition to uraninite, brannerite and uraniferous leu-
ness is not necessarily a function of detrital transport. coxene are important uranium-bearing minerals in many Wit-
The textural interpretation of isolated uraninite grains watersrand reefs (Hallbauer, 1986), but the concentrations and
in the Witwatersrand is particularly difficult because of the proportions of these minerals vary greatly on local and regional
intimate association of uraninite and carbon in the reefs. This scales. Textural studies show that interaction of Ti-bearing
association has variably been ascribed to sedimentological detrital minerals on unconformity surfaces with uranium in
(Minter, 1976; Hallbauer, 1986) or chemical processes (Phil- solution has formed widespread secondary brannerite and ura-
lips and Law, 1997; 2000), and it is now widely accepted that niferous leucoxene (Davidson, 1953; 1957; 1960; Liebenberg,
the carbon is derived from the thermal maturation of migrated 1955; Ramdohr, 1958; Hallbauer, 1986). The migration of ura-
hydrocarbons (e.g., Phillips et al., 1990; Gray et al., 1998; Jol- nium after deposition of the Witwatersrand sediments provides
ley et al., 2004). Round nodules of carbon, known as “flyspeck a clear indication of dissolution from surrounding grains or the
carbon,” are not volumetrically abundant in the Witwatersrand, introduction of uranium by hydrothermal solutions. Similarly,
but they are found in virtually every reef and are intimately the widespread distribution of round uraninite grains in hydro-
related to the distribution of uranium. Although these blebs carbon seams that postdate the deposition of the sediments
have been interpreted as detrital grains (e.g., Hallbauer, 1986), demonstrates that round shapes are not related to sedimentary
it is now believed that they are genetically related to the car- rounding and that uranium has been mobile after deposition of
bon seams and were precipitated around uraniferous minerals the sediments.
by radiolytic polymerization (McCready and Parnell, 1998). In summary, round uraninite grains in hydrocarbon seams
Most Witwatersrand carbon nodules and seams contain high in brittle fractures and along shear zones cannot reflect detri-
uranium concentrations disseminated throughout and imply tal processes and imply substantial uranium mobility after
substantial post-depositional mobility of uranium. deposition. Aggressive dissolution of uraninite by hydrocar-
There are at least two post-depositional processes that could bons together with widespread uranium mobility casts further
account for the intimate association of uraninite with migrated doubt on the significance of any grain shapes. Round uranif-
hydrocarbon material. If the uraninite pre-dates the hydrocar- erous hydrocarbon nodules in granitic rocks surrounding the
bons, then it could precipitate migrating hydrocarbons by radio- basin probably reflect the same alteration and have been dated
lytic polymerization (McCready and Parnell, 1998; England between 2.7 Ga and 2.0 Ga and thus postdate Witwatersrand
et al., 2001). Importantly, this process requires that hydrocar- sedimentation (Klemd, 1999).
bons postdate the uranium minerals, but does not differentiate Even if it is accepted that some or all of the round ura-
between detrital and early hydrothermal uranium mineraliza- ninite grains in the Witwatersrand, in Elliot Lake, and in the
tion. Alternatively, if the uraninite postdates or is synchronous Pilbara Craton are primary detrital minerals, the presence of
with the hydrocarbon-bearing fluid, the carbon may represent uraninite as a detrital mineral in the Indus River (Maynard
a localized site of reduction capable of precipitating uraninite et al., 1991) begs the question of how definitive this criterion
from solution. can really be in constraining the Archean atmosphere. If the
Criteria to differentiate these processes are not clear-cut. In Archean atmosphere truly stabilized uraninite, it is surprising
some cases, the textural evidence suggests that carbon replaced how rare uraninite is in the Archean geological record.
and progressively dismembered pre-existing uraninite grains Given the evidence for post-depositional alteration in
(e.g., Liebenberg, 1955; Smits, 1984; England et al., 2001) the Witwatersrand, Rasmussen and Buick (1999) completely
(Figs. 13A, 13B). However, mesoscopic relationships indicate discard the Witwatersrand ores as a convincing site of detri-
that carbon seams in fractures subparallel to stratigraphy host tal pyrite and uraninite, and focus on heavy mineral assem-
substantial uraninite and are thus incompatible with a simple blages in other largely unmineralized Archean sediments
detrital origin for at least this part of the mineralization (cf. such as those of the Pilbara Craton of Western Australia.
England et al., 2001). In other cases, uraninite and other uranif- These authors have identified pyrite, uraninite and siderite
erous minerals are finely disseminated within the hydrocarbons in Archean sedimentary rocks of inferred detrital origin and
and are probably epigenetic precipitates (Fig. 8F). The process argue that because the sediments are believed to be unaltered,
of nodule formation results in round, discrete uraninite grains they are more likely to reflect the composition of the Archean
that reflect either precipitation of uraninite or disaggregation of atmosphere. However, the presence of sericite in many of
these samples (typically a hydrothermal phase in the Witwa- ball pyrite.” These sulfidic muds have been attributed to a reduc-
tersrand; Law, 1991) and migrated hydrocarbons suggests that ing atmosphere (Hallbauer, 1986) or to exhalative activity around
they have not escaped post-depositional alteration processes, the margins of the Witwatersrand Basin (Hutchinson and Viljoen,
and this study suffers from the same limitations as those out- 1988). The former explanation requires an atmosphere different
lined above for the Witwatersrand. from the present to preserve the pyrite during transport, and fails
to explain the association of the “mud balls” with unconformities.
Possible Evidence for an Oxidizing Archean Atmosphere The hydrothermal activity required for the latter model appears
unlikely given the prevalence of the porous round pyrites in the
The atmosphere debate on the Witwatersrand has focused Witwatersrand Supergroup, the Ventersdorp Contact Reef, and
largely on the current heavy mineral assemblages in the belief the Black Reef at the base of the Transvaal Sequence. Both ideas
that they are detrital. Given the evidence for post-depositional are built on the premise that the Witwatersrand pyrite is detrital.
changes after burial, inferences of minerals existing prior to As shown above, this premise may not be sound and is not inde-
alteration may be more relevant to constraining atmospheric pendently verified.
composition. For example, there is good reason to believe This concept of original Fe-rich pisoliths requires a Pre-
that vital information may be contained within the sediments cambrian atmosphere oxidizing enough to stabilize two forms of
immediately overlying the unconformity. These sediments iron in different parts of the Archean soil horizon with at least
may contain material derived from truncated soil profiles and some zone where ferric iron is abundant. No simple relation-
specifically from the B-horizon in which Fe is concentrated ship between the observed mineralogy and the composition of
in modern profiles (Anand, 1995). Fe-oxyhydroxide concre- the atmosphere is predicted for two reasons. First, atmospheric
tions (e.g., pisoliths) formed in areas of seasonal rainfall, for redox state is only one of several thermodynamic and kinetic
example, are durable, coarse-grained, and have a distinctive constraints on the stability of pyrite and uraninite, and simple
physical appearance (Brewer, 1964). Therefore, they should be assumptions regarding other variables and the whole chain of
identifiable even after considerable transport, metamorphism, weathering, erosion, transport, and depositional effects on indi-
and alteration (Fig. 7). vidual grains are unlikely to reflect any single process reliably
The genesis of the porous, round pyrite can be subdivided (Robinson and Spooner, 1984). For example, detrital uraninite
into models postulating original pyrite and those postulating and pyrite grains are known from modern sediments, albeit in
sulfidation of another mineral. The sulfidation origin for Witwa- minor quantities, and round uraninite and pyrite grains are also
tersrand porous round pyrites was initially proposed by Clem- known from epigenetic deposits. Second, all sedimentary basins
mey (1981) and developed by Phillips and Myers (1989). There are subject to complex post-depositional alteration and diagenesis
is a striking similarity in both morphology and size between the of the original detrital suite by basinal fluids that are commonly,
porous round pyrite and modern Fe-concretions (Fig. 7). Further- but not universally, reducing (Phillips et al., 1990; Morton and
more, the distribution of the concretions in residual conglomer- Hallsworth, 1999). The impact of these fluids must be recognized
ates on unconformities is exactly as predicted from modern ana- and understood before the redox state of the atmosphere at the
logues (Smith and Perdrix, 1982). The mechanical migration and time of deposition can be interpreted.
accumulation of these pisoliths can be predicted from their den- Given the abundant evidence for post-depositional modifica-
sity by analogy with modern sedimentological sorting processes tion of the Witwatersrand detrital assemblage and the evidence for
(Anand, 1995). Oxyhydroxide minerals in the pisoliths would post-depositional alteration, sulfidation, and widespread mobility
be highly susceptible to sulfidation after burial because of their of gold and uranium, we conclude that the mineral assemblage of
high Fe content. In contrast, ferruginous chert has a lower Fe the Witwatersrand is not a reliable indicator of a reducing atmo-
content and generally undergoes only partial alteration to pyrite sphere. In fact, reinterpretation of some porous round pyrites as
(Phillips and Dong, 1993). Textural evidence for sulfidation to sulfidized iron pisoliths may require a relatively oxidizing envi-
form these pyrite grains is equivocal. The distribution of porous ronment to stabilize iron in both its ferric and ferrous forms. The
round pyrite grains is entirely restricted to unconformity surfaces widespread inferred mobility of uranium in the Witwatersrand
near the proximal basin margin, and thus they lie within the phyl- may also require an oxygenated fluid to stabilize U6+ in solution.
losilicate-plus-pyrite alteration envelope associated with Wit- Similar sedimentary environments in the younger rock record are
watersrand mineralization (Phillips and Law, 2000). Unconfor- frequently altered by inflowing oxygenated meteoric waters that
mity surfaces outside the alteration envelope would be required reflect prevailing atmospheric conditions.
to determine the pre-alteration mineralogy. Comparable sulfide
nodules are uncommon elsewhere in the Archean except where Significance of “Old” Isotopic Ages
associated with similar mineralization and are thus unlikely to
reflect a general process related to the Archean atmosphere. A large range of mineral and whole rock ages have been
Traditionally, the origin of the porous round pyrite grains published for the Witwatersrand that reflect pre-, syn-, and post-
has been linked to sulfidic muds accumulated in the intertidal depositional events in the basin (see summaries in Robb and
channel areas of fluvial systems, hence the colloquial name “mud Meyer, 1995 and Robb et al., 1990a). Of particular interest to the
atmosphere debate are “old” ages for gold, pyrite, and uraninite of dubious value. In particular, homogenization and/or second-
that purportedly pre-date sedimentation and thus imply a detrital ary introduction of gold grains, and crosscutting pyrite veins
origin for the mineralization. and overgrowths on round pyrites, are likely to affect isotopic
systematics.
Re-Os Isochron “Ages” for Gold and Pyrite
Kirk et al. (2001; 2002) have recently used the Re-Os iso- U-Pb “Ages” for Uraninite
chron technique in an attempt to directly date Witwatersrand gold The pioneering work of Rundle and Snelling (1977) on the
and pyrite. These authors assumed that pyrite and gold form part geochronology of uraniferous minerals in the Witwatersrand
of the heavy mineral assemblage, and used randomly selected has been widely quoted in the literature in support of the placer
grains from a single sample of the Vaal Reef to construct an iso- model. In their 1977 review of their own and available existing
chron with an implied age of 3010 ± 20 Ma. data, these authors infer two discrete age groups: (1) an older
The Vaal Reef sample is described as “fine-grained quartz- population at 3050 ± 50 Ma that predates deposition of the Wit-
itic conglomerate with angular to sub-rounded horizontally frac- watersrand Basin; and (2) a younger population at 2040 ± 100
tured quartz clasts 5-10 mm in size. The conglomerate is clast Ma that postdates the deposition of the Witwatersrand.
supported and is approximately 10% matrix and 90% clasts. The Rundle and Snelling argue that the data reflect detrital urani-
matrix consists of sericite, fine-grain quartz and carbon seams/ nite derived from a 3050 Ma source followed by a resetting of the
patches. The majority of the pyrite/arsenopyrite, and ~10% of the U-Pb system at 2040 Ma within a closed system with limited sec-
visible gold are within the quartz and sericite matrix, while ~90% ondary introduction of uranium. However, they also note that “it
of the gold, the uraninite and some of the pyrite are confined to would be virtually impossible to distinguish between disturbed
the carbonaceous material” (Kirk et al., 2002). detrital systems and disturbed systems with both detrital and
We interpret the sericitic matrix as part of the phyllosilicate authigenic components.” These ages have been widely reported
alteration assemblage typical of Witwatersrand orebodies. Much as “uraninite” ages (e.g., Robb and Meyer, 1995); however, a
of the gold is located in the carbon seam, whereas the sulfides review of the original paper highlights some fundamental short-
are from the quartzite matrix. Recent studies by several differ- comings in the approach used by Rundle and Snelling (1977) and
ent authors (Parnell, 1996; Gray et al., 1998; Jolley et al., 1999; especially the interpretation of the data by other authors.
2004; Phillips and Law, 2000; England et al., 2001) concluded First, the “ages” were obtained on “portions of the matrix,
that the carbon seams postdate deposition of the sediments and avoiding as far as possible the pebbles, cut in the form of cubes
reflect migration of hydrocarbons during post-depositional alter- ~2cm in size” and thus effectively reflect “bulk rock” analyses
ation. Although the precise paragenetic setting of grains used rather than individual mineral grains. Second, each age group
to construct the isochron remains unknown, it is likely that the reported by Rundle and Snelling has been estimated from a
grains are from different host lithologies, with gold from the series of samples based on their distribution on the Concordia
carbon seam and pyrite from the host conglomerate. In the case diagram, with many samples showing evidence for both lead
of hydrocarbon-hosted gold, textural evidence invariably shows and uranium loss. As a result, individual age estimates are highly
that the gold was introduced after deposition of the hydrocarbon variable. Third, no paragenetic information on uranium minerals
(Ramdohr, 1958; Saager, 1968; Barnicoat et al., 1997; Phillips is reported to provide independent geological constraints on the
and Law, 1997) and is thus not co-genetic with the round pyrite likely validity of the ages. For example, it has now been well
grains as inferred by Kirk et al. (2002). documented that many uraninite grains, particularly those hosted
An important challenge for the Re-Os studies is thus to dem- by carbon seams filling structural dilation zones, cannot be of
onstrate that all grains were part of an isotopically homogeneous detrital origin and must reflect substantial mobility of uranium
initial Os187-Os188 reservoir, which is the foundation of isochron- after burial. Galena is also a common secondary mineral in the
based dating techniques. In practice this generally means that all reefs and associated quartz veins, suggesting an open system on
grains need to be sourced from a grain population from a single the scale of the samples discussed above. Similarly, many sec-
sample all sharing a common genesis. Re-Os dating of round ondary uranium minerals are now well documented in the reefs
pyrite grains from Steyn Reef in the Welkom goldfield by the and form an unknown proportion of the material sampled. Many
same authors clearly reflects this problem and yields imprecise of these minerals are poor hosts for lead that was presumably
“isochron” ages of 3490 ± 900 Ma. There is no reason to believe mobilized during alteration. It seems improbable in the light of
that the individual pyrite grains analyzed are derived from an iso- this petrographic evidence that a closed U-Pb system has oper-
topically homogeneous source and therefore we assign no age ated on either a mineral or whole rock scale.
significance to the “isochron.” In any event, the 3490 Ma age Rundle and Snelling also report data for “uraninite” and
fails to discriminate between detrital and hydrothermal mineral- “thucolite” that are apparently from individual mineral sepa-
ization given the 1800 Ma error. rates. Thucolite refers to “thorium-uranium-carbon-oxygen”-
Post-depositional changes in the composition of gold and bearing material or “carbon nodules” to use the terminology
pyrite grains are now well established for the Witwatersrand in this paper. These ages range from 2760 Ma for a sample
and make the use of gold and pyrite geochemistry for dating from the Dominion Group to 2540 Ma through 2000 Ma for
Witwatersrand samples and are consistently younger than the 14B). For example, Smith and Minter (1980) describe better
host sedimentary rocks (although they were thought to be older grades in quartzites and conglomerates with pyritic foresets and
at the time of the analyses). a general increase in grade from sandstone through conglomerate
In light of these uncertainties and limitations, it is unlikely to carbon seams. Similar associations are widely reported from
that the reported ages accurately reflect the age of uraninite or the Witwatersrand and the implied link between sedimentary
other detrital uranium-bearing minerals. In fact, the younger facies and gold grade has been used to support a placer model for
population reported by Rundle and Snelling could equally have the mineralization.
been used to support a hydrothermal origin for the mineraliza- From a chemical standpoint these data can also be inter-
tion (although these ages also suffer from the limitations out- preted to reflect post-depositional mineralizing processes (Phil-
lined above). lips and Law, 2000):
• an association of better gold grades with round pyrite,
Lithological and Chemical Controls on Witwatersrand reflecting sulfidation of detrital Fe-oxide grains and pre-
Mineralization cipitation of gold transported as sulfur complexes;
• an association of better uranium grades with detrital tita-
Given the close spatial association between gold and uranium nium minerals, reflecting precipitation of uranium by reac-
in Witwatersrand orebodies, it is possible that there is a genetic tion with Ti to form brannerite.
link between the two metals. A common assumption has been The increases in grade could thus reflect either differences
that both minerals are concentrated in response to sedimentary in depositional concentration processes between sandstones
sorting processes (e.g., Smith and Minter, 1980; Hallbauer, 1986; and conglomerates during sedimentation (placer model) or ele-
Smits, 1984; Roscoe and Minter, 1993). These authors point out vated iron and titanium concentrations as detrital phases during
that concentrations of gold and uranium are commonly correlated conglomerate deposition (hydrothermal model). However, given
over several orders of magnitude, and that metal concentrations the secondary origin of the carbon, the elevated grades in carbon-
are also broadly correlated with sedimentary facies (Figs. 14A, rich lithologies are more likely to reflect reduction by organic
A L
B
L
10 L
100
L
L
1 L
10
Au (g/t)
Au (g/t)
L
1.0
L
0.1
Conglomerate
without pyritic foresets Sandstone
with pyritic foresets Carbon seam
0.1
0.01
1 10 100 1000 1 10 100 1000 10 000
U (g/t) U (g/t)
Figure 14. Plots of Au versus uranium for selected Witwatersrand orebodies illustrating the link between lithology, mineralogical associations,
and increasing metal content. (A) Elsburg number 5 Reef and Leader Reef (symbols marked L), Klerksdorp and Welkom goldfields, respectively.
Redrawn from Smith and Minter, 1980. Smith and Minter have interpreted the lithological control on gold grade in terms of a detrital control on
mineralization. Alternatively, the strong association of increased metal concentrations with sulfide-rich facies may reflect a chemical control on
the mineralization (i.e., elevated pyrite in gold-rich samples via sulfidation of detrital iron-bearing minerals). (B) Leader Reef, Welkom goldfield.
Redrawn from Smith and Minter, 1980. Carbon seams are not present at the time of sedimentation, and consistent metal ratios in seams and as-
sociated sedimentary facies may imply a post-depositional control on the mineralization.
Brewer, R., 1964, Fabric and mineral analysis of soils: New York, John Wiley
material and hence hydrothermal ore genesis. These mineralogi- and Sons, 420 p.
cal associations provide strong chemical reasons for the associa- Button, A., and Tyler, N., 1981, The character and economic significance of
tion of mineralization with specific host lithologies. paleoweathering and erosion surfaces in southern Africa: Economic Geol-
ogy, 75th Anniversary Volume, p. 686–709.
A major criticism of the placer model has been its inability to Clemmey, H., 1981, Some aspects of the genesis of heavy mineral assemblages
explain the variety of mineralized sedimentary rock types either in Lower Proterozoic uranium-gold conglomerates: Mineralogical Maga-
on theoretical grounds or by analogy with young placer depos- zine, v. 44, no. 366, p. 399–408.
Clemmey, H., and Badham, N., 1982, Oxygen in the Precambrian atmosphere:
its (Phillips and Law, 2000). This limitation is most obvious in An evaluation of the geological evidence: Geology, v. 10, p. 141–146,
the case of carbon seams that are locally major host lithologies doi: 10.1130/0091-7613(1982)10<141:OITPAA>2.0.CO;2.
for gold and uranium. In contrast, the chemical associations of Coetzee, F., 1965, Distribution and grain size in gold, uraninite, pyrite and
certain other heavy minerals in gold-bearing reefs of the Witwatersrand
gold with iron (pyrite) and carbon, and uranium with carbon Basin: Transactions of the Geological Society of South Africa, v. 68,
and titanium, are clear in all the major orebodies and crosscut p. 61–88.
sedimentary facies. Coward, M.P., Spencer, R.M., and Spencer, C.E., 1995, Development of the
Witwatersrand Basin, South Africa, in Coward, M.P., and Ries, A.C., eds.,
Early Precambrian processes: Geological Society [London] Special Pub-
ACKNOWLEDGMENTS lication 95, p. 243–269.
Davidson, C.F., 1953, The gold-uranium ores of the Witwatersrand: Mining
Magazine (London), v. 88, p. 73–85.
We are grateful to Steve Kesler, Hiroshi Ohmoto, and Davidson, C.F., 1957, On the occurrence of uranium in ancient conglomerates:
Michael Kimberley for the invitation to attend the 2002 sym- Economic Geology and the Bulletin of the Society of Economic Geolo-
posium on the “Evolution of the early atmosphere, hydrosphere, gists, v. 52, p. 668–693.
Davidson, C.F., 1960, The present state of the Witwatersrand controversy: Min-
and biosphere.” We would also like to thank coworkers in this ing Magazine (London), v. 102, p. 84–95, 149–159, 222–229.
field who have been instrumental in shaping our ideas, espe- Davidson, C.F., and Cosgrove, M.E., 1955, On the importance of uraninite
cially Russell Myers, Judy Palmer, Guoyi Dong, and Martin as a detrital mineral: Geological Survey of Great Britain Bulletin, v. 10,
p. 74–80.
Hughes. Richard Spencer was a key person in recognizing the Dimroth, E., and Kimberley, M.M., 1976, Precambrian atmospheric oxygen:
structural controls on hydrocarbon seams during mapping at Evidence in the sedimentary distribution of carbon, sulfur, uranium and
Kinross Mine. Bill Fyfe has provided insightful comments on iron: Canadian Journal of Earth Sciences, v. 13, p. 1161–1185.
Eldridge, C.S., Phillips, G.N., and Myers, R.E., 1993, Sulfides in the Witwa-
the early Earth and its atmosphere. We also thank Rob Hough, tersrand gold fields: New perspectives on old sediments via SHRIMP:
Katy Evans, and Martin Hughes for their constructive comments Geological Society of America Abstracts with Programs, v. 25, no. 6,
on an early draft of this manuscript. Holly Stein and Judy Han- p. 278–279.
Els, B.G., Van den Berg, W.A., and Mayer, J.J., 1995, The Black Reef Quartz-
nah are thanked for discussions on Re-Os and their early review ite Formation in the Western Transvaal: Sedimentological and economic
of this section of the manuscript. We are grateful to Steve Kesler aspects, and significance for basin evolution: Mineralium Deposita, v. 30,
and two anonymous reviewers for insightful and constructive p. 112–123, doi: 10.1007/BF00189340.
England, G., 1999, Constraints on the origin of the Witwatersrand Au-U
reviews that greatly improved the paper. JL acknowledges an deposit, South Africa: Applying new techniques to an old problem, in
Honorary Research Fellowship at the School of Geosciences, Dunphy, J.M., and Hagemann, S.G., eds., Seminar on recent advances in
Monash University. NP acknowledges the support of the School ore genesis, December 1999: Perth, Centre for Strategic Mineral Depos-
its, University of Western Australia, 3 p.
of Geosciences, Monash University, and the School of Earth England, G.E., Rasmussen, B., Krapez, B., and Groves, D.I., 2001, The origin
Sciences, University of Melbourne. of uraninite, bitumen nodules, and carbon seams in Witwatersrand gold-
uranium-pyrite ore deposits, based on a Permo-Triassic analogue: Eco-
nomic Geology and the Bulletin of the Society of Economic Geologists,
REFERENCES CITED v. 96, p. 1907–1920.
Farquhar, J., Bao, H., and Thiemens, M., 2000, Atmospheric influence of
Anand, R.R., 1995, Genesis and classification of ferruginous regolith materials Earth’s earliest sulfur cycle: Science, v. 289, p. 756–758, doi: 10.1126/
in the Yilgarn Craton: Implications for mineral exploration: Townsville, science.289.5480.756.
Australia, Economic Geology Research Unit, James Cook University, Fleet, M.E., MacLean, P.J., and Barbier, J., 1989, Oscillatory-zoned As-bearing
Contribution 54, p. 1–4. pyrite from strata-bound and stratiform gold deposits: An indicator of ore
Armstrong, R.A., Compston, W., Retief, E.A., Williams, I.S., and Welke, H.J., fluid evolution: Economic Geology Monograph, v. 6, p. 356–362.
1991, Zircon ion microprobe studies bearing on the age and evolution of Frimmel, H.E., 1997, Detrital origin of hydrothermal Witwatersrand gold—A
the Witwatersrand triad: Precambrian Research, v. 53, p. 243–266, doi: review: Terra Nova, v. 9, p. 192–197, doi: 10.1046/j.1365-3121.1997.
10.1016/0301-9268(91)90074-K. d01-23.x.
Barnicoat, A.C., Henderson, I.H.C., Knipe, R.J., Yardley, B.W.D., Napier, Frimmel, H.E., and Gartz, V.H., 1997, Witwatersrand gold particle chemistry
R.W., Fox, N.P.C., Kenyon, A.K., Muntingh, D.J., Strydom, D., Winkler, matches model of metamorphosed, hydrothermally altered placer deposits:
K.S., Lawrence, S.R., and Cornford, C., 1997, Hydrothermal gold min- Mineralium Deposita, v. 32, p. 523–530, doi: 10.1007/s001260050119.
eralization in the Witwatersrand Basin: Nature, v. 386, p. 820–824, doi: Frimmel, H.E., Le Roux, A.P., Knight, J., and Minter, W.E.L., 1993, A case
10.1038/386820a0. study of the postdepositional alteration of the Witwatersrand Basal reef
Barnicoat, A.C., Phillips, G.M., Law, J.D.M., Walshe, J.L., Phillips, G.N., and gold placer: Economic Geology and the Bulletin of the Society of Eco-
Fox, N.P.C., 2001, Refuting the irrefutable: A new look at a well-known nomic Geologists, v. 88, p. 249–265.
sample of Witwatersrand gold mineralisation: Townsville, Australia, Eco- Fuller, A.O., 1958, A contribution to the petrology of the Witwatersrand System:
nomic Geology Research Unit, James Cook University, Contribution 59, Transactions of the Geological Society of South Africa, v. 61, p. 10–50.
p. 16–17. Gray, G.J., Lawrence, S.R., Kenyon, A.K., and Cornford, C., 1998, Nature and
Bossière, G., Bonkoungou, I., Peucat, J.-J., and Pupin, J.-J., 1996, Origin and origin of ‘carbon’ in the Archaean Witwatersrand Basin, South Africa:
age of the Paleoproterozoic conglomerates and sandstones of the Tark- Journal of the Geological Society [London], v. 155, p. 39–59.
waian Group in Burkino Faso: West Africa: Precambrian Research, v. 80, Hallbauer, D.K., 1986, The mineralogy and geochemistry of Witwatersrand
p. 153–172, doi: 10.1016/S0301-9268(96)00014-9. pyrite, gold and uranium, and carbonaceous matter, in Anhaeusser, C.R.,
and Maske, S., eds, Mineral deposits of Southern Africa: Johannesburg, Maynard, J.B., Ritger, S.D., and Sutton, S.J., 1991, Chemistry of sands from the
Geological Society of South Africa, p. 731–752. modern Indus River and the Archean Witwatersrand basin: Implications
Hirdes, W., and Saager, R., 1983, The Proterozoic Kimberley reef placer in the for the composition of the Archean atmosphere: Geology, v. 19, p. 265–
Evander gold field, Witwatersrand, South Africa: Gebruder Borntrager, 268, doi: 10.1130/0091-7613(1991)019<0265:COSFTM>2.3.CO;2.
Monograph Series on Mineral Deposits, Berlin, no. 20, 101 p. McCarthy, T.S., 1994, The tectono-sedimentary evolution of the Witwatersrand
Hoefs, J., Nielsen, H., and Schidlowski, M., 1968, Sulfur isotope abundances in Basin with special reference to its influence on the occurrence and char-
pyrite from the Witwatersrand conglomerates: Economic Geology and the acter of the Ventersdorp Contact Reef—A review: South African Journal
Bulletin of the Society of Economic Geologists, v. 63, p. 975–977. of Geology, v. 97, no. 3, p. 247–259.
Holland, H.D., 1984, Chemical evolution of the atmosphere and oceans: Princ- McCready, A.J., and Parnell, J., 1998, A Phanerozoic analogue for Witwa-
eton, New Jersey, Princeton University Press, 582 p. tersrand-type uranium mineralization: Uranium-titanium-bitumen nod-
Holland, H.D., 1994, Early Proterozoic atmospheric change, in Bengtson, S., ules in Devonian conglomerate/sandstone, Orkney, Scotland: Transac-
ed., Early life on Earth: Nobel Symposium 84: New York, Columbia Uni- tions of the Institute of Mining and Metallurgy, v. 107, p. B89–B96.
versity Press, p. 237–244. Meyer, F.M., Tainton, S., and Saager, R., 1990, The mineralogy and geochem-
Holland, H.D., and Rye, R.O., 1997, Evidence in pre-2.2 Ga paleosols istry of small-pebble conglomerates from the Promise Formation in the
for the early evolution of atmospheric oxygen and terrestrial biota: West Rand and Klerksdorp areas: South African Journal of Geology,
Comment and reply: Geology, v. 25, p. 857–859, doi: 10.1130/0091- v. 93, p. 118–134.
7613(1997)025<0857:EIPGPF>2.3.CO;2. Milesi, J.P., Ledru, P., Marcoux, E., Johan, V., Lerouge, C., Sabate, P., Bailly,
Holland, H.D., Kuo, P.H., and Rye, R.O., 1994, O2 and CO2 in the Late L., Respaut, J.P., and Skipwith, P., 2002, The Jacobina Paleoproterozoic
Archaean and Early Proterozoic atmosphere: Mineralogical Magazine, gold-bearing conglomerates, Bahia, Brazil: A ‘hydrothermal shear-res-
v. 58A, p. 424. ervoir’ model: Ore Geology Reviews, v. 19, p. 95–136, doi: 10.1016/
Hutchinson, R.W., and Viljoen, R.P., 1988, Re-evaluation of the gold source S0169-1368(01)00038-5.
in Witwatersrand ores: South African Journal of Geology, v. 91, p. 157– Minter, W.E.L., 1976, Detrital gold, uranium, and pyrite concentrations related
173. to sedimentology in the Precambrian Vaal Reef placer, Witwatersrand,
Jahn, B.M., Bertrand-Satarti, J., Morin, N., and Mace, J., 1990, Direct dating South Africa: Economic Geology and the Bulletin of the Society of Eco-
of stromatolitic carbonates from the Schmidtsdrift Formation (Transvaal nomic Geologists, v. 71, p. 157–176.
Dolomite), South Africa, with implications for the age of the Venters- Minter, W.E.L., 1978, A sedimentological synthesis of placer gold, uranium and
dorp Supergroup: Geology, v. 18, p. 1211–1214, doi: 10.1130/0091- pyrite concentrations in Proterozoic Witwatersrand sediments, in Miall,
7613(1990)018<1211:DDOSCF>2.3.CO;2. A.D., ed., Fluvial sedimentology: Canadian Society of Petroleum Geol-
Jolley, S.J., Henderson, I.H.C., Barnicoat, A.C., and Fox, N.P.C., 1999, Thrust- ogy Memoir 5, p. 801–829.
fracture network and hydrothermal gold mineralization: Witwatersrand Minter, W.E.L., 1999, Irrefutable detrital origin of the Witwatersrand gold and
Basin, South Africa, in McCaffrey, K.J.W., Longergan, L., and Wilkin- evidence of eolian signatures: Economic Geology and the Bulletin of the
son, J.J., eds, Fractures, fluid flow and mineralization: Geological Society Society of Economic Geologists, v. 94, p. 665–670.
[London] Special Publication 155, p. 153–165. Minter, W.E.L., Goedhart, M., Knight, J., and Frimmel, H.E., 1993, Morphol-
Jolley, S.J., Freeman, S.R., Barnicoat, A.C., Phillips, G.M., Knipe, R.J., Pather, ogy of Witwatersrand gold grains from the Basal Reef: Evidence for their
A., Fox, N.P.C., Strydom, D., Birch, M.T.G., Henderson, I.H.C., and detrital origin: Economic Geology and the Bulletin of the Society of Eco-
Rowland, T.W., 2004, Structural controls on Witwatersrand gold mineral- nomic Geologists, v. 88, p. 237–248.
ization: Journal of Structural Geology, v. 26, p. 1067–1086, doi: 10.1016/ Morrison, G.W., Rose, W.J., and Jaireth, S., 1991, Geological and geo-
j.jsg.2003.11.011. chemical controls on the silver content (fi neness) of gold in gold-silver
Kasting, J.F., 1993, Earth’s early atmosphere: Science, v. 259, p. 920–926. deposits: Ore Geology Reviews, v. 6, p. 333–364, doi: 10.1016/0169-
Kasting, J.F., 2001, The rise of atmospheric oxygen: Science, v. 293, p. 819– 1368(91)90009-V.
820, doi: 10.1126/science.1063811. Morton, A.C., and Hallsworth, C.R., 1999, Processes controlling the composi-
Kirk, J., Ruiz, J., Chesley, J., Titley, S., and Walshe, J., 2001, A detrital model tion of heavy mineral assemblages in sandstones: Sedimentary Geology,
for the origin of gold and sulfides in the Witwatersrand Basin based on v. 124, p. 3–29, doi: 10.1016/S0037-0738(98)00118-3.
Re-Os isotopes: Geochimica et Cosmochimica Acta, v. 65, p. 2149–2159, Myers, R.E., Zhou, T., and Phillips, G.N., 1993, Sulphidation in the Witwa-
doi: 10.1016/S0016-7037(01)00588-9. tersrand goldfields: Evidence from the Middelvlei Reef: Mineralogical
Kirk, J., Ruiz, J., Chesley, J., Walshe, J., and England, G., 2002, A major Magazine, v. 57, p. 395–405.
Archean, gold- and crust-forming event in the Kaapvaal Craton: Science, Oberthür, T., and Saager, R., 1986, Delineation of different mineral facies in the
v. 297, p. 1856–1858, doi: 10.1126/science.1075270. Carbon Leader Reef Placer, Carletonville Goldfield, Witwatersrand, and
Klemd, R., 1999, A comparison of fluids causing post-depositional hydrothermal their relation to sedimentology and gold distribution: Geological Society
alteration in Archaean basement granitoids and the Witwatersrand Basin: of South Africa, 21st Congress, Extended abstracts, p. 161–165.
Mineralogy and Petrology, v. 66, p. 111–122, doi: 10.1007/BF01161724. Ohmoto, H., 1996, Evidence in pre-2.2 Ga paleosols for the early evolution of
Krupp, R.K., Oberthür, T., and Hirdes, W., 1994, The early Precambrian atmospheric oxygen and terrestrial biota: Geology, v. 24, p. 1135–1138,
atmosphere and hydrosphere: Thermodynamic constraints from mineral doi: 10.1130/0091-7613(1996)024<1135:EIPGPF>2.3.CO;2.
deposits: Economic Geology and the Bulletin of the Society of Economic Palmer, J.A., 1986, Palaeoweathering in the Witwatersrand and Ventersdorp
Geologists, v. 89, p. 1581–1598. Supergroups [M.Sc. thesis]: Johannesburg, University of the Witwa-
Law, J.D.M., 1991, Alteration and geochemistry of Witwatersrand metasedi- tersrand, 166 p.
ments in the Welkom area: Implications for mineralization [M.Sc. thesis]: Palmer, J.A., Phillips, G.N., and McCarthy, T.S., 1987, The nature of the Pre-
Johannesburg, University of Witwatersrand, 198 p. cambrian atmosphere and its relevance to Archaean gold mineralization,
Law, J.D.M., and Spencer, R.M., 1992, A detailed study of the Kimberley in Ho, S.E., and Groves, D.I., eds, Recent advances in understanding
Reef on portions of Kinross and Winkelhaak Mines: Johannesburg, Precambrian gold deposits: Perth, Geology Department and Extension,
Gencor, Mineral Resources Division, Unpublished Geological Research University of Western Australia, Publication 11, p. 327–340.
Report, 94 p. Palmer, J.A., Phillips, G.N., and McCarthy, T.S., 1989, Paleosols and their
Law, J.D.M., Bailey, A.C., Cadle, A.B., Phillips, G.N., and Stanistreet, I.G., relevance to Precambrian atmospheric composition: Journal of Geology,
1990, Reconstructive approach to the classification of Witwatersrand v. 97, p. 77–92.
‘quartzites’: South African Journal of Geology, v. 93, p. 83–92. Parnell, J., 1996, Phanerozoic analogues for carbonaceous matter in Witwa-
Liebenberg, W.R., 1955, The occurrence and origin of gold and radioactive tersrand ore deposits: Economic Geology and the Bulletin of the Society
minerals in the Witwatersrand System, the Dominion Reef, the Venters- of Economic Geologists, v. 91, p. 55–62.
dorp Contact Reef, and the Black Reef: Transactions of the Geological Phillipe, S., Lancelot, J.R., Clauer, N., and Pacquet, A., 1993, Formation and
Society of South Africa, v. 58, p. 101–227. evolution of the Cigar Lake uranium deposit based on U-Pb and K-Ar iso-
MacLean, P.J., and Fleet, M.E., 1989, Detrital pyrite in the Witwatersrand gold tope systematics: Canadian Journal of Earth Sciences, v. 30, p. 720–730.
fields of South Africa: Evidence from truncated growth banding: Eco- Phillips, G.N., 1988, Widespread fluid infiltration during metamorphism of
nomic Geology and the Bulletin of the Society of Economic Geologists, the Witwatersrand goldfields: Generation of chloritoid and pyrophyllite:
v. 84, p. 2008–2011. Journal of Metamorphic Geology, v. 6, p. 311–332.
Phillips, G.N., and Dong, G., 1993, Chert-plus-pyrite pebbles in Witwatersrand implications for the source of uranium in the Witwatersrand Basin: South
goldfields: Townsville, Australia, Economic Geology Research Unit, African Journal of Geology, v. 93, p. 5–40.
James Cooke University, Contribution 47, 15 p. Robinson, A., and Spooner, E.T.C., 1984, Can the Elliot Lake uraninite-bearing
Phillips, G.N., and Law, J.D.M., 1994, Metamorphism of the Witwa- quartz pebble conglomerates be used to place limits on the oxygen content
tersrand goldfields: A review: Ore Geology Reviews, v. 9, p. 1–31, doi: of the early Proterozoic atmosphere?: Journal of the Geological Society
10.1016/0169-1368(94)90017-5. [London], v. 141, p. 221–228.
Phillips, G.N., and Law, J.D.M., 1997, Hydrothermal origin for Witwatersrand Roscoe, S.M., and Minter, W.E.L., 1993, Pyritic paleoplacer gold and uranium
gold: Society of Economic Geologists Newsletter, v. 31, p. 26–33. deposits, in Kirkham, R.V., Sinclair, W.D., Thorpe, R.I. and Duke, J.M.,
Phillips, G.N., and Law, J.D.M., 2000, Witwatersrand gold fields: Geology, eds., Mineral deposit modelling: Geological Association of Canada Spe-
genesis and exploration: Society of Economic Geologists Reviews, v. 13, cial Paper 40, p. 103–124.
p. 439–500. Rundle, C.C., and Snelling, N.J., 1977, The geochronology of uraniferous min-
Phillips, G.N., and Myers, R.E., 1989, The Witwatersrand goldfields, II: An ori- erals in the Witwatersrand Triad; an interpretation of new and existing
gin for Witwatersrand gold during metamorphism and associated altera- U-Pd age data on rocks and minerals from the Dominion Reef, Witwa-
tion: Economic Geology Monograph 6, p. 598–608. tersrand and Ventersdorp Supergroups: Philosophical Transactions of the
Phillips, G.N., Law, J.D.M., and Myers, R.E., 1990, The role of fluids in the Royal Society of London, v. 286, p. 567–583.
evolution of the Witwatersrand Basin: South African Journal of Geology, Rye, R.O., and Holland, H.D., 1998, Paleosols and the evolution of atmospheric
v. 93, p. 54–69. oxygen: A critical review: American Journal of Science, v. 298, p. 621–
Phillips, G.N., Law, J.D.M., and Myers, R.E., 2001, Is the redox state of the 672.
Archaean atmosphere constrained?: Society of Economic Geologists Saager, R., 1968, Newly observed ore-minerals from the Basal Reef in the
Newsletter, v. 47, no. 1, p. 9–18. Orange Free State goldfield in South Africa: Economic Geology and the
Pretorius, D.A., 1976a, Gold in the Proterozoic sediments of South Africa: Sys- Bulletin of the Society of Economic Geologists, v. 63, p. 116–123.
tems, paradigms, and models, in Wolf, K.H., ed., Au, U, Fe, Mn, Hg, Sb, Saager, R., 1970, Structures in pyrite from the Basal Reef in the Orange Free
W and P deposits: Amsterdam, Elsevier, Handbook of Strata-Bound and State gold-field: Transactions of the Geology Society of South Africa,
Stratiform Ore Deposits, v. 7, p. 1–27. v. 73, p. 29–46.
Pretorius, D.A., 1976b, The nature of the Witwatersrand gold-uranium depos- Smith, N.D., and Minter, W.E.L., 1980, Sedimentological controls of gold and
its, in Wolf, K.H., ed., Au, U, Fe, Mn, Hg, Sb, W and P deposits: Amster- uranium in two Witwatersrand paleoplacers: Economic Geology and the
dam, Elsevier, Handbook of Strata-Bound and Stratiform Ore Deposits, Bulletin of the Society of Economic Geologists, v. 75, p. 1–14.
v. 7, p. 29–88. Smith, R.E., and Perdrix, J.L., 1982, Pisolitic laterite geochemistry for detecting
Pretorius, D.A., 1981, Gold and uranium in quartz-pebble conglomerates: Eco- massive sulphide deposits, in Geochemical exploration in deeply weath-
nomic Geology, 75th Anniversary Volume, p. 117–138. ered terrains: Floreat, Western Australia, CSIRO Report, p. 128–139.
Rasmussen, B., and Buick, R., 1999, Redox state of the Archean atmosphere: Smits, G., 1984, Some aspects of the uranium minerals in the Witwatersrand
Evidence from detrital heavy minerals in ca. 3250–2750 Ma sandstones sediments of the early Proterozoic: Precambrian Research, v. 25, p. 37–
from the Pilbara Craton, Australia: Geology, v. 27, p. 115–118, doi: 59, doi: 10.1016/0301-9268(84)90023-8.
10.1130/0091-7613(1999)027<0115:RSOTAA>2.3.CO;2. Sutton, S.J., Ritger, S.D., and Maynard, J.B., 1990, Stratigraphic control of
Ramdohr, P., 1958, New observations on the ores of the Witwatersrand in South chemistry and mineralogy in metamorphosed Witwatersrand quartzites:
Africa and their genetic significance: Transactions of the Geological Soci- Journal of Geology, v. 98, p. 329–341.
ety of South Africa, v. 61, Annexure, 50 p. Tankard, A.J., Jackson, M.P.A., Eriksson, K.A., Hobday, D.K., Hunter, D.R.,
Reid, A.M., le Roux, A.P., and Minter, W.E.L., 1986, Electron-microprobe and Minter, W.E.L., 1982, Crustal evolution of Southern Africa: New
analyses of Witwatersrand gold particles from the Vaal placer: Geocon- York, Springer-Verlag, 523 p.
gress ’86, Geological Society of South Africa, Johannesburg, Abstracts, Tucker, R.F., 1980, The sedimentology and mineralogy of the Composite Reef
p. 179–181. on Cooke Section, Randfontein Estates Gold Mine, Witwatersrand, South
Reimer, T.O., 1985, Volcanic rocks and weathering in the Early Proterozoic Africa [M.Sc. thesis]: Johannesburg, University of the Witwatersrand,
Witwatersrand Supergroup, South Africa: Geological Survey of Finland 355 p.
Bulletin, v. 331, p. 33–49. Utter, T., 1979, The morphology and silver content of gold from the Upper Wit-
Reimer, T.O., and Mossman, D.J., 1990, Sulfidization of Witwatersrand watersrand and Ventersdorp systems of the Klerksdorp gold field, South
black sands: From enigma to myth: Geology, v. 18, p. 426–429, doi: Africa: Economic Geology and the Bulletin of the Society of Economic
10.1130/0091-7613(1990)018<0426:SOWBSF>2.3.CO;2. Geologists, v. 74, p. 27–44.
Robb, L.J., and Meyer, F.M., 1995, The Witwatersrand Basin, South Africa: von Gehlen, K., 1983, Silver and mercury in single gold grains from the Wit-
Geological framework and mineralisation processes: Ore Geology watersrand and Barberton, South Africa: Mineralium Deposita, v. 18,
Reviews, v. 10, p. 67–94, doi: 10.1016/0169-1368(95)00011-9. p. 529–534, doi: 10.1007/BF00204496.
Robb, L.J., Davis, D.W., and Kamo, S.L., 1990a, U-Pb ages on single detrital Wall, V.J., Mason, R.N., and Hall, G.C., 2004, Hydrothermal alteration and
zircon grains from the Witwatersrand basin, South Africa: Constraints on gold mineralization in the Witwatersrand—When, where and why?: Geo-
the age of sedimentation and on the evolution of granites adjacent to the science Africa 2004, University of the Witwatersrand, Johannesburg,
basin: Journal of Geology, v. 98, p. 311–328. Abstracts, p. 685 - 686.
Robb, L.J., Meyer, F.M., Ferraz, M.F., and Drennan, G.R., 1990b, The distribu-
tion of radioelements in Archaean granites of the Kaapvaal Craton, with MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA


Memoir 198
2006
Evidence from sulfur isotope and trace elements in pyrites for

their multiple post-depositional processes in uranium ores
at the Stanleigh Mine, Elliot Lake, Ontario, Canada
Kosei E. Yamaguchi
Institute for Research on Earth Evolution (IFREE), Japan Agency for Marine-Earth Science and Technology (JAMSTEC),
2-15 Natsushima, Yokosuka, Kanagawa 237-0061, Japan, and NASA Astrobiology Institute
Hiroshi Ohmoto
The Pennsylvania State University, University Park, Pennsylvania 16802, USA
ABSTRACT
The ca. 2.45 Ga pyritic uraniferous quartz-pebble conglomerate (UQC) of the

Matinenda Formation of the Elliot Lake Group, Huronian Supergroup, was used in
this study to investigate the origin of pyrite. A laser-microprobe was used for analysis
of the sulfur isotopic compositions of individual pyrite grains, and an electron-probe
microanalyzer was used for analysis of the trace element compositions of pyrite grains
with overgrowth texture. We found a variation in δ34S values among pyrite crystals (73
analyses) of various size and morphologies that occur in a small (~1 cm3) rock chip:
the total range in δ34S is –9.0‰ to +5.5‰ with respect to CDT (Cañon Diablo Troilite)
with a mean value of +0.6‰ ±2.1‰ (1σ). The widest range of ~15‰ is found among
euhedral pyrite grains whereas variations of ~4‰ to ~6‰ are common in anhedral,
subhedral, and rounded grains of pyrite. These values are in marked contrast to the
δ34S values of pyrite from the Matinenda Formation that were obtained by previous
investigators using bulk-rock sulfur isotope analyses. We found variable concentra-
tions of Co (below detection to 4700 ppm), Ni (to 1900 ppm), and As (to 3400 ppm)
among individual pyrite crystals and within single grains with overgrowth textures.
These elemental concentrations are markedly different between core and overgrowth
parts of pyrite. We demonstrate that the pyrite grains in the Paleoproterozoic UQC
have been isotopically, chemically, and morphologically modified by post-depositional
processes, suggesting that the pyrite grains have undergone multiple generations. The
results of the present study cannot be explained solely by a detrital process. There-
fore, one cannot use the preserved morphology and chemistry of pyrite (and possibly
uraninite) to represent the original features at the time of deposition to support the
hypothesis of an anoxic atmosphere prior to 2.2 Ga.
Keywords: sulfur isotope, pyrite, laser ablation, trace element, Elliot Lake, and sul-
fate reduction.
*[email protected]
Yamaguchi, K.E., and Ohmoto, H., 2006, Evidence from sulfur isotope and trace elements in pyrites for their multiple post-depositional processes in uranium
ores at the Stanleigh Mine, Elliot Lake, Ontario, Canada, in Kesler, S.E., and Ohmoto, H., eds., Evolution of Early Earth’s Atmosphere, Hydrosphere, and Bio-
sphere—Constraints from Ore Deposits: Geological Society of America Memoir 198, p. 143–156, doi: 10.1130/2006.1198(08). For permission to copy, contact
[email protected]. ©2006 Geological Society of America. All rights reserved.
143
144 K.E. Yamaguchi and H. Ohmoto
INTRODUCTION 1981; Ruzicka, 1981, 1988, 1989; Robinson and Spooner, 1982,
1984a, 1984b; Mossman and Farrow, 1992; Frimmel et al., 1993;
Origin of Uraninite and Pyrite and the Rise of Robb and Meyer, 1991, 1995). Some researchers (e.g., Holland,
Atmospheric Oxygen 1984, 1994) have suggested that, after the rise of atmospheric
O2, groundwater-type U deposits began to form by the mixing
The rise of atmospheric O2 is of great importance; it directly of oxic groundwater containing dissolved U and reduced fluid
relates to the chemical evolution of the atmosphere and oceans, containing petroleum-like organic matter (OM) and/or reduced
and the origin and evolution of life on Earth. The current view of minerals such as pyrite. The oldest unequivocally groundwater-
the rise of atmospheric O2 on Earth ca. 2.2–2.0 Ga (e.g., Cloud, type U deposit that has been recognized so far is the 2.0 Ga Oklo
1968; Holland, 1984, 1994; Kasting, 1993; references therein) is U deposit in Gabon (e.g., Gauthier-Lafaye and Weber, 1989).
based on several lines of geological and geochemical observa- The key question that arises here is whether the mode of U min-
tions. These observations include the mode of occurrence and eralization is the same throughout geologic history or different
genesis of uranium ore (e.g., Roscoe, 1969; Dimroth and Kim- before and after 2.2–2.0 Ga.
berley, 1976; Minter, 1976; Phillips and Myers, 1981; Maynard
et al., 1991; Robb and Meyer, 1995; Barnicoat et al., 1997). How-
ever, a consensus among scientists on the timing and mechanism
for the rise of atmospheric oxygen has not been formed, resulting
in vigorous controversy.
Enrichment of uranium and pyrite in uraniferous quartz-
pebble conglomerate (UQC) has been found at many locations in
the world; Witwatersrand, South Africa, and Elliot Lake–Blind
River, Canada (Figs. 1, 2) are two of the best known. The occur-
rence of apparently detrital grains of uraninite and pyrite in UQC
that deposited before 2.2 Ga has been cited as evidence for a
reducing atmosphere, because those minerals are unstable in oxi-
dizing environments (e.g., Grandstaff, 1980). However, contro-
versy continues as to whether such uraninite and pyrite are detri-
tal (Minter, 1976), epigenetic / hydrothermal (Davidson, 1957;
Dimroth and Kimberley, 1976; Phillips et al., 1987; Phillips and
Myers, 1981; Barnicoat et al., 1997), or a hybrid of these two
types (modified placer: Holmes, 1957; Pienaar, 1963; Robert-
son, 1968, 1986; Roscoe, 1969, 1973, 1981; Theis, 1979; Myers,
Figure 1. Arial distribution of the Huronian Supergroup in east On- Figure 2. Generalized stratigraphy of the Huronian Supergroup. Modi-
tario. Modified after Bennett et al. (1991). fied after Young (1991).
Pyrite microanalyses for sulfur isotope and trace element compositions 145
A detrital origin for the uraninite and pyrite in UQC has been sists of volcanic rocks and clastic sedimentary rocks, and con-
based on observed textural features, such as grain size and shape tains economically significant Witwatersrand-type U deposits.
(Ramdohr, 1958, 1969), and related sedimentological criteria The three overlying groups, in ascending stratigraphic order, are
(e.g., Robertson and Steenland, 1960; Roscoe, 1969; Minter, the Hough Lake Group, Quirke Lake Group, and Cobalt Group.
1976; Theis, 1979), intergrain chemical heterogeneity (Grand- Each group has a sedimentary cycle of conglomerate, mudstone,
staff, 1974), and the high Th and rare earth element content of siltstone, and coarse arenite. The Huronian Supergroup contains
the uraninite. Hydraulic sorting, an important sedimentological the world’s oldest well-documented and widespread glaciogenic
criterion to identify the “detrital” hypothesis, depends on the size, rocks (Nesbitt and Young, 1982). Paleosols are preserved at the
shape, and specific gravity of the sediment grains/particles. When base of the Huronian succession. Studies of the Huronian Super-
applying such criteria, it is necessary to assume that the observed group have provided various lines of evidence interpreted as
characteristics of the present (preserved) state of the grains repre- reflecting an increase in oxygen content of Earth’s atmosphere
sent their original state at the time of deposition, or at least repre- during deposition of the Huronian succession, which occurred
sent a modified state that is reasonably traceable to their original ca. 2.5–2.2 Ga (e.g., Hattori et al., 1983a). The tectonic setting of
state. Otherwise, one would be just looking at grains that were the Huronian Supergroup has recently been interpreted in terms
modified in shape and size by post-depositional processes, which of a Wilson cycle of ocean opening and closing. Krough et al.
could lead to an erroneous conclusion for their origin. (1984) and Corfu and Andrews (1986) gave geochronological
constraints on the deposition of the Huronian Supergroup using
Preliminary Observations and Purpose of Study the U-Pb zircon method. They bracketed the deposition of the
Huronian Supergroup between 2450 +25/-10 Ma and 2219.4
During reconnaissance observation of approximately ten ±3.5 Ma.
samples from the pyritiferous UQC from the Matinenda Forma- The Matinenda Formation, the lowermost formation of the
tion of the Paleoproterozoic Huronian Supergroup, we noticed Elliot Lake Group, contains the world-famous pyritiferous UQC
that pyrite grains are not uniform in shape and size at the scale (Figs. 2–4). The Formation is up to 180 m thick in the Elliot Lake
of a thin section. Additionally, there are generally two textures of area (Robertson, 1981) and is mainly composed of cross-bedded
pyrite: porous and non-porous. These preliminary observations
suggest that pyrite grains in the UQC were modified by post-
depositional processes and thus could not be regarded as detri-
tal grains showing original features at the time of deposition.
Therefore, it is necessary to investigate each pyrite grain for its
origin. To detect microscale variation in chemical and isotopic
compositions among individual pyrite grains in the UQC, we uti-
lized micro-analytical methods: S isotope analysis by in situ laser
ablation and trace element analysis by electron-probe microana-
lyzer (EPMA). For trace element analyses, we focused on three
elements: As, Ni, and Co. Arsenic can substitute for S; Ni and
Co can substitute for Fe. Assuming the degree of substitution of
these three trace elements for FeS2 depends on the environments
of mineralization of pyrite, it is expected that pyrite of different
origins will show different trace element characteristics.
GEOLOGICAL SETTING
The geological setting of the Huronian Supergroup has been

described by Young (1991) and Bennett et al. (1991). The Paleo-
proterozoic Huronian Supergroup is exposed in an area ~340 km
E-W and ~300 km N-S, on the northern margin of Lake Huron as
part of the Canadian Shield (Fig. 1). The Huronian Supergroup,
which has been regarded as one of the best-exposed Paleopro-
terozoic successions in the world, consists of sedimentary and
subordinate volcanic rocks unconformably overlying the Archean
basement of the Superior Province.
Figure 3. (A) Scanned digital image of a UQC thin section. (B) X-ray
A generalized stratigraphic column for the Huronian Super- autoradiography; ten-day exposure. Radiogenic parts are indicated in
group is given in Figure 2. The supergroup is subdivided into black. Note close association of sulfide with U-bearing minerals in the
four stratigraphic groups. The Elliot Lake Group, the oldest, con- matrix of the sample.
arkosic sandstones, poorly sorted quartz-pebble conglomerate,

with uraniferous parts near the base (Young, 1991). The arkosic
sandstones consist of poorly sorted quartz and feldspar grains in
a sericite matrix. The Matinenda Formation has three units of
a ~3-m-thick ore horizon, which has several cycles of ~40 cm
conglomerate and sandstone beds separated by ~10-m-thick
quartzite beds. The conglomerate samples are oligomictic and
contain poorly sorted, well-rounded, and tightly packed quartz
pebbles, chert clasts, feldspar, sericite, U-bearing minerals, and
various morphologies of pyrite (Figs. 3, 4). Pyrite and U-bearing
minerals are concentrated in the matrix (Fig. 3), and the content
of pyrite, the commonest sulfide mineral in the UQC, is 6–10%
of the ore (Bennett et al., 1991). There is no major difference in
the mode of occurrence for these minerals among different units
of UQC horizons in the Matinenda Formation. The Matinenda
Formation’s detrital material was produced during intense weath-
ering in a warm humid environment (Nesbitt and Young, 1982),
and its UQC units were deposited in a braided fluvial environ-
ment that developed on Archean basement, with a dominantly
southeasterly paleocurrent direction (Fralick and Miall, 1989).
During the Penokean orogeny (ca. 1.7–1.9 Ga), rocks of the
southern Huronian Supergroup were subjected to low-pressure,
dynamothermal metamorphism, with a maximum grade equiva-
lent to the staurolite zone of the amphibolite facies (Bennett et al.,
1991; Hu et al., 1998). In the Sault Ste. Marie–Elliot Lake area,
the Huronian succession is relatively thin and the metamorphic
grade is low (sub-greenschist facies; Robertson, 1981; Farrow
and Mossman, 1988; Bennett et al., 1991).
SAMPLES
Samples of pyritiferous U-ore for this study were taken in

1995 at the Stanleigh Mine (permanently closed since 1996) in
the Elliot Lake district, Ontario, Canada. The samples belong to
the ca. 2.45 Ga Matinenda Formation (Krough et al., 1984) and
have suffered from minor burial metamorphism of sub-green-
schist facies and little or no modern weathering (Robertson,
1981; Farrow and Mossman, 1988; Bennett et al., 1991). There-
fore, they are suitable for isotope and trace element analysis.
Sulfide (mainly pyrite) and U-bearing minerals are in a matrix of
quartz-pebble conglomerate. Figure 3 shows a high-resolution
digital image of a sample together with its X-ray autoradiogra-
phy (radioactive parts appear in black). Various morphologies
of pyrite (i.e., pyrite with overgrowth texture, euhedral pyrite,
subhedral pyrite, anhedral pyrite, porous pyrite, non-porous
pyrite, fractured pyrite, and non-fractured pyrite; an estimate of
their relative abundance is given later) are observed and a few
Figure 4. Examples of the sample. (A) Photomicrograph showing frac- examples are shown in Figure 4. Pyrite commonly occurs inter-
tured pyrite, rounded pyrite with overgrowth texture, and euhedral py- grown with various other minerals. Classification of pyrite mor-
rite. Uranothorite is sandwiched between pyrite grains. (B) Backscat- phologies as shown above is based on shapes of pyrite grains in
tered electron (BSE) image of pyrite with overgrowth texture (zoom of a polished specimen of the examined UQC sample. Although
the pyrite grain in the upper left corner of A). (C) Photomicrograph of
porous (micro-fractured?) pyrite surrounded by non-porous pyrite. The rounded, non-porous, and non-fractured pyrite grains were not
thin section was dipped in dilute nitric acid for ~1 min to enhance the found in the examined specimen, the scheme of pyrite classifi-
mineral grain boundaries. Scale bars are different among the photos. cation used in this study is basically comparable to that used in
Saager (1981), in which pyrite grains in the UQC of the Witwa- as follows: the accelerating voltage was 25 kV; the beam current
tersrand Basin in South Africa were used. was 15 nA; Kα line was used for As, Ni, and Co; the normal
There is no major difference in the mode of occurrence of beam diameter was 2 μm; and the counting time was kept 120
pyrite (and uraninite) among different units of UQC layers in the seconds. The detection limit was 100–200 ppm for all elements
Matinenda Formation, and on that basis we have assumed that examined. Pyrite grains with overgrowth texture were chosen
the examined 1 cm3 sample is representative of the UQC units for the trace element study by EPMA to examine the differ-
of the Matinenda Formation. This assumption may be supported ences in their chemical composition between the outer over-
from the observation by Meyer et al. (1990), who use similar growth part and inner core part.
UQC samples in the Witwatersrand Supergroup in South Africa
(e.g., Pretorius, 1981), that there is no fundamental chemical dif- RESULTS
ference between pyrites from the various UQC horizons. How-
ever, there is a limitation that we do not know the exact depo- Sulfur Isotope Ratios
sitional setting of the examined 1 cm3 specimen in the braided
stream environment where quartz-pebble conglomerates were A large variation in δ34S values was found among individual
deposited. pyrite grains of various size (~0.1 to ~2.7 mm diameter) and
morphology (euhedral, subhedral, anhedral, or rounded) from
ANALYTICAL METHOD a single ~1 cm3 rock specimen (Fig. 5). The S isotope composi-
tions are summarized in Table 1. Sizes of pyrite crystals were
Laser Microprobe Analysis estimated on the basis of the longest and shortest diameter / edge
measured for each pyrite grain on the polished section (data not
Sulfur isotopic compositions were determined using a shown). We found no clear statistical correlation between the
stable isotope mass spectrometer (VG Prism II) with a Nd- measured crystal size and the δ34S values of pyrite examined.
YAG laser ablation system at the Pennsylvania State Univer- However, negative δ34S values were obtained from relatively
sity (Kakegawa, 1997). Notation for S isotopic values is with small (<200 μm; in many cases euhedral) pyrite crystals. The
respect to CDT (Cañon Diablo Troilite): δ34S = (Rsample/Rstandard δ34S values of 73 data points range from –9.0‰ to +5.5‰ with
– 1) × 1000, where R = 34S/32S. The precision and reproducibil- a mean δ34S value of +0.6‰. Hattori et al. (1983b) analyzed δ34S
ity is ±0.2‰, and laboratory standard SO2 gas, calibrated with values of pyrite in the Matinenda Formation using the conven-
an international standard, was used for data acquisition. The tional bulk-sample method and reported a range from –1.5‰
experimental scheme and conditions are described elsewhere in to +1.2‰ with a mean of +0.2‰ for 75 data points. The range
detail (Crowe et al., 1990; Kelley and Fallick, 1990; Kakegawa, observed in the present study (14.5‰) is much greater than
1997) and therefore are only briefly described here. The pol- that for previously published data (2.7‰) reported by Hattori
ished 1 cm3 sample was placed in a sample chamber and viewed et al. (1983b), whereas the mean values are very similar (only
through a CCD camera. A Nd-YAG laser, whose beam radius 0.4‰ difference). Comparison of histograms for δ34S values of
is ~50 μm, vaporized a pyrite grain, and in some cases its sur- pyrite in the UQC of the Matinenda Formation between this
rounding material (organic matter) in O2 atmosphere, resulting study and those of Hattori et al. (1983b, 1986) is not readily pos-
in a crater with a radius ≤200 μm. The generated gas mixture sible, because Hattori et al. (1983b) presented only the range,
(SO2, CO2, H2O, various hydrocarbons, and excess O2) was then mean, and number of data points they obtained and Hattori et
sent to a graphite-furnace operating at ~950 °C to oxidize com- al. (1986) (see Table 2 of Hattori et al., 1986) presented 19 data
pletely. After cryogenic removal of impurities, SO2 was trans- points instead of 75 data points used in Hattori et al. (1983b).
ferred to the mass spectrometer for isotope ratio measurement. Conventional bulk-rock S isotope analysis requires relatively
Correction of isotope fractionation (1‰) induced during laser large number of samples, and inevitably homogenizes the
ablation was made according to Kakegawa (1997). To moni- potentially variable S isotope compositions of different pyrite
tor the instrumental fractionation of S isotopes, the standard grains. Microanalytical methods utilizing laser and ion beams
SO2 gas was measured frequently for its S isotope composition often reveal true variations in chemical and isotopic compo-
between samples. sitions of geological materials (e.g., Eldridge et al., 1993). We
believe that the range of δ34S values determined by the laser
Electron Microprobe Analysis microprobe method in this study represents the true range of
δ34S values of pyrite in the sample.
Concentrations of As, Ni, and Co in pyrites were measured The S isotopic compositions of individual grains of pyrite
by electron-probe microanalyzer (EPMA; Cameca SX50) at the are separated according to their morphologies, and are shown
Material Research Institute of the Pennsylvania State Univer- in Figure 5B–E. Classification of pyrite morphologies are based
sity. The EPMA is equipped with four wavelength spectrom- on the cross sections of pyrite crystals that appeared on the
eters and one energy dispersive spectrometer. The operating polished surface of the examined rock specimen. Twenty-four
conditions for quantitative analysis of the above elements were pyrite grains are observed as euhedral. Euhedral pyrite has the
Figure 5. Histogram of S isotopic com-

positions of individual grains of pyrite in
the ~1 cm3 rock specimen of the Paleopro-
terozoic UQC, Matinenda Formation, El-
liot Lake Group, Ontario, Canada. Sulfur
isotopic compositions are expressed by
delta notation from the standard CDT. To-
tal number of analyses is 73. Overall dis-
tribution shows a wide range from –9.0‰
to +5.5‰, with a mean value of +0.6 ± 2.1
(1σ)‰. (A) Total analysis. (B) Euhedral
pyrite. (C) Subhedral pyrite. (D) Anhedral
pyrite. (E) Rounded pyrite. (F) Younger
pyrite: combination of (B) and (C). (G)
Older pyrite: combination of (D) and (E).
Hatched column indicates the range in S
isotopic compositions of previous bulk
measurements by Hattori et al. (1983a,
1983b). Note the (previously undetected)
larger spread of S isotopic compositions
revealed by the microanalytical method
those by conventional bulk method.
widest variation in δ34S value, ranging from –9.0‰ to +5.5‰ et al. (1983b) (i.e., –1.5‰ to +1.2‰) is comparable (n = 11) to the
with a mean of +0.2‰. The most positive and negative δ34S val- number that fall within that range (n = 13).
ues of total data points are from euhedral pyrite. The mean δ34S Euhedral pyrite and subhedral pyrite grains are combined
value for euhedral pyrite is less than the mean of the total data into one group termed “younger pyrite,” whereas anhedral
points. Thirteen subhedral pyrite grains show a range in δ34S pyrite is considered “older pyrite.” The histograms of their
value from –3.7‰ to +2.8‰ with a mean of +0.3‰. The mean δ34S values are shown in Figure 5F and 5G. This grouping is
value is similar to that of euhedral pyrite. We use the terms based on interpretation of the δ34S data and trace element data
“anhedral” for pyrite that we cannot group into euhedral, sub- of pyrite grains, which is discussed later in detail.
hedral, or rounded pyrite. The range in δ34S values of anhedral Ratios of the numbers of each type of pyrite grain were
pyrite grains is –3.1‰ to +2.5‰ and the mean value is +0.8‰. used as an approximate estimate of the relative abundance of
This group of pyrite is dominant (data points n = 33) in the each type of pyrite grain. According to this rough estimate,
examined sample of pyritiferous UQC. We found only three euhedral pyrite is ~33%, subhedral pyrite is ~18%, anhedral
grains of rounded pyrite in the examined sample. Their δ34S pyrite is ~45%, and rounded pyrite is ~4%. The older pyrite and
values are +0.2‰, +1.8‰, and +4.0‰, with a mean value of the younger pyrite are found to be nearly equal in abundance
+2.0‰. It should be noted that the mean and median δ34S values (49% and 51%, respectively). From microscopic observation,
for each group of pyrite grains (see Table 2) are all different such estimates for the relative abundance of different types of
from one another, and that the number of euhedral pyrite grains pyrite grains in the examined sample are found to be not much
(n = 24 in total) that fall outside the range observed by Hattori different from those in the other samples obtained from differ-
TABLE 1. SUMMARY OF SULFUR ISOTOPE ANALYSIS

BY LASER MICROPROBE METHOD
Analysis Morphology of G34S Analysis Morphology of G34S
number pyrite crystal (‰) number pyrite crystal (‰)
1 Euhedral +1.82 38 Euhedral +1.83
2 Anhedral +1.09 39 Anhedral –0.75
3 Anhedral +1.85 40 Euhedral +0.39
4 Euhedral –1.11 41 Euhedral +0.95
5 Anhedral +0.98 42 Anhedral +0.59
6 Anhedral +0.94 43 Euhedral –0.31
7 Euhedral +0.61 44 Subhedral –3.72
8 Euhedral –0.61 45 Anhedral +1.26
9 Subhedral –0.02 46 Euhedral +1.75
10 Subhedral +0.89 47 Euhedral +1.03
13 Anhedral –3.05 50 Euhedral +3.05
14 Anhedral +1.76 51 Euhedral +1.63
15 Euhedral +0.99 52 Anhedral +0.49
16 Euhedral +0.94 53 Euhedral –1.82
18 Anhedral 0.00 55 Anhedral +2.46
19 Euhedral –0.34 56 Subhedral –0.18
20 Rounded +0.24 57 Anhedral –1.80
22 Euhedral +0.51 59 Anhedral +1.24
24 Subhedral +0.66 61 Anhedral +1.13
26 Subhedral +2.32 63 Subhedral +0.45
27 Anhedral –0.78 64 Subhedral –0.15
29 Anhedral –3.09 66 Euhedral –8.98
30 Anhedral +0.19 67 rounded +1.78
31 Subhedral +0.29 68 Euhedral +2.53
32 Subhedral –0.80 69 Anhedral –0.11
33 Euhedral v2.13 70 Subhedral +2.59
34 Subhedral –1.64 71 Rounded +4.04
35 Anhedral +1.18 72 Euhedral –0.12
36 Subhedral +2.76 73 Anhedral +2.42
37 Euhedral +5.45
TABLE 2. STATISTICAL ANALYSIS (INCLUDING T-TEST) OF THE SULFUR ISOTOPE

COMPOSITIONS FOR EACH GROUP OF PYRITE TYPES
Sample group All Pyrite type Hattori et al.
Anhedral Euhedral Subhedral Rounded (1983b)
Number of data points (n) 73 33 24 13 3 75
Degree of freedom (d.f. = n-1) 72 32 23 12 2 74
Mean 0.59 0.84 0.24 0.27 2.02 0.2
Median 0.95 1.13 0.78 0.29 1.78 -
Maximum +5.45 +2.46 +5.45 +2.76 +4.04 +1.2
Minimum –8.98 –3.09 –8.98 –3.72 +0.24 –1.5
Variance 4.34 1.97 8.34 3.13 3.65 -
Standard deviation (1V) 2.08 1.40 2.89 1.77 1.91 -
7cal (D = 0.05) 5.48 3.24 4.35 3.05 - -
7table (D = 0.05) 1.99 2.04 2.07 2.18 - -
Note: t cal—t-value calculated from the obtained data sets; t table—t-value found in statistical
textbook.
ent units of UQC in the Matinenda Formation. Such similarity of samples) degrees of freedom (d.f.) is found from the t-test
supports the idea that the examined samples of this study may table to be 1.99 (calculated t-value is 5.48). For anhedral pyrite
be representative of UQC in the Matinenda Formation. grains, t0.05 with d.f. = 32 (n = 33) is 2.04 (calculated t-value is
Because of the apparent similarities between the shape of 3.24). For euhedral pyrite grains, t0.05 with d.f. = 23 (n = 24)
frequency histograms for the obtained data sets and statistical is 2.07 (calculated t-value is 4.35). And for subhedral pyrite
“normal” distribution we statistically analyzed the significance grains, t0.05 with d.f. = 12 (n = 13) is 2.18 (calculated t-value is
of variations in the obtained δ34S values for each group of pyrite 3.05). All the t-values found from the t-test table for all, anhe-
grains. We used the Student’s t-test, and the results of are sum- dral, euhedral, and subhedral pyrite grains (1.99, 2.04, 2.07, and
marized in Table 2. 2.18, respectively) are smaller than the corresponding calcu-
When all samples are considered, the upper 0.05 point (i.e., lated values (5.48, 3.24, 4.35, and 3.05, respectively). Therefore,
α = 0.05) of the t-distribution (t0.05) with 72 ( = n–1; n is number the null hypothesis (i.e., the histogram of the obtained data sets
can be approximated by the normal distribution) is rejected at pyrite, and 100–2300 ppm for core pyrite. Higher Co contents
the level of significance α = 0.05. We suggest that the claim are found in the overgrowth part. More than half of the data
“the histograms of the obtained data sets are statistically differ- points (eight points) for Co content of the overgrowth part are
ent from and not approximated by the normal distribution” is more than 1200 ppm and appear evenly distributed, whereas the
strongly supported by the data. remaining five data points are between 400 and 600 ppm. With
one exception, Co contents are <900 ppm for the core part, and
Trace Element Ratios the mode of these analyses is between 400 and 600 ppm.
In addition to the absolute content of trace elements, their
Trace element data are summarized in Figure 6 and Table As/Ni and Ni/Co ratios are useful for distinguishing between
3. Arsenic concentrations range from 200 to 3400 ppm for the overgrowth and core pyrite (Table 3). For example, the As/Ni
outer overgrowth part of pyrite grains, whereas the inner cores ratio of grain A is <1 for the core part, whereas the overgrowth
range from below detection limit (~100 ppm) to 1000 ppm and part has an As/Ni ratio >1 (grain A). With one exception, the
clearly contain less As. Nickel concentrations are <2000 ppm Ni/Co ratios of grain A also show a clear distinction between
for both types of pyrite. Distribution patterns look similar for the two parts (Ni/Co < 1 for the overgrowth and <1 for the core).
both types of pyrite and it is impossible to distinguish between As/Co ratios do not indicate a clear distinction between the two
the overgrowth part and core part based on their Ni contents. parts. These elemental ratios differ for grains B and C, where
Co concentrations range from 400 to 4700 ppm for overgrowth overgrowth pyrite has generally higher As/Ni ratios and lower
Ni/Co ratios than core pyrite, with some exceptions.
The concentrations of trace elements are different between
various morphologies of pyrite grains. Euhedral pyrite grains
A typically have higher As and Ni concentrations than subhedral,
anhedral, and rounded pyrite. Rounded pyrite grains have lower
As, Ni, and Co concentrations than the other types of pyrite (the
data are not presented in this paper).
DISCUSSION
Origins of S in Pyrite Crystals
To account for the observed variations in the δ34S values

and trace element contents of pyrite in UQC, one may attribute
them to detrital transport of pyrite grains from source rocks that
already had such variations. The source rocks are likely to have
been the Archean basement, granitic and basaltic rocks underly-
ing the Huronian Supergroup. The δ34S values of detrital pyrite,
B if any, are expected to be close to 0‰, because the parental pyrite
crystals in the source rocks typically have δ34S values close to
0‰ (Ohmoto and Goldhaber, 1997). Furthermore, extensive
physical breakdown and mixing of detrital pyrite grains, if any,
during transport in ancient rivers would probably have erased
any potential original variability in the δ34S values and trace ele-
ment contents of pyrite in the source rocks. Thus, variation in
the δ34S values of pyrite as large as 14.5‰—observed in this
study—is probably not caused by the inherited variation of those
in the source rocks.
The similarity of the overall average δ34S values of pyrite in
the UQC between individual analyses (+0.6‰; this study) and
bulk-rock analyses (+0.2‰; Hattori et al., 1983b) may suggest
that the majority of pyrite grains in the UQC are indeed detrital
in origin. However, simple detrital process of source-rock pyrite
Figure 6. Histogram of trace element (As, Ni, and Co) concentrations with δ34S values of 0‰ cannot explain the asymmetrical distri-
of pyrite grains with overgrowth texture in the Paleoproterozoic UQC,
Matinenda Formation, Elliot Lake Group, Ontario, Canada. Elemental bution of pyrite δ34S values (i.e., skewed toward heavier values)
concentration was measured by EPMA. (A) Overgrowth part of pyrite shown in Figure 5A. Furthermore, observation of (1) overgrowth
grains. (B) Inner core part of pyrite grains. textures for pyrite grains, (2) irregularly shaped anhedral pyrite,
TABLE 3. CONCENTRATION OF TRACE ELEMENTS (As, Ni, AND Co)

IN PYRITE GRAINS WITH OVERGROWTH TEXTURES
Sample ID Overgrowth Concentration (ppm) Wt. ratio (ppm/ppm)
or core As Ni Co As/Ni As/Co Ni/Co
Grain A
A-1 Overgrowth 1210 730 1230 1.7 1.0 0.6
A-2 Overgrowth 1080 400 1630 2.7 0.7 0.2
A-3 Overgrowth 3390 810 2790 4.2 1.2 0.3
A-4 Overgrowth 2670 930 4660 2.9 0.6 0.2
A-5 Overgrowth 1030 130 500 7.9 2.1 0.3
A-6 Overgrowth 2660 1930 2420 1.4 1.1 0.8
A-7 Overgrowth 780 540 2150 1.4 0.4 0.3
A-8 Overgrowth 540 310 490 1.7 1.1 0.6
A-9 Overgrowth 680 310 410 2.2 1.7 0.8
A-10 Core 340 1540 580 0.2 0.6 2.7
A-11 Core 680 1280 2290 0.5 0.3 0.6
A-12 Core 440 1330 890 0.3 0.5 1.5
A-13 Core 730 1650 490 0.4 1.5 3.4
A-14 Core 600 1190 830 0.5 0.7 1.4
A-15 Core 600 1820 390 0.3 1.5 4.7
Grain B
B-1 Overgrowth 2250 100 3030 22.5 0.7 0.0
B-2 Core 980 110 520 8.9 1.9 0.2
B-3 Core 340 180 360 1.9 0.9 0.5
Grain C
C-1 Overgrowth 180 330 3450 0.5 0.1 0.1
C-2 Overgrowth 1120 b.d. 550 - 2.0 -
C-3 Overgrowth 1980 160 520 12.4 3.8 0.3
C-4 Core 110 110 410 1.0 0.3 0.3
C-5 Core 910 120 140 7.6 6.5 0.9
C-6 Core b.d. b.d. 270 - - -
Note: b.d.—Below detection (<100 ppm).
and (3) variable grain sizes strongly supports the theory of pyrite ments (tens of permil; e.g., Kaplan and Rittenburg, 1964). This
formation after deposition of the host conglomerate. Similarly, discrepancy may be attributable to environmental conditions for
some authors suggest that textural evidence exists showing some SRB such as (1) elevated rate of BSR induced by generally warm
pyrite grains formed by replacement of rounded pebbles of cherts environments of early Earth (e.g., Ohmoto et al., 1993); (2) abun-
and shales during diagenesis and/or later hydrothermal events dance of readily available OM for SRB (e.g., Canfield, 2001);
(Myers et al., 1993; Phillips and Dong, 1993). These observations and/or (3) low sulfate concentration (e.g., Harrison and Thode,
are fully consistent with our suggestion that non-detrital pyrite is 1958; Boudreau and Westrich, 1984). Remnant OM (algal mat or
present in the UQC (Feather, 1981; Simpson and Bowles, 1981; pyrobitumen) has been found in the matrix of UQC in the Elliot
Roscoe, 1969, 1973, 1981). Lake area (e.g., Nagy and Mossman, 1988, 1992; Mossman et
Activity of sulfate-reducing bacteria (SRB) in anaerobic al., 1993; Nagy, 1993).
environments during transportation of sediments in rivers and/ Thermochemical sulfate reduction (TSR) caused by mixing
or early diagenetic stage of host sediments may be responsible of hydrothermal (reduced and petroleum-bearing) fluids derived
for the observed spread of δ34S values. During bacterial sulfate from a deeper part of the sedimentary basins and sulfate-bear-
reduction (BSR), the lighter S isotope 32S in sulfate is preferen- ing oxic groundwater from the surface could have been respon-
tially reduced using OM to form dissolved sulfide, which then sible for the observed spread in the δ34S values (Kiyosu, 1980;
reacts with reactive Fe to eventually precipitate pyrite (e.g., Ohmoto and Lasaga, 1982; Ohmoto and Goldhaber, 1997). Upon
Berner, 1984). The source of sulfate could have been oxidative reaction with OM or petroleum, which migrated as deep fluids
weathering of sulfide minerals in the source rock region, hydra- at elevated temperature, dissolved sulfate formed by dissolution
tion of volcanic SO2, or oxidation of volcanic H2S followed by of evaporite minerals or oxidation of sulfide minerals (or con-
hydration. The δ34S values of sulfate that formed by oxidation nate seawater) was thermochemically reduced to form pyrite.
of igneous sulfides would have been ~0‰ because of the small This process is virtually the same as for groundwater-type U
kinetic isotopic effect (ΔSO -FeS = δ34SFeS – δ34SSO ) (Fry et al., ores of younger ages (e.g., Gauthier-Lafaye and Weber, 1989).
4 2 2 4
1988; Ohmoto and Goldhaber, 1997). BSR in closed systems The petroleum may have been ubiquitous in Archean and Paleo-
with ΔSO -FeS of –9‰ may explain the result obtained in this study. proterozoic time (Dutkiewicz et al., 1998; Buick et al., 1998).
4 2
This magnitude is in good agreement with the proposed ΔSO -FeS A close association of pyrite with sericite may suggest that both
4 2
value of ~–15‰ to ~–5‰ ca. 2.5 Ga (Kakegawa et al., 1994, formed by high-temperature hydrothermal fluids, which may or
1998; Kakegawa, 1997). However, such ΔSO -FeS is rather small may not be related to the mild metamorphism that occurred dur-
4 2
when compared to that for typical SRB in the modern environ- ing the Penokean orogeny ca. 1.7 Ga (Roscoe, 1969; Willingham
et al., 1985; Bennett et al., 1991) in the Elliot Lake district or to Models for the Formation of Pyrite in the Uranium Ore at
impact-induced metamorphism that occurred during the Sudbury Elliot Lake
event ca. 1.8 Ga (e.g., Krough et al., 1984). Post-depositional flu-
ids passed through the Matinenda Formation during diagenesis Mechanical grinding during physical weathering and trans-
and sub-greenschist facies metamorphism at estimated tempera- port in rivers must have been responsible for cracking / round-
tures of 80–200 °C and 280–350 °C, respectively (Dutkiewicz ing pyrite grains under an O2-poor atmosphere (Fig. 7B) (e.g.,
et al., 2003). These temperature windows overlap with those Holland, 1984), where most pyrite and uraninite are resistant to
that are permissive of TSR (e.g., Ohmoto and Goldhaber, 1997; chemical weathering and therefore survive during weathering and
Machel, 2001). TSR may explain the formational mechanism of transportation in rivers (Fig. 7C1). Under an O2-rich atmosphere,
pyrite, especially pyrite with heavier δ34S values (Ohmoto and the pyrite grains would have been completely decomposed to
Goldhaber, 1997). form sulfate in solution and Fe-oxides, such as they are today
(Fig. 7C2). In a special environment where uplift, erosion, trans-
Origins of Ni, Co, and As in Pyrite Crystals port, and burial are rapid, pyrite may survive as detrital grains
even under an O2-rich atmosphere (Robertson, 1968, 1981, 1986;
Cobalt and nickel have been the most widely applied chemi- Maynard et al., 1991). However, considering the existing sedi-
cal indicators for the environment of pyrite formation (Fleischer, mentological data on the lower Huronian Supergroup, which
1955; Loftus-Hills and Solomon, 1967). Volcanogenic-hydro- favor rather low-energy fluvial environments (e.g., Fralick and
thermal pyrite is generally characterized by a high Co content Miall, 1989) and intense weathering (e.g., Maynard et al. 1991),
with a low Ni/Co (<1) ratio (Hawley and Nichol, 1961; Roscoe, detrital pyrite, if any, would not have survived if the sediments
1965; Kimberley et al., 1980). Nickel is more abundant in mag- were overlain by an O2-rich atmosphere.
matic pyrite (Hawley and Nichol, 1961) and in sedimentary (syn- In subsequent early diagenetic stages (Figs. 7D1, 7D2), dia-
genetic/diagenetic) pyrite than in volcanogenic-hydrothermal genetic pyrite could have formed by BSR as overgrowths on pre-
pyrite (Loftus-Hills and Solomon, 1967). Kimberley et al. (1980) existing pyrite grains or as newly formed euhedral pyrite. Fram-
found that Co is substantially more abundant than Ni in all UQC boidal pyrite may form and, through later aggregation, result in
samples. The Ni/Co ratios of Roscoe (1969) range from 0.3 to the formation of porous pyrite. Some of the overgrown parts of
1.0, whereas the Ni/Co ratios of Kimberley et al. (1980) range pyrite, small isolated non-porous (compact) euhedral/subhedral
from 0.0 to 0.2. These low Ni/Co ratios are comparable to those pyrite crystals, and minor pyrrhotite crystals could have formed
for overgrowth pyrite of the present study, suggesting a volcano- by TSR (Fig. 7F). Anhedral pyrite grains also could have formed
genic-hydrothermal origin. However, some core pyrite also has a by connecting pre-existing pyrite grains with newly precipitated
low Ni/Co ratio (<1), which is comparable to overgrowth pyrite. pyrite by a thermochemical process. Sulfur isotopic evidence sup-
Pyrite in the Archean Samreid Lake VMS (volcanogenic massive ports the above scenario; small isolated euhedral pyrite crystals,
sulfides) deposit near Elliot Lake displays Ni/Co ratios similar which presumably formed by TSR, tend to have δ34S values dif-
to those in the Paleoproterozoic UQC at Elliot Lake (Friedman, ferent from 0‰. Trace element characteristics of the overgrowth
1959). This suggests that pyrite in the UQC was derived (i.e., part of pyrite and of small isolated euhedral pyrite crystals are
detrital) from such Archean volcanic pyrite and/or that the pro- very similar in that As and Co are enriched. However, the core
cess of pyrite formation in such Archean volcanogenic-hydro- part of pyrite with overgrowth textures has relatively depleted
thermal pyrite and in the UQC were the same. concentrations of As and Co compared to the overgrowth part.
Compared with Co and Ni, As is less often used as a After the Penokean orogeny and/or Sudbury impact event,
chemical indicator of the environment of pyrite formation in the dissolution of pyrite grains may have occurred again (Fig.
the UQC in the Elliot Lake area. However, the As content of 7G). Dissolution features of the rim of the overgrown part of
pyrite in this study shows interesting characteristics. Arsenic pyrite suggest dissolution after the formation of pyrite over-
is relatively enriched in the overgrown part of pyrite (gener- growth (Figs. 7E, 7G). Some small isolated euhedral pyrite grains
ally >1000 ppm for overgrowth and <1000 ppm for core). Pre- also have this dissolution feature at their rim. This could probably
vious studies (Gendron et al., 1986; Sullivan and Aller, 1996) have been caused by the infiltration of O2-bearing groundwater
have proposed that in the early diagenesis of modern marine into the conglomerate unit.
sediments, As and Co were mobilized in association with the
(reductive) dissolution of Fe- and Mn-hydroxide in a reducing CONCLUSIONS AND IMPLICATIONS
environment. An enrichment of As in the overgrowth part of
pyrite grains can be explained by the involvement of such an The ca. 2.45 Ga pyritic UQC of the Matinenda Formation
As-bearing reducing fluid, possibly derived from a deeper part of the Elliot Lake Group, Huronian Supergroup, was used in this
of the sedimentary basin. Such process of fluid migration, origi- study to investigate the origin of pyrite. A laser-microprobe was
nally proposed to explain the observed characteristics of the S used for analysis of the S isotopic compositions of individual
isotopic compositions of pyrite in the UQC, may also explain the pyrite grains and EPMA was used to analyze the trace element
observed characteristics of their trace element compositions. compositions of pyrite grains with overgrowth texture. Together
Figure 7. A model for pyrite mineralization in the Paleoproterozoic UQC in the Elliot Lake district, Ontario, Canada. Sizes of the quartz pebbles
and pyrite grains are adjusted for the sake of clarity in the figure (in reality, pyrite grains are by far smaller than quartz pebbles). See text for
explanation of the stages.
with microscopic examinations of the samples, it is demonstrated depends on their morphological features such as shape and size.
that (1) various morphological types of pyrite (euhedral, subhe- However, such sedimentological observation has been based on
dral, anhedral, rounded, overgrowth) are present; (2) S isotopic the preserved state of the pyrite and uraninite in the UQC, which
compositions of the pyrite range from –9.0‰ to +5.5‰; (3) is very unlikely to retain the original characteristics at the time of
euhedral pyrite, which formed after the sedimentation of the host sedimentation of the host rocks. Our results cast doubt on the use
rocks, tends to have large negative or positive δ34S values com- of evidence of “detrital” pyrite (and uraninite by inference) in the
pared to other pyrites; and (4) trace element (As, Ni, and Co) UQC of the Matinenda Formation for a reducing atmosphere ca.
compositions of overgrowths and cores of pyrite grains are dif- 2.45 Ga. Such an interpretation requires a clear demonstration of
ferent. These results suggest that the preserved pyrite grains in morphological, isotopic, and chemical features for truly detrital
the Paleoproterozoic UQC samples had been isotopically, chemi- grains of uraninite and pyrite in the UQC samples that have not
cally, and morphologically overprinted by post-depositional pro- been altered by any later process.
cesses, such as early diagenesis and/or hydrothermal alteration.
Preservation of detrital minerals that are labile under oxida- ACKNOWLEDGMENTS
tive weathering conditions, such as pyrite and uraninite in UQC,
has been widely used by many researchers to argue for a reducing We appreciate G. Bennett for his field assistance, D. Walizer
atmosphere prior to 2.2 Ga. Recognition of such minerals mostly and M. Angelone for their technical assistance, and Y. Watanabe
and K. Hattori for discussion. Detailed and constructive reviews Fralick, P.W., and Miall, A.D., 1989, Sedimentology of the Lower Huronian
by M. Goldhaber, B. Maynard, and S. Kesler were very helpful Supergroup (Early Proterozoic), Elliot Lake area, Ontario, Can-
ada: Sedimentary Geology, v. 63, p. 127–153, doi: 10.1016/0037-
for improving the early manuscript, and therefore greatly appre- 0738(89)90075-4.
ciated. We thank the Stanleigh Mine for access to the under- Friedman, G.M., 1959, The Samreid Lake sulfide deposit, Ontario, an example
ground mine samples. This study was financially supported by of a pyrrhotite-pyrite iron formation: Economic Geology and the Bulletin
of the Society of Economic Geologists, v. 54, p. 265–284.
the NASA Astrobiology Institute (NCC2-1057), the NASA Exo- Frimmel, H.E., Le Roex, A.P., Knight, J., and Minter, W.E.L., 1993, A case
biology Program (NAG5-9089), the National Science Founda- study of the post-depositional alteration of the Witwatersrand Basal Reef
tion (EAR-9706279), and the Japanese Ministry of Education, gold placer: Economic Geology and the Bulletin of the Society of Eco-
nomic Geologists, v. 88, p. 249–265.
Culture, Sports, Science, and Technology (MEXT; #07041081) Fry, B., Ruf, W., Gest, H., and Hayes, J.M., 1988, Sulfur isotope effects associ-
to H.O. ated with oxidation of sulfides by O2 in aqueous solution: Chemical Geol-
ogy, v. 73, p. 205–210.
Gauthier-Lafaye, F., and Weber, F., 1989, The Francevillian (Lower Protero-
REFERENCES CITED zoic) uranium ore deposits of Gabon: Economic Geology and the Bulletin
Barnicoat, A.C., Henderson, I.H.C., Knipe, R.J., Yardley, B.W.D., Napler, Gendron, A., Silverberg, N., Sundby, B., and Lebel, J., 1986, Early diagen-
R.W., Fox, N.P.C., Kenyon, A.K., Muntingh, D.J., Strydom, D., Winskler, esis of cadmium and cobalt in sediments of the Laurentian Trough: Geo-
K.S., Lawrence, S.R., and Cornford, C., 1997, Hydrothermal gold min- chimica et Cosmochimica Acta, v. 50, p. 741–747, doi: 10.1016/0016-
eralization in the Witwatersrand Basin: Nature, v. 386, p. 820–824, doi: 7037(86)90350-9.
10.1038/386820a0. Grandstaff, D.E., 1974, Microprobe analyses of uranium and thorium in ura-
Bennett, G., Dressler, B.O., Robertson, J.A., 1991, The Huronian Supergroup ninite from the Witwatersrand, South Africa, and Blind River, Ontario,
and associated intrusive rocks: Geological Survey of Ontario Special Vol- Canada: Transactions of the Geological Society of South Africa, v. 77,
ume 4-I. p. 291–296.
Berner, R.A., 1984, Sedimentary pyrite formation: An update: Geochi- Grandstaff, D.E., 1980, Origin of uraniferous conglomerate at Elliot Lake,
mica et Cosmochimica Acta, v. 48, p. 605–615, doi: 10.1016/0016- Canada, and Witwatersrand, South Africa: Implications for oxygen in
7037(84)90089-9. the Precambrian atmosphere: Precambrian Research, v. 13, p. 1–26, doi:
Boudreau, B.P., and Westrich, J.T., 1984, The dependence of bacterial sul- 10.1016/0301-9268(80)90056-X.
fate reduction on sulfate concentration in marine sediments: Geochi- Harrison, A.G., and Thode, H.G., 1958, Mechanism of the bacterial reduction
mica et Cosmochimica Acta, v. 48, p. 2503–2516, doi: 10.1016/0016- of sulfate from isotope fractionation studies: Transactions of the Faraday
7037(84)90301-6. Society, v. 54, p. 84–92, doi: 10.1039/tf9585400084.
Buick, R., Rasmussen, B., and Krapez, B., 1998, Archean oil: Evidence for Hattori, K., Campbell, F.A., and Krouse, H.R., 1983a, Sulphur isotope abun-
extensive hydrocarbon generation and migration 2.5–3.6 Ga: American dances in Aphebian clastic rocks: Implications for the coeval atmosphere:
Association of Petroleum Geologists Bulletin, v. 82, p. 50–69. Nature, v. 302, p. 323–326, doi: 10.1038/302323a0.
Canfield, D.E., 2001, Isotope fractionation by natural populations of sulfate- Hattori, K., Krouse, H.R., and Campbell, F.A., 1983b, The start of sulfur oxi-
reducing bacteria: Geochimica et Cosmochimica Acta, v. 65, p. 1117– dation in continental environments: About 2.2 × 109 years ago: Science,
1124, doi: 10.1016/S0016-7037(00)00584-6. v. 221, p. 549–551.
Cloud, P., 1968, Atmospheric and hydrospheric evolution on the primitive Hattori, K., Campbell, F.A., and Krouse, H.R., 1986, Sulphur isotope abun-
Earth: Science, v. 160, p. 729–736. dances in sedimentary rocks; relevance to the evolution of the Precam-
Corfu, F., and Andrews, A., 1986, A U-Pb age for mineralized Nipissing dia- brian atmosphere: Geochemistry International, v. 22, p. 97–114.
base, Gowganda, Ontario: Canadian Journal of Earth Sciences, v. 23, Hawley, J.E., and Nichol, I., 1961, Trace elements in pyrite, pyrrhotite, and
p. 107–112. chalcopyrite of different ores: Economic Geology and the Bulletin of the
Crowe, D.E., Valley, J.W., and Baker, K.L., 1990, Micro-analysis of sulfur-iso- Society of Economic Geologists, v. 56, p. 467–487.
tope ratios and zonation by laser microprobe: Geochimica et Cosmochi- Holland, H.D., 1984, Chemical evolution of the atmosphere and ocean: Princ-
mica Acta, v. 54, p. 2075–2092, doi: 10.1016/0016-7037(90)90272-M. eton, New Jersey, Princeton University Press, 582 p.
Davidson, C.F., 1957, On the occurrence of uranium in ancient conglomerates: Holland, H.D., 1994, Proterozoic atmospheric change, in Bengtson, S., ed.,
Economic Geology and the Bulletin of the Society of Economic Geolo- Early life on Earth: Nobel Symposium 84: New York, Columbia Univer-
gists, v. 52, p. 668–693. sity Press, p. 237–244.
Dimroth, E., and Kimberley, M.M., 1976, Precambrian atmospheric oxygen: Holmes, S.W., 1957, Pronto Mine, in Gilbert, G., ed., Structural geology of
Evidence in the sedimentary distributions of carbon, sulfur, uranium, and Canadian ore deposits, Volume 2: Montreal, Canadian Institute of Mining
iron: Canadian Journal of Earth Sciences, v. 13, p. 1161–1185. and Metallurgy, p. 324–339.
Dutkiewicz, A., Rasmussen, B., and Buick, R., 1998, Oil preserved in fluid Hu, Q., Evensen, N.M., Smith, P.E., and York, D., 1998, A world in a grain of
inclusions in Archean sandstones: Nature, v. 395, p. 885–888, doi: sand; regional metamorphic history from 40Ar/39Ar laser probe analyses of
10.1038/27644. Proterozoic sediments from the Canadian Shield: Precambrian Research,
Dutkiewicz, A., Ridley, J., and Buick, R., 2003, Oil-bearing CO2-CH4-H2O fluid v. 91, p. 287–294, doi: 10.1016/S0301-9268(98)00054-0.
inclusions; oil survival since the Paleoproterozoic after high temperature Kakegawa, T., 1997, Sulfur isotope geochemistry of Archean sedimentary
entrapment: Chemical Geology, v. 194, p. 51–79, doi: 10.1016/S0009- rocks [Ph.D. dissertation]: University Park, Pennsylvania State Univer-
2541(02)00271-1. sity, 163 p.
Eldridge, C.S., Phillips, G.N., and Myers, R.E., 1993, Sulfides in the Witwa- Kakegawa, T., Kawai, H., and Ohmoto, H., 1998, Origins of pyrites in the ~2.5
tersrand Goldfields: New perspectives on gold sediments via SHRIMP: Ga Mt. McRae Shale, the Hamersley District, Western Australia: Geochi-
Geological Society of America Abstracts with Programs, v. 25, p. mica et Cosmochimica Acta, v. 62, p. 3205–3220, doi: 10.1016/S0016-
A278–279. 7037(98)00229-4.
Farrow, C.E., and Mossman, D.J., 1988, Geology of Precambrian paleosols at the Kakegawa, T., Ohmoto, H., Kasahara, Y., Kawai, H., and Hayashi, K., 1994,
base of the Huronian Supergroup, Elliot Lake, Ontario, Canada: Precam- Biogenic and hydrothermal pyrites in Archean shales from the Pilbara-
brian Research, v. 42, p. 107–139, doi: 10.1016/0301-9268(88)90013-7. Hamersley district, Western Australia: Geological Society of America,
Feather, C.E., 1981, Some aspects of Witwatersrand mineralization, with spe- Abstracts with Programs, v. 26, p. A353–A354.
cial reference to uranium minerals: U.S. Geological Survey Professional Kaplan, I.R., and Rittenburg, S.C., 1964, Microbiological fractionation of sul-
Paper 1161, p. Q1–Q23. phur isotopes: Journal of General Microbiology, v. 34, p. 195–212.
Fleischer, M., 1955, Minor elements in some sulfide minerals, in Bateman, Kasting, J.F., 1993, Earth’s early atmosphere: Science, v. 259, p. 920–926.
A.M., ed., Economic Geology, 50th Anniversary Volume, Part 2, p. Kelley, S.P., and Fallick, A.E., 1990, High precision spatially resolved analysis
970–1024. of δ34S in sulfides using a laser extraction technique: Geochimica et Cos-
mochimica Acta, v. 54, p. 883–888, doi: 10.1016/0016-7037(90)90381-T. Phillips, G.N., Myers, R.E., and Palmer, J.A., 1987, Problems with the placer
Kimberley, M.M., Tanaka, R.T., and Farr, M.R., 1980, Composition of mid- model for Witwatersrand gold: Geology, v. 15, p. 1027–1030, doi:
dle Precambrian uraniferous conglomerate in the Elliot Lake–Agnes 10.1130/0091-7613(1987)15<1027:PWTPMF>2.0.CO;2.
Lake area of Canada: Precambrian Research, v. 12, p. 375–392, doi: Pienaar, P.J., 1963, Stratigraphy, petrography, and genesis of the Elliot Lake
10.1016/0301-9268(80)90036-4. Group, Blind River, Ontario, including the uraniferous conglomerate:
Kiyosu, Y., 1980, Chemical reduction and sulfur-isotope effects of sulfate by Geological Survey of Canada Bulletin, v. 83, p. 140.
organic matter under hydrothermal conditions: Chemical Geology, v. 30, Pretorius, D.A., 1981, Gold and uranium in quartz-pebble conglomerate: Eco-
p. 47–56, doi: 10.1016/0009-2541(80)90115-1. nomic Geology, 75th Anniversary Volume, p. 117–138.
Krough, T.E., Davis, D.W., and Corfu, F., 1984, Precise U-Pb zircon and bad- Ramdohr, P., 1958, New observations on the ores of the Witwatersrand in South
deleyite ages for the Sudbury area, in Pye, E.G., Naldrett, A.J., and Gilbin, Africa and its genetic significance: Transactions of the Geological Society
P.E., eds. The geology and ore deposits of the Sudbury structure: Geologi- of South Africa, Annex, v. 61, p. 50.
cal Survey of Ontario Special Volume 1, p. 431–446. Ramdohr, P., 1969, The ore minerals and their intergrowths. Oxford, UK, Per-
Loftus-Hills, G., and Solomon, M., 1967, Cobalt, nickel and selenium in sul- gamon Press, 1174 p.
phides as indicators of ore genesis: Mineralium Deposita, v. 2, p. 228– Robb, L.J., and Meyer, F.M., 1991, A contribution to recent debate concern-
242, doi: 10.1007/BF00201918. ing epigenetic versus syngenetic mineralization processes in the Witwa-
Machel, H.G., 2001, Bacterial and thermochemical sulfate reduction in dia- tersrand Basin: Economic Geology and the Bulletin of the Society of Eco-
genetic settings—Old and new insights: Sedimentary Geology, v. 140, nomic Geologists, v. 86, p. 396–401.
p. 143–175, doi: 10.1016/S0037-0738(00)00176-7. Robb, L.J., and Meyer, F.M., 1995, The Witwatersrand Basin, South Africa:
Maynard, J.B., Rigger, S.D., and Sutton, S.J., 1991, Chemistry of sands from the Geological framework and mineralization processes: Ore Geology
modern Indus River and the Archean Witwatersrand basin: Implications Reviews, v. 10, p. 67–94, doi: 10.1016/0169-1368(95)00011-9.
for the composition of the Archean atmosphere: Geology, v. 19, p. 265– Robertson, D.S., and Steenland, N.C., 1960, On the Blind River uranium ores
268, doi: 10.1130/0091-7613(1991)019<0265:COSFTM>2.3.CO;2. and their origin: Economic Geology and the Bulletin of the Society of
Meyer, F.M., Robb, L.J., Oberthür, T., Saager, R., and Stupp, H.D., 1990, Economic Geologists, v. 55, p. 659–694.
Cobalt, nickel, and gold in pyrite from primary gold deposits and Witwa- Robertson, J.A., 1968, Geology of Township 149 and Township 150, District of
tersrand reefs: South African Journal of Geology, v. 93, p. 70–82. Algoma: Ontario Department of Mines, Geological Report, v. 57, p. 162.
Minter, W.E.L., 1976, Detrital gold, uranium, and pyrite concentrations related Robertson, J.A., 1981, The Blind River uranium deposits: The ores and their set-
to sedimentology in the Precambrian Vaal Reef Placer, Witwatersrand, ting: U.S. Geological Survey Professional Paper 1161-A-BB, p. U1–U23.
South Africa: Economic Geology and the Bulletin of the Society of Eco- Robertson, J.A., 1986, Huronian geology and the Blind River (Elliot Lake)
nomic Geologists, v. 71, p. 157–176. uranium deposits, 7–31, the Pronto Mine, in Evans, E.L., ed., Uranium
Mossman, D.J., and Farrow, C.E.G., 1992, Paleosol and ore-forming pro- deposits of Canada: Canadian Institute of Mining and Metallurgy Special
cesses in the Elliot Lake district of Canada, in Schidlowski, M., Golubic, Paper 33, p. 36–43.
S., Kimberley, M.M., McKirdy, D.M., and Trudinger, P.A., eds. Early Robinson, A., and Spooner, T.C., 1982, Source of the detrital components of
organic evolution: Implications for mineral and energy resources: Berlin, uraniferous conglomerates, Quirke ore zone, Elliot Lake Ontario, Canada:
Springer-Verlag, p. 67–75. Nature, v. 299, p. 622–624, doi: 10.1038/299622a0.
Mossman, D.J., Goodarzi, F., and Gentzis, T., 1993, Characterization of insol- Robinson, A., and Spooner, T.C., 1984a, Postdepositional modification of ura-
uble organic matter from the lower Proterozoic Huronian Supergroup, ninite-bearing quartz-pebble conglomerate from the Quirke ore zone,
Elliot Lake, Ontario: Precambrian Research, v. 61, p. 279–293, doi: Elliot Lake, Ontario, Canada: Economic Geology and the Bulletin of the
10.1016/0301-9268(93)90117-K. Society of Economic Geologists, v. 79, p. 297–321.
Myers, B., 1981, Genesis of uranium-gold pyritic conglomerates: U.S. Geo- Robinson, A., and Spooner, T.C., 1984b, Can the Elliot Lake uranium-bear-
logical Survey Special Paper 1161 p. AA1–AA26. ing quartz-pebble conglomerate be used to place limits on the oxygen
Myers, R.E., Zhou, T., and Phillips, G.N., 1993, Sulphidation in the Witwa- content of the early Proterozoic atmosphere?: Journal of Geology, v. 141,
tersrand Goldfields: Evidence from the Middelvlei Reef: Mineralogical p. 221–228.
Magazine, v. 57, p. 395–405. Roscoe, S.M., 1965, Geochemical and isotopic studies, Noranda and Matta-
Nagy, B., 1993, Kerogens and bitumens in Precambrian uraniferous ore depos- gami area: Canadian Institute of Mining and Metallurgy Bulletin, v. 58,
its: Witwatersrand, South Africa, Elliot Lake, Canada, and the natural fis- p. 965–971.
sion reactors, Oklo, Gabon, in Parnell, J., Kucha, H., and Landis, P., eds., Roscoe, S.M., 1969, Huronian rocks and uraniferous conglomerates: Geologi-
Bitumens in ore deposits: Berlin, Springer-Verlag, p. 286–333. cal Survey of Canada Special Paper 68-40, 205 p.
Nagy, B., and Mossman, D.J., 1988, The nature and origin of kerogens in the Roscoe, S.M., 1973, The Huronian Supergroup, a Paleoaphebian succession
lower Huronian Supergroup, Elliot Lake region, Ontario, Canada: Terra showing evidence of atmosphere evolution, in Young, G.M., ed., Huronian
Cognita, v. 8, p. 219. stratigraphy and sedimentation: Geological Association of Canada Spe-
Nagy, B., and Mossman, D.J., 1992, Stratiform and globular organic matter in cial Paper 12, p. 31–38.
the Lower Proterozoic metasediments at Elliot Lake, Ontario, Canada, Roscoe, S.M., 1981, Temporal and other factors affecting deposition of uranif-
in Schidlowski, M., Golubic, S., Kimberley, M.M., McKirdy, D.M., and erous conglomerate: U.S. Geological Survey Professional Paper 1161-A-
Trudinger, P.A., eds., Early organic evolution: Implications for mineral BB, p. W1–W17.
and energy resources: Berlin, Springer-Verlag, p. 224–231. Ruzicka, V., 1981, Some metallogenic features of the Huronian and post-
Nesbitt, H.W., and Young, G.M., 1982, Early Proterozoic climates and plate Huronian uraniferous conglomerates: U.S. Geological Survey Profes-
motions inferred from major element chemistry of lutites: Nature, v. 299, sional Paper 1161-A-BB, p. V1–V8.
p. 715–717, doi: 10.1038/299715a0. Ruzicka, V., 1988, Geology and genesis of uranium deposits in the early Pro-
Ohmoto, H., and Lasaga, A.C., 1982, Kinetics of reactions between aqueous terozoic: Blind River-Elliot Lake basin, Ontario, Canada, in Recognition
sulfates and sulfides in hydrothermal systems: Geochimica et Cosmochi- of uranium deposits: Proceedings of a Technical Committee Meeting,
mica Acta, v. 46, p. 1727–1745, doi: 10.1016/0016-7037(82)90113-2. London, 18–20 September 1985: Vienna, Austria, International Atomic
Ohmoto, H., Kakegawa, T., and Lowe, D.R., 1993, 3.4-billion-year-old bio- Energy Agency, p. 107–130.
genic pyrites from Barberton, South Africa: Sulfur isotope evidence: Sci- Ruzicka, V., 1989, Conceptual genetic models for important types of uranium
ence, v. 262, p. 555–557. deposits and areas favorable for their occurrences in Canada, in Uranium
Ohmoto, H., and Goldhaber, M.B., 1997, Sulfur and carbon isotopes, in Barnes, resources and geology of North America: Vienna, Austria, International
H.L., ed., Geochemistry of hydrothermal ore deposits (3rd edition): New Atomic Energy Agency, TECDOX 500, p. 49–79.
York, John Wiley and Sons, p. 517–611. Saager, R., 1981, Geochemical studies on the origin of the detrital pyrites in
Phillips, G.N., and Dong, G., 1993, Chert-plus-pyrite pebbles in the Witwa- the conglomerates of the Witwatersrand Goldfields, South Africa: U.S.
tersrand Goldfield: International Geology Review, v. 36, p. 65–71. Geological Survey Professional Paper 1161, p. L1–L17.
Phillips, G.N., and Myers, R.E., 1981, The Witwatersrand gold field, Part II: Simpson, P.R., and Bowles, J.F.W., 1981, Detrital uraninite and pyrite: Are they
An origin for Witwatersrand gold during metamorphism and associated evidence for a reducing atmosphere?: U.S. Geological Survey Profes-
alteration: Economic Geology Monograph 6, p. 311–332. sional Paper 1161, p. S1–S12.
Sullivan, K.A., and Aller, R.C., 1996, Diagenetic cycling of arsenic in Amazon Young, G.M., 1991, Stratigraphy, sedimentology and tectonic setting of the
shelf sediments: Geochimica et Cosmochimica Acta, v. 60, p. 1465–1477, Huronian Supergroup: Geological Association of Canada, Mineralogical
doi: 10.1016/0016-7037(96)00040-3. Association of Canada, Society of Economic Geologists, Joint Annual
Theis, N.J., 1979, Uranium-bearing and associate minerals and their geochemi- Meeting, Toronto, 1991, Field Trip Guidebook B5: Toronto, Geological
cal and sedimentological context, Elliot Lake, Ontario: Geological Sur- Association of Canada, 34 p.
vey of Canada Bulletin, v. 304, p. 50.
Willingham, T.O., Nagy, B., Nagy, L.A., Krinsley, D.H., and Mossman, D.J.,
1985, Uranium-bearing stratiform organic matter in paleoplacers of the
lower Huronian Supergroup, Elliot Lake–Blind River region, Canada:
Canadian Journal of Earth Sciences, v. 22, p. 1930–1944. MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA


Memoir 198
2006
Time constraint for the occurrence of uranium deposits

and natural nuclear fission reactors in the Paleoproterozoic
Franceville Basin (Gabon)
F. Gauthier-Lafaye*
Centre de géochimie de la surface (CGS), UMR 7517, 1 rue Blessig, 67084 Strasbourg Cedex, France
ABSTRACT
Natural fission reactors at the Oklo uranium deposits in Gabon appear to have
formed in a short interval of geologic time during which uranium could migrate to
form deposits and the 235U/238U ratio was still high enough to trigger fission reactions.
At the time of sediment deposition in the ore-hosting Franceville Basin ~2100 m.y. ago,
the oxygen deficient atmosphere would have inhibited uranium dissolution and there-
fore its migration to form deposits. Dissolution and migration of uranium probably
began only during later diagenesis after ca. 2050 Ma, and local reduction reactions in
the presence of hydrocarbons allowed formation of high-grade uranium deposits. At
this time the 235U/238U ratio was still significantly higher than it is today, thus trigger-
ing nuclear fission reactions. Before 2.0 Ga, the 235U/238U ratio was also high enough to
allow fission reactions but no mechanisms were able to produce high-grade uranium
ores. Thus, oxygen in the atmosphere was probably the main factor controlling the
occurrence of natural nuclear fission reactions. This conclusion is in agreement with
earlier suggestions that oxygen contents in atmosphere increased during a “transition
phase” some 2450–2100 m.y. ago.
Keywords: natural fission reactors, Paleoproterozoic, uranium deposit, atmosphere.
INTRODUCTION in iron deposits (BIFs). At this time oxidation/precipitation

of Mn2+ was inhibited by the presence of dissolved Fe2+ (Roy,
Oklo and other high-grade uranium deposits in the Paleo- 2000) and therefore Mn accumulated in seawater. During late
proterozoic Franceville sedimentary basin in Gabon (Fig. 1) are Archean and Paleoproterozoic, the balance of Fe supplied by
the only deposits known to have acted as natural nuclear fis- oceanic hot springs and O2 from photosynthetic production
sion reactors. The Franceville Basin also contains the Moanda was altered, and the Fe2+ concentration in seawater drastically
Mn deposit, the second largest world reserve of manganese. decreased. Then, oxygen was consumed by oxidation and pre-
This paper demonstrates that critical behavior of the uranium cipitation of Mn2+ and, finally, the concentration of O2 in the
deposits in the Franceville Basin could have been achieved atmosphere started to increase when all Mn in seawater pre-
only during a very specific period of time corresponding to cipitated. Afterward pO2 in the atmosphere allowed the oxida-
the initial oxygenation of Earth’s atmosphere. It is generally tion of other elements such as uranium, its remobilization and
admitted that during early Archean, oxygen produced by pho- precipitation, and its concentration in high-grade deposits by
tosynthetic organisms induced Fe2+ oxidation and precipitation redox reactions.
*[email protected]
Gauthier-Lafaye, F., 2006, Time constraint for the occurrence of uranium deposits and natural nuclear fission reactors in the Paleoproterozoic Franceville Basin
(Gabon), in Kesler, S.E., and Ohmoto, H., eds., Evolution of Early Earth’s Atmosphere, Hydrosphere, and Biosphere—Constraints from Ore Deposits: Geological
Society of America Memoir 198, p. 157–167, doi: 10.1130/2006.1198(09). For permission to copy, contact [email protected]. ©2006 Geological Society of
America. All rights reserved.
157
158 F. Gauthier-Lafaye
Figure 1. Location of the Oklo and related deposits in Africa (inset) showing their relation to geological units in the Franceville Basin.
Our data agree with a model suggesting an increase of REGIONAL GEOLOGICAL SETTING AND
oxygen in the atmosphere between 2150 Ma, the time of depo- STRATIGRAPHY
sition of the Francevillian black shales devoid of uranium, and
2050 Ma, which corresponds to the period when uranium was The unmetamorphosed Francevillian Series is exposed in
mobilized in its oxidized state (Gauthier-Lafaye and Weber, the Plateau des Abeilles, Franceville, and Lastoursville intracra-
2003). Holland (1994) estimates that during this period, the tonic basins. These basins represent the continental platform of
pO2 level in the atmosphere increased rapidly from <1% to the more distal Okondja Basin (Weber, 1968; Ledru et al., 1989).
≥15% present atmospheric level. The Gabonese uranium deposits are all located in the Franceville
The possibility that nuclear fission reactions might have Basin and only this basin will be described here. The stratigraphic
occurred in the past was first suggested by Kuroda in 1956 but column of the Franceville Basin has been subdivided by Weber
we had to wait until 1972 to discover that such reactions did (1968) into five formations, which are from bottom to top the FA
indeed occur 1950 m.y. ago in the Oklo uranium deposits of through the FE Formations (Fig. 2)
the Francevillian Series, Gabon (Neuilly et al., 1972). Other
uranium deposits of Proterozoic age in the world (Maas and FA Formation
McCulloch, 1990) do not manifest the characteristic uranium
and rare earth element (REE) isotopic anomalies of natural This formation, which contains all the Francevillian ura-
nuclear fission reactions, suggesting that fission reactions nium deposits, increases from 100 to 1000 m in thickness from
occurred only in the Paleoproterozoic uranium ore deposits of the edge to the central part of the basin. It consists of fluviatile
Gabon. This paper points out the main physical and chemical conglomerates and coarse to medium-grained, poorly sorted
conditions that were necessary to start and to sustain fission sandstones overlain by marine well-sorted sandstones that were
reactions in a natural environment and tries to show that some deposited in tidal to supratidal environments. The FA sandstones
of crucial conditions are related to the age of the uranium of the Franceville Basin are very mature, suggesting that they
deposits. were affected by many cycles of erosion and sedimentation. Such
Time constraint for the occurrence of uranium deposits and natural nuclear fission reactors 159
STANLEY
POOL
FE
FD
Fine grained sandstone

with kaolinite
FC
Chert
FE
Welded tuff
FB2 FD
FD
Pyroclastic
FC Dolomite
FB1
Mn deposit
Carbonated manganese
protore
U Black shales
Sandstones
U
1000 m
FA Radioactive conglomerates
Archean basement
Archean basement U Uranium

Uranium mineralization
mineralisation
Figure 2. Stratigraphic section for the Franceville Basin showing position of uranium mineralization.
multiple cycles allow the formation of heavy mineral deposits in The FA sandstones consist mainly of quartz and muscovite.
sedimentological settings such as fluviatile placers. At Oklo, the Variably altered biotite and feldspar (mostly microcline) are
mineralized sandstones (the uppermost 10 m of the FA Forma- abundant in the lower fluviatile sandstones but rare in the overly-
tion) were deposited in a tidal environment of beaches and very ing marine sandstones. Heavy minerals such as zircon, thorite,
wide bars in which coarse-grained, well-sorted sandstones alter- and monazite are abundant in conglomerates and coarse-grained
nate with micaceous, fine-grained sandstones (Gauthier-Lafaye sandstones. Well-sorted marine sandstones at the top of the FA
1986; Gauthier-Lafaye and Weber 1989). Formation have been cemented extensively by an intense silica.
Three sedimentary facies that are distinguished as red-, A

green-, and black-colored sediments, have been recognized in
FA sandstones (Gauthier-Lafaye and Weber, 1989). The contact
between sediments of different color is clearly discordant to the
bedding, indicating that the coloration is not of sedimentary ori-
gin but was acquired during the diagenesis.
The red color is due to hematite impregnation of the argilla-
ceous matrix; hematite also surrounds the quartz grains between
their detrital boundaries and later silica overgrowths. The argil-
laceous matrix consists mainly of illite, and detrital biotites and
heavy minerals are highly altered. Detrital monazites are altered
to a Th-OH silicate with very low concentrations of U and REE,
whereas altered zircons show several growth zones enriched in
REE, P, Th, and U (with Th/U = 5–10) (Mathieu et al., 2001).
The red-colored sediments also contain dolomite and sulfates
(anhydrite, gypsum, and accessory barite) in the fluviatile and B
tidal environments.
Black sandstones are restricted to the upper parts of the FA
Formation. Their color is mainly due to organic matter consisting
of solidified petroleum (pyrobitumen) that fills primary and sec-
ondary porosity in the sandstone, including fractures (Gauthier-
Lafaye and Weber, 1981; Gauthier-Lafaye 1986; Cortial et al.,
1990; Nagy et al., 1991, 1993). In the primary porosity, organic
matter surrounds detrital quartz grains or is trapped in quartz
overgrowths (Fig. 3). Two types of black sandstones can be rec-
ognized by the type of alteration of the detrital biotites (Gauth-
ier-Lafaye et al., 1989). The first type contains weakly altered
biotites with a composition intermediate between biotite and
chlorite. The matrix consists mainly of iron-rich chlorite. In the
second type of black sandstone, detrital biotites have been altered
to a pale color and resemble muscovite with titanium oxides on
Figure 3. (A) Hydrocarbon trapped in the primary porosity of FA sand-
their surface. The matrix consists of illite without chlorite and the stone. Qd—detrital quartz grain; QII—quartz overgrowth; OM—or-
black color is due to the presence of organic matter in secondary ganic matter. (B) Hydrofracturing of a mineralized sandstone. Black
porosity. The first type of black sandstone is interpreted as sedi- nodules are uranium-bearing hydrocarbon.
ments that were never oxidized, whereas the second type corre-
sponds to sediments that were first oxidized after deposition and
then reduced by fluids associated with oil migration.
Green sediments are usually located between the red and cherts of the FB-C Formation (Fig. 2). The black shales are char-
the black sandstones. They contain highly altered biotites, which acterized by concentrations of organic matter (Corg) that reach a
look like “sandwiches” made of flakes of muscovite and green maximum of 15% (Gauthier-Lafaye and Weber 1989; Cortial
iron-rich chlorite (Gauthier-Lafaye and Weber, 1989). The matrix 1990). The deposition of these black shales may have occurred
contains both illite and green, iron-rich chlorite in varying pro- during a worldwide black shale event (Shunga event of Melezhik
portions. Pyrite accumulations are common in these green sand- et al., 1999) between ca 1.9 and 2.05 Ga. This event is related to
stones. Pyrite corrodes both the detrital quartz grains and their the stromatolite explosion (Semikhatov and Raaben, 1994; Mele-
overgrowths, and is usually associated with green chlorite. These zhik et al., 1997), and stromatolites are well preserved in the FC
pyrite accumulations are interpreted as reduced fronts resulting cherts. Mn-carbonates are a major component near the top of the
from the flow of reducing fluids through oxidized sediments. formation, where they form the protore of the important manga-
nese deposits of Moanda (Weber, 1968, 1997).
FB Formation The uranium content of FB black shale samples located in
various places in the Franceville Basin near and far from ura-
The FB Formation is composed chiefly of black shales that nium deposits ranges between 3.5 ppm (average of 48 samples
reach a thickness of 400–1000 m. On structural highs, mainly at in Gauthier-Lafaye, 1986) and 10.8 (average of six samples in
the edges of the Franceville Basin, these black shales are only Mossman et al., 1998). These values are particularly low when
few meters thick and grade into dolomites and stromatolite-rich compared to those for other black shales compiled by Vine and
Tourtelot (1977), which have an average uranium content of 30 and Weber 1989; Savary and Pagel 1997; Mathieu 1999). The
ppm, and to the USGS Standard Devonian Oil Shale, which has inclusions contain highly saline brines (28.7 wt.% NaCl eq. to 30
a uranium content of 48.8 ppm. wt.% CaCl2 eq.) that are rich in Li, Br, and SO4 (Mathieu et al.,
2001). Optical and electron microscopic properties and chemi-
FC-FD-FE Formations cal compositions of organic matter have been described in detail
by Cortial et al. (1990), Nagy et al. (1993), and Mossman et al.
The FC Formation is 10–40 m thick and consists chiefly (1993, 1998).
of massive dolomite and thick-banded cherts. Stromatolites, Organic matter of the Francevillian Series has reached high
together with organic mats, are well developed in cherts and are levels of thermal maturity; H/C and O/C ratios are low and the
described in detail by Bertrand-Sarfati and Potin (1994). The FD “coke” stage was reached. The true anthracite stage, as defined
Formation consists of black shales, with ignimbrite tuff at the top by using Transmitted Electron Microscopy (TEM), has not been
of the formation. The FE Formation comprises epiclastic sand- detected, thus suggesting the absence of higher-grade metamor-
stones and interlayered shales. The combined FD and FE Forma- phism (Cortial et al. 1990). The main source for the migrated
tions are more than 1000 m thick in the Franceville Basin, and bitumens has been identified as the black shales of the FB Forma-
are probably thicker in the northern part of the Okondja Basin, tion, which were placed below the FA reservoir locally by normal
although this forested area is poorly known. faults. In the upper part of the FA Formation, interlayered shales
with very high organic carbon contents might also have served
INTRUSIONS AND DEFORMATION as source rocks. Hydrocarbons were generated when sediments
of the FB Formation passed through the oil window (Gauthier-
Syenites and pegmatites of the N’Goutou volcanic complex Lafaye, 1986; Mossman, 2001; Mossman et al., 2001).
are interlayered with sediments at the base of FB Formation near
the edge of the Okandja Basin. These rocks have an Rb-Sr age of ISOTOPIC AGES AND EVOLUTION OF THE
2143 ± 143 Ma (Bonhomme et al., 1982). FRANCEVILLE BASIN
Later dykes of dolerite intrude the Franceville Basin, form-
ing two orthogonal systems oriented NNW-SSE and ENE-WSW. Isotopic Ages
These dolerites were dated by the K-Ar method on feldspathic
fractions, giving an age of 970 ± 30Ma (Bonhomme et al., 1982). The Francevillian series rests on Archean basement, which
More recent U-Pb isotopic measurements on zircons from the is composed dominantly of the granitic Chaillu and North
Oklo dyke yield an age of 860 ± 30 Ma (Evins, 2002). Gabon blocks that were emplaced between 3.9 and 3.2 Ga
The Franceville sedimentary succession is only slightly (Caen-Vachette et al., 1988; Ledru et al., 1989). The Francevil-
deformed. Maximum deformation occurs in rocks affected by lian Series and the geological events that affected the sediments
synsedimentary faults activated during a later extensional tectonic have been dated by several geochronometers (Fig. 4). Deposi-
episode. This tectonic event was also responsible for hydrofrac- tion of black shales of the FB Formation has been dated by Rb-
turing of sandstones (Fig. 3) in the upper part of the FA formation Sr isotopic analysis of interlayered syenites of the N’Goutou
(Gauthier-Lafaye 1986). In the Okondja Basin, deformation was volcanic complex (2143 ± 143 Ma) by Bonhomme et al. (1982).
more intense, and the succession is weakly folded. Thereafter, the More recently, new U-Pb isotopic analyses on zircons in ignim-
basin remained remarkably stable until recent uplift resulting in brite tuffs from the FE formation at Bidoudouma have been
its erosion under the present weathering regime. performed (Horie et al., 2004). Results give concordant ages at
2083 ± 6 Ma, which represents the age of deposition of the FD
DIAGENESIS Formation. The time of petroleum generation in the Francev-
ille Basin has been dated by two Sm-Nd isochrons on small
The highest diagenetic level reached by the sediments cor- authigenic clay fractions of the FB Formation. These give ages
responds to the level defined by Winkler (1976) as very low of 2099 ± 115 Ma and 2036 ± 79 Ma (Bros et al., 1992; Stille
grade metamorphism characterized by crystallization of chlorite et al., 1993), which are in good agreement with Pb-Pb dating
and illite, mainly of 1M polytype (Gauthier-Lafaye 1986). 2-M of authigenic clay fractions of the FB black shales (Gauthier-
illite is found locally only at the bottom of the FA Formation. In Lafaye et al., 1996a).
the iron formation, greenalite and stilpnomelane are present; the The age of the uranium deposit is 2050 ± 30 Ma as deter-
persistence of greenalite without minnesotaite or grunerite also mined by classical U-Pb analysis on uraninites (Gancarz, 1978).
indicates a minimal metamorphic grade. The date of the fission reactions has been determined with great
Fluid inclusion studies of quartz overgrowths and carbon- precision by comparing the fluence of the fission reaction to the
ate cements indicate that the maximum temperature reached in amount of fission elements (mainly REE) produced. The result
the Franceville Basin during diagenesis was between 180 and gives an age of 1950 ± 40 Ma (Ruffenach, 1978, 1979; Holliger,
200 °C at a maximum pressure of 1000 ± 200 bars (Openshaw 1988; Naudet 1991). Late diagenesis has been dated from clay
et al., 1978; Gauthier-Lafaye and Weber 1981; Gauthier-Lafaye fractions in the FB Formation using the Rb-Sr method, yielding
Uranite 1950 ± 40 Fission products

from reactors
Uraninite -
2050 ± 30 U-Pb
2099 ± 115
FB black--s hales -
2036 ± 73 Sm-Nd 970 ± 30 K-Ar
Dolerite
N’Goutou 860 ± 30U-Pb on
2143 ± 143
volcanic rocks -
Rb-Sr zircon
(FB. Form.)
Ignimbrite tuff
2083 ± 06
Bidoudouma (FD Form.)
2500 2000 1500 1000 500

Oxidizing
Oxydizing flfluids
uids 106 years
Oxidizing
Oxydizing flfluids
uids
? Around doleritedykes
Around dolerite dikes
Oil and associated fluids
Figure 4. Summary of the main isotopic ages for rocks of the Franceville Basin, the uranium ore deposits, and the natural nuclear fission reactors.
All data are referenced in the text.
an age of 1870 ± 50 Ma (Bonhomme et al., 1982). The doler- sediments, continued during early diagenesis, and ended with
ite dykes were dated by the K-Ar method at 970 Ma ± 30 Ma the extensional tectonic event noted above. During this time,
(Bonhomme et al., 1982). More recently, in situ U-Pb dating of fluids in the FA sandstones and FB black shales escaped from
zircons from the Oklo dike yields an age of 860 ± 30 Ma (Evins, the reservoirs due to compaction of sediments. The chemical
2002). The time of basin uplift is unknown. The only isotopic composition and oxidation state of the fluids that first migrated
measurement that is available is a U-Th age of 76,500 ± 6800 yr into the FA sandstones are not known, but later on, as burial
obtained on secondary torbernite related to the weathering of the progressed, fluids were probably oxidizing as recorded by oxi-
Bangombé uranium deposit (Bros et al., 2000). dation affecting most of the FA sediments and the basal FB
black shales. This oxidation, which occurred at the scale of the
Evolution of the Franceville Basin overall basin, implies recharge by meteoric waters. As burial of
the series increased, organic matter in the FB black shales (and
Between the time of deposition of the sediments and the black shales interlayered in the tidal sediments of the FA Forma-
uplift of the Franceville Basin, two main periods can be rec- tion) underwent maturation and reached the “oil window,” pro-
ognized (Fig. 5). The first period started after deposition of the ducing gas and liquid hydrocarbons. Petroleum then migrated
1 2
Figure 5. Schematic evolution of the

U Franceville Basin during the two main
periods related to formation of the ura-
nium deposits (see text).
FB black shales Oxidized sediments

Oxydized sediments Hydrofracturing Migration of
Tidal environment oxydized fluids
in FA sandstones
Green sediments Overpressured fluids zone Migration of
FA sandstones/Conglomerates
sandstones/Conglomerats hydrocarbon
Archean basement Black sediments U Uranium deposit

into the FA sandstone reservoir, where it was trapped in the pri- origin and is mainly concentrated in galena. Minium and even
mary porosity (Fig. 3). This organic matter does not contain any native lead have been found only in the fission reactors (Gauthier-
uranium, indicating that fluids migrating into the FA reservoir at Lafaye et al., 1996b). Zinc, copper, and molybdenum are present
this time were uranium-free. in minor concentrations. They occur mainly as sulfides (sphaler-
The extensional tectonic phase reactivated ancient faults, and ite, chalcopyrite, chalcocite) in the primary zones and as sulfates
accompanying hydraulic fracturing in the FA sandstone caused a in the weathered zones, although molybdenite is also found in
change of fluid circulation in the basin (Gauthier-Lafaye 1986; wulfenite. Barite is a common accessory mineral except in the
Gauthier-Lafaye and Weber 1989). Fluids responsible for hydro- vanadium-rich zones (Gauthier-Lafaye et al., 1996b).
fracturing came from the FB black shales, which have charac- In all uranium deposits, two types of ores with different ura-
teristics typical of overpressured shales (Osborne and Swarbrick nium contents may be distinguished. In the first type, uranium
1997). Because uranium minerals are associated with the hydrau- is located inside organic matter and its uranium content ranges
lic fractures, it is believed that this tectonic event occurred dur- from 0.1% to 1%. The uranium is then associated with sulfides,
ing the uranium mineralization episode ca. 2050 Ma. Liquid mainly pyrite, chlorite, and calcite. δ13C of this calcite is low (−10
hydrocarbons concentrated mainly in tectonic traps at the top of to –15‰) suggesting that part of its carbon has an organic origin
anticlines and beside normal faults where hydrofracturing was (Gauthier-Lafaye et al., 1989). Barren calcite outside the deposits
intense (Fig. 3). In these areas, petroleum moved into the second- has a δ13C value of 0% to −5‰. In the second type of uranium
ary porosity of the sandstones, and the organic matter has a high ore, the uranium content is high (1%–10%) and uraninite is asso-
uranium content. The migration of hydrocarbon and associated ciated with hematite and illite but not with organic matter. This
reduced fluids coming from the FB black shales was responsible type of ore is located in highly fractured rocks. Ore transitional
for reduction of previously oxidized sediments, forming reduced between these two types exists. This suggests that both types of
fronts that are now represented by the green-sulfide-rich sedi- uranium ores are products of the same oxidation-reduction pro-
ments located between the red and black sandstones and shales cess and that the highly fractured zones have been more perme-
(Fig. 5). ated by the oxidized uranium-bearing fluids.
THE URANIUM DEPOSITS FORMATION OF THE URANIUM DEPOSITS
The Franceville Basin contains five uranium deposits. From During deposition of the FA sandstones and conglomerates,
north to south (Fig. 1) these are Mounana, the first to be dis- reducing conditions prevailed as suggested by their mineralogi-
covered in 1956, Boyindzi, Oklo-Okelobondo, Bangombé, and cal assemblage (fresh biotite in fine non-oxidized sandstones).
Mikouloungou. The Oklo-Okelobondo and Bangombé deposits Uranium-bearing minerals (monazite, thorite, and possibly
contain the famous natural fission reactors. These deposits have detrital uraninite) were therefore stable and could have formed
been mined out with the exception of the Bangombé deposit. deposits similar to the Witwatersrand and Elliot Lake deposits.
Combined production from these deposits amounted to 28,000 However, as the burial of the Franceville Series progressed, oxy-
tons of uranium from 1961 to 1999. gen in the atmosphere evidently increased. This oxygenation of
The uranium deposits of the Franceville Basin are located the atmosphere and hydrosphere is recorded by the primary and
in the upper part of the FA Formation and are associated with secondary oxidation of the FA sandstones that occurred during
tectonic structures that also formed traps for hydrocarbons. In the diagenesis and during the oil migration into the FA reservoir.
non-weathered ore most of the uranium is in the form of uraninite This oxidation event was also responsible for alteration of the
and coffinite, which are closely associated with hydrocarbons uranium-bearing minerals (monazite and possibly uraninite)
occupying secondary porosity of the sandstones. The mineral- and migration of uranium into the petroleum traps where it met
ized organic matter consists of solid pyrobitumen with an atomic reducing conditions favorable for its precipitation.
hydrogen/carbon ratio lower than 0.5 (Cortial et al., 1990). In the field, the relationship between uranium depos-
In the deposits, uranium is associated with calcium, vana- its and oxidation fronts is not conspicuous because the fronts
dium, iron, lead, barium, zinc, copper, and molybdenum (Gauth- were obscured by the later reduction process. The relation can
ier-Lafaye 1986). Up to 6 wt% calcium is disseminated in ura- be seen, however, at Mikouloungou and Otobo and in several
nium-bearing minerals (uraninite). Vanadium contents were very places at Oklo where high-grade uranium ores are developed
high (0.01%–1%) at Mounana, where it was produced, and also in oxidized fractures. The association of uranium with cop-
at Boyindzi and Mikouloungou. At Mounana, vanadium oxides per, molybdenum, and vanadium, which are typical of ura-
(karelianite, montroseite, duttonite, corvusite) and silicate (ros- nium deposits associated with an oxidized front, supports this
coelite) were found in the primary (unweathered) ores, whereas interpretation (Dahlkamp, 1993). The oxidized uranium-bear-
uranium, lead, and barium vanadates occurred in the weathered ing solutions migrated upward through the lower part of the
zones (Geffroy et al., 1964 ; Weber 1968). In the ores with a high FA sandstones, which have been oxidized. Uraninite precipi-
uranium content, hematite and sulfides such as pyrite, marca- tated in tectonic traps when these solutions met hydrocarbons
site, and melnikovite are important locally. Lead is of radiogenic sourced from the FB Formation.
The uranium-bearing fluids were very saline and were and Paris (1978) and in the final report of the European program
trapped at 100–170 °C, at a pressure of 800–1200 bars (Mathieu, “Oklo-Natural Analogue: phase I” (Blanc, 1996).
1999; Mathieu et al., 2001). Some fluids are a mixture between Reactor 9, which is typical, is shown in Figure 6. The size
brines from the FA sandstone reservoir and low Na/Ca fluids typ- of the reactors is variable, reflecting both differences in original
ical of formation fluids associated with hydrocarbons in petro- size and later weathering. The biggest reactor (reactor 2) is a lens
leum fields (Carpenter et al., 1974). If we assume a concentration 12 m long, 18 m deep, and 20–50 cm thick. The reactor at Ban-
of 10 ppb uranium in the solutions, and given that the FA sand- gombé is only 8–10 m long, 1 m wide, and a few centimeters
stones are 1 km thick and have 10% porosity, one can calculate thick, although recent weathering has dissolved a large part of
that 40,000 tons of uranium, the total estimated reserve of the this reactor (Gauthier-Lafaye et al., 2000).
Franceville Basin, corresponds to ten times the water volume The reactor cores consist of a 5–20 cm thick layer of urani-
contained within the FA sandstone pore space. This suggests that nite embedded in clays (illite and chlorite). The uranium content
significant water circulation through the FA reservoir was needed of the core ranges between 40% and 60%. Accessory minerals
to form the uranium deposits and that this water needed to be are mainly sulfides (pyrite and galena), hematite, and phosphates
recharged with O2 in order to maintain its oxidizing potential. (mainly hydroxyapatite). The clay minerals reflect the thermal
The source of the uranium is believed to be detrital heavy gradient during the operation of the reactors (Gauthier-Lafaye et
minerals including uraninite, which accumulated in several 100- al., 1989; Pourcelot and Gauthier-Lafaye, 1999), with Mg-Al-
m-thick conglomerates in the lower FA Formation (Gauthier- chlorite and 2M1-illite close to the core and 1M-illite and Fe-
Lafaye, 1986; Gauthier-Lafaye and Weber, 1989). These con- chlorite at the edge. In some cases, quartz grains, more or less
glomerates contain monazite and thorite, but despite a thorough altered and dissolved, remain in the clays.
search of the FA conglomerates for detrital uraninites, this mineral
has never been found. This may be because the FA sandstones Fission Reactions
passed through the oxidation event that would have dissolved
any uraninite that was present. That oxidation event allowed the Fission reactions started in high-grade uranium ore when the
formation of the uranium deposits and the occurrence of fission uranium content of the sandstones reached 10% (Naudet, 1991).
reactions when a critical mass was reached. Therefore, taking Once fission reactions began, the temperature in the reactor core
into account the chronology of these phenomena, a consider- increased, starting a convective hydrothermal system around the
able increase of the oxygen in the atmosphere evidently occurred reactor. Core temperatures reached 400 °C (Gauthier-Lafaye et
between deposition of sediments and formation of uranium ore al. 1989; Mathieu et al., 2001) and decreased rapidly toward the
deposits. This happened between 2.14 ± 14 Ga and 2.05 ± 30 Ga. edges (thermal gradient around 100 °C/m) with heat transferred
Unfortunately, the large uncertainty in the radiometric data does largely by conduction (Gerard, 1997; Gerard et al. 1997). Hydro-
not allow us to fix the age of the transition stage more precisely. thermal circulation caused the dissolution and migration of 80%
Following formation of the uranium deposits and the fission of the silica from the mineralized sandstones. This decreased the
reactors, diagenesis of the Franceville Series continued until the volume of the sandstones and increased the uranium content of
FA sandstones reached a depth of 4000 m where temperatures the residual layer, which then consisted mainly of clays and ura-
reached 180–200 ºC (Gauthier-Lafaye et al., 1989; Mathieu, nium dioxides. Furthermore, it has been shown that in reactor 2,
1999). This allowed major silicification of the FA sandstones,
which reduced the porosity of the rocks and protected the ura-
nium deposits and the reactors from alteration.
THE NATURAL NUCLEAR FISSION REACTORS
Description of the Reactors
Fifteen natural nuclear fission reactors (hereafter reactors)

were discovered in two uranium deposits of the Franceville
Basin. Fourteen are located in the Oklo-Okelobondo deposit and
one in the small deposit of Bangombé which is 30 km from Oklo-
Okelobondo (Fig. 1). Geological, mineralogical, and geochemical
descriptions of the reactors are given by Gauthier-Lafaye (1986),
Gauthier-Lafaye et al. (1989), and Gauthier-Lafaye et al. (1996b),
as well as in a report that synthesized a recent European study of
fission products in the various reactors (Gauthier-Lafaye et al.,
2000). Other major contributions are reported in the proceedings
of two IAEA international conferences held in Libreville (1975) Figure 6. Cross section of reactor 9 in the Oklo deposit.
up to 50% of the present uranium was introduced into the min- Because the radioactive decay constant for 235U (99.8485 *
eralized layer during the fission reaction, probably by the fluids 10 ) is smaller than that for 238U (1.55125*10−10), the 235U/238U
−10
migrating through the reactors. This process allowed a fivefold ratio changed with time, as shown in Figure 7. Two b.y. ago the
increase in the amount of uranium per unit volume, forming the ratio was similar to that of fuel in PWR (Pressured Water Reac-
reactor core (Gauthier-Lafaye et al., 1989). tor) nuclear plants.
Fluid inclusions show that the hydrothermal fluids produced There is little chance of finding natural fission reactors
by the fission reactions were similar to the diagenetic fluids, younger than 1.9 Ma because the amount of fissile 235U by then
but their temperatures were higher, ranging from 280 to 420 °C was too small to sustain fission reactions in a natural environ-
(Mathieu, 1999; Mathieu et al., 2001). They also contained traces ment. At ages greater than 2.0 Ga, the 235U/238U was sufficient
of CH4, CO2, O2, H2, and H2O2, with the last three species resulting to sustain fission in natural reactors and it was much higher than
from the radiolysis of water in the core of the reactor (Dubessy et this minimum before ca. 2.5 Ga. This is the main argument for
al., 1988; Mathieu et al., 2001). Because H2 diffuses faster than predicting that fission reactors should have been common prior to
O2, very oxidized conditions could have occurred in the reactors, 2.0 Ga. However, no traces of nuclear fission reactors have been
accounting for the precipitation of minium (Pb3O4), which needs found in pre-2.0 Ga uranium deposits such as those in the Witwa-
a high oxidation potential and very low reduced sulfur content tersrand or at Elliot Lake. This means that other conditions must
in the hydrothermal fluids (Savary and Pagel, 1997). During the be important as well. Clearly, a high concentration of uranium is
fission reaction, CH4 and CO2 were released from organic matter of particular importance and this may be achieved only in certain
at high temperatures (Mathieu, 1999). Uraninite depleted in 235U geochemical conditions.
occurs in bitumen nodules, suggesting that organic matter was Fulfillment of conditions 1 and 3 allows a uranium deposit
still mobile during the fission reactions (Nagy et al., 1993) to reach the critical mass required to sustain a fission reaction. If
To start and sustain a fission reaction, the concentration of the physical (porosity) and chemical (content of various minerals
neutron “poisons” in the system must be low (Naudet, 1991). and neutron poisons such as REE, boron, and vanadium) proper-
In geological systems boron, vanadium, and REE are the most ties of the Oklo sandstones are taken into account, a sustained
important poisons. Differences in the vanadium concentration of fission reaction would have occurred at Oklo where the uranium
the various deposits in the Franceville Basin may explain the dis- content was 10% throughout one cubic meter of rock (Naudet,
tribution of the reactors. 1991). In actuality, the uranium content in the reactors was up
to 60%. Such high concentrations in a sedimentary rock indi-
Conditions Necessary for Fission Reactions cate that the uranium was enriched during one or more dissolu-
tion-precipitation episodes and implies that enough oxygen was
The conditions for the occurrence of nuclear fission reac- available in the water to allow the required oxidation-reduction
tions in a geological system have been discussed by Naudet reactions of U to occur.
(1991). Fission reactions occur spontaneously in 238U, resulting
in the production of fast neutrons. If these neutrons are slowed
down, they may induce fission in 235U or 239Pu. Nuclear reactions
can then be sustained. This can happen in a geological system if
at least three conditions are met:
1. The uranium ore must have a high uranium content. This
increases the number and density of spontaneous fission events
in 238U and therefore increases the probability of inducing fission
in 235U, which sustains the fission reaction.
2. The fast neutrons produced by the spontaneous fission of
238
U must be slowed down by migration through water or graph-
ite. Naudet (1991) has shown that water should have been the
main moderator for neutrons in the Oklo reactors, as it must have
been present in the pore space of the sandstones and in clays that
were important mineralogical phases in the reactors. Carbon may
have played an important role in some reactors with abundant
organic matter (Naudet, 1991).
3. The 235U content of the uranium should be high, so that
the probability of a neutron hitting a fissile 235U atom and thus
sustaining the nuclear reaction is also high. This last condition is
satisfied in commercial nuclear plants by increasing the 235U/238U
ratio of the nuclear fuel. At Oklo this condition was met because
235
U was more abundant at that time. Figure 7. Evolution of 235U/238U ratio versus time in Earth.
Cadarache, France, 20–21 May 1999: European Commission, Nuclear

CONCLUSIONS Science and Technology Series, EUR 19137 EN, p. 227–241.
Caen-Vachette, M., Vialette, Y., Bassot, J.P., and Vidal, P., 1988, Apport de la
The mobilization of uranium in the Franceville Basin géochronologie isotopique à la connaissance de la géologie gabonaise:
Chronique Recherche Minière, v. 491, p. 35–54.
occurred during diagenesis ca. 1950 Ma, as the burial of the FA Carpenter, A.B., Trout, M.L., and Pickett, E.E., 1974, Preliminary report on
and FB Formations reached depths of 3000–4000 m. Migration the origin and chemical evolution of lead- and zinc-rich oil field brines in
of uranium was related to oxidation of the whole FA reservoir, Central Mississippi: Economic Geology and the Bulletin of the Society of
alteration of the detrital uranium-bearing minerals, and precipita- Cortial, F., Gauthier-Lafaye, F., Oberlin, A., Lacrampe-Couloume, G., and
tion of iron oxides. This required a complete recharge of oxygen- Weber, F., 1990, Characterization of organic matter associated with
ated waters circulating in the FA reservoir, which would have uranium deposits in the Francevillian Formation of Gabon (Lower Pro-
terozoic): Organic Geochemistry, v. 15, p. 73–85, doi: 10.1016/0146-
required a high level of oxygen in the atmosphere. The oxidized 6380(90)90185-3.
uranium precipitated when it came into contact with hydrocar- Dahlkamp, F.J., 1993, Uranium ore deposits: Berlin, Springer-Verlag, 460 p.
bons concentrated in tectonic traps. Dubessy, J., Pagel, M., Beny, J.M., Christensen, H., Bernard, H., Kosztolanyi,
C., and Poty, B., 1988, Radiolysis evidence by H2–O2 and H2-bearing
The occurrence of oxidation-reduction reactions allowed fluid inclusions in three uranium deposits: Geochimica et Cosmochimica
formation of very high grade uranium ore deposits for the first Acta, v. 52, p. 1155–1167, doi: 10.1016/0016-7037(88)90269-4.
time in Earth history, and this led in turn to nuclear fission reac- Evins, L.Z., 2002, Geochronology of the Oklo and Bangombé fossil natural
reactors: Tracing the effects of geological events [Ph.D. thesis]: Stock-
tions. Younger uranium deposits that formed by similar processes holm, Department of Geology and Geochemistry, Stockholm University,
did not undergo nuclear fission reactions because the abun- Avhandling för Filosofie Doktorexamen no. 313.
dance of U235 was too low after 1950 Ma. Older deposits did not Gancarz, A.J., 1978, U-Pb age (2.05 x 10^9 years) of the Oklo uranium deposit,
in Les réacteurs de fission naturels: Vienna, IAEA, p.513–520.
develop the high grades necessary to induce spontaneous fission Gauthier-Lafaye, F., 1986, Les gisements d’uranium du Gabon et les réacteurs
because oxygenated fluids necessary to form such deposits were d’Oklo: Modèle métallogénique de gites à fortes teneurs du Protérozoique
not available. Thus, the natural nuclear reactors in the Franceville inférieur: Mémoire des Sciences Géologiques, 78: 206 p.
Gauthier-Lafaye, F., and Weber, F., 1981, Les concentration uranifères du
Basin record a unique interval in Earth’s history, of high 235U and Francevillien du Gabon: Leur association avec des gîtes à hydrocarbures
transitional redox conditions, and this accounts for the apparent fossiles du Protérozoïque inférieur: Comptes Rendus de l’Académie des
absence of such remarkable phenomena in older or younger ura- Sciences, series D, v. 292, p. 69–74.
Gauthier-Lafaye, F., and Weber, F., 1989, The Francevillian (Lower Protero-
nium deposits. zoic) uranium ore deposits of Gabon: Economic Geology and the Bulletin
ACKNOWLEDGMENTS Gauthier-Lafaye, F., Weber, F., and Ohmoto, H., 1989, Natural fission reactors
of Oklo: Economic Geology and the Bulletin of the Society of Economic
Geologists, v. 84, p. 2286–2295.
The author wishes to thank the Compagnie des Mines Gauthier-Lafaye, F., Bros, R., and Stille, P., 1996a, Pb-Pb isotope systemat-
d’Uranium de Franceville (COMUF), the Compagnie Générale ics of diagenetic clays: An example from Proterozoic black shales of the
Franceville basin (Gabon): Chemical Geology, v. 133, p. 243–250, doi:
des Matériaux (COGEMA) and the Commissariat à l’Energie 10.1016/S0009-2541(96)00032-0.
Atomique (CEA) for their financial support when working on Gauthier-Lafaye, F., Holliger, P., and Blanc, P.-L., 1996b, Natural fission reac-
the uranium deposits of Gabon and on natural nuclear fission tors in the Franceville basin, Gabon: A review of the conditions and results
of a critical event in a geologic system: Geochimica et Cosmochimica
reactors. Two reviewers, A. Wallace and an anonymous one, are Acta, v. 60, p. 4831–4852, doi: 10.1016/S0016-7037(96)00245-1.
also thanked for constructive reviews. This is EOST/CGS/CNRS Gauthier-Lafaye, F., Ledoux, E., Smellie, J., Louvat, D., Michaud, V., Pérez
contribution n° 2004.602-UMR7517. del Villar, L., Oversby, V., and Bruno, J., 2000, Oklo—Natural analogue
(phase II): Behaviour of nuclear reaction products in a natural environ-
ment: European Commission, Nuclear Science and Technology Series,
REFERENCES CITED Contract FI4W-CT96-0020, Final report EUR 19139 EN, 116 p.
Gauthier-Lafaye, F., and Weber, F., 2003, Natural nuclear fission reactors: time
Bertrand-Sarfati, J., and Potin, B., 1994, Microfossiliferous cherty stromato- constraints for occurrence and their relation to uranium and manganese
lites in the 2000 Ma Franceville Group, Gabon: Precambrian Research, deposits and to the evolution of the atmosphere: Precambrian Research,
v. 65, p. 341–356, doi: 10.1016/0301-9268(94)90112-0. v. 120, p. 81–101, doi: 10.1016/S0301-9268(02)00163-8.
Blanc, P.-L., 1996, Oklo—Natural analogue for a radioactive waste repository Geffroy, J., Cesbron, F., and Lafforgue, P., 1964, Données préliminaires sur les
(phase 1), Volume 1: Acquirement of the project: European Commission, constituants profonds des minerais uranifères et vanadifères de Mounana
Nuclear Science and Technology Series, VI, 123 p. ISBN 92-827-7448- (Gabon): Comptes Rendus de l’Académie des Sciences, series D, v. 259,
1. p. 601–603.
Bonhomme, M., Gauthier-Lafaye, F., and Weber, F., 1982, An example of Gerard, B., 1997, Modélisation 3D des transferts de chaleur et de fluide dans les
Lower Proterozoic sediments: The Francevillian in Gabon: Precambrian formations sédimentaires: Application aux réacteurs d’Oklo (Gabon) [Ph.D.
Research, v. 18, p. 87–102, doi: 10.1016/0301-9268(82)90038-9. thesis]: Nancy, Institut National Polytechnique de Lorraine, 1997, 251 p.
Bros, R., Stille, P., Gauthier-Lafaye, F., Weber, F., and Clauer, N., 1992, Sm-Nd Gerard, B., Royer, J.J., le Carlier, C., Pagel, M., Scius, H., and Gauthier-Lafaye,
isotopic dating of Proterozoic clay material: An example from Francev- F., 1997, 3D modelling of heat and mass transfers around a natural fission
illian sedimentary series, Gabon: Earth and Planetary Science Letters, reactor, in Louvat, D., and von Maravic, H., eds., Proceedings of the first
v. 113, p. 207–218, doi: 10.1016/0012-821X(92)90220-P. joint EC-CEA workshop on the Oklo-natural analogue Phase II project
Bros, R., Roos, P., Andersson, P., Holm, E., Griffault, L., and Smellie, J., 2000, (EUR 18314), Sitges, Spain, June 18-20, 1997: Oklo Phase II workshop,
Mobilization of fission products and 238U-series radionuclides in the Ban- EUR Report Series no. 18314, p. 301–308.
gombé reactor zone: Migration trends and time-scale evaluation, in Lou- Holland, H.D., 1994, Paleoproterozoic atmospheric change, in Bengtson, S.,
vat, D., Michaud, V., and von Maravic, H., eds., OKLO working group ed., Early life on Earth: Nobel Symposium 84: New York, Columbia
proceedings of the third and final EC-CEA workshop on Oklo-phase II, University Press, p. 237–244.
Holliger, P., 1988, Ages U-Pb définis in-situ sur oxydes d’uranium à l’analyseur ral fission reactors, Gabon, Africa: Geology, v. 21, p. 655–658, doi:
ionique: Méthodologie et conséquences géochimiques: Comptes Rendus 10.1130/0091-7613(1993)021<0655:ROOMIT>2.3.CO;2.
de l’Académie des Sciences, series D, v. 307, p. 367–373. Naudet, R., 1991, Oklo: Des réacteurs nucléaires fossiles: Paris, Collection du
Horie, K., Hidaka, H., and Gauthier-Lafaye, F., 2004, U-Pb Zircon geochronol- Commissariat à l’Energie Atomique, 695 p.
ogy of the Franceville series at Bidoudouma, Gabon: Abstracts, Proceed- Neuilly, M., Bussac, J., Frejacques, C., Nief, G., Vendryes, G., and Yvon,
ings, Goldschmidt Conference, 5–11 June 2004, Copenhagen, p. 511 J., 1972, Sur l’existence dans un passé reculé d’une réaction en chaîne
Kuroda, P.F., 1956, On the nuclear physical stability of the uranium minerals: naturelle de fission, dans le gisement d’uranium d’Oklo (Gabon): Comptes
Journal of Chemical Physics, v. 25, p. 781–782, doi: 10.1063/1.1743058. Rendus de l’Académie des Sciences, series D, v. 275, p. 1847–1849.
Ledru, P., Eko N’Dong, J., Johan, V., Prian, J.P., Coste, B., and Haccard, D., Openshaw, R., Pagel, M., and Poty, B., 1978, Phases fluides contemporaines de
1989, Structural and metamorphic evolution of the Gabon orogenic belt: la diagenèse des grès, des mouvements tectoniques et du fonctionnement
Collision tectonics in the Lower Proterozoic: Precambrian Research, 44: des réacteurs nucléaires d’Oklo (Gabon), in Les réacteurs de fission
227–241. naturels: Vienna, IAEA, TC-119/9, p. 265–296.
Maas, R., and McCulloch, M.T., 1990, A search of fossil nuclear reactors in the Osborne, M.J., and Swarbrick, R.E., 1997, Mechanisms for generating over-
Alligator River uranium field, Australia: Constraints from Sm, Gd and Nd pressure in sedimentary basins: A re-evaluation: American Association of
isotopic studies: Chemical Geology, v. 88, p. 301–315, doi: 10.1016/0009- Petroleum Geologists Bulletin, v. 81, p. 1023–1041.
2541(90)90095-O. Pourcelot, L., and Gauthier-Lafaye, F., 1999, Hydrothermal and supergen clays
Mathieu, R., 1999, Reconstitution des paléocirculations fluids et des migra- of the Oklo natural reactors: Conditions of radionuclide release, migra-
tions élémentaires dans l’environnement des réacteurs nucléaires naturels tion and retention: Chemical Geology, v. 157, p. 155–174, doi: 10.1016/
d’Oklo (Gabon) et des argilites de Tournemire France) [Ph.D. thesis]: S0009-2541(98)00194-6.
Nancy, Institut National Polytechnique de Lorraine, 518 p. Roy, S., 2000, Late Archean initiation of manganese metallogenesis: Its signifi-
Mathieu, R., Zetterström, L., Cuney, M., Gauthier-Lafaye, F., and Hidaka, H., cance and environmental controls: Ore Geology Reviews, v. 17, p. 179–
2001, Alteration of monazite and zircon and lead migration as geochemical 198, doi: 10.1016/S0169-1368(00)00013-5.
tracers of fluid paleocirculations around the Oklo-Okélobondo and Ban- Ruffenach, J.C., 1978, Etude des migrations de l’uranium et des terres rares
gombé natural nuclear reaction zones (Franceville basin, Gabon): Chemi- sur une carotte de sondage et application à la détermination de la date des
cal Geology, v. 171, p. 147–171, doi: 10.1016/S0009-2541(00)00245-X. réactions nucléaires, in Les réacteurs de fission naturels: Vienna, IAEA,
Melezhik, V.A., Fallick, A.E., Makarikhin, V.V., and Lubtsov, V.V., 1997, Links TC-119/9, p. 441–471.
between Palaeoproterozoic palaeogeography and rise and decline of stro- Ruffenach J.C, 1979, Les réacteurs nucléaires naturels d’Oklo. Paramètres neu-
matolites: Fennoscandian Shield. Precambrian Research, v. 82, p. 311– troniques, date et durée de fonctionnement, migrations de l’uranium et des
348, doi: 10.1016/S0301-9268(96)00061-7. produits de fission [Ph.D. thesis Sciences]: Université Paris VII, 351 p.
Melezhik, V.A., Fallick, A.E., Medvedev, P.V., and Makarikhin, V.V., 1999, Savary, V., and Pagel, M., 1997, The effects of water radiolysis on local redox
Extreme 13Ccarb enrichment in ca. 2.0 Ga magnesite-stromatolite-dolo- conditions in the Oklo, Gabon, natural fission reactors 10 and 16: Geochi-
mite-red beds association in a global context: A case for the world-wide mica et Cosmochimica Acta, v. 61, p. 4479–4494, doi: 10.1016/S0016-
signal enhanced by a local environment: Earth-Science Reviews, v. 47, 7037(97)00261-5.
p. 1–40, doi: 10.1016/S0012-8252(99)00027-6. Semikhatov, M.A., and Raaben, M.A., 1994, Dynamics of global diversity of
Mossman, D.J., Nagy, B., Rigali, M.J., Gauthier-Lafaye, F., and Holliger, P., Proterozoic stromatolites: 1. Northern Eurasia, China and India: Stratig-
1993, Petrography and paragenesis of organic matter associated with raphy and Geological Correlation, v. 2, p. 10–32.
the natural fission reactors at Oklo, Republic of Gabon: A preliminary Stille, P., Gauthier-Lafaye, F., and Bros, R., 1993, The neodymium isotope sys-
report: International Journal of Coal Geology, v. 24, p. 179–194, doi: tem as a tool for petroleum exploration: Geochimica et Cosmochimica
10.1016/0166-5162(93)90009-Y. Acta, v. 57, p. 4521–4525, doi: 10.1016/0016-7037(93)90502-N.
Mossman, D.J., Gauthier-Lafaye, F., Nagy, B., and Rigali, M., 1998, Geochem- Vine, J.W., and Tourtelot, E.B., 1977, Geochemistry of black shale deposits—A
istry of organic-rich black shales overlying the natural nuclear fission summary report: Economic Geology and the Bulletin of the Society of
reactors of Oklo, Republic of Gabon: Energy Sources, v. 20, p. 521–539. Economic Geologists, v. 65, p. 253–272.
Mossman, D.J., Gauthier-Lafaye, F., and Jackson, S., 2001, Carbonaceous Weber, F., 1968, Une série précambrienne du Gabon: Le Francevillien, sédimen-
substances associated with the Paleoproterozoic natural nuclear fission tologie, géochimie et relation avec les gîtes minéraux associés. Mémoires
reactors of Oklo, Gabon: Paragenesis, thermal maturation and carbon Service Cartes Géologiques d’Alsace-Lorraine, 28: 328 p.
isotopic and trace element composition: Precambrian Research, v. 106, Weber, F., 1997, Evolution of lateritic manganese deposits, in Paquet, H., and
p. 135–148, doi: 10.1016/S0301-9268(00)00129-7. Clauer, N., eds., Soils and sediments: Mineralogy and geochemistry: Ber-
Mossman, D., 2001, Hydrocarbon habitat of the Paleoproterozoic Franceville lin, Springer-Verlag, 369 p.
Series, Republic of Gabon: Energy Sources, v. 23, p. 45–53, doi: 10.108 Winkler, H.G.F., 1976, Petrogenesis of metamorphic rocks: New York,
0/00908310151092137. Springer-Verlag, 334 p.
Nagy, B., Gauthier-Lafaye, F., Holliger, P., Davis, D.W., Mossman, D.J., Lev-
enthal, J.S., Rigali, M., and Parnell, J., 1991, Role of organic matter in
containment of uranium and fissiogenic isotopes at the Oklo natural reac-
tors: Nature, v. 354, p. 472–475, doi: 10.1038/354472a0.
Nagy, B., Gauthier-Lafaye, F., Holliger, P., Mossman, D., Leventhal, J., and
Rigali, M., 1993, Role of organic matter in the Proterozoic Oklo natu- MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA


Memoir 198
2006
Proterozoic sedimentary exhalative (SEDEX) deposits and links to

evolving global ocean chemistry
Timothy W. Lyons*
Anne M. Gellatly
Department of Earth Sciences, University of California, Riverside, California 92521, USA
Peter J. McGoldrick
CODES, ARC Centre of Excellence in Ore Deposits, University of Tasmania, Hobart, TAS 7001, Australia
Linda C. Kah
Department of Earth & Planetary Sciences, University of Tennessee, Knoxville, Tennessee 37996, USA
ABSTRACT
Sedimentary exhalative (SEDEX) Zn-Pb-sulfide mineralization first occurred on a

large scale during the late Paleoproterozoic. Metal sulfides in most Proterozoic deposits
have yielded broad ranges of predominantly positive δ34S values traditionally attributed
to bacterial sulfate reduction. Heavy isotopic signatures are often ascribed to fraction-
ation within closed or partly closed local reservoirs isolated from the global ocean by
rifting before, during, and after the formation of Rodinia. Although such conditions
likely played a central role, we argue here that the first appearance of significant SEDEX
mineralization during the Proterozoic and the isotopic properties of those deposits are
also strongly coupled to temporal evolution of the amount of sulfate in seawater.
The ubiquity of 34S-enriched sulfide in ore bodies and shales and the widespread
stratigraphic patterns of rapid δ34S variability expressed in both sulfate and sulfide
data are among the principal evidence for global seawater sulfate that was increasing
during the Proterozoic but remained substantially lower than today. Because sulfate
is produced mostly through weathering of the continents in the presence of oxygen,
low Proterozoic concentrations imply that levels of atmospheric oxygen fell between
the abundances of the Phanerozoic and the deficiencies of the Archean, which are also
indicated by the Precambrian sulfur isotope record. Given the limited availability of
atmospheric oxygen, deep-water anoxia may have persisted well into the Proterozoic
in the presence of a growing sulfate reservoir, which promoted prevalent euxinia. Col-
lectively, these observations suggest that the mid-Proterozoic maximum in SEDEX
mineralization and the absence of Archean deposits reflect a critical threshold in the
accumulation of oceanic sulfate and thus sulfide within anoxic bottom waters and pore
fluids—conditions that favored both the production and preservation of sulfide miner-
alization at or just below the seafloor. Consistent with these evolving global conditions,
the appearance of voluminous SEDEX mineralization ca. 1800 Ma coincides generally
with the disappearance of banded iron formations—marking the transition from an
early iron-dominated ocean to one more strongly influenced by sulfide availability.
*[email protected]
Lyons, T.W., Gellatly, A.M., McGoldrick, P.J., and Kah, L.C., 2006, Proterozoic sedimentary exhalative (SEDEX) deposits and links to evolving global ocean
chemistry, in Kesler, S.E., and Ohmoto, H., eds., Evolution of Early Earth’s Atmosphere, Hydrosphere, and Biosphere—Constraints from Ore Deposits: Geological
Society of America Memoir 198, p. 169–184, doi: 10.1130/2006.1198(10). For permission to copy, contact [email protected]. ©2006 Geological Society of
America. All rights reserved.
169
170 T.W. Lyons et al.
In further agreement with this conceptual model, Proterozoic SEDEX deposits in

northern Australian formed from relatively oxidized fluids that required reduced
conditions at the site of mineralization. By contrast, the generally more oxygenated
Phanerozoic ocean may have only locally and intermittently favored the formation
and preservation of exhalative mineralization, and most Phanerozoic deposits formed
from reduced fluids that carried some sulfide to the site of ore precipitation.
Keywords: Proterozoic, SEDEX deposits, sulfur isotopes, ocean chemistry, atmo-

spheric oxygenation.
INTRODUCTION within the Proterozoic ocean-atmosphere system. All of these

parameters will be viewed in light of the pronounced 34S enrich-
Sedimentary exhalative (SEDEX) deposits are sulfide-domi- ments that dominate the metal sulfides of most mid-Proterozoic
nated Pb-Zn ore bodies with sphalerite and galena as the prin- SEDEX mineralization. More challenging is an explanation of
cipal ore minerals. Although the specifics of this deposit type the distribution of major SEDEX mineralization in younger sedi-
are complex (Goodfellow et al., 1993; Lydon, 1996; Large et ments. Tectonic influences, specifically the temporally varying
al., 2004), formation through submarine venting of hydrother- extent of continental rifting, are always the essential backdrop to
mal fluids in structurally controlled (rifted) sedimentary basins SEDEX hydrothermal processes. However, in the Phanerozoic,
are universal characteristics. The compilations of Goodfellow et as for the Precambrian, depositional redox may have played a
al. (1993) and Lydon (1996) show us that SEDEX mineraliza- central role in controlling the amount and preservation of metal
tion spans a broad portion of geologic time—ranging from ca. sulfide accumulation on the seafloor.
1800 Ma to possible analogues forming today—however, peak
abundances are well expressed during discrete intervals of the BACKGROUND
mid-Proterozoic (1800–1600 Ma) and late Proterozoic to Paleo-
zoic (600–300 Ma). A clear maximum in ore tonnage is observed SEDEX Deposits
between 1700 and ca. 1600 Ma.
Despite this broad temporal distribution, Lydon (1996, The many details of sedimentary exhalative sulfide deposits,
p. 141) found the lack of early Proterozoic and Archean SEDEX including genetic models, are ably reviewed in Goodfellow et al.
deposits to be enigmatic: (1993), Lydon (1996), and Large et al. (2004) and are only briefly
summarized here. SEDEX deposits are marked by stratiform,
sediment-hosted Zn-Pb-sulfide mineralization, with sphalerite
There does not appear to be any compelling reason, based on current and galena as the principal ore minerals and pyrite typically as the
understanding, why Sedex deposits should not occur in rocks older most abundant sulfide phase. These deposits form in sedimentary
than Middle Proterozoic [late Paleoproterozoic]. The onset of the Mid-
dle Proterozoic does not seem to coincide with any major permanent basins through submarine venting of hydrothermal fluids. SEDEX
change in global climate, ocean water composition, atmosphere com- mineralization is essentially synsedimentary—forming at or just
position, or geotectonic processes. below the seafloor within fault-controlled basins or troughs that
are most often related to major intracratonic or continental mar-
gin rift zones—although mineralization often postdates the most
The purpose of this report is to revisit this assertion within active phase of rifting (Lydon, 1996),
the context of the most recent studies of Precambrian ocean- The deposits range in age from ca. 1800 Ma to pos-
atmosphere chemistry and, in doing so, suggest why such a gap sible modern analogues but cluster in two groups: 1800–1600
is precisely what the models for early Earth predict. It is not our and 600–300 Ma (Fig. 1). Most Proterozoic deposits formed
intent, however, to provide a comprehensive survey of SEDEX between ca. 1700 and ca. 1600 Ma, which, from the standpoint
mineralization or to develop a model that precludes the possi- of ore mass, is the dominant interval of all SEDEX-related Zn
bility of exceptions. Instead, we will suggest that the first-order and Pb mineralization. Ore fluids are generally linked to metal-
patterns, and in particular the absence of Archean deposits and liferous formational waters that were heated within the sedimen-
the abundance of SEDEX mineralization of Proterozoic age, are tary basin under the elevated geothermal conditions of the typi-
completely consistent with a now widely held view of evolving cally extensional tectonic settings. The redox conditions of the
Precambrian ocean chemistry. The essential Proterozoic controls basin lithologies exert an important control on the redox state of
include the first appearance of quantitatively significant sulfate in the metal-transporting fluids (Cooke et al., 2000). These work-
the ocean, the persistence of deep water oxygen deficiency, the ers distinguish two distinct classes of SEDEX Zn-Pb deposits
widespread distribution of bacterial hydrogen sulfide within the termed McArthur-type and Selwyn-type. Metal-transporting
ocean and underlying sediments, and possibly abundant methane fluids for the former appear to have been relatively oxidized
Proterozoic sedimentary exhalative (SEDEX) deposits 171
millions of tonnes
180
Zn + Pb
120
60
0 Figure 1. Age distribution of SEDEX
mineralization as (A) tons of Zn and Pb
B 20 per 100 m.y. interval and (B) number
number of
15 of deposits per 100 m.y. interval (after
deposits
Lydon, 1996). Background for the inter-
10 pretative details and related references
5 are available in the text—including the
0 G.O.E. (Great Oxidation Event).
(Ma) 0 400 800 1200 1600 2000
widespread anoxia/euxinia?
G.O.E. (~2.3 Ga):
second disappearance increasing
major oxidation of BIFs sulfate availability
of Earth’s surface (H2S > Fe)
(sulfate » sulfide), whereas the latter type formed from more possible via BSR alone. A popular hypothesis for the extreme 34S
reduced fluids (sulfide ≥ sulfate). depletions observed in low-temperature sulfide minerals invokes
SEDEX deposits are well known for the large amounts of additional fractionation during bacterial disproportionation of
sulfur they contain, which is present in Proterozoic examples elemental S and other S intermediates, which form through par-
almost exclusively as pyrite, base metal sulfides, and pyrrho- tial oxidation of H2S (Canfield and Thamdrup, 1994; Habicht and
tite. Barite is much more common in the Phanerozoic deposits. Canfield, 2001; cf. Brunner and Bernasconi, 2005).
Because the metal-rich ore fluids would by necessity be compar- The net isotopic fractionation observed between sulfate and
atively sulfide deficient, a substantial portion of the sulfide sul- sulfide reflects both the collective effects of bacterial S reduction
fur must be supplied independently at the site of mineralization. and disproportionation as well as the properties of the sulfate res-
Bacterial reduction of seawater sulfate within anoxic pore fluids ervoir (Zaback et al., 1993). For example, even in the presence
or a euxinic (anoxic and H2S-containing) water column is gener- of large fractionations during BSR with coupled disproportion-
ally considered an essential source of that sulfide. When present, ation reactions and diffusional inputs of sulfate, high δ34S values
the δ34S values of barite are generally consistent with a seawater for sulfate and sulfide can occur in pore-water and water-column
sulfate source, suggesting barite formation by reaction between systems with weak sulfate renewal relative to the rate of bac-
hydrothermal barium and sulfate in ambient seawater. terial consumption. By contrast, low δ34S values—i.e., high net
fractionations—are characteristic of systems where sulfate is
Sulfur Isotopes less limiting, such as in the water column of the Black Sea today
(Lyons, 1997). Anomalously high (strongly positive) δ34S data,
Dissimilatory bacterial sulfate reduction (BSR) typically as described in this report, require strong limitations in sulfate
yields hydrogen sulfide that is strongly depleted in 34S relative availability on either a local or global scale.
to the 34S/32S ratio in the residual parent sulfate. Under pure-cul-
ture laboratory conditions, this sulfide can be depleted by up to Sulfur Geochemical Records of Precambrian Ocean-
40‰–45‰ relative to the sulfate (Chambers et al., 1975; Can- Atmosphere Conditions—An Overview
field, 2001; Detmers et al., 2001). Instantaneous fractionations
of this magnitude are observed during BSR even at very low ini- Relevant details of the Precambrian sulfur cycle are provided
tial sulfate concentrations ranging down to 200 μM, or less than in a recent review paper by Lyons et al. (2004) and in the refer-
1% of the concentration in modern seawater. Below ~200 μM, ences cited within that synthesis. These details are also summa-
rates of sulfate reduction are limited by sulfate transport across rized schematically in Figure 2. Briefly, recent estimates by Shen
the bacterial cell membrane to the internal sites of enzymatically et al. (2001) place the onset of BSR at ca. 3.47 Ga (billion years
catalyzed fractionation. Under these conditions, most of the sul- ago). Despite this possibility, extremely low seawater sulfate
fate entering the cell is reduced, and the net fractionation is mini- concentrations (likely less than 200 μM) throughout the Archean
mized (Canfield, 2001; Habicht et al., 2002). beneath an O2-deficient atmosphere yielded only local examples
The 34S depletions observed in Phanerozoic sedimentary of the large S isotope fractionations that are often diagnostic
sulfides can exceed 60‰, which may be larger than what is of BSR (cf. Ohmoto et al., 1993, Kakegawa et al., 1998, and
Kakegawa and Ohmoto, 1999). These exceptions might record sedimentary sequences (Luepke and Lyons, 2001; Hurtgen et al.,
local, evaporative enrichments in seawater sulfate. By ca. 2.3 Ga, 2002, 2004; Kah et al., 2004; Gellatly and Lyons, 2005). During
sulfate concentrations within seawater increased to levels where the Phanerozoic, by contrast, δ34S fluctuations of similar mag-
fractionations of a few tens of per mil are commonly observed nitude are described by Claypool et al. (1980) for time periods
in bacteriogenic pyrite. This spread in the δ34S data is coincident of 107 to 108 years. By the late Proterozoic, atmospheric oxygen
with the so-called Great Oxidation Event (Holland, 2002; Farqu- levels rose sufficiently—with a corresponding rise in oceanic
har and Wing, 2003; Bekker et al., 2004), when redox conditions sulfate—to support increases in S isotope fractionations between
at Earth’s surface are thought by most workers to have shifted fun- sulfate and sulfide to values typical of the Phanerozoic (i.e., up to
damentally from reducing to oxidizing. At this time, weathering of and exceeding 60‰). Canfield and Teske (1996) attributed this
sulfide minerals exposed on the continents beneath an O2-contain- increased fractionation to evolution among non-photosynthetic
ing atmosphere would have led to appreciable increases in the flux sulfide-oxidizing bacteria, which accelerated the production
of sulfate to the ocean (Fig. 2)—although oxygen concentrations of intermediate S species and thus the disproportionation reac-
were still only a small fraction of present atmospheric levels. tions that can magnify the net isotope effect. However, overall
Although seawater sulfate concentrations rose in the early increases in sulfate concentration would also have supported
Proterozoic, its availability relative to the Phanerozoic ocean larger net fractionations, and a persistence of rapid δ34S vari-
remained low throughout most if not all of the Proterozoic. Evi- ability in the marine sulfate reservoir would make fractionations
dence for low sulfate concentrations—on the order of <5% to between coeval sulfate and sulfide difficult to evaluate (Lyons et
15% of the present-day value of 28–29 mM for the mid-Protero- al., 2004; Hurtgen et al., 2005).
zoic—is seen in the abundance of 34S-enriched pyrite deposited The Archean is generally regarded as a time of low oxygen
during this interval (Shen et al., 2002, 2003; this report) and in and sulfate availability in the ocean. Under these conditions, high
the rapid and large-magnitude S isotope variability observed concentrations of dissolved iron supported extensive deposition
stratigraphically in many Proterozoic sulfide- and sulfate-bearing of banded iron formations (BIFs). Across the Great Oxidation
Dominantly Increasingly
10
Sulfate Concentration
Reducing Oxidized
Environments Environments
8
6
(mM)
Great
4 Oxidation
Event
2
"cap carbonate"
0
ARCHEA N PALEOPROTEROZOIC MESOPROTEROZOIC NEOPROTEROZOIC
Ga 4.0 3.5 3.0 2.5 2.0 1.6 1.0 0.54

ARCHEA N PALE OPROTEROZOIC MESOPROTEROZOIC NEOPROTEROZOIC
40
20
-20
-40
Figure 2. Estimates of sulfate concentration and a schematic of δ34S values for bacteriogenic pyrite from the Archean and Proterozoic. Protero-
zoic sulfate concentrations are based on the Kah et al. (2004) model for S isotope variability as recorded in CAS. Archean estimates reflect the
low sulfate values required to explain the predominance of small bacterial S isotope fractionations observed during this interval (Habicht et al.,
2002). The pyrite δ34S schematic (lower figure) is based on the compilation of Shen et al. (2001). Increasing fractionations between sulfate and
sulfide (Δ34S) during BSR following the G.O.E. are inferred to record sulfate concentrations in the ocean that increased to values in excess of
~200 μM. Canfield and Teske (1996) attributed the increase in Δ34S in the Neoproterozoic to an increasing role by disproportionating bacteria in
the sulfur cycle, although this interpretation has recently been challenged (Hurtgen et al., 2005). Because most of the sulfate in the ocean derives
from continental weathering of sulfide minerals beneath an O2-containing atmosphere, increasing sulfate concentrations in seawater track the
oxygenation of the Proterozoic atmosphere.
Event ca. 2.3 Ga, oxygen in the atmosphere and thus sulfate in Canfield, 1998). Despite this fundamental shift, the redox transi-
the ocean increased to levels that supported widespread marine tion in the marine realm was complicated by the possibility of
BSR. A recent model argues, however, that under the compara- persistent or renewed oxygen deficiency within the isolated, ice-
tively low oxygen conditions of the Proterozoic atmosphere, the covered glacial ocean of the Neoproterozoic snowball Earth epi-
deep ocean may have remained free of oxygen for more than sodes (Hoffman et al., 1998; Hurtgen et al., 2002, 2004).
a billion years beyond the oxidation event (Canfield, 1998).
Canfield argued that within this anoxic environment, and in the SULFUR ISOTOPE TRENDS
presence of a growing oceanic sulfate reservoir, sulfate-reducing
bacteria would have thrived—giving rise to euxinic conditions In an earlier study, Lyons et al. (2000) showed a prepon-
throughout much of the deep Proterozoic water column (also derance of high δ34S values in disseminated bacteriogenic pyrite
Anbar and Knoll, 2002). Recently, Mo isotope data from Protero- extracted from organic-rich shales hosting the Sheep Creek and
zoic black shales have confirmed the possibility of widespread Soap Gulch SEDEX mineralization found within the ca. 1.47 Ga
oxygen deficiency and perhaps euxinia in the Proterozoic ocean lower Belt Supergroup. The massive pyrite within the mineralized
(Arnold et al., 2004). Such conditions may have persisted until intervals showed similar enrichments. On further examination,
the Neoproterozoic when a second major oxygenation of Earth’s strong stratigraphic trends were observed within the host shales,
surface resulted in a widely oxic deep ocean and ultimately the with upsection δ34S decreases of ~45‰ and 12‰ over thick-
appearance of multicellular organisms (Des Marais et al., 1992; nesses of 45 m and 134 m, respectively (Fig. 3). Subsequently,
A B C
0
200
150
Upper Prichard
Strauss and Schieber (1990)

400 200
250
600
Core Depth (m)

300
0 450
-5
10
15
20
25
0
350
Core Depth (m)
400
50 475
450
Core Depth (m)
Middle Prichard
100 500 500
550
525 -5 0 5 10 15 20
0 -5 0 5 10 15 20 25
34S (‰)
200 550
400 575
Prichard
-15 0 15 30 45 60
Lower
34S (‰)
600
800
-5 0 5 10 15 20 25
34S (‰)
Figure 3. Stratigraphic distributions of δ34S data for drill core samples of the (A) Prichard and (B and C) Newland Formations (ca. 1.47 Ga) of the
Belt Supergroup, northwestern United States. Prichard data are from the “main” Belt Basin and represent fine-scale pyrite/pyrrhotite laminae, lenses,
and disseminations within dark-gray to black argillites. The Newland data are from disseminated pyrite within black shales collected within the
Helena Embayment—the easternmost extension of the Belt Basin. The Newland shales are host to massive, SEDEX-type sulfide mineralization in
stratigraphic and spatial proximity to the data shown here. Nevertheless, the low-temperature, bacteriogenic pyrite in these shales and argillites ap-
pears to be unrelated to mineralizing, hydrothermal activity. Further details are available in Lyons et al. (2000) and Luepke and Lyons (2001).
analogous patterns have been found in pyrite and pyrrhotite

from the lower Belt Supergroup away from mineralized regions 18 n=71
in metamorphosed argillitic equivalents. These data show both Aberfeldy 9
+16.6
0
decreasing and increasing stratigraphic trends for δ34S (Fig. 3) 80
Balmat- n=182
(Luepke and Lyons, 2001). Collectively, the isotopic data and the 40
+17.4
Edwards 0 +14.7
relationships between organic carbon and pyrite sulfur in these
10 n=27
sediments were viewed as convincing evidence for marine inputs Jianshengpan 5
+23.4
into the Mesoproterozoic Belt Basin. The extent of marine depo- 0
sition within the Belt Supergroup has been debated for decades 6
n=12
Dongshengmiao 3
(Winston, 1990; Winston and Link, 1993). 0 +27.7
The abundance of 34S-enriched values in the lower Belt 6
n=13
Supergroup was assumed to reflect marine inputs into a highly Tanyaokou 3
0 +35.0
restricted setting that may have been shut off intermittently from 80
the open ocean. At the time of initial publication, the stratigraphic n=428
Sullivan 40
-1.1
trends reproduced in Figure 3 were viewed as evidence for tem- 0
poral variation in the strength of the local marine connection— 14

n=75
Belt 7
with the lowest δ34S values reflecting the most open conditions. 0 +7.3
In subsequent years, we have observed that large and often-sys- 20
n=47
Amjhore 10
tematic δ34S variations over relatively short stratigraphic intervals 0 +9.5
within single sedimentary units are not unusual for Proterozoic 30
n=318
15
sulfide accumulations (Ross et al., 1995; Strauss, 1997, 2002), McArthur River
0
+6.2
including SEDEX mineralization (Carr and Smith, 1977; Smith 40
n=516
et al., 1978; McGoldrick et al., 1999), and that these patterns may Mt. Isa 20
0
+9.1
have global implications (Kah et al., 2004; Lyons et al., 2004). 30
n=134
The pronounced 34S enrichments that dominate the sulfides Dugald River 15
+2.5
0
of the lower Belt Supergroup are also common throughout the 8
n=44
Proterozoic record (Lambert and Donnelly, 1991; Logan et al., Lady Loretta 4
+13.1
0
1995; Canfield, 1998; Canfield and Raiswell, 1999; Gorjan et -15 0 15 30 45
al., 2000; Shen et al., 2002, 2003; Strauss, 2002). Nowhere are δ34S (‰)
these enrichments better expressed than in SEDEX mineraliza-
tion. In Figure 4, we have compiled a large number of δ34S data Figure 4. Summary of δ34S data for sulfides from Proterozoic SEDEX
for SEDEX deposits spanning the Proterozoic. This compilation deposits. Data represent pyrite and the full range of base metal sul-
reflects a range of sulfide minerals, including ore-phase galena fides present, as well as disseminated sulfides within the host sedi-
and sphalerite, as well as sedimentary sulfides (dominantly pyrite) ments. References and location information are as follows: Aberfeldy,
Scotland (Willan and Coleman, 1983); Balmat-Edwards, United States
in host sediments away from the hydrothermal mineralization.
(Whelan et al., 1984); Jianshengpan, Dongshengmiao, and Tanya-
This selection of data is not intended to be a comprehensive sur- okou, China (Ding and Jiang, 2000); Sullivan, Canada (Campbell et
vey of Proterozoic SEDEX ore deposits. Instead, we have tried al., 1978, 1980); Belt Supergroup, United States (Lyons et al., 2000;
to assemble a representative suite of the largest data sets, with an Luepke and Lyons, 2001); Amjhore, India (Guha, 1971; Pandalai et
additional goal of providing broad temporal coverage. The most al., 1991); McArthur River, Australia (Smith and Croxford, 1975; Rye
and Williams, 1981; Eldridge et al., 1993; Shen et al., 2002); Mount
striking observations about Figure 4, and the focus of much of
Isa, Australia (Solomon, 1965; Smith et al., 1978; Andrew et al., 1989;
the discussion that follows, are the wide ranges of δ34S values for Davidson and Dixon, 1992; Painter et al., 1999; McShane, 1996);
each of the deposits and the predominance of 34S enrichments in Dugald River, Australia (Dixon and Davidson, 1996; Davidson and
the ore bodies and host shales. Dixon, 1992); Lady Loretta, Australia, (Carr and Smith, 1977; Scott et
al., 1985). With the exception of the Chinese deposits, the histograms
are stacked in a generally correct stratigraphic order. The Aberfeldy
IMPLICATIONS
deposit is late Neoproterozoic. The Balmat-Edwards, Sullivan, Belt,
and Amjhore deposits fall in a range of 1.3–1.5 Ga. McArthur River,
Stratigraphic Isotope Trends Mount Isa, Dugald River, and Lady Loretta have ages of 1.6–1.7 Ga.
Additional details about age relationships are available in Lyons et al.
The patterns of comparatively rapid δ34S variability (2004) and in the primary references. We note the abundance of 34S-
enriched sulfides present in these deposits. By contrast, pyrite forming
described above for Proterozoic metal sulfides—that is, varia-
in the modern water columns of the Black Sea and Cariaco Basin have
tions of tens of per mil over stratigraphic thicknesses of hundreds δ34S values of –30‰ to –40‰, with Δ34S values of ~–50‰ to –60‰
of meters or less—are also observed in samples of sulfate-S of (Lyons, 1997; Lyons et al., 2003).
Proterozoic age. These patterns have been observed both in gyp-
sum (Fig. 5) (Kah et al., 2001, 2004) and carbonate-associated
Meters
700
Coppermine
Flood Meters
600 Basalt 600
500 500
Figure 5. Left: Stratigraphic distribu-
tion of δ34S data for massive, bedded
400 400 gypsum and carbonate-associated sul-
Gypsum fate (CAS) extracted from dolomite of
CAS the ca. 1.2 Ga Society Cliffs Formation,
300 300
Bylot Supergroup, northeastern Arctic
Canada. Right: CAS data from dolo-
200 200
mite of the 1.3 Ga Dismal Lakes Group,
north-central Arctic Canada (after Kah
et al., 2001, 2004).
100 100
0 0
15 20 25 30 35 40 15 20 25 30 35 40
1.2 Ga Society Cliffs Formation 1.3 Ga Dismal Lakes Group
sulfate (CAS) (Fig. 5) (Hurtgen et al., 2002, 2004; Kah et al., and estimated mid-Proterozoic seawater sulfate concentrations
2004; Gellatly and Lyons, 2005). In contrast to earlier arguments that were only 5% to 15% of present-day levels.
for local controls on the patterns observed in the Belt basin Previous sulfur workers have suggested that upsection
(Lyons et al., 2000; Luepke and Lyons, 2001), the spatially and increases in δ34S observed over tens to hundreds of meters could,
temporally widespread nature of this style of isotopic variability in general, reflect progressive sulfate depletions under closed-
now suggests a global control. Specifically, these δ34S records system sulfate reduction—that is, isotopic evolution behaving as
preserved in sulfate and sulfide, including SEDEX deposits, may a Rayleigh distillation. This explanation is particularly attractive
reflect rapid changes in the isotopic composition of seawater sul- within the likely restricted, rifted tectonic settings that characterize
fate on a global scale. Such patterns of variability are unlike the most SEDEX deposits; however, it fails to explain the upsection
temporally broad, first-order trend of the Claypool et al. (1980) decreases in δ34S that are also observed. Furthermore, analogous
sulfate curve, which has long defined the Phanerozoic paradigm trends have been observed in platform carbonate (CAS) and
for seawater isotopic behavior (also Strauss, 1997; Kampschulte evaporite sequences, where local restriction and progressive 34S
and Strauss, 2004). Almost certainly the “Claypool paradigm” enrichment through local pyrite burial are difficult to imagine
misrepresents the extent of short-term δ34S variability recorded (Kah et al., 2001, 2004).
in younger rocks. We are exploring this possibility and see early Mass balance calculations that weigh sulfate availability
signs of greater complexity in the Paleozoic record (Gill et al., against total sulfide burial within shale hosts provide another
2006; also Kampschulte and Strauss, 2004), and marine barite is important constraint (Fig. 6). The large amounts of metal sul-
revealing the finer texture of δ34S variability during the Meso- fide within SEDEX deposits imply that a substantial fraction of
zoic and Cenozoic (Paytan et al., 1998, 2004). Nevertheless, the the total basinal sulfate/sulfide reservoir is supplying sulfur to
shorter-term δ34S shifts recorded in the barite data are smaller (on localized sites of enrichment. Consequently, sulfur supplied from
the order of 5‰) and longer (roughly 5–10 m.y.) than those of single volumes of local pore water and the immediately overly-
the Proterozoic and likely the Paleozoic. ing water column is typically not adequate to explain the large
A Proterozoic ocean with appreciably lower sulfate concen- amount of ore sulfur present. In shales unaffected by SEDEX pro-
trations, beneath an atmosphere with less oxygen, would be more cesses, however, we can explore stratigraphic isotopic trends in
vulnerable to rapid isotopic variability. A lower mass of sulfate light of simple assumptions about the sulfur source-sink relation-
in the ocean equates to a shorter residence time and thus higher ship. If pyrite in the shale is accumulating across the deep basin
sensitivity to δ34S variation as driven by flux terms such as pyrite floor, and mechanisms are not in place for localized enrichment,
burial, weathering inputs, and volcanogenic sulfur. Kah et al. we can estimate how much pyrite could be generated beneath
(2004) modeled the rate of δ34S variability (dδ34S/dt) observed at a given parcel of seawater. We performed this calculation for a
several Proterozoic localities, using both gypsum and CAS data, 1 m × 1 m area of the water column, assuming closed system
e.g., sulfate values. Under conditions of local isolation, the effects of

•1000 m water depth low oceanic sulfate availability would be exacerbated during sul-
•29 mmoles/L SO42– fate reduction, leading to sulfide δ34S data that would also track
•50% initial sed. porosity
Variables: the oceanic trend with little net fractionation between the parent
•2.7 g/cc detrital sed. density
•water depth sulfate and product sulfide. We have shown that many of the sul-
Calculated stratigraphic fide data are 34S enriched and that their ranges overlap strongly
•initial sulfate conc.
distribution of pyrite deriving with those for the sulfate. Rapid isotopic variability in the ocean,
in seawater from sulfate in a single water in combination with pyrite formation under conditions of limited
•% pyrite in shale column of seawater:
local and global sulfate availability, would yield the δ34S ranges
63 meters @
•sed. rate, strat. thick. 2 wt.% pyrite summarized in Figure 4.
or
~ 6 meters @ 10% modern
seawater sulfate concentration
Controls on 34S Enrichments
(2.9 mmoles/L)
Other workers have suggested that widespread 34S enrich-
Figure 6. Simple mass balance calculation to assess maximum pos- ments in sedimentary pyrite are attributable to comparatively
sible pyrite production from one basin-volume of seawater and specifi- low sulfate concentrations in the Proterozoic ocean. For example,
cally the pyrite’s stratigraphic distribution in a shale host (unaffected Shen et al. (2002) estimated sulfate levels of 0.5–2.4 mM based
by SEDEX mineralization) as predicted from concentrations in the on a model for sulfide accumulation in the late Paleoproterozoic
final rock. Details are provided in the text. deep waters of the McArthur Basin of northern Australia. Shen
et al. (2003), in a focused study of the Mesoproterozoic Roper
Group of the McArthur Basin, invoked a sulfate minimum zone—
a hypothesized interval within a stratified Proterozoic water
behavior for that parcel and quantitative conversion of sulfate to column characterized by exaggerated sulfate deficiencies—to
H2S and sequestration as FeS2 (Fig. 6). In other words, we are explain pronounced 34S enrichments observed within those sedi-
testing the oft-cited argument that systematic upsection increases ments. A Proterozoic sulfate minimum zone, as first suggested by
in δ34S within a black shale might reflect progressive evolution Logan et al. (1995), could have exacerbated the already low sul-
via BSR in a closed system. fate concentrations in the ocean. Hurtgen et al. (2002, 2004) noted
Our simple model assumes that sufficient Fe is present to that 34S enrichments in pyrite and large and rapid δ34S excursions
capture all of the sulfur in the system over the evolution of the recorded in CAS—specifically in response to the Neoproterozoic
closed system. In the example shown, water depth is assumed to glacial episodes—could reflect a suppressed riverine sulfate flux,
be 1000 m. Seawater sulfate concentration is chosen as today’s in combination with nearly complete sulfate reduction within
value (~29 mM). The non-pyrite (detrital sediment) component an isolated, anoxic global ocean (cf. Shields et al., 2004). These
is assumed to have a density of 2.7 g/cc, and the detrital sedi- “snowball Earth” patterns, although generally linked to the over-
ment is assigned an initial porosity of 50%. In this example, all of all low sulfate concentrations of the Proterozoic, are a subset of
the sulfide is precipitated as pyrite—other metal sulfides would the longer-term controls and patterns described here.
modify the results only slightly—and the iron sulfide is assumed Smaller fractionations during bacterial cycling of S would
to precipitate in the original pore space. also facilitate 34S enrichments in Proterozoic sulfides. Specifi-
Under these model conditions, the stratigraphic distribution cally, Canfield and Teske (1996), in a now debated model (e.g.,
of the pyrite resulting from a single water column of seawater Hurtgen et al., 2005), favored reduced fractionations prior to the
can be calculated for any wt.% pyrite in the final sample— Neoproterozoic—arguing that the enhanced 34S depletions dur-
taken as a reasonable black shale value of 2% in this example ing disproportionation reactions had minimal or no influence
(Fig. 6)—yielding accumulation over a stratigraphic thickness prior to further increases in Earth surface oxidation late in the
of only 63 m. This thickness is insufficient to explain the strati- Proterozoic (cf. Johnston et al., 2005). Regardless, Canfield and
graphic extents of many progressive 34S enrichments observed in Raiswell (1999), Strauss (2002), Shen et al. (2003), and others
the shale record (e.g., Fig. 3), and a shallower water column or have shown that fractionations at least as large as those possible
lower original seawater sulfate concentration could reduce the by BSR alone, and thus light δ34S values, are also observed dur-
stratigraphic estimates significantly. For example, our estimate ing the mid-Proterozoic, which suggests sulfate concentrations of
of ~10% modern sulfate concentrations for the mid-Proterozoic >200 μM (Habicht et al., 2002).
would reduce the thickness estimates by a factor of ten, thus add- Some diagenetic models suggest that extreme 34S enrich-
ing further credibility to arguments for global rather than local ments in sulfide analogous to those described here (e.g., Fig. 4)
controls on observed δ34S trends in the shales. may require sulfate concentrations of less than 1 mM (Habicht et
If the estimates for low sulfate in the Proterozoic ocean are al., 2002; see Canfield, 2004). Because most pyrite in marine sys-
correct, the stratigraphic trends preserved within the metal sul- tems forms early, in the surface sediments or in the water column,
fides of shales and ore deposits could be tracking global seawater δ34S values tend to be 34S depleted in Phanerozoic sediments.
δ34S values commonly range between –20‰ and –40‰ in many within sedimentary systems support AOM and thus anomalously
modern euxinic sediments and Phanerozoic black shales (Sage- high subsurface rates of BSR and comparatively high associ-
man and Lyons, 2003). For example, although the modern Black ated δ34S values for hydrogen sulfide and pyrite (Aharon and Fu,
Sea is only weakly linked to the Mediterranean Sea, the balance 2000, 2003; Arvidson et al., 2004; M. Formolo, 2006, personal
between rates of net sulfate reduction and sulfate replenishment, commun.). It is interesting to note that pyrite δ34S values as high
in combination with pyrite that forms primarily in the water as those found commonly in the Proterozoic are best developed
column, yields typical open-system, 34S-depleted pyrite that is today where subsurface, AOM-driven secondary sulfur is over-
almost 60‰ lighter than the coeval seawater sulfate (Calvert et printing freshwater sediments (Jørgensen et al., 2004). Despite
al., 1996; Lyons, 1997; Wilkin and Arthur, 2001). a lack of direct evidence, high methane availability may have
The 1 mM sulfate maximum is lower by roughly a factor of supported high rates of Proterozoic BSR, which, in combination
two than the mid-Proterozoic predictions summarized here (e.g., with lower overall sulfate concentrations, basin restriction, and
Kah et al., 2004). Although not a big difference, these results sug- strong hydrogen sulfide retention, would have favored high δ34S
gest that SEDEX deposits and Proterozoic sediments in general, values for the metal sulfides.
through quantitative sulfate reduction and efficient hydrogen sul- In modern deep-water, cold seep environments, bacterial
fide retention as metal sulfides, may exceed typical, more-recent mats of sulfide-oxidizing Beggiatoa are often abundant, reflect-
early diagenetic environments in their ability to generate extreme ing the high levels of methane and associated BSR that typify the
34
S enrichments. In modern marine sediments, as much as 95% near-surface environment immediately below the mats (Larkin
of hydrogen sulfide generated is reoxidized (Jørgensen et al., et al., 1994; Zhang et al., 2005). Such mats form today chemo-
1990), in part because of limited supplies of reactive iron (Can- autotrophically—deriving energy from oxidation of the copious
field et al., 1992). Under lower sulfate/sulfide conditions, iron sulfide formed during AOM. In addition to their abundances
limitation becomes less of a factor, particularly if iron and other of pyrite, modern cold seep and gas hydrate localities are also
sulfide-reactive metals are available in great abundance—sug- known for enhanced, often extensive barite deposition (Dickens,
gesting anomalously high burial efficiencies for Proterozoic H2S 2001; Greinert et al., 2002; Torres et al., 2003).
(Hurtgen et al., 2005). Also, the very high sedimentation rates Organic biomarkers consistent with sulfide-oxidizing bacte-
expected in the isolated rift settings and the lack of Proterozoic ria were recently reported from inter-ore beds at the giant Protero-
bioturbation, regardless of benthic O2 levels, would have a pro- zoic McArthur River Zn-Pb SEDEX deposit (Logan and Hinman,
nounced effect on sulfate transport, reactive carbon availability, 2001), and a diverse microfossil assemblage is preserved in black
and sulfide retention within the sediments and could enhance 34S chert nodules intimately associated with Zn-Pb mineralization
enrichment (Hurtgen et al., 2005). Furthermore, rates of BSR (Oehler and Logan, 1977). Additional petrographic evidence for
associated with SEDEX mineralization may have been anoma- a microbial role in sulfide formation in the Proterozoic deposits
lously high. comes from a diversity of macro- and micro-textures in iron and
base metal sulfides (McGoldrick, 1999). In particular, “crinkly-
The Possible Role of Methane wavy” laminated pyrite beds, which are ubiquitous in the north-
ern Australian deposits and the eastern Belt Basin, are interpreted
Reduced concentrations of sulfate in the early ocean would as pyritized microbial mats (Schieber, 1986, 1990; McGoldrick,
have favored the production and preservation of methane (Habi- 1999; McGoldrick et al., 1999) (Fig. 7).
cht et al., 2002). Sulfate-reducing bacteria compete with metha-
nogens for metabolizable organic compounds, and anaerobic Other Evidence for Low Proterozoic Sulfate
oxidation of methane (AOM) occurs through a consortium of
microorganisms, with BSR playing a central role (Boetius et Low sulfate concentrations in the Proterozoic ocean are fur-
al., 2000). Also, methanogens are anaerobes, and methanotropic ther suggested by other independent evidence: (1) a comparative
bacteria readily oxidize methane in the presence of oxygen. scarcity of massive bedded gypsum deposits prior to the mid
Conditions of low oxygen, in combination with limited sulfate to late Proterozoic (Kah et al., 2004); (2) the generally lower
availability, would support a methane-rich ocean-atmosphere amounts of barite associated with Proterozoic SEDEX deposits
system (Pavlov et al., 2000; Habicht et al., 2002), which may relative to Phanerozoic examples (Lydon, 1996)—although bar-
have persisted, albeit at a reduced level, throughout most of the ium availability also plays an important role (Cooke et al., 2000);
Proterozoic (Pavlov et al., 2003). The possibility of high levels (3) possibly the low CAS concentrations in Proterozoic carbon-
of atmospheric methane during the Precambrian was indirectly ates (Hurtgen et al., 2002, 2004; Pavlov et al., 2003; Gellatly and
advanced through the paleosol study of Rye et al. (1995), which Lyons, 2005); and (4) perhaps even the widespread occurrence of
has been challenged recently by Ohmoto et al. (2004). dolomite in Precambrian rocks compared to younger sediments.
Although a connection to the Proterozoic sulfur geochem- Finally, carbon-sulfur relationships summarized in Strauss (2002)
istry is highly speculative, we also note that the most extreme suggest marine deposition during the Paleo- and Mesoprotero-
34
S-enrichments found in modern marine sediments occur in zoic that was characterized by lower concentrations of seawater
association with high fluxes of methane. High fluxes of methane sulfate than in the latest Precambrian and Phanerozoic. Lyons et
Figure 7. Core photos showing microbial mat-like textures preserved in pyrite and base metal sulfide ore from SEDEX deposits (Schieber, 1990;
McGoldrick, 1999). (A) Unpolished slabbed core showing crinkly laminations in bedded pyrite from the late Paleoproterozoic Lady Loretta
Zn-Pb-Ag deposit, northwest Queensland (width of core is ~45 mm). (B) Polished slabbed core from late Paleoproterozoic Grevillea Zn-Pb-Ag
prospect (northwest Queensland) with crinkly laminated pyrite (width of sample ~20 mm). (C) Polished slabbed core of laminated high-grade
Zn-Pb sulfide ore from the Lady Loretta deposit (width of sample ~20 mm); note two prominent crinkly laminated pyrite bands (py) and partly
disrupted laminated base metal sulfide and pyrite domains (bms). (D) slabbed core from the Mesoproterozoic Sheep Creek Cu-Co, Montana
(width of sample ~55 mm); note compactional draping of pyrite laminae around early-formed barite laths.
al. (2000) described analogous high C/S ratios for the shales of this from a simple model that required only estimates of oxygen
the lower Belt Supergroup (cf. Hurtgen et al., 2005). concentrations in the Proterozoic atmosphere and respiratory
oxygen demand in the ocean—but without direct evidence. He
Paleoredox further suggested that despite the relatively low sulfate condi-
tions, the combination of a growing marine sulfate reservoir in
Proterozoic seawater sulfate concentrations, while low, the presence of deep-water anoxia may have led to widespread
were at least a factor of ten higher than those present during the BSR and thus ocean-scale euxinia. Subsequent to Canfield’s
Archean, as would be expected with the increased oxidation at initial prediction, Mesoproterozoic euxinia was shown by Shen
Earth’s surface beginning ca. 2.3 Ga. This extreme sulfur limita- et al. (2003), Poulton et al. (2004), and Brocks et al. (2005) to
tion alone may explain the absence of Archean SEDEX mineral have occurred at least on the scale of an individual basin and per-
deposits. Nonetheless, despite this fundamental shift in redox haps more generally throughout the ocean (Arnold et al., 2004).
(Holland, 2002; Farquhar and Wing, 2003; Bekker et al., 2004), Canfield (1998) extrapolated his model predictions for eux-
a growing body of evidence is suggesting that oxygen deficiency inia to suggest that BIFs may have disappeared ca. 1.8 Ga not
may have persisted in the deep ocean for as long as 1.5 b.y. fol- because of comprehensive oxygenation of the ocean but rather
lowing the Great Oxidation Event. Canfield (1998) first predicted through a buildup of H2S in the water column, which would have
modulated Fe solubility through pyrite formation. As such, the conduits for possible methane and other hydrocarbon migration
1.8 Ga threshold would represent the time when increasing sul- would have been ideal for enhancing metal availability and per-
fur availability overcame the Fe fluxes to the ocean, including haps hydrogen sulfide production. A close relationship between
hydrothermal inputs (Isley and Abbott, 1999; Abbott and Isley, SEDEX-type mineralization and ancient hydrocarbon seeps or
2001; cf. Lowell and Keller, 2003). Consequently, it may not be vents has been suggested by others (e.g., Johnson et al., 2004).
fortuitous that SEDEX deposits first appear roughly when BIFs
disappear—with both phenomena marking the generally wide- Comparisons with Younger Deposits
spread availability of H2S. Also of relevance, Condie et al. (2001)
argued for a prominent maximum in Precambrian black shale Temporal selectivity of SEDEX mineralization is, to some
deposition during a window from 2000 to 1700 Ma, which they extent, biased by a few very large and well-studied deposits. Sim-
linked at least partially to superplume activity. ilarly, the number of SEDEX deposits for a given time interval
An additional and essential benefit of widespread anoxia may be less meaningful than the total mass of mineralization.
is the enhanced preservation of sulfide minerals residing on or Nevertheless, peak periods of SEDEX mineralization seem clear
near the seafloor. It is well known that sulfides deposited on the and include both Proterozoic and Paleozoic time intervals (Fig. 1)
seafloor today along mid-ocean ridges are rapidly oxidized off (Lydon, 1996). The most straightforward conclusion from this
axis beneath the oxic water column (Alt et al., 1989; Alt, 1994; distribution is that the Proterozoic onset of large-scale SEDEX
Edwards et al., 2003). mineralization almost certainly coincides with the early to mid-
Proterozoic accumulation of sulfate and sulfide within the global
Paleotectonics and the Importance of Local Controls ocean system. Although sulfate was increasing over this interval,
the predominance of 34S-enriched sulfides suggests that sulfate
Although tectonic conditions must also be favorable for the remained well below modern levels.
production of hydrothermal, mineralizing fluids, high levels of The biggest complication in an otherwise consistent story
rift-related heat flow are certainly not unique to the sites and time is that Phanerozoic deposits are also known for their 34S enrich-
windows of SEDEX ore deposition. Intracratonic and continental ments—for example, Red Dog, Rammelsberg, and the deposits
margin, fault-controlled basins and troughs are common in rift of the Selwyn Basin, all of which are Paleozoic in age (Anger et
zones throughout geologic history, and patterns of rifting are often al., 1966; Goodfellow, 1987; Eldridge et al., 1988; Jennings and
intimately related to supercontinent breakup. It may be relevant, King, 2002) (Fig. 8). Given this, it would be easy to de-empha-
however, that the mid-Proterozoic maximum in SEDEX miner- size the importance of global controls in favor of local param-
alization coincides with a proposed peak in superplume activity eters, such as rift-related basin restriction with associated anoxia
and thus oceanic hydrothermal fluxes (Isley and Abbott, 1999; and local sulfate deficiencies. Although the abundant barite asso-
Condie et al., 2001; see also Barley and Groves, 1992). Isolated, ciated with Paleozoic deposits suggests that sulfate was increas-
rift basins may also have favored the development of local anoxia, ing in the ocean, our recent work (Gill et al., 2006) and the stud-
although the importance of this is less clear for the Proterozoic ies of Horita et al. (2002), Lowenstein et al. (2003), Brennan et
deposits when deep-water oxygen deficiency may have been al. (2004), and Canfield (2004) suggest that sulfate increased but
common. Finally, rift-associated hydrothermal activity and fault remained low relative to the modern ocean well into the Paleo-
Figure 8. Summary of δ34S data for selected

Phanerozoic SEDEX deposits. Data repre-
sent pyrite and base metal sulfides. Refer-
ences and location information are as fol-
lows: Anvil District, Canada (Shanks et al.,
1987); Howards Pass, Canada (Goodfellow
and Jonasson, 1987); Rammelsberg, Germa-
ny (Anger et al., 1966); Meggen, Germany
(Buschendorf et al., 1963); Red Dog, Unit-
ed States (Kelley et al., 2004); Reocin and
Troya, Spain (Velasco et al., 1994).
zoic. As for the Proterozoic, restricted, rift basin settings during sistent, primary observations have emerged, which provide a
the Phanerozoic would have intensified the conditions of com- framework for further exploring the origins and distributions of
paratively low global oceanic sulfate availability. SEDEX mineralization:
The predominance of 34S-enriched metal sulfides of Phanero- 1. The absence of volumetrically significant SEDEX
zoic SEDEX deposits highlights even more clearly the necessity for mineralization during the Archean can be attributed to
high rates of BSR in combination with anomalously high efficiency limitations in sulfide availability, which ultimately reflect
in sulfide mineralization. Again, high, fault-controlled fluxes of seawater sulfate concentrations beneath an oxygen-defi-
methane and other hydrocarbons would drive high microbial pro- cient atmosphere that were perhaps less than 1% of those
duction of hydrogen sulfide at or below the sediment-water inter- present today.
face, and associated hydrothermal activity would supply the metals 2. Following the fundamental shift to a more oxidizing atmo-
necessary for sulfide mineralization. Most Phanerozoic SEDEX sphere ca. 2.3 Ga, sulfate fluxes to the ocean increased
deposits are of the “Selwyn-type” (Cooke et al., 2000), and there- significantly through continental weathering of sulfides
fore the ore-forming fluids carried metals and reduced S together. exposed in the presence of oxygen. Despite increased
Mixing this ore-fluid sulfide with sedimentary sulfide formed at the atmospheric availability, oxygen may have remained
sites of ore deposition may produce S isotope patterns that are inher- scarce in the deep ocean for at least another billion years.
ently more complex than those seen in “McArthur-type” deposits. This deep ocean condition was at least partially a conse-
Mechanisms of SEDEX mineralization may intrinsically quence of atmospheric oxygen levels that remained far
lead to strong local sulfate deficiencies and thus high 34S enrich- below those present today. Increasing sulfate combined
ments—regardless of the age. Zaback et al. (1993) and Jørgensen with an O2-poor ocean could have spawned widespread
et al. (2004) remind us, however, that “closed-system” isotopic euxinia throughout the Proterozoic ocean. The disappear-
behavior often marks settings where sulfate resupply is occur- ance of banded iron formations ca. 1.8 Ga may mark the
ring—but via fluxes that are overwhelmed by high rates of con- attainment of a critical threshold in H2S accumulation in
sumption by BSR. This model, for example, helps explain large the ocean.
amounts of 34S-enriched sulfide, as are observed in SEDEX 3. Given conditions that were tectonically favorable to hydro-
deposits, particularly if H2S retention is high (Fig. 4). Today, thermal sourcing of metals, the appearance of SEDEX
such high rates of BSR are observed in association with AOM, mineralization ca. 1.8 Ga likely marks the ubiquity of H2S
particularly where other easily metabolized hydrocarbons are in the marine system—including oxygen-deficient bottom
also released through seeps (Formolo et al., 2004). One could waters—which favored the production and preservation
speculate that the fault control that is central to SEDEX mineral- of metal sulfide mineralization on the seafloor.
ization plays a major role as a conduit for hydrocarbon migration 4. The common occurrence of 34S enrichments within
in these organic-rich, shale-dominated settings. SEDEX sulfides is viewed as a combined product of
The temporal and spatial distribution of Phanerozoic SEDEX globally low seawater sulfate, high rates of bacterial sul-
mineralization almost certainly reflects the more strongly oxidizing fate reduction, and locally enhanced sulfate limitations
conditions of the ocean, which would influence both H2S availabil- under the partially closed rifted marine basins in which
ity and preservation of minerals precipitated at or near the seafloor. SEDEX mineralization occurs. Limited sulfate availabil-
As a result, younger deposits may be limited to areas and episodes of ity may also be recorded in the lower amounts of barite
basin-scale anoxia (Goodfellow and Jonasson, 1984; Turner, 1992), associated with Proterozoic deposits.
or they may have required an oxygen-shielding cap, either through 5. Despite the possibility of a widely sulfidic deep ocean,
barite deposition or mineralization below the sediment-water inter- the high rates of sulfate reduction expressed in the 34S
face. The drop-off in abundance of Proterozoic deposits following enrichments must, to an appreciable extent, reflect local
the peak at 1800–1600 Ga is more difficult to explain. Perhaps, in bacterial activity. Enhanced local hydrocarbon fluxes
contrast to arguments for persistent euxinia, conditions were already and/or globally pervasive methane could have supported
too oxidizing to facilitate widespread bottom-water sulfide produc- these high rates.
tion and enhanced metal sulfide preservation. Proterozoic deposits 6. Following a mid-Proterozoic peak in SEDEX mineral-
are relatively lacking in barite, as would be expected given the com- ization, the generally waning occurrence of these depos-
paratively low sulfate availability in the ocean. Thus, any protection its throughout the latter part of the Proterozoic and only
from oxygen offered by a barite cap would be minimal. intermittent Phanerozoic occurrences may reflect increas-
ing oxygenation of deep seawater and corresponding
SUMMARY decreases in ambient H2S. Under these conditions, basins
with restricted circulation and local bottom-water anoxia
Our goal here was to explore the first-order temporal pat- become increasingly important.
terns observed for SEDEX mineralization in light of the most 7. Paleozoic SEDEX sulfides commonly bear 34S enrich-
recent models for early ocean-atmosphere chemistry. Despite ments analogous to those observed in the Proterozoic.
the many unanswered questions, a number of internally con- The enrichments are partially a product of the local
reservoir properties—specifically inhibited seawater/sul- Arnold, G.L., Anbar, A.D., Barling, J., and Lyons, T.W., 2004, Molybdenum
isotope evidence for widespread anoxia in mid-Proterozoic oceans: Sci-
fate exchange with the open ocean. However, as for the ence, v. 304, p. 87–90, doi: 10.1126/science.1091785.
Proterozoic, lower global seawater sulfate availability Arvidson, R.S., Morse, J.W., and Joye, S.B., 2004, The sulfur biogeochemistry
relative to today and high rates of bacterial sulfate reduc- of chemosynthetic cold seep communities, Gulf of Mexico, U.S.A: Marine
Chemistry, v. 87, p. 97–119, doi: 10.1016/j.marchem.2003.11.004.
tion are suggested. It is an oversimplification to imagine Barley, M.E., and Groves, D.I., 1992, Supercontinent cycles and the distri-
that restricted marine conditions a priori favor extreme bution of metal deposits through time: Geology, v. 20, p. 291–294, doi:
sulfate limitations. For example, Paleozoic deposits show 10.1130/0091-7613(1992)020<0291:SCATDO>2.3.CO;2.
Bekker, A., Holland, H.D., Wang, P.-L., Rumble, D., III, Stein, H.J., Hannah,
large amounts of associated barite. Furthermore, purely J.L., Coetzee, L.L., and Beukes, N.J., 2004, Dating the rise of atmospheric
closed system behavior is inconsistent with large amounts oxygen: Nature, v. 427, p. 117–120, doi: 10.1038/nature02260.
of metal sulfide formation. Instead, we emphasize the Boetius, A., Ravenschlag, K., Schubert, C.J., Rickert, D., Widdel, F., Gieseke,
A., Amann, R., Jørgensen, B.B., Witte, U., and Pfannkuche, O., 2000, A
importance of anomalously high rates of bacterial reduc- methane microbial consortium apparently mediating anaerobic oxidation
tion and H2S retention under conditions of limited but not of methane: Nature, v. 407, p. 623–626, doi: 10.1038/35036572.
Brennan, S.T., Lowenstein, T.K., and Horita, J., 2004, Seawater chemistry and
isolated sulfate supply. the advent of biocalcification: Geology, v. 32, p. 473–476, doi: 10.1130/
G20251.1.
ACKNOWLEDGMENTS Brocks, J.J., Love, G.D., Summons, R.E., Knoll, A.H., Logan, G.A., and
Bowden, S.A., 2005, Biomarker evidence for green and purple sulphur
bacteria in a stratified Palaeoproterozoic sea: Nature, v. 437, p. 866–870.
Support for this project was provided by NSF grants EAR- Brunner, B., and Bernasconi, S.M., 2005, A revised isotope fractionation model
9596079 and EAR-9725538 (Lyons and Kah). The authors for dissimilatory sulfate reduction in sulfate reducing bacteria: Geochi-
mica et Cosmochimica Acta, v. 69, p. 4759–4771.
benefited from many conversations with A. Anbar, D. Canfield, Buschendorf, Fr., Nielsen, H., Puchelt, H., and Ricke, W., 1963, Schwefel-
M. Hurtgen, and J. Luepke. D. Winston and G. Zieg are thanked isotopen-untersuchungen am pyrit-sphalerit-baryt-lager Meggen/Lenne
for introducing the authors to the rocks of the Belt Supergroup. (Deutschland) und an verschiedenen Devon-evaporiten: Geochimica et
Cosmochimica Acta, v. 27, p. 501–523.
Research into the Proterozoic Zn-Pb-Ag deposits of north- Calvert, S.E., Thode, H.D., Yeung, D., and Karlin, R.E., 1996, A stable isotope
ern Australia has been supported by grants from the Austra- study of pyrite formation in the Late Pleistocene and Holocene sediments
of the Black Sea: Geochimica et Cosmochimica Acta, v. 60, p. 1261–
lian Research Council (ARC) and Australian Mineral Industry 1270, doi: 10.1016/0016-7037(96)00020-8.
Research Association (AMIRA) to McGoldrick. Frank Corsetti Campbell, F.A., Ethier, V.G., Krouse, H.R., and Both, R.A., 1978, Isotopic
and Barry Maynard provided insightful reviews, and we thank composition of sulfur in the Sullivan orebody, British Columbia: Eco-
nomic Geology and the Bulletin of the Society of Economic Geologists,
Steve Kesler and Hiroshi Ohmoto for their many efforts in v. 73, p. 246–268.
assembling this volume. Campbell, F.A., Ethier, V.G., and Krouse, H.R., 1980, The massive sulfide zone;
Sullivan orebody: Economic Geology and the Bulletin of the Society of
REFERENCES CITED Canfield, D.E., 1998, A new model for Proterozoic ocean chemistry: Nature,
v. 396, p. 450–453, doi: 10.1038/24839.
Abbott, D., and Isley, A., 2001, Oceanic upwelling and mantle-plume activity: Canfield, D.E., 2001, Isotope fractionation by natural populations of sulfate-
Paleomagnetic tests of ideas on the source of the Fe in early Precambrian reducing bacteria: Geochimica et Cosmochimica Acta, v. 65, p. 1117–
iron formations, in Ernst, R.E., and Buchan, K.L., eds., Mantle plumes: 1124, doi: 10.1016/S0016-7037(00)00584-6.
Their identification through time: Geological Society of America Special Canfield, D.E., 2004, The evolution of the Earth surface sulfur reservoir: Amer-
Paper 352, p. 323–339. ican Journal of Science, v. 304, p. 839–861.
Aharon, P., and Fu, B., 2000, Microbial sulfate reduction rates and sulfur and Canfield, D.E., and Raiswell, R., 1999, The evolution of the sulfur cycle: Amer-
oxygen isotope fractionations at oil and gas seeps in deepwater Gulf ican Journal of Science, v. 299, p. 697–723.
of Mexico: Geochimica et Cosmochimica Acta, v. 64, p. 233–246, doi: Canfield, D.E., and Teske, A., 1996, Late Proterozoic rise in atmospheric oxy-
10.1016/S0016-7037(99)00292-6. gen concentration inferred from phylogenetic and sulphur-isotope stud-
Aharon, P., and Fu, B., 2003, Sulfur and oxygen isotopes of coeval sulfate-sulfide ies: Nature, v. 382, p. 127–132, doi: 10.1038/382127a0.
in pore fluids of cold seep sediments with sharp redox gradients: Chemi- Canfield, D.E., and Thamdrup, B., 1994, The production of 34S-depleted sul-
cal Geology, v. 195, p. 201–218, doi: 10.1016/S0009-2541(02)00395-9. fide during bacterial disproportionation of elemental sulfur: Science,
Alt, J.C., 1994, A sulfur isotopic profile through the Troodos ophiolite, v. 266, p. 1973–1975.
Cyprus: Primary composition and the effects of seawater hydrothermal Canfield, D.E., Raiswell, R., and Bottrell, S., 1992, The reactivity of sedimen-
alteration: Geochimica et Cosmochimica Acta, v. 58, p. 1825–1840, doi: tary iron minerals toward sulfide: American Journal of Science, v. 292,
10.1016/0016-7037(94)90539-8. p. 659–683.
Alt, J.C., Anderson, T.F., and Bonnell, L., 1989, The geochemistry of sulfur Carr, G.R., and Smith, J.W., 1977, A comparative isotopic study of the Lady
in a 1.3 km section of hydrothermally altered oceanic crust, DSDP Hole Loretta zinc-lead-silver deposit: Mineralium Deposita, v. 12, p. 105–110,
504B: Geochimica et Cosmochimica Acta, v. 53, p. 1011–1023, doi: doi: 10.1007/BF00204509.
10.1016/0016-7037(89)90206-8. Chambers, L.A., Trudinger, P.A., Smith, J.W., and Burns, M.S., 1975, Fraction-
Anbar, A.D., and Knoll, A.H., 2002, Proterozoic ocean chemistry and evolu- ation of sulfur isotopes by continuous cultures of Desulfovibrio desulfuri-
tion: A bioinorganic bridge?: Science, v. 297, p. 1137–1142, doi: 10.1126/ cans: Canadian Journal of Microbiology, v. 21, p. 1602–1607.
science.1069651. Claypool, G.E., Holser, W.T., Kaplan, I.R., Sakai, H., and Zak, I., 1980, The age
Andrew, A.S., Heinrich, C.A., Wilkins, R.W.T., and Patterson, D.J., 1989, Sul- curves of sulfur and oxygen isotopes in marine sulfate and their mutual
fur isotope systematics of copper ore formation at Mount Isa, Australia: interpretations: Chemical Geology, v. 28, p. 199–260, doi: 10.1016/0009-
Economic Geology and the Bulletin of the Society of Economic Geolo- 2541(80)90047-9.
gists, v. 84, p. 1614–1626. Condie, K.C., Des Marais, D.J., and Abbott, D., 2001, Precambrian super-
Anger, G., Nielsen, H., Puchelt, H., and Ricke, W., 1966, Sulfur isotopes in the plumes and supercontinents: A record in black shales, carbon isotopes,
Rammelsberg ore deposit (Germany): Economic Geology and the Bul- and paleoclimates?: Precambrian Research, v. 106, p. 239–260, doi:
letin of the Society of Economic Geologists, v. 61, p. 511–536. 10.1016/S0301-9268(00)00097-8.
Cooke, D.R., Bull, S.W., Large, R.R., and McGoldrick, P.J., 2000, The impor- Guha, J., 1971, Sulfur isotope study of the pyrite deposit of Amjhore, Shahbad
tance of oxidized brines for the formation of Australian Proterozoic strati- District, Bihar, India: Economic Geology and the Bulletin of the Society
form sediment-hosted Pb-Zn (Sedex) deposits: Economic Geology and of Economic Geologists, v. 66, p. 326–330.
the Bulletin of the Society of Economic Geologists, v. 95, p. 1–17. Habicht, K.S., and Canfield, D.E., 2001, Isotope fractionation by sulfate-
Davidson, G.J., and Dixon, G.H., 1992, Two sulphur isotope provinces deduced reducing natural populations and the isotopic composition of sulfide
from ores in the Mount Isa eastern succession, Australia: Mineralium in marine sediments: Geology, v. 29, p. 555–558, doi: 10.1130/0091-
Deposita, v. 27, p. 30–41, doi: 10.1007/BF00196078. 7613(2001)029<0555:IFBSRN>2.0.CO;2.
Des Marais, D.J., Strauss, H., Summons, R.E., and Hayes, J.M., 1992, Carbon Habicht, K.S., Gade, M., Thamdrup, B., Berg, P., and Canfield, D., 2002, Cali-
isotope evidence for the stepwise oxidation of the Proterozoic environ- bration of sulfate levels in the Archean Ocean: Science, v. 298, p. 2372–
ment: Nature, v. 359, p. 605–609, doi: 10.1038/359605a0. 2374, doi: 10.1126/science.1078265.
Detmers, J., Brüchert, V., Habicht, K.S., and Kuever, J., 2001, Diversity of Hoffman, P.F., Kaufman, A.J., Halverson, G.P., and Schrag, D.P., 1998, A Neo-
sulfur isotope fractionations by sulfate-reducing prokaryotes: Applied proterozoic snowball Earth: Science, v. 281, p. 1342–1346, doi: 10.1126/
and Environmental Microbiology, v. 67, p. 888–894, doi: 10.1128/ science.281.5381.1342.
AEM.67.2.888-894.2001. Holland, H.D., 2002, Volcanic gases, black smokers, and the Great Oxidation
Dickens, G.R., 2001, Sulfate profiles and barium fronts in sediment on the Event: Geochimica et Cosmochimica Acta, v. 66, p. 3811–3826, doi:
Blake Ridge: Present and past methane fluxes through a large gas hydrate 10.1016/S0016-7037(02)00950-X.
reservoir: Geochimica et Cosmochimica Acta, v. 65, p. 529–543, doi: Horita, J., Zimmermann, H., and Holland, H.D., 2002, Chemical evolution of
10.1016/S0016-7037(00)00556-1. seawater during the Phanerozoic: Implications from the record of marine
Ding, T.-P., and Jiang, S.-Y., 2000, Stable isotope study of the Langshan evaporites: Geochimica et Cosmochimica Acta, v. 66, p. 3733–3756, doi:
polymetallic mineral district, Inner Mongolia, China: Resource Geology, 10.1016/S0016-7037(01)00884-5.
v. 50, p. 25–38. Hurtgen, M.T., Arthur, M.A., and Halverson, G.P., 2005, Neoproterozoic sulfur
Dixon, G.H., and Davidson, G.J., 1996, Stable isotope evidence for thermo- isotopes, the evolution of microbial sulfur species, and the burial effi-
chemical sulfate reduction in the Dugald River (Australia) strata-bound ciency of sulfide as sedimentary pyrite: Geology, v. 33, p. 41–44, doi:
shale-hosted zinc-lead deposit: Chemical Geology, v. 129, p. 227–246, 10.1130/G20923.1.
doi: 10.1016/0009-2541(95)00177-8. Hurtgen, M.T., Arthur, M.A., and Prave, A.R., 2004, The sulfur isotope
Edwards, K.J., McCollom, T.M., Konishi, H., and Buseck, P.R., 2003, Seafloor composition of carbonate-associated sulfate in Mesoproterozoic to
bioalteration of sulfide minerals: Results from in situ incubation studies: Neoproterozoic carbonates from Death Valley, California, in Amend,
Geochimica et Cosmochimica Acta, v. 67, p. 2843–2856, doi: 10.1016/ J.P., Edwards, K.J., and Lyons, T.W., eds., Sulfur biogeochemistry—
S0016-7037(03)00089-9. Past and present: Geological Society of America Special Paper 379,
Eldridge, C.S., Compston, W., Williams, I.S., Both, R.A., Walshe, J.L., and p. 177–194.
Ohmoto, H., 1988, Sulfur isotope variability in sediment-hosted massive Hurtgen, M.T., Arthur, M.A., Suits, N., and Kaufman, A.J., 2002, The sulfur
sulfide deposits as determined using the ion microprobe SHRIMP: I. An isotopic composition of Neoproterozoic seawater sulfate: Implications for
example from the Rammelsberg orebody: Economic Geology and the a snowball Earth?: Earth and Planetary Science Letters, v. 203, p. 413–
Bulletin of the Society of Economic Geologists, v. 83, p. 443–449. 429, doi: 10.1016/S0012-821X(02)00804-X.
Eldridge, C.S., Williams, N., and Walshe, J.L., 1993, Sulfur isotope variabil- Isley, A.E., and Abbott, D.H., 1999, Plume-related mafic volcanism and the
ity in sediment-hosted massive sulfide deposits as determined using the deposition of banded iron formation: Journal of Geophysical Research,
ion microprobe SHRIMP: II. A study of the H.Y.C. Deposit at McArthur v. 104, p. 15461–15477, doi: 10.1029/1999JB900066.
River, Northern Territory, Australia: Economic Geology and the Bulletin Jennings, S., and King, A.R., 2002, Geology, exploration history and future
of the Society of Economic Geologists, v. 88, p. 1–26. discoveries in the Red Dog district, western Brooks Range, Alaska, in
Farquhar, J., and Wing, B.A., 2003, Multiple sulfur isotopes and the evolution Cooke, D., and Pongratz, J., eds., Giant ore deposits: Characteristics, gen-
of the atmosphere: Earth and Planetary Science Letters, v. 213, p. 1–13, esis and exploration: Hobart, Australia, Centre for Ore Deposit Research,
doi: 10.1016/S0012-821X(03)00296-6. CODES Special Publication 4, p. 151–158.
Formolo, M.J., Lyons, T.W., Zhang, C., Kelley, C., Sassen, R., Horita, J., Johnson, C.A., Kelley, K.D., and Leach, D.L., 2004, Sulfur and oxygen iso-
and Cole, D.R., 2004, Quantifying carbon sources in the formation of topes in barite deposits of the western Brooks Range, Alaska, and impli-
authigenic carbonates at gas hydrate sites in the Gulf of Mexico, v. 205, cations for the origin of the Red Dog massive sulfide deposits: Economic
p. 253–264. Geology and the Bulletin of the Society of Economic Geologists, v. 99,
Gellatly, A.M., and Lyons, T.W., 2005, Trace sulfate in mid-Proterozoic carbon- p. 1435–1448.
ates and the sulfur isotope record of biospheric evolution: Geochimica et Johnston, D.T., Wing, B.A., Farquhar, J., Kaufman, A.J., Strauss, H., Lyons,
Cosmochimica Acta, v. 69, p. 3813–3829. T.W., Kah, L.C., and Canfield, D.E., 2005, Active microbial sulfur dispro-
Gill, B.C., Lyons, T.W., and Saltzman, M.R., 2006, Parallel, high-resolution carbon portionation in the Mesoproterozoic: Science, v. 310, p. 1477–1479.
and sulfur isotope records of the evolving Paleozoic marine sulfur reservoir: Jørgensen, B.B., Bang, M., and Blackburn, T.H., 1990, Anaerobic mineraliza-
Palaeogeography, Palaeoclimatology, Palaeoecology (in press). tion in marine sediments from the Baltic Sea-North Sea transition: Marine
Goodfellow, W.D., 1987, Anoxic stratified oceans a source of sulphur is sedi- Ecology Progress Series, v. 59, p. 39–54.
ment-hosted stratiform Zn-Pb deposits (Selwyn Basin, Yukon, Canada): Jørgensen, B.B., Böttcher, M.E., Lüschen, H., Neretin, L.N., and Volkov, I.I., 2004,
Chemical Geology, v. 65, p. 359–382. Anaerobic methane oxidation and a deep H2S sink generate isotopically
Goodfellow, W.D., and Jonasson, I.R., 1984, Ocean stagnation and ventilation heavy sulfide in Black Sea sediments: Geochimica et Cosmochimica
defined by δ34S secular trends in pyrite and barite, Selwyn Basin, Yukon: Acta, v. 68, p. 2095–2118, doi: 10.1016/j.gca.2003.07.017.
Geology, v. 12, p. 583–586, doi: 10.1130/0091-7613(1984)12<583: Kah, L.C., Lyons, T.W., and Chesley, J.T., 2001, Geochemistry of a 1.2 Ga
OSAVDB>2.0.CO;2. carbonate-evaporite succession, northern Baffin and Bylot islands: Impli-
Goodfellow, W.D., and Jonasson, I.R., 1987, Environment of formation of the cations for Mesoproterozoic marine evolution: Precambrian Research,
Howards Pass (XY) Zn-Pb deposit, Selwyn Basin, Yukon, in Morin, J.A., v. 111, p. 203–234, doi: 10.1016/S0301-9268(01)00161-9.
ed., Mineral deposits of Northern Cordillera: Canadian Institute of Min- Kah, L.C., Lyons, T.W., and Frank, T.D., 2004, Low marine sulphate and pro-
ing and Metallurgy, Special Volume 37, p. 19–50. tracted oxygenation of the Proterozoic biosphere: Nature, v. 431, p. 834–
Goodfellow, W.D., Lydon, J.W., and Turner, R.J.W., 1993, Geology and genesis 838, doi: 10.1038/nature02974.
of stratiform sediment-hosted (SEDEX) zinc-lead-silver sulphide depos- Kakegawa, T., and Ohmoto, H., 1999, Sulfur isotope evidence for the origin
its: Geological Association of Canada Special Paper 40, p. 201–251. of 3.4 to 3.1 Ga pyrite at the Princeton gold mine, Barberton Green-
Gorjan, P., Veevers, J.J., and Walter, M.R., 2000, Neoproterozoic sulfur-isotope stone Belt, South Africa: Precambrian Research, v. 96, p. 209–224, doi:
variation in Australia and global implications: Precambrian Research, 10.1016/S0301-9268(99)00006-6.
v. 100, p. 151–179, doi: 10.1016/S0301-9268(99)00073-X. Kakegawa, T., Kawai, H., and Ohmoto, H., 1998, Origins of pyrite in the
Greinert, J., Bollwerk, S.M., Derkachev, A., Bohrmann, G., and Suess, E., ~2.5 Ga Mt. McRae Shale, the Hamersley District, Western Australia:
2002, Massive barite deposits and carbonate mineralization in the Deru- Geochimica et Cosmochimica Acta, v. 62, p. 3205–3220, doi: 10.1016/
gin Basin, Sea of Okhotsk: Precipitation processes at cold seep sites: S0016-7037(98)00229-4.
Earth and Planetary Science Letters, v. 203, p. 165–180, doi: 10.1016/ Kampschulte, A., and Strauss, H., 2004, The sulfur isotopic evolution of
S0012-821X(02)00830-0. Phanerozoic seawater based on the analysis of structurally substituted
sulfate in carbonates: Chemical Geology, v. 204, p. 255–286, doi: Economic Geology and the Bulletin of the Society of Economic Geolo-
10.1016/j.chemgeo.2003.11.013. gists, v. 72, p. 1393–1409.
Kelley, K.D., Leach, D.L., Johnson, C.A., Clark, J.L., Fayek, M., Slack, J.F., Ohmoto, H., Kakegawa, T., and Lowe, D.R., 1993, 3.4-billion-year-old bio-
Anderson, V.M., Ayuso, R.A., and Ridley, W.I., 2004, Textural, compo- genic pyrites from Barberton, South Africa: Sulfur isotope evidence: Sci-
sitional, and sulfur isotope variations of sulfide minerals in the Red Dog ence, v. 262, v. 555–557.
Zn-Pb-Ag deposits, Brooks Range, Alaska: Implications for ore forma- Ohmoto, H., Watanabe, Y., and Kumazawa, K., 2004, Evidence from massive
tion: Economic Geology and the Bulletin of the Society of Economic siderite beds for a CO2-rich atmosphere before ~1.8 billion years ago:
Geologists, v. 99, p. 1509–1532. Nature, v. 429, p. 395–399, doi: 10.1038/nature02573.
Lambert, I.B., and Donnelly, T.H., 1991, Atmospheric oxygen levels in the Pre- Painter, M.G.M., Golding, S.D., Hannan, K.W., and Neudert, M.K., 1999, Sedi-
cambrian: A review of isotopic and geological evidence: Palaeogeography, mentologic, petrographic and sulfur isotope constraints on fine-grained
Palaeoclimatology, Palaeoecology, v. 97, p. 83–91, doi: 10.1016/0031- pyrite formation at Mount Isa Mine and environs, Northwest Queensland,
0182(91)90184-S. Australia: Economic Geology and the Bulletin of the Society of Economic
Large, R., McGoldrick, P., Bull, S., and Cooke, D., 2004, Proterozoic strati- Geologists, v. 94, p. 883–912.
form sediment-hosted zinc-lead-silver deposits of northern Australia, in Pandalai, H.S., Changkakoti, A., Krouse, H.R., and Gunalan, N., 1991, The
Deb, M., and Goodfellow, W.D., eds., Sediment hosted lead-zinc sulphide relationship between carbon, sulfur and pyritic iron in the Amjhore
deposits; attributes and models of some major deposits in India, Australia Deposit, Bihar, India: Economic Geology and the Bulletin of the Society
and Canada: New Delhi, Narosa Publishing House, p. 1–23. of Economic Geologists, v. 86, p. 862–869.
Larkin, M.J., Aharon, P., and Henk, M.C., 1994, Beggiatoa in microbial mats at Pavlov, A.A., Hurtgen, M.T., Kasting, J.F., and Arthur, M.A., 2003, Meth-
hydrocarbon vents in the Gulf of Mexico and Warm Mineral Springs, Flor- ane-rich Proterozoic atmosphere?: Geology, v. 31, p. 87–90, doi:
ida: Geo-Marine Letters, v. 14, p. 97–103, doi: 10.1007/BF01203720. 10.1130/0091-7613(2003)031<0087:MRPA>2.0.CO;2.
Logan, G.A., and Hinman, M.C., 2001, Biogeochemistry of the 1640 Ma Pavlov, A.A., Kasting, J.F., Brown, L.L., Rages, K.A., and Freedman,
McArthur River (HYC) lead-zinc ore and host sediments, Northern Terri- R., 2000, Greenhouse warming by CH4 in the atmosphere of early
tory, Australia: Geochimica et Cosmochimica Acta, v. 65, p. 2317–2336, Earth: Journal of Geophysical Research, v. 105, p. 11981–11990, doi:
doi: 10.1016/S0016-7037(01)00599-3. 10.1029/1999JE001134.
Logan, G.A., Hayes, J.M., Hieshima, G.B., and Summons, R.E., 1995, Termi- Paytan, A., Kastner, M., Campbell, D., and Thiemens, M.H., 1998, Sulfur iso-
nal Proterozoic reorganization of biogeochemical cycles: Nature, v. 376, topic composition of Cenozoic seawater sulfate: Science, v. 282, p. 1459–
p. 53–56, doi: 10.1038/376053a0. 1462, doi: 10.1126/science.282.5393.1459.
Lowell, R.P., and Keller, S.M., 2003, High-temperature seafloor hydro- Paytan, A., Kastner, M., Campbell, D., and Thiemens, M.H., 2004, Seawater
thermal circulation over geologic time and Archean banded iron for- sulfur isotope fluctuations in the Cretaceous: Science, v. 304, p. 1663–
mations: Geophysical Research Letters, v. 30, no. 7, p. 1391, doi: 1665, doi: 10.1126/science.1095258.
10.1029/2002GL016536. Poulton, S.W., Fralick, P.W., and Canfield, D.E., 2004, The transition to a sul-
Lowenstein, T.K., Hardie, L.A., Timofeeff, M.N., and Demicco, R.V., 2003, phidic ocean ~1.84 billion years ago: Nature, v. 431, p. 173–177, doi:
Secular variation in seawater chemistry and the origin of calcium chloride 10.1038/nature02912.
basinal brines: Geology, v. 31, p. 857–860, doi: 10.1130/G19728R.1. Ross, G.M., Bloch, J.D., and Krouse, H.R., 1995, Neoproterozoic strata of
Luepke, J.J., and Lyons, T.W., 2001, Pre-Rodinian (Mesoproterozoic) super- the southern Canadian Cordillera and the isotopic evolution of seawa-
continental rifting along the western margin of Laurentia: Geochemi- ter sulfate: Precambrian Research, v. 73, p. 71–99, doi: 10.1016/0301-
cal evidence from the Belt-Purcell Supergroup: Precambrian Research, 9268(94)00072-Y.
v. 111, p. 79–90, doi: 10.1016/S0301-9268(01)00157-7. Rye, D.M., and Williams, N., 1981, Studies of the base metal sulfide deposits at
Lydon, J.W., 1996, Sedimentary exhalative sulphides (SEDEX), in Eckstrand, McArthur River, Northern Territory, Australia. III: The stable isotope geo-
O.R., Sinclair, W.D., and Thorpe, R.I., eds., Geology of Canadian mineral chemistry of the H.Y.C., Ridge, and Cooley deposits: Economic Geology
deposit types: Geological Survey of Canada, Geology of Canada, v. 8, and the Bulletin of the Society of Economic Geologists, v. 76, p. 1–26.
p. 130–152. Rye, R., Kuo, P.H., and Holland, H.D., 1995, Atmospheric carbon dioxide con-
Lyons, T.W., 1997, Sulfur isotopic trends and pathways of iron sulfide forma- centrations before 2.2 billion years ago: Nature, v. 378, p. 603–605, doi:
tion in upper Holocene sediments of the anoxic Black Sea: Geochi- 10.1038/378603a0.
mica et Cosmochimica Acta, v. 61, p. 3367–3382, doi: 10.1016/S0016- Sageman, B.B., and Lyons, T.W., 2003, Geochemistry of fine-grained sediments
7037(97)00174-9. and sedimentary rocks, in Mackenzie, F.T., ed., Sediments, diagenesis,
Lyons, T.W., Werne, J.P., Hollander, D.J., and Murray, R.W., 2003, Contrast- and sedimentary rocks: Amsterdam, Elsevier, Treatise on Geochemistry,
ing sulfur geochemistry and Fe/Al and Mo/Al ratios across the last oxic- v. 7: p. 115–158.
to-anoxic transition in the Cariaco Basin, Venezuela: Chemical Geology, Schieber, J., 1986, The possible role of benthic microbial mats during the
v. 195, p. 131–157, doi: 10.1016/S0009-2541(02)00392-3. formation of carbonaceous shales in shallow mid-Proterozoic basins:
Lyons, T.W., Kah, L.C., and Gellatly, A.M., 2004, The Precambrian sulphur Sedimentology, v. 33, p. 521–536.
isotope record of evolving atmospheric oxygen, in Eriksson, P.G., et al, Schieber, J., 1990, Pyritic shales and microbial mats: significant factors in
eds., The Precambrian Earth: Tempos and events: Developments in Pre- the genesis of stratiform Pb-Zn deposits of the Proterozoic: Mineralium
cambrian geology: Amsterdam, Elsevier, p. 421–440. Deposita, v. 25, p. 7–14.
Lyons, T.W., Luepke, J.J., Schreiber, M.E., and Zieg, G.A., 2000, Sulfur geo- Scott, K.M., Smith, J.W., Sun, S.-S., and Taylor, G.F., 1985, Proterozoic copper
chemical constraints on Mesoproterozoic restricted marine deposition: deposits in NW Queensland, Australia: Sulfur isotopic data: Mineralium
Lower Belt Supergroup, northwestern United States: Geochimica et Cos- Deposita, v. 20, p. 116–126, doi: 10.1007/BF00204322.
mochimica Acta, v. 64, p. 427–437, doi: 10.1016/S0016-7037(99)00323-3. Shanks, W.C., Woodruff, L.G., Jilson, G.A., Jennings, D.S., Modene, J.S., and
McGoldrick, P.J., 1999, Northern Australian SEDEX deposits: Microbial oases Ryan, B.D., 1987, Sulfur and lead isotope studies of stratiform Zn-Pb-Ag
in Proterozoic seas, in Mineral deposits: Processes to processing: Pro- deposits, Anvil Range, Yukon: Basinal brine exhalation and anoxic bot-
ceedings of the 5th Biennial SGA Meeting and the 10th Quadrennial tom-water mixing: Economic Geology and the Bulletin of the Society of
IAGOD Symposium, London, 22–26 August 1999: Rotterdam, A.A. Economic Geologists, v. 82, p. 600–634.
Balkema, v. 2, p. 885–888. Shen, Y., Buick, R., and Canfield, D.E., 2001, Isotopic evidence for microbial
McGoldrick, P.J., Dunster, J., and Aheimer, M., 1999, New sedimentological, sulphate reduction in the early Archaean era: Nature, v. 410, p. 77–81,
geochemical and textural observations from the Lady Loretta deposit: doi: 10.1038/35065071.
Implications for ore genesis, in Holm, O., Pongratz, J., and McGoldrick, Shen, Y., Canfield, D.E., and Knoll, A.H., 2002, Middle Proterozoic ocean
P., eds., Basins, fluids and Zn-Pb ores: Hobart, Australia, Centre for Ore chemistry: Evidence from the McArthur Basin, northern Australia: Amer-
Deposit Research, CODES Special Publication 2, p. 49–58. ican Journal of Science, v. 302, p. 81–109.
McShane, M.B.J., 1996, Stratigraphy and mineralisation, Bernborough, Mount Shen, Y., Knoll, A.H., and Walter, M.R., 2003, Evidence for low sulphate and
Isa South, northwest Queensland [Postgraduate thesis: Brisbane, Univer- anoxia in a mid-Proterozoic marine basin: Nature, v. 423, p. 632–635,
sity of Queensland, 86 p. doi: 10.1038/nature01651.
Oehler, J.H., and Logan, R.G., 1977, Microfossils, cherts, and associated min- Shields, G., Kimura, H., Yang, J., and Gammon, P., 2004, Sulphur isotopic
eralization in the Proterozoic McArthur (H.Y.C.) lead-zinc-silver deposit: evolution of Neoproterozoic-Cambrian seawater: New francolite-bound
sulphate δ34S data and a critical appraisal of the existing record: Chemical Whelan, J.F., Rye, R.O., and deLorraine, W., 1984, The Balmat-Edwards zinc-
Geology, v. 204, p. 163–182, doi: 10.1016/j.chemgeo.2003.12.001. lead deposits-synsedimentary ore from Mississippi Valley-type fluids:
Smith, J.W., and Croxford, N.J.W., 1975, An isotopic investigation of the Economic Geology and the Bulletin of the Society of Economic Geolo-
environment of deposition of the McArthur mineralization: Mineralium gists, v. 79, p. 239–265.
Deposita, v. 10, p. 269–276, doi: 10.1007/BF00207885. Wilkin, R.T., and Arthur, M.A., 2001, Variations in pyrite texture, sulfur iso-
Smith, J.W., Burns, M.S., and Croxford, N.J.W., 1978, Stable isotope studies of tope composition, and iron systematics in the Black Sea: Evidence for
the origins of mineralization at Mount Isa. I: Mineralium Deposita, v. 13, Late Pleistocene and Holocene excursions of the O2–H2S redox transition:
p. 369–381, doi: 10.1007/BF00206570. Geochimica et Cosmochimica Acta, v. 65, p. 1399–1416, doi: 10.1016/
Solomon, P.J., 1965, Investigations into sulphide mineralization at Mount Isa, S0016-7037(01)00552-X.
Queensland: Economic Geology and the Bulletin of the Society of Eco- Willan, R.C.R., and Coleman, M.L., 1983, Sulfur isotope study of the Aberfeldy
nomic Geologists, v. 60, p. 737–765. barite, zinc, lead deposit and minor sulfide mineralization in the Dalradian
Strauss, H., 1997, The isotopic composition of sedimentary sulfur through time: metamorphic terrain, Scotland: Economic Geology and the Bulletin of the
Palaeogeography, Palaeoclimatology, Palaeoecology, v. 132, p. 97–118, Society of Economic Geologists, v. 78, p. 1619–1656.
doi: 10.1016/S0031-0182(97)00067-9. Winston, D., 1990, Evidence for intracratonic, fluvial and lacustrine settings of
Strauss, H., 2002, The isotopic composition of Precambrian sulphides—Seawa- Middle and Late Proterozoic basins of western U.S.A., in Gower, C.F.,
ter chemistry and biological evolution, in Altermann, W., and Corcoran, Rivers, T., and Ryan, B., eds., Mid-Proterozoic Laurentia-Baltica: Geo-
P.L., eds., Precambrian sedimentary environments: A modern approach to logical Association of Canada Special Paper 38, p. 535–564.
ancient depositional systems: Oxford, Blackwell Science, International Winston, D., and Link, P.K., 1993, Middle Proterozoic rocks of Montana, Idaho
Association of Sedimentologists Special Publication 33, p. 67–105. and eastern Washington, in Reed, J.C., Jr., et al., eds., Precambrian: Con-
Strauss, H., and Schieber, J., 1990, A sulfur isotope study of pyrite genesis: The terminous U.S.: Boulder, Colorado, Geological Society of America, Geol-
mid-Proterozoic Newland Formation, Belt Supergroup, Montana: Geo- ogy of North America , v. C-2, p. 487–517.
chimica et Cosmochimica Acta, v. 54, p. 197–204. Zaback, D.A., Pratt, L.M., and Hayes, J.M., 1993, Transport and reduction of
Torres, M.E., Bohrmann, G., Dubé, T.E., and Poole, F.G., 2003, Formation of sulfate and immobilization of sulfide in marine black shales: Geology, v. 21,
modern and Paleozoic stratiform barite at cold methane seeps on conti- p. 141–144, doi: 10.1130/0091-7613(1993)021<0141:TAROSA>2.3.CO;2.
nental margins: Geology, v. 31, p. 897–900, doi: 10.1130/G19652.1. Zhang, C.L., Huang, Z., Cantu, J., Pancost, R.D., Brigmon, R.L., Lyons, T.W.,
Turner, R.J.W., 1992, Formation of Phanerozoic stratiform sediment-hosted and Sassen, R., 2005, Lipid biomarkers and carbon-isotope signatures of
zinc-lead deposits: Evidence for the critical role of ocean anoxia: Chemi- a microbial (Beggiatoa) mat associated with gas hydrates in the Gulf of
cal Geology, v. 99, p. 165–188, doi: 10.1016/0009-2541(92)90037-6. Mexico: Applied and Environmental Microbiology, v. 71, p. 2106–2112,
Velasco, F., Herrero, J.M., Gil, P.P., Alvarez, L., and Yusta, I., 1994, Mississippi doi: 10.1128/AEM.71.4.2106-2112.2005.
Valley-type, sedex, and iron deposits in Lower Cretaceous rocks of the
Basque-Cantabrian Basin, northern Spain, in Fontbote, L., and Boni, M.,
eds., Sediment-hosted Zn-Pb Ores: Special Publication 10, Society for
Geology Applied to Ore Deposits: Springer-Verlag, Berlin-Heidelberg,
p. 246–270. MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA


Memoir 198
2006
Precambrian Mississippi Valley–type deposits: Relation to changes

in composition of the hydrosphere and atmosphere
Stephen E. Kesler
Martin H. Reich*
Department of Geological Sciences, University of Michigan, Ann Arbor, Michigan 48109, USA
ABSTRACT
We have evaluated the temporal distribution of Mississippi Valley-type (MVT)

Zn-Pb deposits with special attention to the nature and number of deposits of Pre-
cambrian age. Our evaluation is based on the widely used model for MVT mineral-
ization involving metal-bearing brines that lack reduced S and that deposit sulfides
only where they encounter a reservoir of sulfide or where sulfate in the metal-bearing
brine is reduced to sulfide. For MVT systems of this type, basins with abundant sul-
fate would be most favorable for development of MVT mineralization because these
would allow transport of metals in sulfate-rich brines and deposition of metals in
areas where the sulfate was reduced. Because abundant sulfate requires abundant
atmospheric oxygen, the distribution of MVT deposits through time might reflect
compositional changes in Earth’s atmosphere, especially the suggested Great Oxida-
tion Event (GOE).
A compilation of new data for the Bushy Park-Pering district in the Transvaal
Supergroup of South Africa, the world’s oldest known MVT province, and published
information on other Precambrian MVT deposits in the Ediacara, Berg Aukas/
Abenab, Gayna River, Warrabarty, Nanisivik, Kamarga (Century), McArthur River
(Coxco), Ramah, and Esker districts shows that they are generally similar in geologic
setting and mineralogy to those in Phanerozoic rocks. Fluid inclusions in some Neo-
proterozoic deposits, including Berg Aukas/Abenab, Gayna River, Warrabarty, and
Nanisivik, record higher temperatures and salinities than found in most Phanerozoic
deposits, possibly reflecting igneous activity or a more proximal basinal setting dur-
ing Precambrian time. Fluid inclusion leachate data for several Precambrian MVT
deposits suggest that their parent brines formed by evaporation of seawater, and S
isotope compositions indicate that the S was derived largely from coeval seawater sul-
fate. Comparisons of data from all deposits show no evidence for a gradual increase
in temperature or salinity backward through time, such as might be caused by higher
heat flow during early stages of Earth history, although the magnitude of this effect
might be lost in the uncertainty of most fluid inclusion measurements. These observa-
tions confirm that MVT deposits reflect the chemistry of their source basins, which
are as old as 2.6 Ga. No MVT deposits or suitable host rocks of an older age are
known.
Precambrian MVT deposits do differ from their Phanerozoic analogues in
the magnitude of mineralization. Precambrian deposits and districts formed at an
*Reich also at: Departamento de Geología, Facultad de Ciencias Físicas y Matemáticas, Universidad de Chile, Santiago, Chile.
Kesler, S.E., and Reich, M.H., 2006, Precambrian Mississippi Valley–type deposits: Relation to changes in composition of the hydrosphere and atmosphere, in
Kesler, S.E., and Ohmoto, H., eds., Evolution of Early Earth’s Atmosphere, Hydrosphere and Biosphere—Constraints from Ore Deposits: Geological Society of
America Memoir 198, p. 185–204, doi: 10.1130/2006.1198(11). For permission to copy, contact [email protected]. ©2006 Geological Society of America.
All rights reserved.
185
186 S.E. Kesler and M.H. Reich
estimated rate of 5.5 per billion years versus a significantly larger rate of ~60 per
billion years for Phanerozoic deposits, and the Phanerozoic deposits are consider-
ably larger. Furthermore, the transition from low-magnitude, Precambrian-type to
high-magnitude, Phanerozoic-type MVT mineralization took place at the beginning
of Cambrian time rather than at the 2.3 Ga GOE. This appearance of widespread
MVT mineralization is closer to the time at which sulfate concentrations in the world
ocean are estimated to have reached present-day levels. Although these conclusions
are subject to considerable uncertainty because of the limited number of Precam-
brian deposits, the lack of an increase in the frequency of MVT mineralization at the
GOE suggests that widespread MVT mineralization requires higher levels of sulfate
than could have been provided by this event, or that the appearance of sulfate in the
ocean was considerably delayed. Finally, the presence of MVT deposits in basins that
formed considerably before the GOE suggests that local sulfate concentrations were
available at even early points in Earth’s history.
Keywords: Mississippi Valley-type deposit, MVT, basin, brine, Pb-Zn.
INTRODUCTION in the composition of the atmosphere and hydrosphere. Among

subsurface environments, sedimentary basins are most likely to
Mineral deposits have an uneven distribution through time reflect the surface environment because their rocks originated at
that has been related to long-term changes in global heat flow, the surface. The most widespread hydrothermal mineral deposits
tectonism, and compositions of the atmosphere and oceans that formed by sedimentary hydrothermal systems are sandstone
(Meyer, 1988; Barley and Groves, 1992). This uneven tempo- and unconformity U deposits, which contain redox-sensitive U,
ral distribution is particularly distinct for deposits that formed at and sedex, Irish-type, and Mississippi Valley-type (MVT) Zn-Pb
Earth’s surface, such as uranium-bearing conglomerates, later- deposits, which contain redox-sensitive S. Elsewhere in this vol-
ites, evaporites, and iron formations, all of which contain redox- ume, reviews for two of these deposit types, sandstone-U (Gauth-
sensitive elements that might reflect increasing oxygen contents ier-Lafaye) and sedex Zn-Pb (Lyons et al.) suggest that their
in the atmosphere and sulfate contents in the hydrosphere (Hol- temporal distribution reflects increasing atmospheric oxygen and
land, 1984). marine sulfate, respectively. Whereas sedex Zn-Pb deposits form
Ore deposits that formed in the deep subsurface also show on the seafloor and sandstone-U deposits form in subsurface
temporal variations that might be due, at least in part, to changes zones dominated by meteoric water, MVT Zn-Pb deposits form
at somewhat greater depths in zones dominated by formation
waters. These deposits also show a distinct age distribution, with
JURASSIC
fewer Precambrian representatives (Fig. 1). Here, we review the
characteristics of these Precambrian MVT deposits with the goal
TRIASSIC of determining whether this pattern reflects a change in the oxy-
PERMIAN gen content of the atmosphere and sulfate content of the oceans.
CARBONIFEROUS
GEOLOGICAL SETTING AND GEOCHEMICAL
DEVONIAN CHARACTERITICS OF MVT DEPOSITS
SILURIAN
MVT deposits are part of a continuum of Zn-Pb deposits,
ORDOVICIAN
also including sedex and Irish-type, that form in craton margin
CAMBRIAN and intracratonic sedimentary basins (Sangster, 1990). Although
NEOPROTEROZOIC all three deposit types have similar mineralogy, textures, host
rocks, and geology settings differ considerably (Fig. 2). At one
MESOPROTEROZOIC
end of the spectrum are MVT deposits, which consist of coarse-
PALEOPROTEROZOIC grained sphalerite, galena, and other minerals (Fig. 3A) that fill
ARCHEAN breccia-type porosity in dolomitized limestone along the margins
of sedimentary basins (Anderson and Macqueen, 1988). MVT
N=0 5 10 15
sulfides contain primary fluid inclusions with temperatures of
Figure 1. Age distribution for host rocks for MVT deposits, modified from 75–200 °C and salinities of 10–30 equivalent wt.% NaCl that are
Leach and Sangster (1993) to include additional Precambrian deposits. interpreted to be basinal brines that were expelled from deeper
Precambrian Mississippi Valley–type deposits 187
Figure 2. Highly schematic model for

formation of MVT, sedex, and Irish-
type Zn-Pb deposits in sedimentary
basins. Dashed arrows show flow paths
for mineralizing solutions. The model
is not meant to indicate that these three
deposits can form at the same time or
that all of them form exclusively in an
extensional environment.
parts of adjacent basins (Fig. 2). At the other end of the continuum
are sedex deposits, which consist largely of fine-grained, well
layered Zn-Pb and other sulfides (Fig. 3B) in black shales along
rifted margins of basins (Scott, 1997). Although sedex sulfides do
not provide much useful fluid inclusion data, other lines of evi-
dence suggest that their parent fluids were slightly hotter and less
saline than MVT fluids, that they were a mixture of basinal brine
and seawater, and that they were driven at least partly by heat
from mafic dikes that intruded rift-margin faults (Fig. 2). Irish-
type deposits contain both layered and coarse-grained ores and
appear to have formed largely along faults that cut rocks between
the seafloor on which sedex deposits formed and the deeper res-
ervoirs in which MVT deposits formed (Fig. 2).
We have focused on MVT deposits in this study because
they form at greatest depth in the subsurface and are most likely
to reflect the chemical composition of their host sedimentary
basins. Most MVT deposits consist entirely or dominantly of
galena and sphalerite with sparry dolomite, calcite, and local
quartz and pyrite or marcasite. Some MVT deposits, which con-
sist dominantly of fluorite, have been considered to form a dis-
tinct group in recent surveys and are not included in this summary
(Sangster, 1990; Leach and Sangster, 1993). Similarly, barite is
present in only a few deposits and, where present, it is not coeval
Figure 3. (A) Typical MVT ore showing brecciated blocks of dolostone

(BX) surrounded by a matrix of sparry dolomite and sphalerite (D/S) (New
Market mine, Tennessee). (B) Typical sedex ore showing deformed layers
of sphalerite (S) and galena (G) (Sullivan mine, British Columbia).
with galena and sphalerite and thus does not provide a useful tion and transport. In general, biogenic and diagenetic sulfides
constraint on the composition of the metal-bearing MVT fluid. have lower δ34S values than sulfides formed by thermochemical
Other metals that are found in small amounts locally, including reduction of seawater sulfate, and some MVT deposits, such as
copper, cobalt, arsenic, and silver, are not sufficiently systematic the large Tri-State district, contain sulfides with low δ34S values
in their distribution to classify MVT deposits further, and most that might be of this origin (Ohmoto and Rye, 1979; Ohmoto and
were derived largely from nearby wall rocks (Burstein et al., Goldhaber, 1997; Leach and Sangster, 1993). Although the exis-
1992). Thus, insights into the processes that formed most MVT tence of numerous possible pathways makes it more difficult to
deposits are provided largely by constraints on the behavior of identify the actual source of S, and the presence of districts such
Pb, Zn, and S, and it is these metals that we must look to for as Tri-State show that alternative pathways operated locally, they
information on the redox state of hydrothermal fluids in their host do not negate the generalization that high basinal sulfate contents
sedimentary basins. favor formation of MVT deposits.
Experimental studies show that chloride complexes are effec- Other possible complications to the model used here involve
tive ligands for Zn and Pb in low-temperature, basinal brines and solubility constraints. For instance, Ba cannot be transported
that bisulfide and organic complexes are less effective, particu- in the presence of significant dissolved sulfate (Blount, 1977),
larly in the presence of abundant dissolved Ca and Na (Giordano, and the presence of barite in MVT deposits has been cited as
1985; Barrett and Anderson, 1988; Sicree and Barnes, 1996). The evidence that MVT brines lacked sulfate. However, as noted
presence of reduced (sulfide) S in MVT brines greatly reduces the above, barite is actually absent from most MVT deposits and,
solubility of chloride-complexed Zn and Pb (Anderson, 1983). where present, it was not deposited at the same time as galena
The most widely applied model for formation of MVT deposits and sphalerite. A more important complication involves the pos-
is based on this relation and involves metal-bearing brines that sibility that acid MVT brines might carry metals and reduced S
lack reduced S and that deposit MVT sulfides only where they together, thus obviating the need for sulfate-rich environments
encounter a reservoir of sulfide or where sulfate in the metal- noted above (Sverjensky, 1981). Ohmoto et al. (1990) have sug-
bearing brine is reduced to sulfide (Anderson and Macqueen, gested that MVT districts such as the Upper Mississippi Valley,
1988; Anderson, 1991; Leach and Sangster, 1993). where sulfides have uniform δ34S values and equilibrium frac-
According to this “Anderson model,” basinal environments tionation, formed from fluids that contained both metals and
containing only reduced S could not form MVT deposits because sulfide, which requires that the fluid was either unusually hot or
their low-temperature brines would not transport sufficient met- acidic. Wall-rock alteration in most MVT deposits does not pro-
als. Conversely, basinal environments that lack S would allow vide evidence of highly acid brines, however, and maintenance
extensive migration of metals in low-temperature brines, but of acidity in brines hosted by carbonate rocks requires high levels
would not have the capacity to deposit MVT sulfides. The most of CO2, which have been documented in only a few MVT depos-
favorable basinal environments for formation of MVT deposits its (Haynes et al., 1989). If this exception does apply, it is most
would be those containing sulfate-bearing evaporites and brines likely for MVT districts that formed at high temperatures, such
and local accumulations of sulfide formed from this evaporite as the Upper Mississippi Valley district (McLimans et al., 1980;
sulfate. All other things being equal, the MVT-forming capac- Ohmoto et al., 1990). As is apparent from this list of complica-
ity of such sedimentary basins should increase with increasing tions, the Anderson model probably is less clearly applicable to
amounts of sulfate that could be reduced locally to cause ore Irish-type and sedex deposits, and they are not included in this
deposition. Possible mechanisms for sulfate reduction in Precam- discussion (Sangster, 1990).
brian sedimentary settings include near-surface bacterial action In addition to general acceptance of the Anderson model,
and deeper thermochemical sulfate reduction, as is the case in use of Precambrian MVT deposits as possible sensors of oxygen
modern settings (Ohmoto and Goldhaber, 1997; Habicht et al., and sulfate contents of the atmosphere and hydrosphere requires
2002). Because basinal brines form largely by evaporation of that (1) their geology and geochemistry be similar to Phanero-
seawater or dissolution of marine evaporites (Hanor, 1987), their zoic MVT deposits to which the Anderson model was originally
sulfate contents should reflect the composition of their source applied, and (2) their age be constrained by geologic, paleomag-
ocean and the oxygen content of its coeval atmosphere. netic, or isotopic observations. In the following sections, we
Although the Anderson model provides a useful framework review Precambrian MVT deposits from these two perspectives.
for evaluating possible relations between MVT ore formation
and compositional evolution of the atmosphere and hydrosphere, GEOLOGIC FEATURES OF PRECAMBRIAN MVT
it is not without complexities. Ohmoto and Goldhaber (1997) DEPOSITS
have described eight pathways by which seawater sulfate can
be fixed as sulfide in ore deposits. Whereas some of these path- MVT deposits that are hosted by Precambrian rocks and to
ways involve single-stage reduction of sulfate by biogenic, ther- which comparisons might be made include, in order of increasing
mochemical, or more direct inorganic processes, others involve host rock age, Ediacara, Berg Aukas/Abenab, Gayna River, War-
multiple steps such as bacterial reduction of seawater sulfate dur- rabarty, Nanisivik, McArthur River (Coxco), Kamarga, Esker,
ing diagenesis, formation of pyrite, and its subsequent dissolu- Ramah, and Bushy Park-Pering-Zeerust (Fig. 4). Geological fea-
GAYNA ESKER NANISIVIK PRECAMBRIAN MVT levels above those typical of Phanerozoic MVT deposits (Frim-
RIVER AND RELATED
DEPOSITS mel et al., 1996).
The Gayna River district, which is hosted by passive mar-
RAMAH
gin, platform carbonates of the Little Dal Group in Canada, is
more typical of Phanerozoic MVTs in both composition and
COXCO geologic setting (Table 1). The Little Dal Group is distinguished
by its extensive stromatolites with internal structures typical of
Precambrian reefs and by the presence of evaporites, including
MORRO AGUDO- WARRABARTY local salt casts (Hardy, 1979; Narbonne and Aitken, 1995; Turner
VAZANTE BERG
AUKAS
ZEERUST
et al., 2000). MVT mineralization is hosted by sedimentary and
BUSHYBUSHY
PARK PARK – KAMARGA solution-collapse breccias that contain sphalerite and galena
PERING EDIACARA with sparry dolomite and local barite with snow-on-the-roof and
colloform textures typical of Phanerozoic MVTs (Hardy, 1979;
Figure 4. Location of Precambrian and related MVT deposits discussed
in this paper. Hewton, 1982).
A final deposit that might belong in this group is Warrab-
arty in the Patterson orogen of Western Australia (Anderson et
al., 2001, 2002). Warrabarty consists of Zn-rich mineralization
tures of these deposits and districts are summarized in Table 1 in lower greenschist facies, metamorphosed carbonaceous dolos-
and discussed briefly in the next section. tones, and limestones of the Broadhurst Formation, a part of the
Some deposits of possible MVT origin are not included in Meso-Neoproterozoic Throssell Group. Mineralization consists
this survey because of inadequate data, because they appear to largely of sphalerite and galena with lesser pyrite, pyrobitumen,
be transitional in character, or because metamorphic and other and sparry dolomite in breccias, veins, and disseminations in
overprints obscure their original features. For instance, Neopro- dolomitized wall rock (Smith, 1996).
terozoic platform carbonate sequences in Brazil, including the
Una (Irecê Basin) and Bambui (São Francisco Basin) Groups MVT Deposits in Mesoproterozoic Basins
that originally covered an area of more than 300,000 km2, host
Pb-Zn mineralization of uncertain type. The best-known depos- MVT districts are scarce in rocks of Mesoproterozoic age,
its, including Morro Agudo, Vazante, and related deposits in the the only good example being the Nanisivik district (Fig. 4).
São Francisco Basin, have been classified as MVT, sedex, and Although commonly classified as MVT, Nanisivik is somewhat
Irish-type in different studies, but most recent descriptions sug- unusual in form and mineralogy (Table 1). The ore-hosting Soci-
gest that they are least like MVTs and most like sedex (Iyer et ety Cliffs Formation is a platform carbonate sequence consist-
al., 1992; Hitzman et al., 1995; Kyle and Misi, 1997; Misi et ing largely of stromatolitic to massive dolostone with gypsum
al., 2005). Similarly, the Balmat-Edwards district in New York horizons that are among the oldest known extensive evaporite
has a general geologic setting typical of MVT deposits, including deposits (Kah et al., 2001). Solution-collapse and karst breccias
association with evaporites, but it is so highly metamorphosed are widespread in the upper part of the Society Cliffs Formation,
that original textures necessary to determine its origin have been but debate persists about their relation to MVT mineralization
largely obliterated (Whelan et al., 1984, 1990). (Jackson and Ianelli, 1981; Olson, 1984; Ford, 1986). The ore
zone is unusually sulfide-rich and ranges in texture from alternat-
MVT Deposits in Neoproterozoic Basins ing layers of pyrite, sphalerite, sparry dolomite, and galena that
fill open spaces to massive sulfides showing evidence of multiple
Two of the four possible Neoproterozoic MVT districts, stages of wall-rock replacement (Olson, 1984; Ghazban et al.,
Ediacara and Berg Aukas/Abenab, have unusual metal contents. 1990). Large iron sulfide bodies surround the Pb-Zn ore bodies
The Ediacara district, which is in the central Flinders Ranges of and make up most of the district (Arne et al., 1991; Sutherland
South Australia, is actually hosted by Cambrian shelf carbonates and Dumka, 1995). Iron-rich MVT mineralization forms from
of the Ajax Limestone (Table 1), but is part of a dominantly Pro- hot brines at temperatures of 200 °C or more, which is consistent
terozoic sedimentary sequence (Christie-Blick et al., 1995). Min- with fluid inclusion evidence discussed below and might account
eralization consists of galena and pyrite with minor chalcopyrite for the unusual tube-like form of the main ore zone (St. Marie et
and sphalerite and rare tetrahedrite and pearceite and has higher al., 2001).
silver grades than most Phanerozoic MVTs (Drew and Both,
1984; McFarlane and Bone, 1994). The Berg Aukas/Abenab dis- MVT Deposits in Paleoproterozoic Basins
trict, which is in the foreland thrust belt of the Damara orogen of
Namibia, consists of galena, sphalerite, and sparry dolomite with Probable MVT districts in Paleoproterozoic rocks, includ-
trace amounts of pyrite, tetrahedrite, enargite, and chalcopyrite, ing Coxco (McArthur River), Kamarga (Century), Ramah, and
as well as vanadium, silver, germanium, gallium, and cadmium at Esker, differ somewhat from Phanerozoic MVTs, especially in
TABLE 1. GEOLOGICAL CHARACTERISTICS OF PRECAMBRIAN MVT DEPOSITS COMPILED FROM SOURCES IDENTIFIED IN THE TEXT
Deposit Ediacara Berg Aukas Gayna River Warrabarty Nanisivik Coxco
(McArthur River)
Host Rocks
Basin Setting Passive Margin Rift Margin Passive Margin Transtensional Rift Rift Margin Rift Margin
Formation/Subgroup Ajax Abenab Broadhurst Society Cliffs Emmeruga, Mara
Group Otavi Little Dal Throssell Uluksan McArthur
Supergroup MacKenzie Mtns. Bylot
Age Cambrian 0.75 Ga 0.78–1.08 Ga 0.71–1.08 Ga ca. 1.25 Ga 1.60–1.68 Ga
Possible Paleoaquifer Pound Qtzt Nosib Gp Katherine Fm. Coolbro Sandstone Eqaluik Gp Twallalh Gp
Evidence for Evaporites no No Yes No Yes Yes
Deformation Minor Minor Minor Moderate Minor Minor
Metamorphism Not significant Not significant Not significant Mid-greenschist Not significant Not significant
Mineral Deposits
Area of District (km2) 30 1500 200 Unknown 600 600
Tonnage 29 Mt small ~50 Mt Unknown 13.3 Mt Small
Grade ~1% Pb 5% Pb+ Zn 3–6% Zn+Pb 9.7% Zn, 0.9% Pb 2.5% Zn, 0.5% Pb
Additional Information ~35 Mt Fe sulfide
Ore Setting styl, bx diss, bz bx, repl bx, vn, repl Possible bx bx
Major Minerals gal, py, qz, cpy sph, gal, dol sph, gal, dol sph, gal, dol sph, gal, dol, py/mar sph, gal, py/mar, dol
Minor Minerals cpy, sph, tet, pear, dol bar cpy cpy, bn, tet, bar, qz
Unusual Elements Ag, As V, Ag, Ge, Ga, As None Cu
Bitumen No Not reported Yes Yes Yes
Age of Mineralization Cambrian(?) 0.75 Ga 0.78–1.08 Ga possible 0.84 Ga 1.095 Ga 1.64–1.68 Ga
Evidence for Age Geologic Geologic Geologic Model Pb isotope Paleomagnetic Geologic
Deposit Kamarga Ramah Esker Pering Bushy Park

Host Rocks
Basin Setting Rift(?) Passive Margin Passive Margin Passive Margin Passive Margin
Formation/Subgroup Gunpowder Creek Reddick Bight Rocknest Campbellrand Campbellrand
Group McNamara Ramah Coronation Ghaap Ghaap
Supergroup Transvaal Transvaal
Age 1.653 Ga ca. 2.0 Ga ca. 1.9 Ga 2.64 to 2.43 Ga 2.64 to 2.43 Ga
Possible Paleoaquifer Torpedo Creek Fm. Roswell Harbour Odjick Fm Makwassie Qz Makwassie Qz Phry
(?) Fm. Phry
Evidence for Evaporites Yes No Yes No No
Deformation Minor Extensive Extensive Minor Minor
Metamorphism Not significant Greenschist facies Low grade Not significant Not significant
Mineral Deposits
Area of District (km2) 2(?) 1000 60 2400 ?
Tonnage 50 Mt No information 80 M 18 Mt 10 Mt
Grade 3% Zn+Pb No information 4.7% Pb+Zn 3.6% Zn, 0.6% Pb 5% Zn, 0.6% Pb
Additional Information (Includes 10 Mt 10%)
Ore Setting vn, bx, diss bx, repl diss, repl, bx bx bx
Major Minerals py, sph, gal sph, gal, dol gal, sph, dol sph, gal, dol sph, gal, dol
Minor Minerals cpy, jor, fl py, qz, fs cpy qz, cpy cpy, py
Unusual Elements As, Sb Cu Cu Cu
Bitumen Yes Yes No Yes Not reported
Age of Mineralization Probably 1.6X Ga Possibly 1.86 Ma 1.90–1.84 Ma pre 1.98 Ga Pre 1.98 Ga
Evidence for Age Geologic Geologic Geologic Geologic Geologic
Note: Mineralogy: bar—barite, bn—bornite, cpy-chalcopyrite, ccy—chalcedony, dol—dolomite, fs—feldspar, fl—fluorite, gal—galena, jor—jordanite,
mar—marcasite, pear—pearcite, po—pyrrhotite, py—pyrite, qz—quartz, sph—sphalerite, tet—tetrahedrite Form of Ore -bx—breccia, diss—disseminated,
repl—replacement, styl—stylolite.
their apparently closer association with possibly coeval sedex contain layers of acicular, radiating carbonate fans (Coxco nee-
mineralization. These relations are best displayed in the “Carpen- dles) thought to have been precipitated as aragonite (Winefield,
taria zinc belt” of Australia, which contains both the McArthur 2000). Mineralization at Coxco is hosted by a karst system that
River-Coxco and Kamarga districts. formed during uplift and weathering of domal stromatolites and
MVT mineralization at Coxco is hosted by the latest Paleo- fine-grained dolomite of the Mara Dolomite Member, and is very
proterozoic McArthur (Coxco) and McNamara (Kamarga) similar to Phanerozoic MVTs (Walker et al., 1983). It contains a
Groups in north-central Australia. Both districts contain much first stage of colloform sphalerite, galena, and pyrite-marcasite
better known sedex deposits, HYC in the McArthur Group associated with abundant organic matter, which was covered by
and Century in the McNamara Group (Plumb et al., 1990). The detritus from the overlying Lynnot Formation and then followed
McArthur River district and immediately adjacent area contains by a second stage of crosscutting, coarse-grained pyrite-marca-
MVT mineralization at Coxco and possibly at Ridge and Cooley site, sphalerite, galena, sparry dolomite, and minor bitumen in
(Williams, 1978; Walker et al., 1983; Selley et al., 2001; Gar- veins and dolomite breccias of probable tectonic origin. Miner-
ven and Bull, 2000). Overlying middle and upper McArthur alization at Cooley and Ridge is similar to the second stage at
Group platform dolostones, which host MVT mineralization, Coxco, but includes late chalcopyrite, tetrahedrite, and bornite,
which are more abundant than in most Phanerozoic MVT depos- vaal Basin, which are known somewhat confusingly as the Gri-
its (Williams, 1978). quatown West and Transvaal Basins.
The McNamara district hosts probable MVT mineraliza- The Griquatown West Basin contains the Pering and Bushy
tion at Kamarga, which consists of pyrite and sphalerite in veins, Park Zn-Pb deposits. The two deposits are sufficiently far apart
breccias, disseminations, and massive replacements in the Gun- to be considered separate districts, although each district contains
powder Creek Formation (Jones, 1986). The Gunpowder Creek only one important deposit. Pering and surrounding prospects are
Formation contains abundant textures typical of evaporites that largely in breccias of probable karst origin that are aligned along
have undergone replacement and is thought to have been depos- fracture systems (Wheatley et al., 1986b; Kruger et al., 2001).
ited in a sabkha environment (Jones et al., 1999). Mineralization at Pering includes early fine-grained, colloform
The Esker district near the Arctic coast in Canada is hosted sphalerite and subordinate galena and chalcopyrite associated
by passive margin shelf sediments of the Rocknest Formation, with sparry dolomite and hydrocarbons, and a final stage of
which include abundant silica pseudomorphs after halite, gyp- coarse-grained sphalerite, galena, sparry dolomite, quartz, and
sum, and possibly anhydrite (Grotzinger, 1986a, 1986b, 1986c). calcite (Greyling et al., 2001). Bushy Park, the more northerly
MVT mineralization, consisting largely of sparry dolomite, of the two districts, is also hosted by breccias of probable col-
sphalerite, and galena with minor chalcopyrite, is in ore-matrix lapse origin and consists largely of coarse-grained sphalerite,
breccias and disseminations in a regionally extensive stromato- sparry dolomite, and galena (Wheatley et al., 1986a; Martini et
litic reef zone, and several possibly coeval Cu-Co-Pb-Zn sedex al., 1995; Baugaard et al., 2001; Schaefer et al., 2001).
showings are in the underlying Odjick Formation (Rhondacorp The Transvaal Basin contains the Zeerust district and scat-
2002; Wachowiak, 2001). Paragenetic relations indicate that tered smaller deposits that surround the Bushveld Complex,
MVT mineralization followed regional dolomitization, silicifica- which intruded the Transvaal sequence ca. 2.06 Ga (Altermann
tion, and sparry dolomitization (Wachowiak, 2001; Wachowiak and Nelson, 1998; Eriksson and Altermann, 1998). Mineraliza-
et al., 1997, 1998). tion at Zeerust consists dominantly of fluorite in stratabound
The Ramah district in northeastern Labrador, Canada, is zones and breccia bodies and the smaller deposits consist largely
the most strongly deformed and metamorphosed of the deposits of sphalerite and galena with variable amounts of fluorite (Mar-
reviewed here, and this limits the certainty with which it can be tini, 1976; Roberts et al., 1993; Martini et al., 1995; Poetter,
placed in the MVT class. The district is hosted by shelf carbon- 2001). As discussed in the next section these deposits are thought
ates of the Reddick Bight Formation, part of the Paleoproterozoic to have formed as part of a large hydrothermal system around the
Ramah Group (Morgan, 1975; Knight and Morgan, 1977, 1981; Bushveld Complex. In view of their high fluorine content and
Mengel et al., 1991). These rocks were deformed and metamor- igneous association, they are not included in this survey.
phosed during the 1.86 Ga Torngat orogeny, reaching green-
schist facies in the area of the MVT prospects (Korstgard et al., Age of Precambrian MVT Deposits
1987; Mengel et al., 1991; Mengel and Rivers, 1994; Scott and
Gauthier, 1996; Hayashi et al., 1997). Mineralization is hosted Geological relations provide the only constraints on the
by breccias that are cemented by sparry dolomite, pyrite, sphal- age of mineralization in most of the Precambrian MVT districts
erite, galena, quartz, and calcite, as well as carbonaceous material discussed here. At Ediacara, mineralization took place during
thought to have been derived from bitumen (Archibald, 1992; formation of the Neocambrian–Cambrian sedimentary sequence
Archibald and Wilton, 1994; Wilton et al., 1993, 1994). Feldspar (Drew and Both, 1984). At Berg Aukas/Abenab, mineralization is
gangue is present locally, although it might have formed during related to 0.75 Ga rifting along the northern margin of the Damara
later remobilization, as is indicated by crosscutting veins with Basin (Frimmel et al., 1996). Age relations for Warrabarty are
MVT-like mineralogy. particularly poorly defined. No measurements are available on
the deposit and the Throssell Group, with which mineralization
MVT Deposits in Archean Basins is roughly coeval, is limited only by 1.08 Ga granites that it over-
lies unconformably and a post-ore 0.71 Ga metamorphic event
With the exception of Gayna River and Nanisivik, the depos- (Blockley and Myers, 1990; Smith, 1996). Common Pb isotope
its described so far are not particularly compelling examples of models suggest that mineralization was synchronous with forma-
MVT mineralization. MVT deposits in the Transvaal Supergroup tion of sediment-replacement copper deposits in the area (Nifty)
of South Africa, however, are remarkably similar to Phanerozoic ca. 0.84 Ga (Smith, 1996).
deposits. The Neoarchean Ghaap and Chuniespoort Groups in At Gayna River, main-stage mineralization is younger than
the lower part of the Transvaal sequence consist largely of car- the 0.78 Ga dikes that cut it (Hewton, 1982). At Coxco (McAr-
bonate platform sediments with abundant stromatolitic reefs and thur River), early mineralization was deposited while karst zones
overlying banded iron formation and contain the Pering, Bushy in the Reward Dolomite were being filled by clastic sediment
Park, Zeerust, and other smaller MVT districts (Beukes, 1987; of the overlying Lynott Formation, which constrains it to an age
Altermann and Nelson, 1998; Eriksson and Altermann, 1998). of ca. 1.64 Ga, and second stage mineralization at Coxco took
These deposits are found in two remnants of the original Trans- place after deposition of the Lynott Formation (Walker et al.,
1983; Page et al., 2000). No age constraints are recognized for coexist with ore minerals at Bushy Park yield an age of 2.145 ±
mineralization at Cooley and Ridge, but they and Coxco should 0.007 Ga, which is the best age estimate available at this time for
be roughly coeval with the 1.69 Ga enclosing sedimentary rocks Bushy Park-Pering mineralization (Schaefer, 2002).
if MVT mineralization is part of the HYC sedex system (Plumb In the Transvaal Basin, some Zn-Pb MVT deposits might
et al., 1990). The age of mineralization at Kamarga has not been have formed ca. 2.35 Ga, as indicated by relations at Gena-
measured; its suggested relation to the Century sedex deposit dendal where mineralization in Chuniespoort carbonate rocks
constrains it to an age only slightly less than that of the 1.67 forms a feeder for Zn-rich shale at the base of the overlying Pre-
Ga McNamara Group host rocks (Plumb et al., 1990; Jones et toria Group (Martini, 1990; Eriksson et al., 2001a). However,
al., 1999). Ages at Esker and Ramah are also not well known. fluorite mineralization at Zeerust yields a Sm-Nd isochron age
At Esker, MVT mineralization is cut by faults related to the of 2.06 Ga that is essentially the same as the Bushveld Com-
earliest phase of compressional deformation in the Coronation plex, which dominated fluid migration in the Transvaal Basin
Group, making it probably post-1.90 Ga and pre-1.84 Ga in age (Kesler et al., 2003).
(Hoffman and Bowring, 1984; Bowring and Grotzinger, 1992;
Wachowiak, 2001). Breccia-hosted ore at Ramah is cut by struc- GEOCHEMISTRY OF PRECAMBRIAN MVT
tures related to the Torngat orogen, which formed during colli- DEPOSITS
sion of the Nain and Rae terranes ca. 1.86 Ga (Wilton et al., 1993;
Archibald, 1992). Fluid Inclusion Temperatures and Geochemistry
Paleomagnetic and isotopic measurements supplement
information from geologic relations at Nanisivik, where a dia- Temperatures of 75–200 °C are commonly cited for Pha-
base dike interpreted to cut ore has been correlated with the 0.72 nerozoic MVT deposits with most of the variation related to
Ga Franklin dike event (Olson, 1984; Heaman et al., 1992; Pehrs- position in the flow path relative to the source basin or possible
son and Buchan, 1999; Symons et al., 2000). Rb-Sr isotope com- sources of additional heat (Leach and Sangster, 1993; Rowan et
positions of sphalerite, dolomite, and leachates from Nanisivik al., 2001). Significantly higher heat flow in Precambrian time
do not fall on an isochron but are consistent with an age between probably produced a hotter basinal environment. Estimates based
ca. 0.75 and 1.25 Ga (Christensen et al., 1993). A paleomagnetic on the history of global heat flow and resulting geotherms (Pol-
pole on recrystallized dolomite around the ore zone corresponds lack and Chapman, 1977; Pollack, 1997) indicate that continental
to an age of ca. 1.095 Ga on the North American apparent polar temperatures during Paleoproterozoic time were probably ~100
wander path (Symons et al., 2000). Although we have accepted °C higher at depths of ~10 km, a level typical of the base of a
a Mesoproterozoic age for Nanisivik here, some organic matter thick sedimentary basin. A more conservative estimate of 50 °C
in the deposit appears to reflect a younger age (Gize, 1986), the is used here, assuming shallower source depths for mineralizing
age of the Franklin dikes has been challenged, and Ar-Ar mea- brines, lower temperatures in active sedimentary basins, and
surements on MVT-related orthoclase yield Ordovician ages published estimates of ocean Precambrian temperatures (Knauth,
(Sherlock et al., 2004). Until these contradictory observations 2005). This effect is shown in Figure 5 as a gradually increasing
are resolved, inclusion of Nanisivik in this compilation should be increment to the Phanerozoic temperature range. Increased ocean
regarded as tentative. temperatures could have facilitated generation of basinal brines
In the Transvaal Supergroup deposits of South Africa, the with salinities above the 10%–30% level typical of Phanero-
age of MVT mineralization is constrained by geologic relations zoic deposits, although it is more difficult to estimate this effect
and isotopic measurements. Mineralization is found largely in because of the range of solution compositions observed in MVT
porosity related to an extensive karst system that developed on deposits. An approximation based on the increase in salinity with
the Ghaap and Chuniespoort rocks in one or more intervals dur- temperature in the NaCl-H2O system (Sourirajan and Kennedy,
ing Paleoproterozoic time. At Bushy Park-Pering, the age of ore- 1962) results in salinities of ~33% for late Archean brines with
related karsting limits mineralization to pre-2.1 Ga (Martini et temperatures of 250 °C. This effect is shown as a similar continu-
al., 1995). The Kalkdam and Katlani prospects, which are hosted ous decrease in inclusion salinities though Precambrian time in
by Ventersdorp lavas that underlie the Transvaal sequence in the Figure 6.
Pering area, have similar Pb isotope compositions and yield a Homogenization temperatures for fluid inclusions, largely
Rb-Sr isochron age of 1.977 Ga, which has been interpreted to in sphalerite, from Precambrian MVT deposits are generally
indicate that mineralizing fluids were expelled during deforma- more complex, with higher temperatures and more CO2 and
tion of the 2.0 Ga Kheis Belt on the western edge of the Kaapvaal locally CH4, than their Phanerozoic counterparts (Fig. 5). Edia-
craton (Duane et al., 1991; Kruger et al., 2001). Relatively high cara, the youngest deposit, has homogenization temperatures
87
Sr/86Sr ratios for calcite, dolomite, and sphalerite at Pering and typical of Phanerozoic MVT deposits (Drew and Both, 1984),
Bushy Park require that the brine contacted evolved, potassium- but temperatures for Berg Aukas (100–210 °C), Gayna River
rich rock outside the ore-hosting carbonate sequence, possibly (156–231 °C), Warrabarty (165–245 °C for gray-stage sphal-
during the Kheis event (Kruger et al., 2001; Schaefer et al., 2001; erite), and Nanisivik (87 to above 300 °C) are progressively
Kesler et al., 2003). Ar-Ar measurements on illite thought to higher (Misiewicz, 1988; Carriere and Sangster, 1992; Olson,
FLUID INCLUSION HOMOGENIZATION TEMPERATURES
EDIACARA
DECREASING AGE
L. Cambrian
OF DEPOSITS AND
GAYNA HOST ROCKS
0.16
RIVER
0.78-1.08 Ga
0.08
0
BERG AUKAS
~0.75Ga
0.12 NANISIVIK
0.08 ~1.25 Ga
0.04
0
0.50 COXCO II
0.40 1.60-1.68 Ga
0.30
0.20
0.10
0
KAMARGA – 1.65 Ga
ESKER - ~1.9 Ga
RAMAH - ~2.0 Ga
0.16 BUSHY
BUSHY
PARK
0.12
~2.1 Ga
0.08
0.04
0
0 100 200 300 400
TEMPERATURE (ºC)
ºC)
Figure 5. Homogenization temperatures of fluid inclusions in sphalerite from Precambrian MVT deposits compiled from sources discussed in the
text. Arrows show range of values where data are not available to plot a histogram. Shaded rectangle in background shows range of temperatures
typical of Phanerozoic MVT deposits (Leach and Sangster, 1993), and darker shaded triangle shows increased temperatures that might result
from higher Precambrian heat flow as discussed in the text.
1984; McNaughton and Smith, 1986). Highest temperatures at print that might have corrupted the homogenization temperature
Nanisivik, which are slightly above 300 °C, have been attributed record (Archibald, 1992). At Kamarga, fluid inclusions in sphal-
to reheating by the diabase dike that cuts ore, but temperatures erite have homogenization temperatures of 270–320 °C and are
of 250 ºC have been interpreted as primary, perhaps reflecting accompanied by CO2-rich vapor inclusions that are not as clearly
higher heat flow for the host rift-margin basin (Olson, 1984; related to a post-ore overprint (Jones et al. 1999).
McNaughton and Smith, 1986). Fluid inclusions in Archean-hosted Zn-rich MVT deposits
This trend of increasing homogenization temperatures with are more typical of those in Phanerozoic-hosted MVT deposits
increasing age does not continue for Mesoproterozoic and older with the exception of high-temperature, gas-rich inclusions of
deposits, however. Inclusions in sphalerite from second-stage uncertain origin that are present in some of the older deposits.
mineralization at Coxco (McArthur River) homogenize at 100– Reconnaissance observations that we have made on sphalerite
170 °C and those at Esker homogenize at ~100–150 ºC, both of and sparry dolomite from Bushy Park yielded homogenization
which are typical of Phanerozoic MVT deposits. Homogeniza- temperatures of 77–120 °C for primary and pseudosecondary
tion temperatures for sphalerite at Ramah fall in the same range aqueous inclusions and 130–195 °C for secondary aqueous
(120–180 °C), but inclusions in quartz and dolomite extend to inclusions, all in sphalerite (Table 2). Secondary aqueous inclu-
320 °C and contain CO2, probably reflecting a metamorphic over- sions from sparry dolomite homogenized at temperatures of
DECREASING AGE
OF DEPOSITS AND
FLUID INCLUSION SALINITIES HOST ROCKS
0.4 GAYNA
RIVER
0.2 0.78-1.08 Ga
0
WARRABARTY ~0.71-1.08 Ga
NANISIVIK ~1.25 Ga
0.3
0.2 COXCO II
1.60-1.68 Ga
0.1
0
KAMARGA – 1.65 Ga
ESKER - ~1.9 Ga
RAMAH - ~2.0 Ga
0.3
BUSHY
0.2 PARK
0.1 ~2.1 Ga
0
0 10 20 30 40
EQUIVALENT WEIGHT PERCENT NaCl
Figure 6. Salinity of fluid inclusions from Precambrian MVT deposits compiled from sources discussed in the text. Arrows show range of values
where data are not available to plot a histogram. Shaded rectangle in the background shows the range of salinities typical of Phanerozoic MVT
deposits (Leach and Sangster, 1993), and darker shaded triangle shows increased salinities that might result from higher temperatures caused by
higher Precambrian heat flow, as discussed in the text.
TABLE 2. SUMMARY OF FLUID INCLUSION MEASUREMENTS

FOR THE BUSHY PARK DISTRICT, SOUTH AFRICA
Sample no. Mineral Type Homogenization Freezing temperature Equivalent
temperature (°C) wt% NaCl
(°C)
BP-01-A Sphalerite (L+V) 184–195 –9 to –8 12–13
BP-03 Sphalerite (L+V) 129–192 –11 to –8 12–15
BP-13 Dolomite (L+V) 150–187 –11 to –5 8–15
150–190 °C, slightly higher than temperatures of 100–175 °C Freezing temperatures (and salinities estimated from them)
reported by Wheatley et al. (1986b) for gangue carbonate at Pering. for fluid inclusions in sphalerite from some of the Precambrian
Schaefer et al. (2001) reported a similar range of homogenization MVT deposits extend to values in and above the high end of the
temperatures (90–168 °C) for Bushy Park sphalerite and dolomite, range typical of Phanerozoic MVT deposits (Fig. 6). Whereas the
with no systematic difference between inclusions in different min- range of freezing temperatures for inclusions at Gayna (–12 to
erals. They also reported primary vapor-rich inclusions containing −24 °C) falls in the center of the Phanerozoic range, those from
variable proportions of CO2 and CH4 from Bushy Park, and used Coxco II (McArthur River) (−22 to –28 °C) and Esker (−18 to
these inclusions to determine a pressure correction of ~50 °C for the –24 °C) are nearer the high end of the range (Walker et al., 1983;
homogenization temperatures. Greyling et al. (2001) reported tem- Wachowiak et al., 1997, 1998; Carriere and Sangster, 1992;
peratures of 157–210 °C for Pering based on intersecting isochors Smith, 1996). Freezing temperatures from quartz and dolomite at
in aqueous and carbonic inclusions. Ediacara (−23 to –28 °C) are also near the top of the Phanerozoic
MVT range. Similar patterns are seen in the deposits for which TABLE 3. COMPOSITION OF FLUID INCLUSION LEACHATES
only salinities are quoted. Inclusions in sphalerite and dolomite FROM THE BUSHY PARK DISTRICT, SOUTH AFRICA
Sample no. Mineral Na/Br Cl/Br Na/K Ca/Mg
from Warrabarty have salinities of 15–26 total salt and contain (atomic) (atomic) (atomic) (atomic)
significant amounts of Ca (Smith, 1996); inclusions in quartz and BP-1 Sphalerite 196 336 17.3 6.4
Sphalerite 204 335
dolomite from Ramah have salinities of 15–29 equivalent wt% Dolomite 171 333
NaCl (D. Wilton, 2002, written commun.), and inclusions from BP-2 Sphalerite 204 331
sphalerite and dolomite at Berg Aukas have average salinities of BP-3 Sphalerite
Dolomite
216
188
336
347
23 equivalent wt% NaCl (Misiewicz, 1988). All of these are in Dolomite 192 361
the upper part of the Phanerozoic MVT range (Fig. 6). Inclusion BP-6 Sphalerite
Sphalerite
204
197
321
303
16.2
16.4
7.3
7.9
fluids at Nanisivik have even higher salinities, with first-melting Sphalerite 188 313
(eutectic) temperatures of –50 °C or more, unusually low last- BP-8
BP-10
Sphalerite
Sphalerite
212
184
336
305
melting temperatures of –25 to –45 °C, and estimated salinities of Dolomite 167 308
24–35 equivalent wt% NaCl (McNaughton and Smith, 1986). BP-13 Sphalerite 163 305
Sphalerite 161 308
Our reconnaissance freezing measurements on primary and Sphalerite 150 285
pseudosecondary inclusions in Bushy Park sphalerite indicate the Note: Analyses carried out at University of Michigan using method
described by St. Marie et al. (2000).
presence of significant Ca or Mg in the fluids, and yield final-
melting temperatures of –1 to –20 °C, corresponding to salinities
of up to 22 equivalent wt% NaCl (Fig. 6). Secondary inclusions
700
in sphalerite and sparry dolomite, which are not included in Fig- Fluid Inclusion Leachate Compositions SW
ure 6, have final melting temperatures of –3 to –14 °C, corre-
600
sponding to salinities of ~5–17 equivalent weight percent NaCl.
Schaefer et al. (2001) reported a similar range of salinities from
Cl/Br (atomic)
500 Seawater
27 to 1 equivalent wt% NaCl for inclusions at Bushy Park and Evaporation
suggested that it reflected the presence of a high salinity ore fluid 400
Curve
that mixed with meteoric water to form fluids with intermediate
salinities, which is in agreement with our observations. Greyling
Cl
300
et al. (2001) reported a similar large range of salinities (but not Sphalerite
freezing temperatures) for Pering, extending from only a few 200
Bushy Park
equivalent wt% NaCl to values as high as 50%, reflecting abun- Nanisivik
dant Ca and Mg in the inclusions. Sparry Dolomite
100 Bushy Park
Na-Cl-Br compositions of fluid inclusion leachates from Berg Aukas
most Phanerozoic MVT minerals fall on or near the seawater 0
evaporation line in plots of Na/Br versus Cl/Br, suggesting that 0 100 200 300 400 500 600
their source brines formed by evaporation of seawater (Kesler Na/Br (atomic)
et al., 1995, 1996; Viets et al., 1996; Chi and Savard, 1997; St.
Figure 7. Na/Br versus Cl/Br diagram comparing composition of fluid
Marie and Kesler, 2000). Of the Precambrian MVT deposits and inclusion leachates from Precambrian MVT deposits with the compo-
districts included in this study, Na-Cl-Br leachate data are avail- sition of evaporated modern seawater (line). Data from this study for
able for Nanisivik (Viets et al., 1996) and Berg Aukas (Chetty and Bushy Park, from Chetty and Frimmel (2000) for Berg Aukas, and
Frimmel, 2000) and are reported here for Bushy Park (Table 3). from Viets et al. (1996) for Nanisivik.
Two leachates from sphalerite at Nanisivik and one from sparry
dolomite at Berg Aukas plot just above the seawater evapora-
tion line, with Nanisivik farther along the evaporation trend (Fig.
7). Leachates from Bushy Park sphalerite and sparry dolomite applied here (Fig. 8). Best agreement between MVT sulfides and
plot in about the same location as Berg Aukas, also just above rock sulfate is seen at Nanisivik, where δ34S values of 21‰–31‰
the seawater evaporation line (Fig. 7). The similarity in position for MVT sulfides are almost identical to δ34S values of 22‰–32‰
of Precambrian and Phanerozoic MVT deposits in Figure 7 sug- for evaporite gypsum from the Society Cliffs Formation (Olson,
gests that the Na-Cl-Br composition of seawater and MVT brine- 1984; Ghazban et al., 1990). Ghazban et al. (1990) showed that
forming process were similar throughout most of Earth history. δ34S values of the sulfides in Nanisivik ore could be accounted
for by deposition from a fluid containing S with δ34S values of
Sulfur Isotope Geochemistry 26 ± 1‰ and suggested that the S was derived from seawater,
and more recent data of Kah et al. (2001) show that it could have
Sulfides in most of the Precambrian MVT deposits have high come directly from Society Cliffs gypsum. Unusually low δ13C
δ34S values that approach those of sulfate in coeval evaporites or values of 6‰ to –12‰ for sparry dolomite in the Nanisivik ore
seawater, a pattern that is consistent with the Anderson model assemblages suggest that the sulfate was reduced by reaction
DECREASING AGE
SULFUR ISOTOPE COMPOSITIONS OF DEPOSITS AND
HOST ROCKS
EDIACARA - L. Cambrian
BERG AUKAS ~0.75Ga
WARRABARTY ~0.71-1.08 Ga
0.8
0.6 NANISIVIK ~1.25 Ga
0.4
0.2
0
0.6
0.4
COXCO II - 1.60-1.68 Ga
0.2
0
COXCOI >1.68 Ga
0.4
0.2
0
KAMARGA – 1.65 Ga
RAMAH - ~2.0 Ga
0.4 BUSHY PARK ~2.1 Ga
0.2
0
0 10 20 30 40
δ34S (‰)
Figure 8. Histograms showing isotopic composition of sulfur in sphalerite and galena from Precambrian MVT deposits. Arrows show range of
values where data are not available to plot a histogram. Shaded rectangles show isotopic composition of sulfate in seawater coeval with host
rocks for the deposits, based on sources discussed in the text.
with organic matter, providing strong support for the Anderson tope values are available for sulfate in the sedimentary host rocks,
model (Ghazban et al., 1990). although δ34S values estimated for seawater sulfate of Paleopro-
Sulfides at Coxco (McArthur River) have δ34S values signif- terozoic age range from ~15‰ to 20‰, which falls within the
icantly lower than coeval seawater. δ34S values of 1.3‰–21.7‰ range of MVT sulfide values (Fig. 8). Sulfur isotope data are
are lower than the 20‰–32‰ estimated for sulfate in the McAr- not available for ore at Esker, but trace sulfate in the Rocknest
thur Group from analysis of trace sulfate in carbonate rocks and Formation has δ34S values of ~12‰–25‰ (Ueda et al., 1991).
barite that probably replaced sedimentary sulfates (Walker et al., Sulfur isotope analyses have not been published for the Gayna
1983; Bottomley et al., 1992). Second-stage sulfides at Coxco deposits. δ34S values for gray-stage sphalerite at Warrabarty
have lower values of 0.9‰–16.1‰, and sphalerite, galena, and range from 1.5‰ to 20.4‰, a somewhat larger range than seen
pyrite from the Cooley and Ridge deposits have even lower in the other deposits, but have a “distinct mode” at 11‰–14‰
δ34S values that are similar to those in the HYC sedex deposit (Smith, 1996). Uncertainty about the age of Warrabarty mineral-
(Walker et al., 1983; Rye and Williams, 1981). A possible source ization limits the degree to which it can be compared with coeval
of S with the necessary low δ34S values is pyrite or H2S with δ34S seawater sulfate, although the best comparison is probably with
values near 0‰ in nearby Tawallah Group black shales (Shen et that for Berg Aukas.
al., 2002). δ34S values obtained in this study for Bushy Park sulfides
Sulfides of possible MVT origin from Ramah have δ34S val- (Table 4) range from 15.2‰ to 16.6‰ for galena and 15.8‰–
ues of ~8–34‰ (D. Wilton, 2002, written commun.). No S iso- 19.7‰ with one unusually high value of 24.1‰ for sphalerite
TABLE 4. SULFUR ISOTOPE ANALYSES OF MINERALS More detailed grade-tonnage and frequency comparisons,
FROM THE BUSHY PARK DISTRICT, SOUTH AFRICA however, indicate that the two groups differ. The grade-tonnage
Sample Mineral ¥34S (‰)
BP-13-GA Galena 15.2
plot for Precambrian and Phanerozoic MVT deposits (Fig. 9) can
BP-UEO-3-GA Galena 16.6 be divided into four quadrants at a grade of 7% Pb+Zn and a
BP-UEO-1-SP Sphalerite 15.8
BP-13-SP Sphalerite 17.3
tonnage of 10 million that results in a nearly equal number of
BP-UEO-3-SP Sphalerite 17.3 Phanerozoic deposits in all four quadrants (29%, 24%, 26%, and
21% for NW, NE, SE, and SW quadrants, respectively). Precam-
BP-UEO-8-SP Sphalerite 24.1 brian deposits are distributed much differently among the quad-
Note: Analyses carried out at Queen's University in the
laboratory of T.K. Kyser.
rants, however, with none (0%) in the NW, 13% in the NE, 62%
in the SE, and 25% in the SW (Fig. 9). Viewed only from the
perspective of grade, only 13% of the Precambrian deposits fall
(Fig. 8). Schaefer et al. (2001) reported lower δ34S values of above the 7% division compared with 53% of the Phanerozoic
4.8‰–8.5‰ for sphalerite and galena and –9.7‰–26.7‰ for deposits. However, 75% of the Precambrian MVTs have more
diagenetic(?) pyrite from Bushy Park. Sulfate evaporites are not than 10 million tons compared with only 50% of the Phanerozoic
known in the Transvaal sequence, but δ34S values of 13‰–17‰ deposits. Thus, Precambrian deposits are generally lower grade
have been reported for trace sulfate in carbonate rocks from the but higher tonnage, which should mean that they have generally
Malmani Subgroup (Buchanan and Rouse, 1982, in Strauss, similar metal contents. This is confirmed by the fact that ~50% of
1993), similar to the range estimated by Canfield and Raiswell the Precambrian deposits plot above the 1 million ton diagonal in
(1999) for late Archean seawater. This range is very similar to Figure 9, compared with ~41% of the Phanerozoic deposits. Con-
most of our values for sulfides at Bushy Park (Table 4), suggest- sidering all of the uncertainties involved, this suggests that the
ing that MVT sulfide was derived largely from coeval seawater amount of metal moved by Precambrian and Phanerozoic MVT-
sulfate. The lower δ34S values observed for Bushy Park sphalerite forming systems was about the same, but that fewer Precambrian
by Schaefer et al. (2001) probably reflect additions of S from systems made ore.
diagenetic pyrite or related sources. Frequency comparisons indicate further than there were
not many Precambrian systems. Figure 1 shows that ~11 MVT
SIGNIFICANCE OF PRECAMBRIAN MVT deposits and districts formed during Precambrian time versus at
DEPOSITS TO EARLY EARTH ATMOSPHERE AND least 30 during Phanerozoic time. In view of the large difference
HYDROSPHERE COMPOSITIONS in duration of these time periods, this amounts to a very differ-
ent indicated “MVT-formation rate” of ~5.5 per billion years for
Comparison of Precambrian and Phanerozoic MVT Precambrian deposits and districts versus 60 per billion years
deposits for Phanerozoic deposits and districts. If Nanisivik is not
Our results show that Precambrian and Phanerozoic MVT

deposits share many similarities. Deposits of both ages are along
100
the margins of passive-margin and rifted basins, and are hosted PHANEROZOIC
DEPOSITS
largely by platform carbonate rocks containing extensive reefs. 50 PRECAMBRIAN
With the exception of Ramah and Esker, most of the Precambrian DISTRICTS PHANEROZOIC
10
PRECAMBRIAN
10
MVT deposits are in undeformed sequences or foreland thrust

6
5
10
belts, as are most of their Phanerozoic counterparts. Both Pre-

Pb + Zn (wt%)
10
7
NANISIVIK 8
cambrian and Phanerozoic deposits have generally similar and ~1.25 Ga
simple mineralogy. Where ore mineralogy is more complex and 10 GAYNA RIVER
0.78-1.08 Ga
10
includes elements such as copper, silver, and cobalt, these are

4
WARRABARTY
interpreted to have resulted from local features that contaminated 5
0.71-1.08 Ga ESKER
1.9 Ga
the ore fluid (Frimmel et al., 1996; Wu et al., 1997). Fluid inclu-
sion observations, including homogenization temperatures and BUSHY
COXCO PARK
salinities, are also generally similar for deposits of both ages, and PERING KAMARGA
1.61-1.68 Ga ~2.1 Ga ~2.1 Ga? 1.65 Ga
Precambrian exceptions seem to be just that rather than indica- 1
tions of systematic temporal changes. For instance, only Gayna 0.1 1.0 10 100 1000
River and Nanisivik, among the unmetamorphosed Precambrian ORE TONNAGE (TONNES/106 )
MVT deposits discussed here, have significantly higher tempera-
tures than those of Phanerozoic deposits (which probably reflect Figure 9. Grade-tonnage plot comparing MVT deposits and districts
of Phanerozoic and Precambrian age. Data for Phanerozoic deposits
more proximal basinal sources or later dike-related heating), and and districts from Leach and Sangster (1993) and data for Precambrian
even older Precambrian deposits such as Coxco and Bushy Park deposits and districts from references cited here. Grade-tonnage for
have temperatures typical of Phanerozoic deposits (Fig. 5). Warrabarty shown as a range estimated from Smith (1996).
Precambrian in age, as noted above, the Precambrian MVT- nental growth rates are themselves matters of significant debate
formation rate would be even lower. The age of ore-hosting (Condie, 2000), but even those based on episodic growth produce
sedimentary units varies greatly for both Precambrian and Pha- relatively smooth cumulative curves for the change in total crust
nerozoic rocks, although the significance of this variation is com- volume with time. As can be seen in Figure 10, growth rates of
plicated by uncertainty about the exact age of mineralization. this type are not similar to our estimate of the formation of MVT
For instance, Cambrian rocks contain a significant majority of deposits through Precambrian time. This is not surprising in view
deposits that probably formed later in Paleozoic time (Leach et of the other important factors that control shelf carbonate devel-
al., 2001). Within the Precambrian, MVT formation appears to opment, especially sea level and climate (Walker et al., 2002).
have been greatest during late and early Proterozoic time, with In the absence of global data, more detailed comparisons of
only Nanisivik from the intervening period. These complications shelf carbonate volumes must be confined to specific areas. One
can be minimized by comparing the abundance of deposits in of the largest areas of early Precambrian carbonate sedimenta-
0.5-b.y. groups that bracket the range of possible depositional tion was the Transvaal-West Griqualand-Kanye Basins of the
ages for most of the deposits. As seen in Figure 10, a plot of this Kaapvaal Craton in South Africa, where platform sediments cov-
type shows a steep rise in MVT formation for the last 0.5 b.y. of ered an area of ~600,000 km2 (Beukes, 1987). If carbonates of the
Earth history and a much smaller and not greatly changeable rate Kaapvaal Craton correlate with those of the Jeerinah Formation
during earlier (Precambrian) time. in the Pilbara Craton of Australia to form the Vaalbara terrane, as
Frequency comparisons of this type should be normalized suggested by Cheney (1996), this platform probably covered at
to the amount of favorable carbonate-bearing sedimentary basins least an additional 100,000 km2. MVT deposits in this sequence
that remain from each of the time periods of interest, thus taking can be compared with two important shelf sequences in North
into account both formation of appropriate host rocks and their America: (1) Cambrian–Ordovician platform sediments of the
preservation during later events. For instance, carbonate rocks Appalachian Basin (USA) with an area of at least 100,000 km2,
are scarce in most early Archean greenstone belts and related and (2) Devonian platform sediments of the Lennard shelf (Aus-
sedimentary basins; significant volumes of carbonate sediment tralia) with an area of ~50,000 km2 (Kesler, 1996; Vearncombe
began to form only by ca. 2.7–2.4 Ga when favorable shelf et al., 1996). MVT mineralization is widespread in the Vaalbara
environments developed for the first time (Eriksson et al., 1998; of South Africa but absent from correlative rocks in Austra-
2001a, 2001b). Thus, the lower frequency for Precambrian MVT lia (Blockley and Myers, 1990). In the Appalachian province,
deposits might simply reflect a lack of suitable shelf carbonates MVT mineralization is found in all areas, although it is most
in which ore could have formed. Although quantitative data are abundant in the south and progressively less abundant north-
not available on the change in shelf carbonates through time, they ward (Kesler, 1996). In the Lennard shelf, MVT mineralization
should be related to continental growth rates because shelf envi- is found largely in two areas, at the north and south ends of the
ronments formed on the margins of flooded continents. Conti- belt (Vearncombe et al., 1996).
35 100
CRUSTAL GROWTH
30
NUMBER OF MVT DEPOSITS OR DISTRICTS
CUMULATIVE CRUSTAL GROWTH (PERCENT)
30 PRECAMBRIAN MARINE
SULFATE - OHMOTO
SULFATE CONCENTRATION ((Mm)
25 Figure 10. Number of MVT deposits and

districts hosted by rocks of Phanerozoic
and Precambrian age compiled from
20 data of Leach and Sangster (1993) and
this paper, plotted in 0.5-b.y. intervals.
Crustal growth curve estimated from
15 data of Condie (2000) and competing
estimates of marine sulfate from Kah et
al. (2004) and Ohmoto (2004). Fluctua-
10
tions in sulfate concentration (Lowen-
PRECAMBRIAN
stein et al., 2003) of the Paleozoic ocean
MARINE SULFATE - KAH
5 are not shown.
MVT
DEPOSITS
0 0 0
0 0.25
0-0.5 0.75
0.5-1.0 1.25
1.0-1.5 1.75
1.5-2.0 2.25
2.0-2.5
AGE (BILLIONS OF YEARS)
The density of MVT mineralization in these Precambrian Sulfate in the ocean is linked, in turn, to the composition of the
and Phanerozoic basins differs greatly, however. Looking only atmosphere and, especially, its oxygen content.
at reserves and production for deposits and districts that have Two competing histories have been suggested for the evolu-
been or probably will be mined, the Vaalbara province con- tion of the ocean and coexisting atmosphere. One history holds
tains ~1 million tons of Zn compared with ~4 million tons for that the oxygen content of the atmosphere increased by as much
the Appalachian province and 4 million tons for the Lennard as an order of magnitude from levels of less than 10−2 present
shelf. If MVT Zn-Pb and F-Ba districts, rather than deposits atmospheric level (PAL) during the Great Oxidation Event
and prospects, are counted, the Vaalbara contains approxi- (GOE) ca. 2.3 Ga (Holland, 1999; 2002; Bekker et al., 2004).
mately six districts compared with at least 25 for the Appala- Prior to the GOE, sulfate concentrations in the ocean are esti-
chian province and about six for the much less extensive Len- mated to have been less than 1% of present levels and the main
nard shelf. Thus, when viewed in terms of the areas, Phanero- source of sulfate was disproportionation or photochemical oxi-
zoic shelf sequences clearly contain more MVT ore than their dation of volcanic SO2 and oxidation of H2S by cyanobacteria
Precambrian analogues. (Hattori and Cameron, 1986; Canfield and Raiswell, 1999; Far-
If the scarcity of Precambrian MVT deposits cannot be quhar et al., 2000; Habicht et al., 2002; Holland, 2002; Farquhar
attributed to a lack of reef carbonate rocks, perhaps it is the and Wing, 2003). After the GOE, sulfate concentrations in the
porosity of the carbonates that limited Precambrian MVT ocean are thought to have increased gradually through Protero-
mineralization. Low porosity would have limited the volume zoic time, reaching present levels of 10–30 mM only near the
of MVT minerals that could be deposited by invading fluids, end of Proterozoic time (Shen et al., 2002; Kah et al., 2004). The
and therefore the grade of the deposit. This might be suspected alternative view holds that there was no GOE and the oxygen
because Phanerozoic-age carbonate reefs were built by a vari- content of the atmosphere and the sulfate content of the oceans
ety of organisms, whereas Precambrian reefs are almost exclu- have fluctuated between 10 and 30 mM since at least Neoar-
sively stromatolitic (Webb, 2001; Petrov and Semikhatov, 2001; chean time (Ohmoto, 1997, 2004).
Grotzinger, 1986a, 1986b, 1986c; Hoffman, 1989; Ricketts and The two competing marine sulfate histories are shown in
Donaldson, 1989). Systematic comparisons of porosity in stro- Figure 10, where it can be seen that our curve for the change
matolitic reefs versus reefs of other types are not available, but in MVT mineralization through time nearly parallels the GOE-
anecdotal information that can be gleaned from literature on type curve. Similarity of the MVT and GOE-type seawater sul-
oil fields does not offer much encouragement for this line of fate curves suggests that sulfate concentrations similar to those
thought. Recent studies of stromatolitic reef rocks of a range in the modern ocean are required for efficient MVT mineraliza-
of ages include descriptions of oil and MVT mineralization, as tion, and that a scarcity of sulfide-rich traps derived from this
well as porosity estimates of 8%–11% (Counter, 1993; White- sulfate limited the formation of MVT deposits in Precambrian
sell, 1995; Kuznetzov, 1997; Grotzinger, 2000; Lemon, 2000; time. Evaporative brines derived from low-sulfate seawater were
Osmond, 2000). Porosities at and below the low end of this probably capable of dissolving and transporting metals in much
range are sufficient to produce a deposit containing 7% Pb+Zn the same amounts observed in Phanerozoic basins. However,
regardless of the relative abundances of galena and sphalerite. low sulfate concentrations in basinal waters would have limited
These comparisons argue against the tempting possibility the amount of sulfide that could have formed from such brines
that the scarcity of Precambrian MVTs can be attributed solely (or their rare evaporites), whether by bacterial or thermochemi-
to a lack of suitable Precambrian carbonate rocks. Furthermore, cal reduction. Local accumulations of sulfate-rich seawater, such
by including in this compilation several Precambrian MVT dis- as those proposed for Archean and Neoproterozoic time, might
tricts that have been explored but not produced, we have also have been the necessary precursors to sulfide-rich traps for early
minimized the likelihood that less extensive exploration in MVT ore formation (Cameron, 1982, 1983; Kasting, 1991). In
Precambrian terranes can account for their lower MVT endow- that context, it is interesting to note that some Precambrian MVT
ment. Thus, we conclude that Precambrian MVT deposits are deposits, especially Nanisivik and Kamarga, were close to evap-
indeed less abundant than Phanerozoic deposits and must seek orite sequences, a factor that might have led to higher sulfate and
a reason beyond local geologic factors for this difference. derived sulfide concentrations.
Similarity of these curves does not exclude other possible
Relation between Atmosphere-Hydrosphere Compositions controls on MVT genesis. Christensen et al. (1997) and Leach
and MVT Mineralization et al. (2001) have pointed out that Phanerozoic MVT deposits
formed largely during Devonian and Permian time, possibly
With no geologic reasons to account for the global MVT his- because of long-distance brine migration during episodic com-
tory shown in Figure 10, the pattern likely reflects changes in the pressive or extensional tectonic events (Bradley and Leach,
composition of the atmosphere and hydrosphere. As noted at the 2003; Kesler et al., 2004). In addition, the sulfate content of the
outset of this study, the efficiency of MVT mineralization should world ocean during Phanerozoic time varied considerably in
be directly related to the sulfate content of basinal fluids and, response to large-scale continent aggregation and disaggrega-
by extension, the ocean in which the sediments were deposited. tion, although it did not reach levels as low as those estimated
for Paleoproterozoic and Neoarchean time (Horita et al., 2002; volcaniclastic sediments: Sedimentary Geology, v. 120, p. 225–256, doi:
10.1016/S0037-0738(98)00034-7.
Lowenstein et al., 2003). All of these factors operate at tim- Anderson, G.M., 1983, Some geochemical aspects of sulphide precipitation in
escales considerably smaller than the 500 Ma interval used in carbonate rocks, in Kisvarsanyi, G., Grant, S.K., Pratt, W.P., and Koenig,
Figure 10, however, and it appears that they were masked by a J.W., eds., Proceedings, International Conference on Mississippi Valley-
Type Lead-Zinc Deposits: Rolla, University of Missouri-Rolla, p. 61–76.
longer-term increase in the sulfate content of the ocean during Anderson, G.M., 1991, Organic maturation and ore precipitation in Southeast
Precambrian time. Missouri: Economic Geology and the Bulletin of the Society of Economic
Although the similarity of the MVT and GOE-type sul- Geologists, v. 86, p. 909–926.
Anderson, G.M., and Macqueen, R.W., 1988, Ore deposit models—6: Missis-
fate curves agrees with the “GOE-type history” for Earth’s sippi Valley-type lead-zinc deposits: Geoscience Canada Reprint Series,
early atmosphere and hydrosphere, perplexing complications v. 3, p. 79–90.
remain. In particular, the Neoarchean–Paleoproterozoic Trans- Anderson, B.R., Gemmell, J.B., and Berry, R.F., 2001, The geology of the Nifty
copper deposit, Throssell Group, western Australia: Implications for ore
vaal Basin, which is significantly older than the GOE, gener- genesis: Economic Geology and the Bulletin of the Society of Economic
ated MVT deposits very similar to those in Phanerozoic rocks. Geologists, v. 96, p. 1535–1565.
Perhaps these deposits, which probably formed ca. 2.0 Ga, Anderson, B.R., Gemmell, J.B., and Nelson, D.R., 2002, Lead isotope evolu-
tion of mineral deposits in the Proterozoic Throssell Group, western Aus-
took advantage of GOE-related sulfate. Or, maybe they did not tralia: Economic Geology and the Bulletin of the Society of Economic
form by the Anderson-type model advocated here. Neither of Geologists, v. 97, p. 897–911.
these alternatives seems geologically reasonable, however, and Archibald, S., 1992, The geology, mineralization, geochemistry and metallog-
eny of the Paleoproterozoic Ramah Group, northern Labrador [M.Sc. the-
it is much more likely that the basinal brines that formed Bushy sis]: St. John’s, Memorial University of Newfoundland, 76 p.
Park and Pering were part of the Transvaal sedimentary sys- Archibald, S.M., and Wilton, D.H.C., 1994, A lower Proterozoic (ca. 2.0 by)
tem and therefore representative of Neoarchean–Paleoprotero- MVT deposits from the Ramah Group, northern Labrador: Geological
Association of Canada Program with Abstracts, v. 19, p. 4.
zoic conditions. Thus, enough sulfate to form MVT deposits Arne, D.C., Curtis, L.W., and Kissin, S.A., 1991, Internal zonation in a car-
was almost certainly available, even at this early stage of Earth bonate-hosted Zn-Pb-Ag deposit, Nanisivik, Baffin Island, Canada: Eco-
history. Whether this was sulfate that formed because of local nomic Geology and the Bulletin of the Society of Economic Geologists,
v. 86, p. 699–717.
environmental factors or these deposits are the sole remaining Barley, M.E., and Groves, D.I., 1992, Supercontinent cycles and the distri-
MVT signal of a more sulfate-rich early Earth remains to be bution of metal deposits through time: Geology, v. 20, p. 291–294, doi:
determined. 10.1130/0091-7613(1992)020<0291:SCATDO>2.3.CO;2.
Barrett, T.J., and Anderson, G.M., 1988, The solubility of sphalerite and galena
Finally, although basin-hosted deposits are most likely to in 1–5 m NaCl solutions to 300°C: Geochimica et Cosmochimica Acta,
reflect atmosphere and ocean compositions, our results provide v. 52, p. 813–820, doi: 10.1016/0016-7037(88)90353-5.
encouragement for study of other subsurface deposits. Deeper Baugaard, W.D., Gregg, J.M., and Ahler, B., 2001, Stratigraphy and diagenesis
of the Paleoproterozoic Bushy Park zinc-lean deposit, South Africa: Geo-
deposits related to igneous activity might also be useful. For logical Society of America Abstracts with Programs, v. 33, p. 337.
instance, the oxidation state of magmas controls the form of Bekker, A., Holland, H.D., Wang, P.-L., Rumble, D., III, Stein, J.H., Coetzee,
S in the magma, which controls, in turn, whether metals form L.L., and Beukes, N.J., 2004, Dating the rise of atmospheric oxygen:
Nature, v. 427, p. 117–121, doi: 10.1038/nature02260.
immiscible sulfides in the magma or separate into a vapor Beukes, N.J., 1987, Facies relations, depositional environments and diagenesis
phase that could form hydrothermal ore deposits. Although in a major Early Proterozoic stromatolitic carbonate platform to basinal
current estimates suggest that fO2 of the mantle has remained sequence, Campbellrand Subgroup, Transvaal Supergroup, Southern
Africa: Sedimentary Geology, v. 54, p. 1–46, doi: 10.1016/0037-
within a single log-unit of present values (Canil, 1997; Delano, 0738(87)90002-9.
2001; Holland, 2002), even these small differences can produce Blockley, J.G., and Myers, J.S., 1990, Proterozoic rocks of the Western Austra-
large changes in the SO2/H2S ratio of magmas and associated lian Shield—Geology and mineralization, in Hughes, F.E., ed., Geology
of the mineral deposits of Australia and Papua New Guinea: Melbourne,
hydrothermal solutions (Core, 2004). Australasian Institute of Mining and Metallurgy, p. 607–615
Blount, C.W., 1977, Barite solubilities and thermodynamic quantities up to
ACKNOWLEDGMENTS 300°C and 1400 bars: American Mineralogist, v. 62, p. 942–957.
Bottomley, D.J., Veizer, J., Nielsen, H., and Moczydlowska, M., 1992, Iso-
topic composition of disseminated sulfur in Precambrian sedimentary
We are grateful to Derek Wilton, Nawojka Wachowiak, rocks: Geochimica et Cosmochimica Acta, v. 56, p. 3311–3322, doi:
Bruce Ehlers, Peter McGoldrick, and Bruce Gemmell for 10.1016/0016-7037(92)90307-5.
Bowring, S.A., and Grotzinger, J.P., 1992, Implication of new chronostratigra-
sharing information with us on Ramah, Esker, Bushy Park, phy for the tectonic evolution of Wopmay Orogen, northwest Canadian
Kamarga, and Warrabarty, respectively. The manuscript has shield: American Journal of Science, v. 292, p. 1–20.
been improved by conversations with B.H. Wilkinson and Bradley, D.C., and Leach, D.L., 2003, Tectonic controls of Mississippi Valley-
type lead-zinc mineralization in orogenic forelands: Mineralium Depos-
Linda Kah, and challenging reviews by H.L. Barnes, M.B. ita, v. 38, p. 652–667, doi: 10.1007/s00126-003-0355-2.
Goldhaber, and H. Ohmoto. Burstein, I.B., Shelton, K.L., Hagni, R.D., and Brandom, R.T., 1992, Mobiliza-
tion of copper by MVT fluids penetrating igneous basement rocks; sulfur
isotope studies of the Precambrian Boss-Bixby Fe-Cu deposits, Southeast
REFERENCES CITED Missouri: Geological Society of America Abstracts with Programs, v. 24,
no.7, p.234.
Altermann, W., and Nelson, D.R., 1998, Sedimentation rates, basin analysis and Cameron, E.M., 1982, Sulphate and sulphate reduction in early Precambrian
regional correlation of three Neoarchean and Paleoproterozoic sub-basins oceans: Nature, v. 296, p. 145–148, doi: 10.1038/296145a0.
of the Kaapvaal craton, as inferred from precise U-Pb zircon ages from Cameron, E.M., 1983, Evidence from early Proterozoic anhydrite for sulphur
isotopic partitioning in Precambrian oceans: Nature, v. 304, p. 54–56, doi: Farquhar, J., Bao, H., and Thiemens, M., 2000, Atmospheric influence of
10.1038/304054a0. Earth’s earliest sulfur cycle: Science, v. 289, p. 756–758, doi: 10.1126/
Canfield, D.E., and Raiswell, R., 1999, The evolution of the sulfur cycle: Amer- science.289.5480.756.
ican Journal of Science, v. 299, p. 697–723. Ford, D.C., 1986, Genesis of paleokarst and strata-bound zinc-lead sulfide
Canil, D., 1997, Vanadium partitioning and the oxidation sate of Archean kom- deposits in Proterozoic dolostone, northern Baffin Island, Canada—A dis-
atiite magmas: Nature, v. 389, p. 842–845, doi: 10.1038/39860. cussion: Economic Geology and the Bulletin of the Society of Economic
Carriere, J.J., and Sangster, D.F., 1992, Preliminary studies of fluid inclusions in Geologists, v. 81, p. 1562–11566.
sphalerite, quartz, and dolomite from Gayna River MVT deposit, North- Frimmel, H.D., Deane, J.G., and Chadwick, P.J., 1996, Pan-African tectonism
west Territories: Geological Survey of Canada Report 92-01A, p. 47–53. and the genesis of base metal sulfide deposits in the northern foreland
Cheney, E.S., 1996, Sequence stratigraphy and plate tectonic significant of of the Damara orogen, Namibia, in Sangster, D.F., ed., Carbonate-hosted
the Transvaal succession of southern Africa and its equivalent in West- lead-zinc deposits: Society of Economic Geologists Special Publication
ern Australia: Precambrian Research, v. 79, p. 3–24, doi: 10.1016/0301- 4, p. 204–217.
9268(95)00085-2. Garven, G., and Bull, S.W., 2000, Groundwater flow modeling of sedex-type
Chetty, D., and Frimmel, H.E., 2000, The role of evaporites in the genesis of ore genesis in the McArthur Basin, northern Australia: Geological Society
base metal sulphide mineralization in the Northern Platform of the Pan- of America Abstracts with Programs, v. 32, no. 7, p. 60–61.
African Damara Belt, Namibia: Geochemical and fluid inclusion evidence Ghazban, F., Schwarcz, H.P., and Ford, D.C., 1990, Carbon and sulfur isotope
from carbonate wall-rock alteration: Mineralium Deposita, v. 35, p. 364– evidence for in situ reduction of sulfate, Nanisivik lead-zinc deposits,
376, doi: 10.1007/s001260050247. Northwest Territories, Baffin Island, Canada: Economic Geology and the
Chi, G., and Savard, M.M., 1997, Sources of basinal and Mississippi Valley- Bulletin of the Society of Economic Geologists, v. 85, p. 360–375.
type mineralizing brines; mixing of evaporated seawater and halite-disso- Giordano, T.H., 1985, A preliminary evaluation of organic ligands and metal-
lution brine: Chemical Geology, v. 143, p. 121–125, doi: 10.1016/S0009- organic complexing in Mississippi Valley-type ore solutions: Economic
2541(97)00096-X. Geology and the Bulletin of the Society of Economic Geologists, v. 80,
Christensen, J.N., Halliday, A.N., Kesler, S.E., and Sangster, D.F., 1993, Fur- p. 96–106.
ther evaluation of the Rb-Sr dating of sphalerite; the Nanisivik Precam- Gize, A.P., 1986, The development of a thermal mesophase in organic matter
brian MVT deposit, Baffin Island, Canada: Geological Society of Amer- from high-temperature ore deposits, in Dean, W.E., ed., Organics and ore
ica Abstracts with Programs, v. 25, p. 471. deposits: Proceedings, Denver Regional Exploration Geologists Society
Christensen, J.C., Halliday, A.N., and Kesler, S.E., 1997, Rb-Sr dating of sphal- Symposium, Wheatridge, Colorado, p. 137–150.
erite and the ages of Mississippi Valley-type Pb-Zn deposits: Society of Greyling, L.N., Huizenga, J.M., and Gutzmer, J., 2001, A review of Missis-
Economic Geologists Special Publication 4, p. 527–535. sippi Valley-type deposits with emphasis on the Paleoproterozoic Per-
Christie-Blick, N., Dyson, I.A., and von der Borch, C.C., 1995, Sequence stra- ing Pb-Zn deposit, South Africa: Johannesburg, Economic Geology
tigraphy and the interpretation of Neoproterozoic history: Precambrian Research Institute, University of the Witwatersrand, Information Cir-
Research, v. 73, p. 3–26, doi: 10.1016/0301-9268(94)00096-A. cular no. 356, 32 p.
Condie, K.C., 2000, Episodic continental growth and supercontinents: After- Grotzinger, J.P., 1986a, Cyclicity and paleoenvironmental dynamics, Rock-
thoughts and extensions: Tectonophysics, v. 322, p. 153–162, doi: nest platform, northwest Canada: Geological Society of America Bul-
10.1016/S0040-1951(00)00061-5. letin, v. 97, p. 1208–1231, doi: 10.1130/0016-7606(1986)97<1208:
Core, D.P., 2004, Oxygen and sulfur fugacities of granitoids: Implications for CAPDRP>2.0.CO;2.
ore-forming processes [Ph.D. dissertation]: Ann Arbor, University of Grotzinger, J.P., 1986b, Evolution of early Proterozoic passive-margin carbon-
Michigan, 396 p. ate platform, Rocknest Formation, Wopmay Orogen, Northwest Territo-
Counter, K.J., 1993, Diagenesis of the Lower Ordovician Ogdensburg Forma- ries, Canada: Journal of Sedimentary Petrology, v. 56, p. 831–847.
tion, northern New York and southeastern Ontario; evidence for a hydro- Grotzinger, J.P., 1986c, Upward shallowing platform cycles: A response to 2.2
thermal “hot spot,” in Program with Abstracts, American Association of billion years of low-amplitude, high-frequency (Milankovitch band) sea-
Petroleum Geologists and Society of Economic Paleontologists and Min- level oscillations: Paleoceanography, v. 1, p. 403–416.
eralogists Annual Meeting, New Orleans, p. 88. Grotzinger, J.P., 2000, Facies and paleodevelopmental setting of thrombolite-
Delano, J.W., 2001, Redox history of the Earth’s interior since ~3900 Ma: stromatolite reefs, terminal Proterozoic Nama Group (ca. 550–543 Ma),
Implications for prebiotic molecules: Origins of Life and Evolution of the central and southern Namibia: Communications of the Geological Survey
Biosphere, v. 31, p. 311–341, doi: 10.1023/A:1011895600380. of South West Africa/Namibia, v. 12, p. 221–233.
Drew, G.J., and Both, R.A., 1984, The carbonate-hosted silver-lead deposits of Habicht, K.S., Gade, M., Thamdrup, B., Berg, P., and Canfield, D.E., 2002, Cal-
the Ediacara mineral field, South Australia; petrological, fluid inclusion ibration of sulfate levels in the Archean ocean: Science, v. 298, p. 2372–
and sulphur isotope studies: Australian Journal of Earth Sciences, v. 31, 2375, doi: 10.1126/science.1078265.
p. 177–201. Hanor, J.S., 1987, Origin and migration of subsurface sedimentary brines:
Duane, M.J., Kruger, F.J., Roberts, P.J., and Smith, C.B., 1991, Pb and Sr iso- Tulsa, Oklahoma, Society of Economic Paleontologists and Mineralogists
topes and origin of Proterozoic base metal (fluorite) and gold deposits, Short Course, v. 21, 247 p.
Transvaal Sequence, South Africa: Economic Geology and the Bulletin of Hardy, J.L., 1979, Stratigraphy, brecciation and mineralization, Gayna River,
the Society of Economic Geologists, v. 86, p. 1491–1505. N.W.T. [M.Sc. thesis}: Toronto, University of Toronto, 64 p.
Eriksson, P.G., Condie, K.C., Tirsgaard, H., Mueller, W.U., Altermann, W., Hattori, K., and Cameron, E.M., 1986, Archean magmatic sulphate: Nature,
Miall, A.D., Aspler, L.B., Catuneanu, O., and Shiarenzelli, J.R., 1998, v. 319, p. 45–47, doi: 10.1038/319045a0.
Precambrian clastic sedimentation systems: Sedimentary Geology, v. 120, Hayashi, K., Fujisawa, H., Holland, H.D., and Ohmoto, H., 1997, Geochem-
p. 5–53, doi: 10.1016/S0037-0738(98)00026-8. istry of ~1.9 Ga sedimentary rocks from northeastern Labrador, Canada:
Eriksson, P.G., and Altermann, W., 1998, An overview of the geology of the Geochimica et Cosmochimica Acta, v. 61, p. 4115–4437, doi: 10.1016/
Transvaal Supergroup dolomites (South Africa): Environmental Geology, S0016-7037(97)00214-7.
v. 36, p. 179–188, doi: 10.1007/s002540050334. Haynes, F.M., Beane, R.L., and Kesler, S.E., 1989, Simultaneous transport of
Eriksson, P.G., Altermann, W., Catuneanu, O., van der Merwe, R., and Bumby, metal and reduced sulfur, Mascot-Jefferson City zinc district, East Ten-
A.J., 2001a, Major influences on the evolution of the 2.67–2.1 Ga Trans- nessee: Evidence from fluid inclusions: American Journal of Science,
vaal basin, Kaapvaal craton: Sedimentary Geology, v. 141–142, p. 205– v. 289, p. 994–1038.
231, doi: 10.1016/S0037-0738(01)00075-6. Heaman, L.M., LeCheminant, A.M., and Rainbird, R.H., 1992, Nature and tim-
Eriksson, P.B., Martins-Neto, M.A., Nelson, D.R., Aspler, L.B., Chiarenzelli, ing of Franklin igneous events, Canada: Implications for a Late Protero-
J.R., Catuneanu, O., Sarkar, S., Altermann, W., and Rautenbach, C.J., de zoic mantle plume and the break-up of Laurentia: Earth and Planetary Sci-
W., 2001b, An introduction to Precambrian basins: Their characteristics ence Letters, v. 109, p. 117–131, doi: 10.1016/0012-821X(92)90078-A.
and genesis: Sedimentary Geology, v. 141–142, p. 1–35. Hewton, R.S., 1982, Gayna River: A Proterozoic Mississippi Valley-type zinc-
Farquhar, J., and Wing, B.A., 2003, Multiple sulfur isotopes and the evolution lead deposit, in Hutchinson, R.W., Spence, C.D., and Franklin, J.M., eds.,
of the atmosphere: Earth and Planetary Science Letters, v. 213, p. 1–13, Precambrian sulphide deposits: H.S. Robinson Memorial Volume: Geo-
doi: 10.1016/S0012-821X(03)00296-6. logical Association of Canada Special Paper 25, p. 667–700.
Hitzman, M.W., Thorman, C.H., Romagna, G., Oliviera, R.F., Dardenne, M.A., Korstgard, J., Ryan, B., and Wardle, R., 1987, The boundary between Protero-
and Drew, L.J., 1995, The Morro Agudo Zn-Pb deposits, Minas Gerais, zoic and Archean crustal blocks in central West Greenland and northern
Brazil; a Proterozoic Irish-type carbonate-hosted sedex-replacement Labrador, in Park, R.G., and Tarney, J., eds., Evolution of the Lewisian
deposit: Geological Society of America Abstracts with Programs, v. 27, and comparable Precambrian high-grade terrains: Geological Society
p. 408. [London] Special Publication 27, p. 247–259.
Hoffman, P.F., 1989, Pethei reef complex (1.9 Ga), Great Slave Lake, N.W.T., Knauth, L.P., 2005, Temperature and salinity history of the Precambrian ocean:
in Geldsetzer, H.H., James, N.P., and Tebbutt, G.E., eds., Reefs: Canada Implications for the course of microbial evolution: Paleogeography,
and adjacent areas: Canadian Society of Petroleum Geologists Memoir Paleoclimatology, Paleoecology, v. 219, p. 53–69.
13, p. 38–48. Kruger, F.J., Duane, M.J., Turner, A.M., Whitelaw, H.T., and Verhagen, B.T.,
Hoffman, P.F., and Bowring, S.A., 1984, Short-lived 1.9 Ga continental margin 2001, Paleohydrology of c. 2Ga old Mississippi Valley-type Pb-Zn
and its destruction, Wopmay orogen, northwest Canada: Geology, v. 12, deposit, South Africa: Radiogenic and stable isotope evidence: Johan-
p. 68–72, doi: 10.1130/0091-7613(1984)12<68:SGCMAI>2.0.CO;2. nesburg, Economic Geology Research Institute, University of the Witwa-
Holland, 1984, The chemical evolution of the atmosphere and oceans: Princ- tersrand, Information Circular no. 355, 22 p.
eton, New Jersey, Princeton University Press, 582 p. Kuznetzov, V.G., 1997, The role of karst in the formation of Yurubchen-Tok-
Holland, H.D., 1999, When did the Earth’s atmosphere become oxic? A reply: homsk field, Siberian Craton, Russia: American Association of Petroleum
Geochemical News, v. 100, p. 20–22. Geologists and Society of Economic Paleontologists and Mineralogists,
Holland, H.D., 2002, Volcanic gases, black smokers, and the Great Oxidation v. 6, p. 65.
Event: Geochimica Cosmochimica Acta, v. 66, p, 3811–3826. Kyle, J.R., and Misi, A., 1997, Origin of Zn-Pb-Ag sulfide mineralization
Horita, J., Zimmermann, H., and Holland, H.D., 2002, Chemical evolution of within upper Proterozoic phosphate-rich carbonate strata, Irece basin,
seawater during the Phanerozoic; implications from the record of marine Bahia, Brazil: International Geology Review, v. 39, p. 383–399.
evaporites: Geochimica et Cosmochimica Acta, v. 66, p. 3733–3756, doi: Leach, D.L., Bradley, D.C., Lewchuk, M., Symons, D.T.A., Brannon, J., and
10.1016/S0016-7037(01)00884-5. de Marsily, G., 2001, Mississippi Valley-type lead-zinc deposits through
Iyer, S.S., Hoefs, J., and Krouse, H.R., 1992, Sulfur and lead isotope geochem- geological time: Implications from recent age-dating research: Minera-
istry of galenas from the Bambui Group, Minas Gerais, Brazil; implica- lium Deposita, v. 36, p. 711–740, doi: 10.1007/s001260100208.
tions for ore genesis: Economic Geology and the Bulletin of the Society Leach, D.L., and Sangster, D.F., 1993, Mississippi Valley-type deposits, in
of Economic Geologists, v. 87, p. 437–443. Kirkham, R.V., Sinclair, W.D., Thorpe, R.I., and Duke, J.M., eds., Min-
Jackson, G.D., and Ianelli, T.R., 1981, Rift-related cyclic sedimentation in the eral deposit modeling: Geological Association of Canada Special Paper
Neohelikian Borden basin, northern Baffin Island: Canada Geological 40, p. 289–314.
Survey Paper 81-10, p. 269–302. Lemon, N.M., 2000, A Neoproterozoic fringing stromatolite reef complex,
Jones, D.A., 1986, The Kamarga deposit: Stratabound zinc-lead mineralisation Flinders Ranges, South Australia: Precambrian Research, v. 100, p. 109–
in the middle Proterozoic McNamara Group, northwest Queensland [Ph. 120, doi: 10.1016/S0301-9268(99)00071-6.
D. dissertation]: Armidale, Australia: University of New England, 147 p. Lowenstein, T.K., Hardie, L.A., Timofeeff, M.N., and Demicco, R.V., 2003,
Jones, D.J., Bull, S., and McGoldrick, P., 1999, The Kamarga deposit: A large, Secular variation in seawater chemistry and the origin of calcium chloride
low grade, stratabound zinc resource in the Proterozoic ‘Carpentaria Zinc basinal brines: Geology, v. 31, p. 857–860, doi: 10.1130/G19728R.1.
Belt’ of northern Australia in Stanley, C.J., ed., Mineral deposits: Pro- Martini, J.E.J., 1976, The fluorite deposits in the dolomite series of the Marico
cesses to processing: Rotterdam, A.A. Balkema, p. 873–876. district, Transvaal, South Africa: Economic Geology and the Bulletin of
Kah, L.C., Lyons, T.W., and Chesley, J.T., 2001, Geochemistry of a 1.2 Ga the Society of Economic Geologists, v. 71, p. 625–635.
carbonate-evaporite succession, northern Baffin and Bylot Islands; impli- Martini, J.E.J., 1990, The Genadendal Zn-Pb-Mn mineralization, South Africa:
cations for Mesoproterozoic marine evolution: Precambrian Research, A possible Early Proterozoic alkaline hydrothermal system: Economic
v. 111, p. 203–234, doi: 10.1016/S0301-9268(01)00161-9. Geology and the Bulletin of the Society of Economic Geologists, v. 85,
Kah, L.C., Lyons, T.W., and Frank, T.D., 2004, Low marine sulphate and pro- p. 1172–1185.
tracted oxygenation of the Proterozoic biosphere: Nature, v. 431, p. 834– Martini, J.E.J., Eriksson, P.G., and Snyman, C.P., 1995, The Early Proterozoic
838, doi: 10.1038/nature02974. Mississippi Valley-type Pb-Zn-F deposits of the Campbellrand and Mal-
Kasting, J.E., 1991, Box models for the evolution of atmospheric oxygen: An mani Subgroups, South Africa: Mineralium Deposita, v. 30, p. 135–145,
update: Global and Planetary Change, v. 5, p. 125–131, doi: 10.1016/0921- doi: 10.1007/BF00189342.
8181(91)90133-H. McFarlane, K., and Bone, Y., 1994, Mississippi Valley-type lead-zinc miner-
Kesler, S.E., 1996, Appalachian Mississippi Valley-type deposits: Paleoaqui- alization at Donkey Bore and Old Wirrealpa Springs, central Flinders
fers and brine provinces, in Sangster, D.F., ed., Carbonate-hosted lead- Range, South Australia: Geological Society of Australia Abstracts, v. 37,
zinc deposits: Society of Economic Geologists Special Publication 4, p. p. 278.
29–57. McLimans, R.K., Barnes, H.L., and Ohmoto, H., 1980, Sphalerite stratigraphy
Kesler, S.E., Appold, M.S., Walter, L.M., Martini, A.M., Huston, T.J., and of the Upper Mississippi Valley zinc-lead district, southwest Wisconsin:
Kyle, J.R., 1995, Na-Cl-Br systematics of Mississippi Valley-type brines: Economic Geology and the Bulletin of the Society of Economic Geolo-
Geology, v. 23, p. 641–644, doi: 10.1130/0091-7613(1995)023<0641: gists, v. 75, p. 351–361.
NCBSOM>2.3.CO;2. McNaughton, K., and Smith, T.E., 1986, A fluid inclusion study of sphalerite
Kesler, S., Chesley, J.T., Christensen, J.N., Hagni, R.D., Heijlen, W., Kyle, J.R., and dolomite from the Nanisivik lead-zinc deposit, Baffin Island, North-
Muchez, P., Misra, K.C., and van der Voo, R., 2004, Tectonic controls of west Territories, Canada: Economic Geology and the Bulletin of the Soci-
Mississippi Valley-type lead-zinc mineralization in organic forelands; dis- ety of Economic Geologists, v. 81, p. 713–720.
cussion: Mineralium Deposita, v. 39, p. 512–514, doi: 10.1007/s00126- Mengel, F., and Rivers, T., 1994, Metamorphism of pelitic rocks in the Paleo-
004-0422-3. proterozoic Ramah group, Saglek area, northern Labrador: Mineral reac-
Kesler, S.E., Gleason, J.D., Smith, C.N., Ahler, B.A., and Taylor, D.R., 2003, tions, P-T conditions and influence of bulk composition: Canadian Min-
Age and provinces of Precambrian MVT mineralization, Transvaal eralogist, v. 32, p. 781–801.
Supergroup, South Africa: Geological Society of America Abstracts with Mengel, R., Rivers, T., and Reynolds, P., 1991, Lithotectonic elements and
Programs, v. 35, p. 234. tectonic evolution of Torngat Orogen, Saglek Fiord, northern Labrador:
Kesler, S.E., Martini, A.M., Appold, M.S., Walter, L.M., Huston, T.J., and Canadian Journal of Earth Sciences, v. 28, p. 1407–1423.
Furman, F.C., 1996, Na-Cl-Br systematics of fluid inclusions from Mis- Meyer, C. 1988, Ore deposits as guides to the geologic history of the Earth:
sissippi Valley-type deposits, Appalachian basin: Constraints on solute Annual Review of Earth and Planetary Sciences, v. 16, p. 147–171, doi:
origin and migration paths: Geochimica et Cosmochimica Acta, v. 60, 10.1146/annurev.ea.16.050188.001051.
p. 225–233, doi: 10.1016/0016-7037(95)00390-8. Misi, A., Sundaram, S.S.I., Coelho, C.E.S., Tassinari, C.C.G., Franca-Rocha,
Knight, I., and Morgan, W.C., 1977, Stratigraphic subdivision of the Aphebian J.S., Cunha, I. de A., Gomes, A.S.R., de Oliveira, T.F., Teixeira, J.B.G.,
Ramah Group, northern Labrador: Geological Survey of Canada Paper and Filho, V.M.C., 2005, Sediment-hosted lead-zinc deposits of the Neo-
77-15, p. 1–31. proterozoic Bambuí Group and correlative sequences, São Francisco Cra-
Knight, I., and Morgan, W.C., 1981, The Aphebian Ramah Group, northern ton, Brazil: A review and a possible metallogenic evolution model: Ore
Labrador: Geological Survey of Canada Paper 81-10, p. 313–330. Geology Reviews, v. 26, p. 263–304.
Misiewicz, J.E., 1988, The geology and metallogeny of the Otavi Mountain Rowan, E.L., Goldhaber, M.B., and Hatch, J.R., 2001, The role of regional
Land, Damara Orogen, SWA/Namibia with particular reference to the fluid flow in the Illinois basin’s thermal history: Constraints from fluid
Berg Aukas Zn-Pb-V deposit—A model of ore genesis [M.Sc. thesis]: inclusions and maturity of Pennsylvanian coals: American Association of
Grahamstown, South Africa, Rhodes University, 143 p. Petroleum Geologists Bulletin, v. 86, p. 257–277.
Morgan, W.C., 1975, Geology of the Precambrian Ramah Group and basement Rye, D.M., and Williams, N., 1981, Studies of the base metal sulfide deposits at
rocks in the Nachvak Fiord-Saglek Fiord area, north Labrador: Geologi- McArthur River, Northern Territory, Australia III: The stable isotope geo-
cal Survey of Canada Paper 74-54, p. 1–42. chemistry of the H.Y.C., Ridge, and Cooley deposits: Economic Geology
Narbonne, G.M., and Aitken, J.D., 1995, Neoproterozoic of the Mackenzie and the Bulletin of the Society of Economic Geologists, v. 76, p. 1–26.
Mountains, northwestern Canada, in Knoll, A.H., and Malcolm, W., eds., St. Marie, J., and Kesler, S.E., 2000, Iron-rich and iron-poor Mississippi Valley-
Neoproterozoic stratigraphy and Earth history: Precambrian Research, v. type mineralization, Metaline district, Washington: Economic Geology
73, p. 101–121. and the Bulletin of the Society of Economic Geologists, v. 95, p. 1091–
Ohmoto, H., 1997, When did the Earth’s atmosphere become oxic?: Geochemi- 1106.
cal News, v. 93, p. 12–27. St. Marie, J., Kesler, S.E., and Allen, C.R., 2001, Origin of iron-rich Missis-
Ohmoto, H., 2004, The Archean atmosphere, hydrosphere and biosphere, in sippi Valley-type deposits: Geology, v. 29, p. 59–62, doi: 10.1130/0091-
Eriksson, P.G., Altermann, W., Nelson, D.R., Mueller, W.U., and Catune- 7613(2001)029<0059:OOIRMV>2.0.CO;2.
anu, O., eds., The Precambrian Earth: Tempos and events: Developments Sangster, D.F., 1990, Mississippi Valley-type and sedex lead-zinc deposits: A
in Precambrian Geology, v. 12, p. 361–388. comparative examination: Transactions of the Institution of Mining and
Ohmoto, H., and Goldhaber, M.B., 1997, Sulfur and carbon isotopes, in Barnes, Metallurgy, section B, v. 99, p. B21–B42.
H.L., ed., Geochemistry of hydrothermal ore deposits (3rd edition): New Schaefer, M., 2002, Paleoproterozoic Mississippi Valley-Type Pb-Zn Deposits
York, John Wiley and Sons, p. 517–612. of the Ghaap Group, Transvaal Supergroup in Griqualand West, South
Ohmoto, H., Kaiser, C.J., and Geer, K.A., 1990, Systematics of sulphur iso- Africa [Ph.D. dissertation]: Johannesburg, Rand Afrikaans University,
topes in recent marine sediments and ancient sediment-hosted base metal 367 p.
deposits, in Herbert, H.K. and Ho, S.E., eds., Stable isotopes and fluid Schaefer, M., Gutzmer, J., and Beukes, N.J., 2001, Genesis of carbonate-hosted
processes in mineralization: Geology Department and University Exten- Pb-Zn deposits in the Later Archean Transvaal Supergroup, Northern
sion, University of Western Australia, Publication 23, p. 70–120. Cape Province, South Africa, in Piestrzykinski, A., ed., Mineral deposits
Ohmoto, H., and Rye, R.O., 1979, Isotopes of sulfur and carbon, in Barnes, at the beginning of the 21st century: Proceedings of the Biennial SGA-
H.L., ed., Geochemistry of hydrothermal ore deposits (2nd edition): New SEG Meeting: Lisse, Swets and Zeitlinger Publishers, p. 177–180.
York, John Wiley and Sons, p. 509–567. Scott, S.D., 1997, Submarine hydrothermal systems and deposits, in Barnes,
Olson, R.A., 1984, Genesis of paleokarst and strata-bound zinc-lead sulfide H.L., ed., Geochemistry of hydrothermal ore deposits (3rd edition): New
deposits in a Proterozoic dolostone, northern Baffin Island, Canada: Eco- York, John Wiley and Sons, p. 797–876.
nomic Geology, v. 79, p. 056–1103. Scott, D.J., and Gauthier, G., 1996, Comparison of TIMS (U-Pb) and laser abla-
Osmond, J.C., 2000, West Willow Creek field; first productive lacustrine stro- tion microprobe ICP-MS (Pb) techniques for age determination of detrital
matolite mound in the Eocene Green River Formation, Uinta Basin, Utah: zircons from Paleoproterozoic metasedimentary rocks from northeastern
Mountain Geologist, v. 37, p. 157–170. Laurentia, Canada, with tectonic implications: Chemical Geology, v. 131,
Page, R.W., Jackson, M.W., and Krassay, A.A., 2000, Constraining sequence p. 127–142, doi: 10.1016/0009-2541(96)00030-7.
stratigraphy in North Australian basins: SHRIMP U-Pb zircon chronology Selley, D., Winefield, P., Bull, S.W., Scott, R.J., and McGoldrick, P.J., 2001,
between Mt. Isa and McArthur River: Australian Journal of Earth Sci- Sub-basins, depositional cycles and the tectono-sedimentary setting of the
ences, v. 47, p. 431–459, doi: 10.1046/j.1440-0952.2000.00797.x. HYC Zn-Pb-Ag deposit: Geological Society of America Abstracts with
Pehrsson, S.J., and Buchan, K.L., 1999, Borden dykes of Baffin Island, N.W.T.: Programs, v. 33, p. 270–271.
A Franklin U-Pb baddeleyite age and a paleomagnetic reinterpretation: Shen, Y., Canfield, D.E., and Knoll, A.H., 2002, Middle Proterozoic ocean
Canadian Journal of Earth Sciences, v. 36, p. 65–73, doi: 10.1139/cjes- chemistry: Evidence from the McArthur Basin, northern Australia: Amer-
36-1-65. ican Journal of Science, v. 302, p. 81–109.
Petrov, P.Y., and Semikhatov, M.A., 2001, Sequence organization and growth Sherlock, R.L., Lee, J.K.W., and Cousens, B.L., 2004, Geologic and geochro-
patterns of late Mesoproterozoic stromatolites reefs: An example from the nologic constraints on the timing of mineralization at the Nanisivik zinc-
Burovaya Formation, Turukhansk Uplift, Siberia: Precambrian Research, lead Mississippi Valley-type deposit, northern Baffin Island, Nunavut,
v. 111, p. 257–281, doi: 10.1016/S0301-9268(01)00163-2. Canada: Economic Geology and the Bulletin of the Society of Economic
Plumb, K.A., Ahmad, M., and Wygralak, A.S., 1990, Mid-Proterozoic basins Geologists, v. 99, p. 279–293.
of north Australian craton—Regional geology and mineralisation, in Sicree, A.A., and Barnes, H.L., 1996, Upper Mississippi Valley district ore-
Hughes, F.E., ed., Geology of the mineral deposits of Australia and Papua fluid model: The role of organic complexes: Ore Geology Reviews, v. 11,
New Guinea: Melbourne, Australasian Institute of Mining and Metal- p. 105–131, doi: 10.1016/0169-1368(95)00018-6.
lurgy, p. 881–902. Smith, S.G., 1996, Geology and geochemistry of the Warrabarty carbonate-
Poetter, P.A., 2001, Origin and exploration potential of sedimentary-hosted hosted Zn-Pb prospect, Paterson orogen, Western Australia [Ph.D. dis-
Mississippi Valley-Type F-Zn-Pb deposits in the Neoarchean platform sertation]: Hobart, University of Tasmania, 162 p.
dolostones of the Transvaal Supergroup, Northwestern Province, South Sourirajan, S., and Kennedy, G.S., 1962, The system H2O-NaCl at elevated
Africa [abs.]: Muenster, WWU Muenster, Diplom thesis, (http://general. temperatures and pressures: American Journal of Science, v. 260, p. 115–
rau.ac.za/geology/PPM/Abstracts.html). 141.
Pollack, H.N., 1997, Thermal characteristics of the Archean, in de Wit, M.H., Strauss, H., 1993, The sulfur isotope record of Precambrian sulfates: New data
and Ashwal, L.D., eds., Greenstone belts: Oxford, UK, Clarendon Press, and a critical evaluation of the existing record: Precambrian Research,
p. 223–232. v. 63, p. 225–246, doi: 10.1016/0301-9268(93)90035-Z.
Pollack, H.N., and Chapman, D.S., 1977, On the regional variation of heat flow, Sutherland, R.A., and Dumka, D., 1995, Geology of the Nanisivik mine,
geotherms, and lithospheric thickness: Tectonophysics, v. 38, p. 279–296, N.W.T., Canada, in Misra, K., ed., Carbonate-hosted lead-zinc-fluo-
doi: 10.1016/0040-1951(77)90215-3. rite-barite deposits on North America: Society of Economic Geologists
Rhondacorp 2002, www.rhondacorp.com, 30 June 2002. Guidebook Series, v. 22, p.4–18.
Ricketts, B.D., and Donaldson, J.A., 1989, Stromatolite reef development on Sverjensky, D.A., 1981, The origin of a Mississippi Valley-type deposit in the
a mud-dominated platform in the Middle Precambrian Belcher Group of Viburnum Trend, southeast Missouri: Economic Geology and the Bulletin
the Hudson Bay, in Geldsetzer, H.H., James, N.P., and Tebbutt, G.E., eds., of the Society of Economic Geologists, v. 76, p. 1848–1872.
Reefs: Canada and adjacent areas: Canadian Society of Petroleum Geolo- Symons, D.T.A., Symons, T.B., and Sangster, D.F., 2000, Paleomagnetism of
gists Memoir 13, p. 113–119. the Society Cliffs dolostone and the age of the Nanisivik zinc deposits,
Roberts, P.J., Gize, A.P., Duane, M.J., and Verhagen, B., 1993, Precambrian Baffin Island, Canada: Mineralium Deposita, v. 35, p. 672–682, doi:
hydrocarbon residues associated with Mississippi Valley-type mineraliza- 10.1007/s001260050270.
tion in the Transvaal Sequence, South Africa: South African Journal of Turner, E.C., Narbonne, G.M., and James, N.P., 2000, Framework composition
Geology, v. 96, p. 57–60. of early Neoproterozoic calcimicrobial reefs and associated microbialites,
Mackenzie Mountains, N.W.T., Canada, in Grotzinger, J.P., and James, Wheatley, C.J.V., Whitfield, G.G., Kenny, K.J., and Birch, A., 1986b, The
N.P., eds., Carbonate sedimentation and diagenesis in the evolving Pre- Pering carbonate-hosted zinc-lead deposit, Griqualand West, in Anhae-
cambrian world: Society for Sedimentary Geology Special Publication usser, C.R., and Maske, S., eds., Mineral deposits of southern Africa, p.
67, p. 179–205. 867–874.
Ueda, A., Cameron, E.M., and Krouse, H.R., 1991, 34S-enriched sulphate in Whelan, J.F., Rye, R.O., and deLoraine, W.F., 1984, The Balmat-Edwards zinc-
the Belcher Group, N.W.T., Canada: Evidence for dissimilatory sulphate lead deposits; synsedimentary ore from Mississippi Valley-type fluids:
reduction in the early Proterozoic ocean: Precambrian Research, v. 49, Economic Geology and the Bulletin of the Society of Economic Geolo-
p. 229–233, doi: 10.1016/0301-9268(91)90034-8. gists, v. 79, p. 239–265.
Vearncombe, J.R., Chisnall, A.W., Dentith, M.C., Dorling, S.L., Rayner, M.J., Whelan, J.F., Rye, R.O., deLoraine, W.F., and Ohmoto, H., 1990, Isotopic geo-
and Holyland, P.W., 1996, Structural controls on Mississippi Valley-type chemistry of the mid-Proterozoic evaporite basin; Balmat, New York:
mineralization, the southeast Lennard shelf, Western Australia: Society of American Journal of Science, v. 290, p. 396–424.
Economic Geologists Special Publication 4, p. 74–95. Whitesell, T.C., 1995, Diagenetic features associated with sequence boundar-
Viets, J.G., Hofstra, A.H., and Emsbo, P., 1996, Solute compositions of fluid ies in upper Miocene carbonate strata, Las Negras, Spain [M.Sc. thesis]:
inclusions in sphalerite from North American and European Mississippi Lawrence, University of Kansas, 292 p.
Valley-type ore deposits: Ore fluids derived from evaporated seawater, in Williams, N., 1978, Studies of the base metal sulfide deposits at McArthur
Sangster, D.F., ed., Carbonate-hosted lead-zinc deposits: Society of Eco- River, Northern Territory, Australia I: The Cooley and Ridge deposits:
nomic Geologists Special Publication 4, p. 465–482. Economic Geology and the Bulletin of the Society of Economic Geolo-
Wachowiak, N.M., 2001, Sediment-hosted Pb-Zn (Esker) /Cu mineralization, gists, v. 73, p. 1005–1035.
~1.89 Ga Rocknest Platform, Northwest Territories/Nunavut, Canada; Wilton, D.H.C., Archibald, S.A., Churchill, R.A., Phillips, R.D., and Saunders,
characteristics, paragenetic analysis and controls [M.Sc. thesis]: Toronto, J.K., 1993, Metallogenic potential of the lower Proterozoic supracrustal
University of Toronto, 176 p. sequence developed on the Archean Nain Craton, northern Labrador, in
Wachowiak, N.M., Gummer, P.K., Grotzinger, J.P., and Spooner, E.T.C., 1998, Davies, J.F., Gibson, H.L., and Whitehead, R.S., eds., Third annual field
Approximately 1.89 Ga MVT/SSC mineralization, Esker deposit, Rock- conference of the Geological Society of CIM, Bathurst, NB, Canada,
nest carbonate platform, Wopmay orogen, N.W.T., Canada: Geological p. 23–41.
Society of America Abstracts with Programs, v. 30, p. 369. Wilton, D.H.C., Archibald, S.M., Hussey, A.M., and Butler, R.W., Jr., 1994,
Wachowiak, N.M., Gummer, P.K., Pope, M., and Grotzinger, J.P., 1997, Strat- Metallogeny of the Ramah Group: Discovery of a new Pb-Zn explora-
abound Zn-Pb mineralization to the Esker horizon, Paleoproterozoic tion target, northern Labrador: Current Research (1994) Newfoundland
Rocknest Formation, Wopmay orogen, Canada, in Program and Abstracts, Department of Mines and Energy, Geological Survey Branch, Report 94-
NWT Geoscience Forum 25th Anniversary, Yellowknife: Yellowknife, 1, p. 415–428.
Canada, Indian and Northern Affairs Canada, p. 96–98. Winefield, P.R., 2000, Development of late Paleoproterozoic aragonitic sea-
Walker, L.J., Wilkinson, B.H., and Ivany, L.C., 2002, Continental drift and Pha- floor cements in the McArthur Group, northern Australia, in Grotzinger,
nerozoic carbonate accumulation in shallow-shelf and deep-marine set- J.P., and Noel, J.P., eds., Carbonate sedimentation and diagenesis in the
tings: Journal of Geology, v. 110, p. 75–87, doi: 10.1086/324318. evolving Precambrian world: Society for Sedimentary Geology Special
Walker, R.N., Gulson, B., and Smith, J., 1983, The Coxco deposit—A Pro- Publication 67, p. 145–159.
terozoic Mississippi Valley-type deposit in the McArthur River District, Wu, Y., Hagni, R.D., and Paarlberg, N.L., 1997, Silver distribution in iron sul-
Northern Territory, Australia: Economic Geology and the Bulletin of the fides at the Buick and Brushy Creek mines, Viburnum Trend, Southeast
Society of Economic Geologists, v. 78, p. 214–249. Missouri, in Sangster, D.F., ed., Carbonate-hosted lead-zinc deposits:
Webb, G.E., 2001, Biologically induced carbonate precipitation in reefs through Society of Economic Geologists Special Publication 14, p. 577–586.
time, in Stanley, G.D., ed., The history and sedimentology of ancient reef
systems: Topics in Geobiology, v. 17, p. 25–37.
Wheatley, C.J., Friggens, P.J., and Dooge, F., 1986a, The Bushy Park carbon-
ate-hosted zinc-lead deposit, Griqualand West, in Anhaeusser, C.R., and
Maske, S., eds., Mineral deposits of southern Africa, p. 891–900. MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA


Memoir 198
2006
Superheavy S isotopes from glacier-associated sediments of the

Neoproterozoic of south China: Oceanic anoxia or sulfate limitation?
Liu Tie-bing*
Institute of Geology, Chinese Academy of Sciences, P. O. Box 634, Beijing 100029, China
J. Barry Maynard*
Department of Geology, University of Cincinnati, P. O. Box 210013, Cincinnati Ohio 45221, USA
John Alten
Department of Geology, Miami University, 114 Shideler Hall, Oxford, Ohio 45056, USA
ABSTRACT
Black shales and Mn carbonates interbedded with glacial deposits from the Neo-
proterozoic of southern China exhibit extremely heavy values of pyrite S isotopes
that may reflect the peculiar environment of Earth at this time. δ34S averages +30‰
at Tanganshan and +44‰ at Xiangtan, compared with typical values of 0‰ to +5‰
found in younger deposits. Furthermore there is no distinction between the shales
and the Mn carbonate ores in the Neoproterozoic, unlike the younger deposits, which
show much lighter δ34S in the shales than in the Mn ores (the spread is 25‰). Most
other chemical parameters are very similar to both the younger Mn deposits and
those from the Paleoproterozoic. The exception is the rare earth elements (REE).
All Neoproterozoic Fe ores and most Neoproterozoic Mn ores lack the positive Eu
anomaly that characterizes Archean and Paleoproterozoic Fe-Mn accumulations.
On the other hand, Neoproterozoic Mn deposits have positive Ce anomalies on North
American Shale Composite (NASC) normalized plots, in contrast to other MnCO3
ores. The ΣREE is also higher than in other Mn deposits, but lower than in modern
deep-sea crusts.
Sulfide S values in all Neoproterozoic shales tend to be exceptionally variable
and to often show much heavier values than can be found in marine strata from
the Phanerozoic. Therefore the anomalous δ34S values we observed reflect peculiar
conditions in the world oceans at this time rather than purely local effects. Times of
enrichment of seawater sulfate in 34S do not correspond to periods of glaciation, so
the likely cause of the S isotope patterns is not worldwide glaciation, but a generally
low level of dissolved sulfate S in the Neoproterozoic oceans that allowed modest
increases in the amounts of S removed as pyrite to drive down the oceanic S reservoir
enough to produce strong Rayleigh reservoir effects. The abundance worldwide of
Sturtian-age Mn and Fe deposits indicates an increase in Fe flux to the oceans that
would have been sufficient to depress SO42- levels severely and to result in residu-
al dissolved S extremely enriched in 34S. REE evidence indicates that most of this
enhanced Fe and Mn flux came from diagenetic remobilization of detrital oxides
*E-mail: [email protected]; [email protected]
Liu, T.-B., Maynard, J.B., and Alten, J., 2006, Superheavy S isotopes from glacier-associated sediments of the Neoproterozoic of south China: Oceanic anoxia or
sulfate limitation? in Kesler, S.E., and Ohmoto, H., eds., Evolution of Early Earth’s Atmosphere, Hydrosphere, and Biosphere—Constraints from Ore Deposits:
Geological Society of America Memoir 198, p. 205–222, doi: 10.1130/2006.1198(12). For permission to copy, contact [email protected]. ©2006 Geological
Society of America. All rights reserved.
205
206 Liu Tie-bing, J.B. Maynard, and J. Alten
rather than from ridge-crest hydrothermal systems, in contrast to the Paleoprotero-

zoic banded iron formations. Rapid introduction of lateritic soil residues to restricted
basins by low-latitude glaciation could have provided the needed excess Fe and Mn
to drive this system.
Keywords: Neoproterozoic, glaciation, REE, S isotopes.
INTRODUCTION deposits (Maynard, 1991). In a study of the relationship of Mn

ore genesis to black shale sedimentation, Liu (1988, 1990) com-
The Neoproterozoic was a time of profound change in pared the whole-rock carbon-sulfur geochemistry of one of these
Earth’s oceans and biosphere. One peculiar manifestation of Neoproterozoic deposits, Xiangtan, with younger deposits and
these changes is the reappearance of iron formations in abun- found the patterns to be similar. Subsequent work on the sulfur
dance in the sedimentary record, for example Rapitan in Can- isotopic composition of the Chinese Neoproterozoic deposits,
ada (Yeo, 1981), the iron-rich beds of the Chuos Formation in however, has shown them to contain extraordinarily heavy S
Namibia (Breitkopf, 1988), and the Braemar Ironstone of South (Tang and Liu, 1999; Li et al., 1999).
Australia (Lottermoser and Ashley, 2000). Accompanying these Is this super-heavy sulfur peculiar to the Neoproterozoic
iron formations are some of the world’s most important Mn ore of south China? Is it somehow related to the Mn-Fe mineraliza-
deposits, e.g., Urucum in Brazil (Urban et al. 1992), and in all tion? Is it somehow related to glaciation? To address these ques-
cases the ore bodies are found in close association with glacial tions we have undertaken detailed geochemical analysis of two
110° E
ICE-COVER,
EROSION
RY
N DA
U
BO Changsha
Minle
Minle (Taojiang)
(Taojiang)
T
IF
Tanganshan
Tanganshan
Xiangtan
Xiangtan
R
N Figure 1. Location of the Xiangtan, Tanganshan,

OSITIO
EP and Minle Neoproterozoic Mn deposits in south
D China and the Ordovician Taojiang deposit.
E
N
IAL-M ARI
LAC
G
Shanghai
30°
Changsha
HUNAN
25° N
26°
N
22° N Hong Kong
50 100 km
SOUTH CHINA
SEA
Redrawn from Wang and Li, 2003
110° E 114° 118°
Superheavy S isotopes from glacier-associated sediments 207
of these deposits, Xiangtan and Tanganshan. In this paper we stratigraphic marker in south China. Thus there are two diamic-
show how isotopic, trace element, and rare earth element (REE) tite intervals separated by a black shale with Mn carbonate beds.
data constrain models for the genesis of Mn-Fe deposits of this Ages are sparse, but a tuff bed in the Liantuo Formation has
time period and help understand the Neoproterozoic atmosphere- yielded one date of 748 ± 12 Ma from single-zircon U-Pb dating,
ocean system. and the Datangpo/Minle Formation has been dated by Rb-Sr at
728 ± 27 Ma (Li et al., 2000). These ages and sequence strati-
Geological Setting graphic interpretations led Wang and Li (2003, p. 155) to sug-
gest that the Nantuo diamictites are correlative with the Sturtian
There are 11 large to medium-sized commercial Mn opera- glacial deposits of Australia. If so, the Datangpo Mn mineraliza-
tions in Neoproterozoic host rocks of south China distributed tion is comparable to the Braemar ironstone, which is interbed-
throughout the Hubei, Guizhou, Sichuan, and Hunan Provinces ded with the glacial deposits (Gorjan et al., 2000). The overlying
(Xu et al., 1990; Fan et al., 1992; Li et al., 1999). We studied Doushantuo carbonates, however, have been dated to 584 ± 26
Xiangtan and Tanganshan in Hunan province (Fig. 1). The ores Ma by Lu-Hf and 599.3 ± 4.2 Ma by Pb-Pb methods (Barfod et
are contained within black shales that are overlain and underlain al., 2002), which suggests that the Nantuo glacial deposits cor-
by glacial deposits (Fig. 2). The regional stratigraphic section respond to the later Marinoan glaciation episode. Dobrzinski et
comprises, from the base, the Liantuo Formation sandstones, the al. (2004), however, argued that both the Chunmu and the Nan-
Gucheng or Chunmu sandstones and diamictites, the Datangpo tuo diamictites are Sturtian, on the basis of the 728 Ma date for
Formation (Minle Formation of some authors), comprising the intervening Datangpo Formation. Paleomagnetic data are
shales and Mn carbonates, and, at the top, the Nantuo Forma- also ambiguous. Rui and Piper (1997) assigned the Datangpo
tion diamictites. The Nantuo is overlain abruptly by carbonates Formation to the Sturtian, whereas Macouin et al. (2004) favored
of the Doushantuo Formation and this contact is an important a Marinoan age. Both Shen (2002) and Macouin et al. (2004)
argued for a Marinoan age based on carbon isotope excursions,
but either major glacial episode could produce this effect. Thus
there are two conflicting age interpretations. If the Marinoan age
is correct, then perhaps the Chunmu tillite is a Sturtian equivalent
and the Datangpo Formation represents a relatively long time
period. Alternatively, the Chunmu and Nantuo tillites could both
be phases of Sturtian glaciation and the Datangpo was rapidly
deposited during a brief interglacial episode.
The reconstruction of Xu et al. (1990) shows the Datangpo
Formation lying within a fault-bounded marine trough between
the Yangtze Block to the northwest and the Hunan Block to the
southeast. Continental diamictites occupied the margin of the
Yangtze Block during Datangpo deposition, whereas glacio-
marine strata are found on the Hunan Block. Thus deposition was
in a narrow fjord-like rift valley. Urban et al. (1992) proposed
a similar ice-covered long, narrow basin as the setting for the
Urucum deposits of Brazil. Eyles and Januszczak (2004a, 2004b)
have likewise emphasized the rift setting of Neoproterozoic gla-
cial deposits as key to their genesis.
ANALYTICAL METHODS
Samples were collected from mine passageways and were

ground in a tungsten carbide ring mill. They were analyzed for
major elements by X-ray fluorescence (XRF) and for trace ele-
ments by inductively coupled plasma–mass spectrometry (ICP-
MS) supplemented by hydride-atomic absorption (AA) for As
and Se and by LECO elemental analyzer for C and S. XRF anal-
yses were done at the University of Cincinnati using a Rigaku
3070 spectrometer. Loss-on-ignition residues of the samples
were fused with lithium metaborate. The resulting glass beads
Figure 2. Stratigraphic setting of the Mn mineralization in the Xiang- were then ground and pressed into disks with polyvinyl alcohol
tan deposit (after Liu, 1988). binder. Calibration was done by empirical multiple regression
against a set of U.S. Geological Survey and National Institute same procedures for a set of NIST standards of Mn and Fe-rich
of Standards and Technology (NIST) high-Mn standards treated rocks and a representative suite of ore samples from Paleoprotero-
in the same way. Reproducibility is ±2%. ICP-MS and AA were zoic and Phanerozoic Mn deposits. Table 2 gives trace element
performed by XRAL Laboratories of Toronto, Canada. Precision results and Table 3 gives REE results from the same samples and
is about ±5% for most elements with detection limits generally standards. Many of the NIST standards do not have trace and
of 0.1–1 ppm for trace elements and 0.05–0.1 for REE. LECO rare earth elements included in their documentation, so we pres-
analyses were performed on whole-rock samples and on acid- ent values for these standards for the first time. Mamatwan is the
insoluble residues at Indiana University. Precision was ±1% for largest Paleoproterozoic deposit and comprises a significant frac-
C and ±5% for S. tion of the world’s Mn reserves. Molango is the largest Mesozoic
Sulfur was extracted from the samples by a method origi- deposit, and Taojiang, although much smaller, is representative
nally developed by Canfield et al. (1986) and modified by the of Paleozoic deposits. For each deposit, we have selected repre-
biogeochemistry laboratory at Indiana University (Bruchert, sentative ore samples, either Mn-rich or Fe-rich, and some host
1998). First, sample powders are extracted for 24 h with meth- rocks. Extensive trace element and REE data for these deposits
ylene chloride in a soxhlet apparatus. This procedure removes have not previously been published.
native sulfur and bitumen-bound sulfur. Extracting the bitumen Inspection of the tabulated results for the ores and shales in
also removes organic coatings from sulfide mineral grains and the Chinese deposits shows a continuum of compositions between
makes them easier to dissolve in subsequent steps. Any extracted high and low Al end members. The data set shows a strong posi-
native sulfur is reprecipitated as copper sulfide on copper shot in tive correlation between % Mn and % Ccarbonate (r2 = 0.74) and a
a receiving flask. The S-coated copper shot is then transferred to negative correlation between % Mn and both % Corg (r2 = −0.77)
a closed reaction system where it is treated with 6N HCl under and %Al2O3 (r2 = −0.71). No other elements show a positive cor-
a nitrogen stream and heated to 90 °C. This process converts relation to % Mn and there are no correlations between any of the
the copper sulfide to H2S, which is bubbled into a AgNO3 trap whole rock parameters and δ34S.
where the S is precipitated as Ag2S. Next 0.5–2.0 g of the MeCl- Thus the highest-grade ore samples from the Chinese
extracted samples are transferred to the same reaction system. deposits are low in Al and Corg, but high in Ccarb. We have used
Heating the sample to 90 °C in 6N HCl releases S contained in 3% Al2O3 as the upper cutoff to separate ore samples from sub-
monosulfides (FeS, galena, sphalerite) and also in pyrrhotite. ore grade, but mineralized, samples. In Figure 3 we compare
This sulfide converts to H2S in the acid environment, which is the composition of this end-member ore type to an average
again bubbled into a AgNO3 trap. This S is customarily desig- MnCO3 ore as calculated from the representative ores given
nated as the acid-volatile fraction (AVS). The sulfur in pyrite and in Tables 1 and 2, giving equal weight to ores of each of the
marcasite, FeS2, occurs in the S1- form and must be reduced to three ages. Two groups of elements show enrichment in the
the S2- form to be converted to H2S gas. This reduction is accom- Chinese ores: one group comprising P and Mo, which are prob-
plished using chromous chloride in 6N HCl, again at 90 °C, fol- ably related to the high organic carbon content of the ores, and
lowed by precipitation of Ag2S from the H2S gas. We designate another comprising Ti, Y, Zr, and Nb, which are elements typi-
this S as the chrome-reducible fraction (CrRS). The Xiangtan cally associated with heavy minerals. Al is identical in the two
and Tanganshan samples yielded only pyrite S, except for two data sets, so these excess resistate elements are not associated
samples from the ore dump at Tanganshan that had large amounts with clays. Pb is also relatively high in the Chinese ores and Ca
of native S. We interpret this native S as a weathering product of and Ta are depleted.
pyrite and have not investigated surface samples further because Because the Chinese ores are hosted by black shales, we have
of the large impact of weathering, even with relatively brief peri- used the geostandard SDO-1 as a comparison. The Chinese black
ods of exposure. Sulfur isotopic ratios were determined at Indi- shales are very similar to this standard (Fig. 4). Compared with
ana University using the Finnigan MAT 252 mass spectrometer SDO-1, Mn is somewhat enriched in the Chinese host shales, so
described by Studley et al. (2002). Mn mineralization is distributed at low grade in the host rocks, a
pattern that is also seen at Molango. Mo and U are low compared
RESULTS to SDO-1, which is probably related to much higher organic car-
bon in the standard (10% compared to ~3%–4% for the Chinese
Whole Rock Chemistry shales). Thus, just as for the ores, the host rocks appear to have
normal chemistries except for their S isotopic values.
The Chinese Mn ores have extreme S isotope values, and For the REEs, some differences can be seen between the
might be expected to also have significant anomalies in other Chinese material and other ores (Fig. 5). Almost all of the older
chemical properties. However, whole rock data for the Mn ores and younger Mn deposits in Table 3 have negative Ce anoma-
and for the host black shales are similar to values reported from lies. Both the Chinese ore deposits by contrast display posi-
other Mn deposits and for representative unmineralized black tive anomalies. Yang et al. (1999) also reported a positive Ce
shales. In Table 1 we give major element results for the Xiangtan anomaly for one sample from the correlative Songtao deposit.
and Tanganshan ores and shales plus analytical data using the For Eu, the samples from other deposits all have positive
TABLE 1. MAJOR ELEMENTS, C AND S FOR Mn AND Fe ORES AND RELATED STANDARD REFERENCE MATERIALS
Sample name Fe Mn Ti Si Al Ca K P Mg Na LOI total Ccarb Corg S G34S pyrite
Analysis unit % % % % % % % % % % % % % % % ‰
oxide oxide oxide oxide oxide oxide oxide oxide oxide oxide oxide oxide
Standards
NIST-25d Mn ore 4.16 66.9 0.13 2.52 5.52 0.11 1.40 0.30 0.43 0.03 14.7 96.1 0.36
NIST-26 Fe ore 79.5 0.14 0.07 6.10 0.96 2.50 0.02 0.09 3.33 0.00 4.46 97.2 0.03
NIST-27b Fe ore 93.9 0.17 0.02 1.31 0.62 0.08 0.05 0.09 0.05 0.01 0.39 96.7 0.02
NIST-28a Fe ore 93.8 0.53 0.02 1.50 0.76 0.06 0.04 0.07 0.05 0.00 3.00 99.8 0.02
NIST-69b Bauxite 7.49 0.11 1.90 13.4 48.8 0.15 0.09 0.14 0.10 0.02 27.2 99.4
NIST-88b Dolomite 0.27 0.02 0.02 1.13 0.36 30.1 0.16 0.02 21.0 0.03 47.0 100.1
NIST-97b Flint clay 1.13 0.01 2.43 41.2 40.0 0.04 0.91 0.07 0.22 0.09 13.9 100.0
NIST-690 Fe ore 93.2 0.23 0.02 3.71 0.18 0.20 0.00 0.05 0.20 0.00 0.06 97.9 0.01
NIST-1413 Al sand 0.39 0.03 0.10 81.7 9.55 0.77 3.77 0.07 0.07 1.63 1.55 99.6 0.05
VL-1 Laterite 35.8 0.07 3.15 1.16 37.4 0.03 0.09 0.16 0.03 0.01 22.4 100.3 0.29
Tanganshan
TG-3 Mn ore 7.74 27.9 0.56 30.6 6.99 1.90 2.12 0.21 1.93 0.06 19.9 99.9 3.60 3.88 4.45 31.41
TG-4 Mn ore 2.91 59.5 0.05 0.50 0.24 4.70 0.02 0.30 2.51 0.09 31.6 102.4 10.16 2.45 1.32 23.78
TG-5 Mn ore 4.59 56.4 0.30 6.03 2.54 2.90 0.04 0.26 1.12 0.07 26.1 100.4 7.19 3.29 2.60 29.59
TG-6 Shale 6.34 12.5 0.74 50.9 9.68 5.24 0.55 0.24 3.63 0.10 10.9 100.8 1.25 4.02 1.76
TG-7 Shale 4.02 1.57 0.81 56.2 15.9 0.85 9.18 0.16 1.90 0.13 8.97 99.7 0.00 4.26 3.53 29.70
TG-8 Shale 7.22 6.99 0.68 50.0 11.9 2.03 4.98 0.34 3.64 0.20 12.6 100.5 0.60 4.34 5.24 31.07
TG-9 Shale 9.46 20.6 0.49 33.4 6.82 8.40 0.21 0.61 1.46 0.34 17.7 99.6 3.85 3.48 7.08 30.48
Xiangtan
XT-0.3 Shale 5.24 13.2 0.65 42.0 11.6 1.38 3.72 0.37 1.43 0.21 20.2 100.1 1.82 0.94 4.56 48.44
XT-0.4 Mn ore 2.79 26.9 0.47 39.1 8.47 1.26 2.76 0.26 1.28 0.14 17.1 100.5 4.45 0.91 2.46 41.46
XT-0.6 Mn ore 2.18 60.1 0.10 3.13 0.85 1.68 0.52 0.21 1.92 0.04 30.5 101.3 9.43 0.69 0.77 57.58
XT-1.3 Mn ore 2.2 49.0 0.15 17.0 0.80 2.06 0.31 0.12 2.18 0.04 26.4 100.2 8.07 0.55 0.80 47.36
XT-2.4 Mn ore 4.67 39.8 0.54 15.0 5.01 2.05 1.34 0.48 1.43 0.21 29.0 99.5 6.49 2.07 3.14 36.82
XT-3.7 Mn ore 4.25 34.5 0.21 22.8 5.26 2.83 0.65 0.37 1.97 0.21 26.9 99.9 6.41 1.58 2.71 42.86
XT-4.2 Shale 3.35 0.15 0.71 60.0 14.2 0.10 4.23 0.08 1.37 0.04 16.9 101.2 0.00 5.08 3.42
XT-4.3 Dolomite 0.31 0.29 0.03 20.1 1.33 24.6 0.39 0.11 15.7 0.00 37.2 100.0 10.12 4.27 0.96 15.36
XT-4.5 Mn ore 3.25 25.7 0.47 31.9 8.25 2.54 2.21 0.21 1.61 0.10 23.7 100.0 4.93 2.76 2.40 43.65
XT-6.1 Shale 8.34 16.3 0.99 34.2 9.68 1.33 3.44 0.20 1.50 0.00 23.9 100.0 2.36 2.99 9.09 39.26
XT-7.2 Shale 2.43 0.18 0.72 64.1 15.3 1.47 4.51 0.06 1.65 0.05 11.0 101.5 0.00 4.14 3.34 60.93
XT-8.1 Shale 2.87 0.3 0.72 62.9 14.2 1.45 4.41 0.07 1.56 0.04 13.2 101.7 0.00 3.47 2.94 53.05
Other Mn and Fe Ores
SO-5100 Mamatwan Fe formation 37.1 15.7 0.15 18.5 2.58 5.6 0.86 0.04 3.36 0.10 15.0 98.8 0.06
SO-5101 Mamatwan Mn ore 6.55 42.8 0.01 5.78 0.18 18.9 0.01 0.02 2.96 0.03 20.2 97.4 0.03
SO-5108 Hotazel Fe formation 35.0 0.16 0.01 67.1 0.17 0.09 0.01 0.04 0.07 0.02 0.86 103.5 0.01
SO-5206 Molango shale 4.64 2.56 0.29 33.3 6.84 17.2 1.32 0.10 6.70 0.18 27.5 100.5 2.54
SO-5207 Molango Mn ore 13.5 35.9 0.09 10.8 2.59 2.64 0.01 0.13 8.23 0.01 26.6 100.4 0.27
SO-5209 Molango Mn ore 11.0 36.3 0.07 12.0 2.79 2.19 0.00 0.14 8.79 0.00 27.4 100.7 0.17
SO-5301 Taojiang Mn ore 1.02 30.0 0.01 11.2 0.51 23.0 0.22 0.05 1.26 0.02 31.9 99.2 0.33
SO-5302 Taojiang Mn ore 1.04 38.2 0.02 4.59 0.63 16.9 0.27 0.05 2.34 0.02 34.8 98.8 0.06
TABLE 2. TRACE ELEMENTS
Sample name Ag As Ba Bi Cs Cu Ga Mo Nb Ni Pb Rb Sb Se Sn Sr Ta Te Th Tl U V Y Zn Zr
Analysis unit ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm
Standards
NIST-25d Mn ore 4 32 2000 0.7 0.8 401 22 10 4 481 36 19 5.2 0.3 5 163 <0.5 0.2 4.2 5.2 1.3 104 16 706 18
NIST-26 Fe ore 5 2 23.8 <0.1 0.1 5 2 <2 4 7 <5 0.8 0.2 <0.1 118 16 <0.5 0.2 0.8 <0.5 0.6 32 7 <10 4
NIST-27b Fe ore 4 2.4 49.9 <0.1 0.2 42 2 <2 2 19 21 0.6 1.3 0.1 6 12 1.2 0.3 0.4 <0.5 0.3 39 35 26 2
NIST-28a Fe ore 4 12 33.2 <0.1 0.6 16 4 2 3 17 25 1.3 0.3 <0.1 3 26 <0.5 0.2 0.4 <0.5 3.6 104 21 18 2
NIST-69b Bauxite 5 14 94.8 1.2 0.2 19 70 16 918 13 47 3.6 0.9 1.8 13 133 33 0.4 95 <0.5 12 145 63 15 2800
NIST-88b Dolomite 9 0.8 21.5 <0.1 0.2 <5 <1 <2 2 9 <5 3.3 <0.1 0.1 <1 69 <0.5 0.1 0.5 <0.5 0.2 <5 8.9 <10 <2
NIST-97b Flint clay 2 3 178 0.9 3.4 24 63 8 45 35 49 34 0.8 0.3 11 91 2.8 0.3 37 <0.5 7.9 254 37 72 450
NIST-690 Fe ore 3 4 21.7 <0.1 <0.1 13 2 <2 3 11 <5 0.4 0.2 <0.1 <1 4.1 <0.5 0.1 0.3 <0.5 0.1 37 3.6 <10 2
NIST-1413 Al sand 3 0.3 1040 <0.1 2.3 6 9 <2 4 7 224 119 0.5 0.2 <1 212 <0.5 0.1 2.5 0.5 1 10 6.5 21 107
VL-1 Laterite 2 2.3 21 <0.1 <0.1 115 46 <2 15 8 10 0.8 0.4 0.2 3 1.8 0.8 <0.1 14 <0.5 1.4 880 3.6 171 173
Tanganshan
TG-3 Mnore 3 121 296 0.5 3.5 83 19 91 24 109 61 56 15 4.2 4 44 1 0.5 11 0.5 2.4 273 28 22 87
TG-4 Mnore 3 30 48.7 0.3 0.2 <5 3 4 15 27 31 0.9 5.8 1.1 3 93 <0.5 0.2 1 <0.5 0.5 29 33 20 26
TG-5 Mnore 9 66 51.4 0.7 0.3 16 9 9 35 63 104 1.7 22 2.8 10 46 <0.5 0.2 5.5 <0.5 1.3 131 48 18 40
TG-6 Shale 3 61 100 0.3 4.2 136 18 101 23 149 90 21 46 1.9 8 86 1.5 0.2 18 1.4 5.2 348 59 41 148
TG-7 Shale 3 13 1510 0.1 5.6 43 20 19 18 71 36 215 10 2.4 4 328 1.1 <0.1 17 1.6 3.5 173 59 53 238
TG-8 Shale 3 61 1090 0.2 3.8 62 18 79 19 61 25 94 5 3.5 4 207 1.4 0.3 13 0.7 4 269 51 131 169
TG-9 Shale 3 92 69.8 0.3 1.1 99 10 45 18 77 66 7 18 3.5 5 174 0.7 0.2 12 <0.5 3.3 160 57 76 90
Xiangtan
XT-0.3 Shale 5 152 810 0.2 12.2 52 23 10 19 52 111 106 10.8 1 5 111 0.8 0.2 13.7 3.6 3.53 180 46.1 61 157
XT-0.4 Mnore 3 106 436 0.3 8.5 38 19 14 12 38 44 65.6 7.7 1 4 59.3 0.8 0.2 9.8 2.1 2.45 128 32.1 109 96
XT-0.6 Mnore 2 37 260 <0.1 1 9 12 5 14 14 8 16.2 2.7 0.5 5 84.4 <0.5 0.2 1.2 <0.5 0.71 53 23.9 45 29
XT-1.3 Mnore 5 30.7 257 <0.1 0.7 13 13 8 21 33 8 11 2.9 0.4 3 118 <0.5 0.2 1.2 0.5 0.72 52 41.9 487 30
XT-2.4 Mnore 3 252 1140 0.4 5.2 55 20 38 23 85 23 35.8 11.1 1.2 4 366 0.6 0.6 12.4 5.2 5.79 137 51.5 38 97
XT-3.7 Mnore 3 144 5130 0.2 2.1 32 15 25 11 34 17 17.1 4.9 0.6 3 321 0.5 0.3 6.9 2.9 2.58 73 50.6 5 66
XT-4.2 Shale 4 119 1070 0.2 15.1 44 24 3 15 54 71 121 2.2 3.7 3 36.2 0.9 0.4 12.9 2 3.71 65 31.6 69 205
XT-4.3 Dolomite 4 24.1 418 0.1 1.9 34 8 7 2 28 38 8.4 0.9 2.8 <1 1140 <0.5 0.3 1.4 <0.5 3.63 112 25.5 21 134
XT-4.5 Mnore 2 92.4 15200 0.2 5 39 21 17 12 38 14 53.5 2.6 1.9 3 169 0.5 0.4 8.8 2.1 2.63 65 46.9 44 80
XT-6.1 Shale 3 >500 686 0.4 8.7 91 22 62 49 174 105 88.8 30.8 3 5 95.3 1 1.2 17.1 5.7 5.24 240 42 32 166
XT-7.2 Shale 2 101 522 0.2 12.4 43 21 <2 16 43 22 135 3.1 4.2 4 30.9 1 0.3 14.1 2.7 2.75 64 33.2 39 205
XT-8.1 Shale 2 80.2 481 0.2 4.5 37 21 <2 13 55 189 130 1.3 3.1 <1 39.6 0.8 0.3 14.4 0.7 2.79 68 42.6 51 190
Other Mn Ores
SO-5100 Mamatwan Fe fmn <1 160 1.2 <5 6 <2 2 8 6 23.9 <1 62.8 0.9 2 <0.5 0.18 12 10.1 47 15.2
SO-5101 Mamatwan Mn ore <1 286 0.2 <5 10 <2 <1 17 6 1 <1 169 1.4 0.4 <0.5 0.16 <5 4.9 9 2.8
SO-5102 Mamatwan Mn ore <1 467 <0.1 <5 8 <2 <1 14 <5 <0.2 <1 226 0.7 0.4 <0.5 0.12 <5 6 27 2.5
SO-5103 Mamatwan Mn ore <1 89.8 <0.1 <5 8 <2 <1 18 <5 <0.2 <1 62.5 0.7 0.4 <0.5 0.23 6 5.1 159 2.5
SO-5104 Mamatwan Mn ore <1 598 <0.1 <5 9 <2 <1 25 <5 <0.2 <1 126 0.7 0.3 <0.5 0.12 <5 5 99 1.8
SO-5108 Hotazel Fe fmn <1 8.9 0.1 <5 2 <2 1 8 10 0.6 <1 3.7 7.4 0.2 <0.5 0.07 13 3.4 <5 2.4
SO-5206 Molango shale <1 233 4.7 25 9 16 4 131 16 56 <1 331 1.4 3.9 7.9 3.07 67 19.9 287 82.9
SO-5207 Molango Mn ore <1 65.2 0.1 6 9 2 2 40 6 <0.2 <1 41.2 0.5 1.4 <0.5 1.46 56 8.2 54 12.9
SO-5209 Molango Mn ore <1 24.3 0.1 8 9 <2 1 79 6 <0.2 <1 38.6 <0.5 1.4 <0.5 1.8 77 9 41 16.3
SO-5301 Taojiang Mn ore <1 403 0.5 11 6 <2 <1 95 13 7.7 <1 1370 1 0.7 <0.5 0.23 12 11.3 120 3.1
SO-5302 Taojiang Mn ore <1 465 0.7 31 8 <2 <1 96 16 10.2 <1 974 1.9 0.7 <0.5 0.11 25 9.6 115 2.1
TABLE 3. RARE-EARTH ELEMENTS
Sample name La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu SumREE Ce/Ce* Eu/Eu*
Analysis unit ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm nasc nasc
Standards 57 58 59 60 62 63 64 65 66 67 68 69 70 71
NIST-25d Mnore 18.8 56.5 4.74 18.4 3.8 0.95 3.16 0.49 3.3 0.59 1.67 0.25 1.4 0.21 114 1.30 1.20
NIST-26 Feore 4.7 9.7 1.17 4.9 0.8 0.31 1.05 0.17 1.02 0.22 0.63 0.1 0.6 0.07 25 0.90 1.49
NIST-27b Feore 11 25.6 3.7 17.7 5.5 3.42 7.29 1.58 8.8 1.44 3.71 0.55 2.7 0.45 93 0.87 2.37
NIST-28a Feore 10.5 13.7 2.07 9.3 1.9 0.66 2.6 0.45 3.3 0.65 1.97 0.28 1.6 0.25 49 0.64 1.30
NIST-69b bauxite 74.7 252 10.8 32.7 5.6 0.81 5.45 1.16 7.2 2.09 8.01 1.4 10.5 1.5 414 1.93 0.64
NIST-88b dolomite 5 4.6 0.85 2.9 0.6 0.13 0.54 0.11 0.74 0.16 0.43 0.05 0.3 0.05 16 0.49 1.00
NIST-97b flintclay 21.5 37.8 3.99 13.3 2.6 0.68 3.99 0.78 6.4 1.35 4.23 0.68 4.9 0.67 103 0.89 0.93
NIST-690 Feore 1.4 1.9 0.26 1.1 0.2 0.1 0.32 0.06 0.45 0.09 0.28 0.05 0.3 <0.05 7 0.69 1.74
NIST-1413 high-Alsand 8.2 13.8 1.68 6.7 1.2 0.47 0.99 0.16 0.95 0.17 0.53 0.09 0.5 0.08 36 0.81 1.89
VL-1 laterite 2.2 13.7 0.48 1.6 0.6 0.19 0.63 0.12 0.75 0.2 0.53 0.11 0.8 0.14 22 2.90 1.36
Tanganshan
TG-3 Mnore 34 82.7 7.35 29.2 5.4 1.5 5.37 0.84 6.4 1.19 3.73 0.55 3.4 0.59 182 1.14 1.22
TG-4 Mnore 29.1 92.5 7.54 31.1 6.8 2.2 7.19 1.17 6.88 1.42 4.23 0.5 2.8 0.33 194 1.36 1.38
TG-5 Mnore 34.8 105 8.36 34.5 8.4 2.49 7.85 1.31 9.1 1.67 4.84 0.66 4.2 0.52 224 1.34 1.35
TG-6 shale 61.6 155 14 59.8 9.9 1.95 11.5 2.09 14.5 2.82 8.32 1.08 7.5 1.18 351 1.15 0.80
TG-7 shale 54.9 111 12.8 51.3 10.9 1.52 10 1.55 11.5 2.19 5.59 0.76 5 0.76 280 0.91 0.64
TG-8 shale 42 88.8 10.1 43.1 9.3 2.14 10.2 1.53 10.5 1.81 5.36 0.73 4.7 0.72 231 0.94 0.96
TG-9 shale 54.4 170 14 58.6 12.3 2.68 12.6 1.84 12.9 2.31 6.41 0.87 5.2 0.72 355 1.34 0.95
Xiangtan
XT-0.3 shale 54.3 125 12.4 48.4 9.8 1.68 8.49 1.33 8.9 1.71 5.29 0.79 5.4 0.68 284 1.05 0.81
XT-0.4 Mn ore 32.1 75 7.79 32.6 7.3 1.11 6.46 1.13 7.16 1.25 3.7 0.56 3.6 0.48 180 1.03 0.71
XT-0.6 Mn ore 23 69 6 23.7 4.5 0.9 4.89 0.78 5.03 1.15 2.9 0.4 2.6 0.4 145 1.28 0.84
XT-1.3 Mn ore 29 89.8 7.41 31.9 6.5 1.38 6.22 0.97 6.67 1.3 3.65 0.51 2.9 0.36 189 1.33 0.95
XT-2.4 Mn ore 48.2 127 11.5 44.2 8.3 1.58 7.61 1.33 8.06 1.86 5.9 0.86 5.2 0.82 272 1.17 0.87
XT-3.7 Mn ore 43.4 134 11.8 50.1 11.6 2.32 10.2 1.7 11.3 2.19 5.88 0.92 4.8 0.57 291 1.29 0.94
XT-4.2 shale 41.7 69.5 8.76 32.6 7.2 1.35 6.08 0.91 5.8 1.14 3.45 0.5 3.2 0.51 183 0.79 0.90
XT-4.3 dolomite 8.3 15.4 1.71 7.4 2.3 0.57 2.75 0.46 3.01 0.72 2.17 0.33 1.8 0.3 47 0.89 1.00
XT-4.5 Mn ore 39.5 100 9.68 40.7 9 1.2 9.1 1.33 8.53 1.74 5.2 0.72 4.2 0.63 232 1.11 0.58
XT-6.1 shale 81.6 199 15.6 55.7 9.9 1.72 8.56 1.36 8.1 1.62 5.04 0.74 5.5 0.77 395 1.21 0.82
XT-7.2 shale 46 77.9 9.57 36.7 7.4 1.18 6.1 0.91 5.91 1.2 3.95 0.58 3.5 0.49 201 0.81 0.77
XT-8.1 shale 42.5 71 9.28 36.3 7.8 1.65 7.52 1.15 7.12 1.42 4.29 0.61 4.4 0.67 196 0.78 0.95
Other Mn Ores
SO-5100 Mamatwan Fe formation 4.1 5.2 0.56 2.5 0.5 0.16 0.74 0.12 1.01 0.24 0.87 0.14 0.9 0.16 17 0.75 1.15
SO-5101 Mamatwan Mn ore 2.5 3 0.41 1.9 0.3 0.13 0.49 0.07 0.5 0.12 0.34 0.06 0.4 0.09 10 0.65 1.49
SO-5102 Mamatwan Mn ore 3.5 3.5 0.52 2.4 0.4 0.2 0.7 0.1 0.63 0.16 0.49 0.06 0.5 0.07 13 0.57 1.66
SO-5108 Hotazel Fe formation 2.3 3.1 0.26 1 0.1 0.07 0.29 0.03 0.38 0.08 0.32 0.03 0.3 0.06 8 0.87 1.80
SO-5206 Molango shale 23.3 37 4.9 20.3 3.8 0.84 4.16 0.62 3.39 0.69 2 0.26 1.8 0.26 103 0.75 0.93
SO-5207 Molango Mn ore 10.4 16.5 2.17 8.7 1.6 0.41 1.72 0.26 1.37 0.28 0.76 0.09 0.7 0.18 45 0.76 1.09
SO-5209 Molango Mn ore 13.2 19.7 2.53 10.3 1.8 0.44 2.02 0.31 1.6 0.33 0.94 0.13 0.8 0.11 54 0.74 1.01
SO-5301 Taojiang Mn ore 11.4 23.9 1.73 7.3 1.6 0.77 2.12 0.32 1.92 0.36 0.94 0.14 0.8 0.11 53 1.17 1.84
SO-5302 Taojiang Mn ore 13.9 20.4 2.06 9.4 1.8 0.86 2.19 0.31 1.8 0.36 1.06 0.14 0.9 0.14 55 0.83 1.90
1.5
Nb
Log ppm/ppm other ores
1.0 Zr Pb
Ti
P Y Mo
0.5
0.0
Figure 3. End-member MnCO3 ores from
-0.5 the Chinese Neoproterozoic deposits com-
Ta pared to the average of typical MnCO3
Ca Sr
-1.0 ores from other deposits.
-1.5
-2.0
0 10 20 30 40 50 60 70 80 90 100
Atomic number
2.5
Mn
2.0
ppm/ppm USGS SDO1
1.5
1.0
Ca
0.5 Figure 4. Comparison of compositions of
0.0
host shales and the black shale standard
SDO1.
-0.5
-1.0 Mo
U
-1.5
0 10 20 30 40 50 60 70 80 90 100
Element
10.00
5.00
TG
ppm/ppm NASC
1.00
XT
Ta o j i a n g Figure 5. Most Mn carbonate ores exhibit
a positive Eu anomaly and a negative Ce
0.50
anomaly. Both Chinese deposits have prom-
Molango inent positive Ce anomalies and Xiangtan
(XT) has a negative Eu anomaly whereas
0.10 Tanganshan (TG) has the typical positive
Eu anomaly.
Mamatwan
0.05
Ce Eu
0.01
56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72
Element
Eu anomalies, normalized to NASC. For the Chinese deposits, Mo available for near-shore deposition. Dobrzinski et al. (2004)
Tanganshan also has a positive anomaly, whereas all of the sam- have performed a similar paleoredox analysis using S/C, U/Th,
ples from the Xiangtan deposit have distinct negative Eu anom- Cd, Mo, and Ce* as indicators. Their focus was largely on the
alies. The La/Yb ratio is less than 1.0 for both of the Chinese glacial marine units, which they assigned to oxic depositional
deposits, whereas the ratio is substantially greater than 1.0 for conditions. The interglacial interval, however, was more reduc-
the others. On the other hand, the ΣREE is considerably higher in ing, falling at the border between oxic and suboxic or between
the Neoproterozoic deposits. These differences in REE chemistry suboxic and anoxic, depending on the indicator.
are not related to detrital contamination by heavy minerals; there
is no correlation between ΣREE and either P2O5 or Zr content Isotopes
of the ores, so xenotime or zircon are not controlling the REE
chemistry of the ores. Therefore the signature is mostly carried The Chinese Neoproterozoic Mn ores are radically different
by the Mn minerals themselves and is a fundamental character of in S isotope chemistry from other Mn ores. Figure 8 shows the
the deposits. distribution of S isotopes in MnCO3 ores and host black shales
In studies of black shales like the host rocks to these depos- from Phanerozoic deposits. Note that the Phanerozoic shales
its, a common question is the degree of anoxia in the basin dur- yield broadly negative δ34S values with a mode at about –25‰.
ing deposition. Several ways to estimate this property have been
proposed. Among the most common are degree of pyritization
(Raiswell et al., 1988), C-S plots, and Mo content (Crusius et al.,
1996). Liu (1988) measured degree of pyritization (DOP) in the 10
shales hosting a number of Mn deposits and reported for Xiang- Shale

8
tan that footwall shales average 0.70, the ores average 0.49, and Ore
the hanging wall shales average 0.71. For comparison, Raiswell
Total S %
6
et al. (1988) suggested 0.45 as the boundary between fully oxic
and suboxic environments and 0.75 as the boundary between 4
suboxic and fully anoxic. The Xiangtan ores, using this scale, l M a ri n
e
N o rm a
were deposited under oxic to suboxic conditions, whereas their 2
host shales were deposited under conditions transitional between

0
suboxic and anoxic. Comparable data are not available for Tan- 0.0 1.0 2.0 3.0 4.0 5.0 6.0
ganshan. Organic C %
In Figure 6 we show a plot of organic C versus total S for
Xiangtan and Tanganshan. The data show a pattern typical of Figure 6. C vs. S plot shows most samples of both ore and shale plot
above the normal marine line and have a poor correlation between the
anoxic-euxinic conditions, in which there is a poor correlation
two variables. Both features of the plot are consistent with anoxic con-
between S and Corg, with S falling mostly between 2% and 4%. ditions in the basin of deposition.
Thus the C-S relations indicate more reducing conditions than
does DOP.
Mo/Al ratios are commonly used to assess the extent of Mo
enrichment in shales (e.g., Warning and Brumsack, 2000). For
the Chinese shale samples, the Mo/Al (ppm metal/ % metal) ratio 8
at Tanganshan averages 12, whereas at Xiangtan it averages only Host Shale

3. For comparison, the ratio is to 16–20 for modern euxinic sedi- 6 Ore
ments and <1 for average shale. A comparison of Mo versus Corg
(Fig. 7) shows no correlation, contrary to the situation seen in most
ppp Mo
ds
modern sediments, which show an excellent correlation. Further- 4 se
ic
more, the Mo contents are almost all much lower than for modern ox
an
e rn
euxinic sediments. The DOP and C/S patterns suggest that the od
2 M
host shales for the Chinese Mn deposits were formed under tran-
sitional suboxic-anoxic conditions and that the degree of anoxia
was greater at Tanganshan than at Xiangtan. However, Mo levels 0
are somewhat lower in these shales than we would expect from 0.0 1.0 2.0 3.0 4.0 5.0 6.0
the pattern seen in modern sediments. Euxinic basins sequester Organic C %

large amounts of Mo as MoS2 in bottom sediments (Barling and
Figure 7. Mo in both ores and host shales does not correlate with the
Anbar, 2004). Perhaps the low Mo in these Neoproterozoic black amount of organic carbon, unlike the situation in Phanerozoic shales
shales is related to widespread oceanic anoxia, with a higher flux and modern sediments, and the amount of Mo is less than is found in
of Mo to deep-water sediments than today, which resulted in less modern sulfidic sediments.
14
12
Shale
tinction between the ores and the host shales. Xiangtan and Tan-
ganshan deposits have significantly different δ34S, although both
Number of Samples
10 Ore
are far more enriched in 34S than other deposits shown in Table
8
1. The average for Xiangtan is +44.3‰, whereas Tanganshan is
6 somewhat less enriched at +29.3‰. For comparison, Tang and
4 Liu (1999) reported shale δ34S values that average +53.5‰ and
MnCO3 ores that average +49.3‰ for Minle. Li et al. (1999)
2
reported two shale and two ore samples from Xiangtan that aver-
0
-50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 age +61.2‰ and +52.1‰ respectively, and two shales and four
δ34S ore samples from Songtao that average +45.6‰ and +54.4‰,
respectively. Thus not only are the Neoproterozoic ores peculiar
Figure 8. S isotope distribution of pyrite in Phanerozoic Mn ores and in
their host shales. Note that the ores are ~30‰ heavier than the shales. in their high S values, they are also peculiar in the coincidence of
the shale and MnCO3 values, which indicates a difference in ore-
forming conditions between the Neoproterozoic and the younger
deposits. The Paleoproterozoic Kalahari ore samples were also
This distribution is typical of syngenetic pyrite formed by bacte- analyzed for S isotopes, but the amount of S was too low for reli-
rial sulfate reduction in fine-grained sediments (e.g., see Ohmoto able measurement.
and Goldhaber, 1997, Fig. 11.17). The MnCO3 ores, on the other We did not measure C isotopes in our samples, but there are
hand, have a more irregular distribution with a distinctly higher literature data from nearby deposits. Tang and Liu (1999), Li at
mode at about +5‰. Okita (1992) explained this heavier S as al. (1999), and Yang et al. (1999) report 24 analyses of carbon-
resulting from formation of pyrite deeper in the sediment than is ate carbon, mostly from the Minle and Songtao deposits. The
normally the case during the diagenesis of marine sediments. The average δ13C value is –8.7‰ with a narrow standard deviation of
large volume of Mn in these sediments poises the Eh at values 1.9. This value is similar to the average of –9.1‰ for the Kala-
too high for sulfate reduction. Until all Mn oxide is converted to hari deposits (Gutzmer, 1996) and –13.1‰ for Molango (Okita,
Mn carbonate, Eh stays high and no sulfides form. Thus pyrite 1987). Thus the Chinese deposits fall within the range for other
formation takes place farther from the sediment-water interface MnCO3 ores for this parameter.
than usual. This greater depth reduces diffusive exchange with
the bulk seawater, resulting in complete reduction of a smaller DISCUSSION
sulfate reservoir and thus smaller amounts of pyrite with heavier
δ34S than normal. General Model for Mn Ore Genesis
In Figure 9 we show our data plus previous results from
similar Chinese deposits reported by Li et al. (1999) for the In modern sedimentary basins, Mn shows a strong tendency
Xiangtan and Songtao deposits and by Tang and Liu (1999) for to be enriched around the margins of areas with deeper, anoxic
the Minle deposit. The Neoproterozoic ores and their host shales bottom waters. The Baltic and the Black Seas exhibit this behav-
are uniformly shifted to very enriched δ34S values, with no dis- ior (see, e.g., Sternbeck and Sohlenius, 1997). The cause is the
great insolubility of the Fe sulfide, pyrite, compared to the Mn
sulfide, alabandite. Fe and Mn are transported into the basin of
20 deposition as coatings of Fe2O3 or MnO2 on detrital particles and
Shale are released as soluble Fe2+ or Mn2+ to the pore waters of the sedi-
16
ment during diagenesis through bacterial reactions (see e.g., Pot-
Number of Samples
Ore
12 ter et al., 2005, p. 138). The dissolved Fe is quickly incorporated

into pyrite, but the Mn diffuses up into the overlying water mass.
8 As a consequence, Fe is vanishingly low in the deep portions of
anoxic basins, whereas dissolved Mn is much higher. Both sub-
4 stances form insoluble oxides and so are absent in solution in the
shallow, oxygenated water mass. There is a peak in dissolved Mn
0
-50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 just beneath the redox interface that reflects the redissolution of
δ34S MnO2 particles that form in the shallow water and sink through
the interface (Fig. 10).
From this cycling pattern of Mn in modern basins, a general
Figure 9. S isotope distribution of pyrite in Neoproterozoic Mn ores of
South China and their host shales. Note lack of distinction between the model of Mn ore genesis has been developed in which Mn is sol-
two lithologies and the very heavy values compared to Phanerozoic ubilized from deep-water sediments in anoxic basins and repre-
analogs. cipitated around the margins of these basins at the point where
Concentration (ppb)
O2
100 300 500 MnO2 MnO2 MnO2 MnO2
-200
redoxcline
MnO 2 O 2 zone Mn2+ Mn2+ Mn2+
0
Depth relative to anoxic interface (m)
t MnCO3
H 2S en
e d im
hs
ri c
Fe ic-
an
org
FeS2 FeS2
-500 Mn 2+
Figure 11. Schematic model for development of MnCO3 by early dia-

H 2 S zone genetic reaction of Mn oxide with sedimentary organic matter.
-1000
around the margins of the basin. Where the clastic sedimentation
rate is low, significant Mn accumulations can develop. Although
these accumulations start as oxides, they are usually preserved
as carbonates in the rock record. Reaction with organic matter
-1500 in the sediment converts the primary Mn oxide to secondary Mn
carbonate, which is depleted in 13C as a result of derivation of a
Figure 10. Distribution of Fe and Mn in the Black Sea, a modern an-
oxic basin that shows Mn concentration at the redox boundary. (Modi-
portion of the carbon from decaying organic matter.
fied from Force and Maynard, 1991, Fig. 11.2) A key observation supporting this model is a strong corre-
lation between Mn contents in the rocks and C isotopes (Okita
and Shanks, 1992). The production of the MnCO3 mineralization
requires the consumption of large amounts of organic matter and
the redox interface impinges on the seafloor. This model was first most likely occurred during early diagenesis, when bacterial pro-
articulated by Force and Cannon (1988) from their observations cesses are most effective. The process can be represented sche-
of Mn distributions in modern sediments and facies analysis of matically by the reaction:
several ancient Mn deposits. Subsequently the model has been
developed in some detail based on stable isotopic studies of Pha- 2MnO2 + CH2O + HCO3- = 2MnCO3 + H2O + OH- . (1)
nerozoic deposits, particularly Molango in Mexico (Okita, 1987;
Okita et al. 1988; Maynard et al., 1990; Okita 1992; Okita and From this relationship, about one-half the carbon in Mn carbon-
Shanks, 1992). See also reviews by Force and Maynard (1991), ates is derived from organic matter, one-half from seawater. For
who emphasized the ancient record and favored a dominant role marine organic matter δ13C is −30‰ to −20‰, whereas seawater
for basin geometry, and by Calvert and Pedersen (1996), who is close to 0‰. Thus a diagenetic Mn carbonate should have val-
emphasized the modern record and argued for a dominant role of ues of −15‰ to −10‰, close to the observed values of −13‰ to
surface-water productivity in controlling Mn distribution. −9‰ quoted above.
The process of Mn enrichment begins with the precipitation At the same time that the Mn is oxidizing the organic matter,
of Mn oxides within the water column at the interface between it also attacks any Fe sulfide in the sediment (Aller and Rude,
oxidizing and reducing conditions, usually a halocline. Most of 1988; Schippers and Jørgensen, 2001). Pyrite is nearly ubiqui-
the precipitated Mn simply redissolves as it passes downward tous in marine sediments because of the reaction between detrital
through the water column, unless the seafloor is shallow enough Fe oxides and seawater sulfate. In Mn-rich sediments, however,
to intercept the redox interface (Fig. 11). This phenomenon pro- this process is blocked because any precursor FeS that forms is
duces what might be called a “manganese compensation depth.” quickly destroyed:
Below this depth, which occurs at about –200 m in the Black
Sea, MnO2 particulates dissolve while settling through the water FeS + 4.5MnO2 + 4H2O = FeOOH + 4.5Mn2+
column and none reach the bottom. Above this depth, solid + SO42- + 7OH-. (2)
MnO2 is stable and does not dissolve as it sinks. Consequently,
the deep-water sediments are low in Mn, sediments close to the Reaction (2) predicts that Mn ore deposits should be very low
compensation depth have a strong enrichment in Mn oxide par- in S, as is observed in most occurrences other than those in the
ticles, and shallow-water sediments are low in Mn. Thus there is Neoproterozoic. Also the minor amount of pyrite that does form
a critical depth for Mn enrichment that produces a “bathtub ring” should be relatively heavy isotopically. As mentioned above, this
prediction of heavy S is based on the requirement that any pyrite was everywhere anoxic, pyrite formation in sediments or at the
that forms be relatively late, forming after all of the Mn oxide ridges removed most of the sulfur, but left plenty of Fe to accu-
has been converted to MnCO3. Therefore the degree of contact mulate in the water column as soluble Fe2+. This situation is the
with the overlying seawater reservoir of sulfate S will be limited, reverse of the modern oceans, where oxic bottom waters keep
sulfate reduction will go to completion, and the small amount of most of the detrital Fe insoluble as the oxide and quickly convert
sulfide that does form will be isotopically close to its parent sul- any soluble Fe2+ in vent fluids into insoluble Fe3+. As a conse-
fate, in contrast to normal pyrite in black shale, which is highly quence, dissolved S is in great excess over dissolved Fe, and the
depleted in 34S. Subsequent work has shown that this model has concentration of SO42- is high.
broad applicability to Mn ore deposits. See for example Nyame The precipitation of most of the seawater reservoir of sul-
(1998) on the Nsuta deposit of Ghana, and Tsikos (1999) for the fur as sulfide resulted in the sequestration of large amounts of
Hotazel deposit in the Kalahari Mn field of South Africa. 32
S in sediments and the accumulation of residual 34S-enriched S
in the water column. On melting of the ice, there is overturn of
Ore Genesis in the Neoproterozoic—The Snowball the oceans and this Fe- and Mn-laden deep water wells up onto
Earth Model shallow platforms and deposits the Rapitan-type iron formations
characteristic of the Neoproterozoic (Maynard, 1991) and large
The Neoproterozoic deposits appear chemically to be simi- deposits of Mn oxide. Because this upwelling deep water con-
lar to their older and younger counterparts, except for different tains sulfate strongly enriched in 34S, the shallow-water diage-
S isotope and REE patterns. The similarity in carbon isotopes netic pyrite produced from it is similarly very heavy.
and trace element chemistries suggests that the MnCO3 formed Although in broad terms this model provides a satisfactory
during early diagenesis from a Mn oxide precursor, just as in explanation for a wide array of observations, there are some sig-
other deposits. How was this Mn oxide deposited in such large nificant discrepancies:
amounts and what accounts for the super heavy S in the associ- • The Fe and Mn deposits are interbedded with the glacial
ated pyrite? deposits rather than succeeding them.
An appealing hypothesis has been put forward by Gorjan et • The ores formed in narrow rifts rather than on open
al. (2000; see also Gorjan et al., 2003) in which they related iso- shelves.
topic behavior and the abundance of Fe and Mn mineralization to • Sulfur isotopic compositions are typically heavy through-
turnover after melting of glacial ice on a “snowball” Earth (Fig. out the Proterozoic and excursions appear to be unrelated
12). The concept of a frozen Earth in the Neoproterozoic (Kirsch- to glacial episodes.
vink, 1992) explains many of the biological, sedimentological, • The REE evidence from the associated iron formations is
and geochemical anomalies seen at this time and accordingly incompatible with a totally anoxic deep ocean.
has received much attention. In the Gorjan et al. (2000, 2003)
version of this model, the ocean became totally anoxic under its An Ice-covered Rift Model for Ore Genesis in the
ice cover, which led to a buildup of dissolved Mn and Fe in the Neoproterozoic
bottom water. Because the flux to the oceans of detrital Fe from
the continents and hydrothermal Fe from ridge crests is much These aspects of the Mn-Fe mineralization in the Neo-
greater than the flux of dissolved SO42- and because the seafloor proterozoic suggest that the snowball Earth model needs to be
refined. The hypothesis that best explains these additional obser-
vations is a “partial snowball Earth,” with ice-covered continents
and marginal seas, but with an ice-free open ocean at low lati-
THAW tudes. See Eyles and Januszczak (2004a, 2004b), Poulsen et al.
12CO 12CO (2002), and Poulsen (2003) for useful discussions of full versus
2
MUD P 2
sea level partial ice cover from a field and from a modeling basis. Three
O2 AEROBIC principal lines of evidence support a partial, or “soft” snowball
ANAEROBIC model for mineralization: stratigraphic sequence, stable isotopic
values, and REE compositions.
Fe2+ 34SO4-2 H12CO-3
P O S T- G L A C I A L Stratigraphic Setting of Mineralization
The interglacial position of the ores suggests that the snow-
Gorjan et al. 2000 ball Earth model of Fe-Mn deposition needs to be modified to
allow the main episode of mineralization to occur during periods
Figure 12. “Snowball Earth” model for Neoproterozoic Fe-Mn depos- of partial melting of the ice cover accompanied by oxidation of
its. Rapid melting of ice at the end of the event pushes Fe and Mn rich
deep water onto shelves to precipitate as oxides. Residual sulfate left in surface waters. Multiple glacial advances and retreats are hard to
the oceans is very heavy and produces a pulse of heavy S in sediments reconcile with a totally frozen Earth. Such fluctuations are, how-
at the end of glaciation. ever, compatible with a model of individual ice-covered marine
basins that could have been time transgressive, as proposed by Glaciations

Eyles and Januszczak (2004a, 2004b). 500
The dominance of glacial-marine rift environments is also
more consistent with partial ice cover. The Neoproterozoic Fe Marinoan
and Mn deposits are uniquely associated with narrow rifts on
Age (Ma)
continental crust (Maynard, 1991); none are found on stable Sturtian
shelves of continental margins. The “hard snowball” explanation
for the Fe and Mn mineralization would predict deposition as
widespread sheets on continental shelves (see Gorjan et al., 2003,
p. 95), much like the famous Pennsylvanian black shales of the
U.S. midcontinent described by Heckel (1991), which are related 1000
to maximum flooding associated with deglaciation.
-20 0 +20 +40
Sulfur Isotopic Values Throughout the Neoproterozoic Median δ34 S (%0, CDT) after Strauss (1997)
The recent compilations of sulfide S isotopic data by Strauss
Figure 13. Approximate isotopic curve of sulfide S for the Neopro-
(1997) and by Canfield and Raiswell (1999) show that heavy sul- terozoic. The line shown is the median of a very wide spread of values
fur has been characteristic of both sulfides and sulfates during and so has a higher degree of uncertainty than sulfate S curves for the
much of the Proterozoic and is not correlated to the glacial epi- Phanerozoic. Note the lack of correspondence of heavy S to the end of
sodes. Another important aspect of the Proterozoic S record is that glaciation. The curves actually suggest a dramatic decrease in δ34S in
the spread between sulfate and sulfide values is much less than the post-glacial oceans.
in Phanerozoic rocks, with several sulfide analyses lying higher
than contemporaneous seawater sulfate. Strauss (1997) sug-
gested that these patterns could result from a low-sulfate ocean in
which the amount of pyrite formed at the sediment-water inter- for Neoproterozoic S change. They identified four major excur-
face was limited by the amount of sulfate in the overlying water sions to heavy S: one is slightly above the Sturtian glacials and
rather than by the amount of organic carbon, which is the case in reaches +40‰; the second is 150 m higher in the section and has
modern marine sediments (e.g., Canfield, 2001). A low-sulfate the greatest departure, to +50‰; the third is midway between the
ocean would have been prone to periodic drawdowns of sulfate Sturtian and Marinoan glacial episodes and reaches +35‰; and
concentration during times of higher Fe flux to the ocean, which the final excursion is 100 m above the Marinoan equivalents and
would have resulted in increased pyrite formation. The isotopic reaches +40‰. Thus there are several episodes of development
composition of the residual sulfate S in seawater would then have of heavy S and only the first one can be convincingly related to
spiked to very high values. glaciation. Their data also shows that the sulfate δ34S values are
Determining a worldwide curve for Neoproterozoic S iso- comparable to the sulfide values, indicating that sulfate reduction
tope values comparable to the familiar curves for the Phanero- within sediments must have gone essentially to completion dur-
zoic has been difficult because of the scarcity of primary sulfate ing this time interval. Hurtgen et al. (2002) and Canfield (2004)
minerals from this time period and the high variability in sulfide interpreted these patterns to indicate generally low SO42- concen-
values for rocks of a given age. We have made an approximate trations in seawater at this time. Canfield (2004) has suggested
sulfide S curve by plotting the medians of the values reported by values as low as 200–300 μM, which he attributed to a much
Strauss (1997) for each time interval. As shown in Figure 13, the greater flux of sedimentary S back to the mantle via subduction
median values show large fluctuations, but are invariably heavy of pyrite-rich deep-sea sediments during the Proterozoic than in
compared to Phanerozoic values. There do seem to be periods the modern oceans.
within the Neoproterozoic that experienced spikes to extremely We conclude that the occurrence of extremely heavy δ34S
heavy sulfide S, but these do not match the main glacial episodes. values of pyrite in the Chinese Mn deposits is not related directly
Instead there seems to be a sharp fall in δ34S following each gla- to glaciation but to generally low concentrations of dissolved
ciation, the converse of the prediction from the “hard snowball” SO42- in the Neoproterozoic world ocean that made seawater sub-
model. Perhaps these swings indicate periods of increased deep- ject to rapid and severe swings in its S content. Excess iron, which
water circulation that produced well-oxygenated oceans with must have been present judging from the abundance of Rapitan-
higher SO42- concentrations. type iron formations, would have driven SO42- concentrations
A finer-scale sulfate S curve is available from the work of to very low values through formation of more pyrite (Canfield
Hurtgen et al. (2002), who measured trace sulfate held in car- and Raiswell, 1999), and this could account for the uniformly
bonate minerals from the Neoproterozoic of Namibia. Their sam- extremely high δ34S values found. The combination of high Fe
pling starts above the Sturtian glacial strata and thus does not flux and a low SO42- ocean is what produced the S isotopic signa-
cover the interval corresponding to the Chinese Mn deposits, but ture of the Chinese Mn deposits. We further conclude from our
their results do provide important constraints on the mechanisms survey of literature data that these very heavy values occurred
repeatedly at many times and in many places in the Neoprotero- Normally, these vent-sourced REE are immediately scav-
zoic and are hence more likely to have resulted from world ocean enged by Fe oxides precipitating around the vents (Mitra et al.,
effects rather than from peculiar chemistry in isolated basins. 1994). If, however, the ridge-crest hydrothermal systems vent
into oxygen-free bottom water, this precipitation of Fe oxides and
REE Evidence for an Oxidizing World Ocean scavenging of REE will not occur, and both the Fe and the vent-
signature REE can be concentrated in bottom waters and carried
The nature and source of this Fe flux can be reconstructed into shallower waters to be precipitated far from their source. The
from the behavior of REE in the Mn deposits and in their iron evolution of REE patterns with age suggests that the Archean
formation cousins. Iron formations have distinctive REE patterns and Paleoproterozoic iron formations had Fe dominantly sourced
for each of the three main periods of iron mineralization (Fig. from ridge-crest vents discharging into anoxic seawater, whereas
14). Both the Archean Algoma-type and Paleoproterozoic Supe- the Neoproterozoic deposits received their Fe from seawater with
rior-type iron formations have pronounced positive Eu anomalies a relatively minor hydrothermal component. The absence of a
on NASC normalized plots. Notice also in Table 3 that all of the positive Eu anomaly in the Neoproterozoic deposits indicates
NIST iron-ore samples, which are from the Lake Superior region, that they were deposited during periods of general oxidation of
have positive anomalies. This anomaly is conspicuously absent oceanic bottom water. Dobrzinski et al. (2004) came to the same
in the Neoproterozoic Rapitan-type deposits (Derry and Jacob- conclusion using a variety of paleoredox indicators.
sen, 1990; Klein and Beukes, 1993; and Bau and Möller, 1993; Neodymium isotopes also suggest a waning hydrothermal
Klein and Ladeira, 2004). influence on Fe deposits through time. Jacobsen and Pimentiel-
Positive Eu anomalies are associated with mid-ocean ridge Klose (1988) reported that Archean and Paleoproterozoic iron
vent fluids (see, e.g., Cocherie et al., 1994, on Red Sea sediments). formations have εNd values similar to the mantle, whereas the
Destruction of calcic plagioclase in the oceanic crust leads to a Neoproterozoic Urucum iron formations are similar to modern
release of excess Eu to hydrothermal solutions. Because of the seawater. The average values are +2.7 for Algoma-type IF, +1.0
high temperatures required for the removal of Eu from plagio- for Superior-type, and –2.9 for a single determination from Uru-
clase, Eu release is confined to axial vents. The results of Mich- cum (reported in Derry and Jacobsen, 1990).
ard and Albarède (1986) suggest that temperatures greater than Eu anomalies in Mn deposits do not present such a clear
350 °C are required. Furthermore, at lower temperatures, there picture (Maynard, 2004). Note in Table 3 that virtually all Mn
seems to be little release of any of the REE seawater. For exam- deposits, regardless of age, have positive Eu anomalies. This may
ple, Wheat et al. (2002) studied low-temperature hydrothermal indicate that a significant hydrothermal contribution is present in
springs from the Juan de Fuca ridge and found them to be net Mn ores throughout geologic time, and furthermore that the bot-
sinks instead of sources of REE to seawater. It also seems likely tom water into which the volcanic solutions exhaled was anoxic,
that most Mn release is from the axial vents: Murton et al. (1999) as in the Force and Cannon (1988) model. The Chinese deposits
were able to account for all of the Mn release from a 50-km seg- are mixed, with the three Tanganshan ore samples each having a
ment of the Mid-Atlantic Ridge by flux from the Broken Spur positive anomaly whereas all Xiangtan ore samples have a nega-
vent field. tive anomaly. Neoproterozoic Mn deposits in the Urucum district
1.00
EU/EU*
Urucum 1.01
Neoproterozoic
ppm/ppm NASC
1.32
Hamerseley
Paleoproterozoic
0.10 Figure 14. Europium anomalies in iron for-
mations decrease with decreasing age, indi-
Vermillion cating a diminishing contribution from vol-
Archean
canic-hydrothermal sources to the deposits.
3.71
Ce Eu Data of Derry & Jacobsen 1990

0.01
56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72
Element
of Brazil have uniformly negative anomalies (Graf et al., 1994), of REE. Mn was then exported to shallow water to precipitate
similar to seawater. The Xiangtan and Urucum deposits may have at the oxic/anoxic interface, while some Fe precipitated in deep
a much lower hydrothermal component than the other Mn ores or water as pyrite, thereby removing most of the dissolved SO42- in
perhaps the bottom water in their basins was sufficiently oxic to the basin. This process left residual sulfur strongly enriched in
permit formation of Fe oxides close to the vents, scavenging all 34
S to be incorporated in the shallow-water deposits. Thus the
of the vent-derived REE. super-heavy values we find are not the direct result of glaciation
The behavior of Ce may provide some clues to this situation. but are an indirect result through the impact of glaciation on Fe
Ce shows a much stronger association with the Mn component behavior (Fig. 15). Support for a lateritic source of the Fe and
of Fe-Mn accumulations than to the Fe component (DeCarlo, Mn comes from the very low Na2O content of the host shales,
1991). Fleet et al. (1983) has suggested that the relative size of which averages 0.14% at Tanganshan and 0.07% at Xiangtan,
the negative Ce anomaly on NASC-normalized plots has a linear compared with typical shales, which have ~1% Na2O (Li, 2000,
relationship to the ratio of hydrothermal- to hydrogenous-sourced table VI-4). Dobrzinski et al. (2004) used the CIA index of Nes-
material in the Mn accumulation. In the modern oceans, the bitt and Young (1982) to characterize the weathering state of the
hydrothermal end member has a strongly negative anomaly with source area for the Neoproterozoic deposits of south China. They
a Ce/Ce* value of ~0.46 (Hein et al., 1996, their table 8 corrected calculated an average CIA of 62 for the lower diamictite, 71 for
for geometric calculation of Ce*). Hydrogenous Mn accumula- the interglacial unit, and 65 for the upper glacial unit. For com-
tions, sourced only from seawater, have a strongly positive Ce parison, modern glacial marine sediments of the Scotia Sea aver-
anomaly with a value of ~1.6 (data of Usui and Someya, 1997). age only 55 (data of Diekmann, et al. 2000), whereas modern
The rate of deposition accounts for much of this difference (May- soils average 72 (Maynard, 1992). Thus the Chinese Neoprotero-
nard, 2004). The hydrothermal deposits form relatively quickly zoic glacial section shows higher degrees of weathering than the
and preserve the original REE signature of the water, whereas modern, and the Datangpo Formation was sourced from deeply
the hydrogenous Mn nodules form extremely slowly and, in the weathered material.
modern oxygen-rich deep sea, catalyze the oxidation of Ce3+ to
Ce4+, producing significant Ce enrichments and high Ce/Ce* Unresolved Problems
values. Almost all Fe and Mn ores in Table 3 show negative Ce
anomalies, indicating an appreciable hydrothermal contribution This model of Neoproterozoic Fe-Mn mineralization raises
into anoxic bottom waters. The Chinese deposits depart from several questions that cannot be answered with the data presently
this trend, being uniformly positive like the modern Mn nodules, available:
suggesting at least mildly oxidizing conditions for the Neopro- 1. Why did extensive Fe and Mn mineralization not occur
terozoic and slower deposition rates than in other Mn deposits. during the second glacial episode of the Neoproterozoic?
Slower deposition is also consistent with the higher ΣREE seen The available literature indicates that all the sizable occur-
in the Chinese deposits (Table 3). rences are Sturtian. Are there mis-correlated Marinoan-
The combination of the Eu and Ce data indicate that Archean age deposits or are they truly absent? Could the continents
and Paleoproterozoic Fe and Mn deposits are dominantly hydro- have lacked an extensive lateritic weathering mantle at the
thermal in the source of metals and REE and that the ocean time of the Marinoan glaciations? Perhaps the Marinoan
basins worldwide were dominantly anoxic at the time of their glacial episode was truly global whereas the Sturtian was
formation. The Neoproterozoic Fe deposits, and possibly the Mn only partial.
deposits, received a minor hydrothermal contribution and were
deposited during a time of global oxygenation of the oceans. In
the Phanerozoic, iron formations are confined to the immediate
vicinity of hydrothermal vents, whereas Mn is far-traveled. Each
Mn deposit in Table 3 seems to have had a variable hydrother-
mal component, but each was associated with a large but isolated
basin with anoxic bottom water (e.g., Maynard et al., 1990; Force
and Maynard, 1991).
If seafloor hydrothermal processes were supplying only lim-
ited amounts of Fe and Mn to the Neoproterozoic ore deposits,
what was the source of the high metal fluxes? We suggest that
low-latitude glaciation scraped off tropical soils highly enriched Figure 15. Interglacial model for Fe and Mn mineralization in the Neo-
in Fe and Mn in the form of oxides rather than silicates and thus proterozoic. Mineralization occurs by upwelling of deep water onto
highly susceptible to diagenetic remobilization. This lateritic shallow shelves during interglacial transgression. Mid-ocean ridges
are intermittently exposed to oxidizing water, which precipitates Fe
material was then dumped into small ocean basins that had lim- oxides and REE from vent fluids. The Mn and Fe in the deposits are
ited communication with the open ocean and so could become sourced dominantly from lateritic weathering residues deposited in
anoxic at depth without receiving the mid-ocean ridge signature narrow marine rifts by low-latitude glaciers.
2. What causes the variation in Eu anomaly in Mn ores, but Calvert, S.E., and Pedersen, T.F., 1996, Sedimentary geochemistry of manga-
not in the Fe ores? Does it reflect more localized sources nese: Implications for the environment of formation of manganiferous
black shales: Economic Geology and the Bulletin of the Society of Eco-
for Mn than for Fe? Does it reflect preferential sorption of nomic Geologists, v. 91, p. 36–47.
Eu to Mn oxides under certain conditions? Canfield, D.E., 2001, Isotope fractionation by natural populations of sulfate-
3. What causes the other S isotope excursions in the Neo- reducing bacteria: Geochimica et Cosmochimica Acta, v. 65, p. 1117–
1124, doi: 10.1016/S0016-7037(00)00584-6.
proterozoic that are not associated with Fe-Mn deposits? Canfield, D.E., 2004, The evolution of the Earth surface sulfur reservoir: Amer-
Are these times of higher Corg input? Could they be times ican Journal of Science, v. 304, p. 839–861.
of evaporite deposition with drawdown of sulfate levels? Canfield, D.E., and Raiswell, R.W., 1999, The evolution of the sulfur cycle:
American Journal of Science, v. 299, p. 697–723.
Could they be times of enhanced Fe flux from mid-ocean Canfield, D.E., Raiswell, R.W., Westrich, J.T., Reaves, C.M., and Berner, R.A.,
ridges? Could they reflect times of greater global anoxia 1986, The use of chromium reduction in the analysis of reduced inorganic
in which more of this mid-ocean ridge flux goes to pyrite sulfur in sediments and shales: Chemical Geology, v. 54, p. 149–155, doi:
10.1016/0009-2541(86)90078-1.
rather than to Fe oxide? Cocherie, A., Calvez, J.Y., and Oudin-Dunlop, E., 1994, Hydrothermal
activity as recorded by Red Sea sediments: Sr-Nd isotopes and REE
Resolving these questions will require detailed geochemi- signatures: Marine Geology, v. 118, p. 291–302, doi: 10.1016/0025-
cal analyses of a range of Neoproterozoic deposits. In particu- 3227(94)90089-2.
lar, the analyses of small amounts of sulfate δ34S contained in Crusius, J., Calvert, S., Pedersen, T., and Sage, D., 1996, Rhenium and molyb-
denum enrichments in sediments as indicators of oxic, suboxic, and sul-
carbonates (e.g., Kampschulte and Strauss, 2004) need to be fidic conditions of deposition: Earth and Planetary Science Letters, v. 145,
extended to older rocks and to other sections, and more REE p. 65–78, doi: 10.1016/S0012-821X(96)00204-X.
data are needed. DeCarlo, E.H., 1991, Paleoceanographic implications of rare earth elements
variability within a Fe-Mn crust from the central Pacific Ocean: Marine
Geology, v. 98, p. 449–467, doi: 10.1016/0025-3227(91)90116-L.
CONCLUSIONS Derry, L.A., and Jacobsen, S.B., 1990, The chemical evolution of Precam-
brian seawater: Evidence from REEs in banded iron formations: Geo-
chimica et Cosmochimica Acta, v. 54, p. 2965–2977, doi: 10.1016/0016-
The Neoproterozoic Mn deposits of China were deposited 7037(90)90114-Z.
in a restricted, linear basin during partial retreat of the Sturtian Diekmann, B., Kuhn, G., Rachold, V., Abelmann, A., Brathauer, U., Fütterer,
glaciers. Their exceptionally heavy δ34S values can be explained D., Gersonde, R., and Grobe, H., 2000, Terrigenous sediment supply
in the Scotia Sea (Southern Ocean): Response to Late Quaternary ice
by an excess of Fe flux over S flux to these restricted basins. dynamics in Patagonia and on the Antarctic Peninsula: Palaeogeography,
Eu anomalies in associated Fe ores indicate that this Fe flux Palaeoclimatology, Palaeoecology, v. 162, p. 357–387, doi: 10.1016/
was dominated by diagenetic rather than hydrothermal sources, S0031-0182(00)00138-3.
Dobrzinski, N., Bahlburg, H., Strauss, H., and Zhang, Q.-R., 2004, Geochemi-
whereas Ce anomalies in Mn ores suggest relatively oxidizing cal climate proxies applied to the Neoproterozoic glacial succession
conditions in the bottom water at the site of Mn deposition. A on the Yangtze platform, south China, in Jenkins, G.S., McMenamin,
possible source of enhanced diagenetic flux of Fe and Mn is rapid M.A.S., McKay, C.P., and Sohl, L., eds., The extreme Proterozoic: Geol-
ogy, geochemistry and climate: Washington D.C., American Geophysical
deposition of lateritic soil residues scraped off by low-latitude Union, Geophysical Monograph 146, p. 13–32.
glaciers into sub-oxic or anoxic bottom waters of a rift basin. The Eyles, N., and Januszczak, N., 2004a, ‘Zipper-rift’: A tectonic model for Neo-
easily remobilized Fe and Mn oxides in this residue would have proterozoic glaciations during breakup of Rodinia after 750 Ma: Earth-
Science Reviews, v. 65, p. 1–73, doi: 10.1016/S0012-8252(03)00080-1.
accumulated in low-oxygen bottom water and then precipitated Eyles, N., and Januszczak, N., 2004b, Interpreting the Neoproterozoic gla-
under more oxidizing conditions around the edges of the basin cial record: The importance of tectonics, in Jenkins, G.S., McMenamin,
during sea-level rise in interglacial periods. M.A.S., McKay, C. P., and Sohl, L., eds., The extreme Proterozoic: Geol-
ogy, geochemistry and climate: Washington D.C., American Geophysical
Union, Geophysical Monograph 146, p. 125–168.
REFERENCES CITED Fan, D.-L., Liu, T.-B., and Ye, J., 1992, The process of formation of manganese
carbonate deposits hosted in black shale series: Economic Geology and
Aller, R.C., and Rude, P.D., 1988, Complete oxidation of solid phase sul- the Bulletin of the Society of Economic Geologists, v. 87, p. 1419–1429.
fide by manganese and bacteria in anoxic marine sediments: Geo- Fleet, A.J., Boström, K., Laubier, L., and Smith, K.L., 1983, Hydrothermal and
chimica et Cosmochimica Acta, v. 52, p. 751–765, doi: 10.1016/0016- hydrogenous ferro-manganese deposits; do they form a continuum? The
7037(88)90335-3. rare earth element evidence, in P.A. Rona, ed., Hydrothermal processes at
Barfod, G.H., Albarède, F., Knoll, A.H., Xiao, S., Télouk, P., Frei, R., and seafloor spreading centers: New York, Plenum Press, p. 535–555.
Baker, J., 2002, New Lu-Hf and Pb-Pb age constraints on the earliest ani- Force, E.R., and Cannon, W.F., 1988, A depositional model for shallow-marine
mal fossils: Earth and Planetary Science Letters, v. 201, p. 203–212. manganese deposits around black-shale basins: Economic Geology and
Barling, J., and Anbar, A.D., 2004, Molybdenum isotope fractionation during the Bulletin of the Society of Economic Geologists, v. 83, p. 83–117.
adsorption by manganese oxides: Earth and Planetary Science Letters, Force, E.R., and Maynard, J.B., 1991, Manganese: Syngenetic deposits on
v. 217, p. 315–329, doi: 10.1016/S0012-821X(03)00608-3. the margins of anoxic basins, in Force, E.R., Eidel, J.J., and Maynard,
Bau M., and Möller, P., 1993) Rare earth element systematics of the chemically J.B., eds., Sedimentary and diagenetic mineral deposits: A basin analysis
precipitated component in Early Precambrian iron formations and the approach to exploration: El Paso, Texas, Society of Economic Geologists,
evolution of the terrestrial atmosphere-hydrosphere-lithosphere system: Reviews in Economic Geology, v. 5, p. 147–159.
Geochimica et Cosmochimica Acta, v. 57, p. 2239–2249. Gorjan, P., Veevers, J.J., and Walter, M.R., 2000, Neoproterozoic sulfur-isotope
Breitkopf, J.H., 1988, Iron formations related to mafic volcanism and ensialic variations in Australia and global implications: Precambrian Research,
rifting in the Damara Orogen, Namibia: Precambrian Research, v. 38, v. 100, p. 151–179, doi: 10.1016/S0301-9268(99)00073-X.
p. 111–130, doi: 10.1016/0301-9268(88)90087-3. Gorjan, P., Walter, M.R., and Swart, R., 2003, Global Neoproterozoic (Stur-
Bruchert, V., 1998, Early diagenesis of sulfur in estuarine sediments: The role tian) post-glacial sulfide-sulfur isotope anomaly recognized in Namibia:
of sedimentary humic and fulvic acids: Geochimica et Cosmochimica Journal of African Earth Sciences, v. 36, p. 89–98, doi: 10.1016/S0899-
Acta, v. 62, p. 1567–1586, doi: 10.1016/S0016-7037(98)00089-1. 5362(03)00002-2.
Graf, J.L., O’Connor, E.A., and Van Leeuwen, P., 1994, Rare earth element chemistry, v. 7: Amsterdam, Elsevier-Pergamon, p.289–308.
evidence of origin and depositional environment of Late Protero- Maynard, J.B., Okita, P.M., May, E.D., and Martinez-Vera, A., 1990, Palaeo-
zoic ironstone beds and manganese-oxide deposits, SW Brazil and SE geographic setting of late Jurassic manganese mineralization in the Mol-
Bolivia: Journal of South American Earth Sciences, v. 7, p. 115–133, doi: ango District, Mexico, in Parnell, J., Ye, L.-J., and Chen, C.-M., eds.,
10.1016/0895-9811(94)90003-5. Sediment hosted mineral deposits: Boston, Blackwell Scientific, Interna-
Gutzmer, J., 1996, Genesis and alteration of the Kalhari and Postmasburg tional Association of Sedimentologists Special Publication 11, p. 17–30.
manganese deposits, Griqualand West, South Africa [Ph.D. dissertation]: Michard, A., and Albarède, F., 1986, The REE content of some hydrother-
Johannesburg, Rand Afrikaans University. mal fluids: Chemical Geology, v. 55, p. 51–60, doi: 10.1016/0009-
Heckel, P.H., 1991, Thin, widespread Pennsylvanian black shales of Midcon- 2541(86)90127-0.
tinent North America, a record a cyclic succession of widespread pyc- Mitra, A., Elderfield, H., and Greaves, M.J., 1994, Rare earth elements in sub-
noclines in a fluctuating epeiric sea, in Dennison, J.M., and Ettensohn, marine hydrothermal fluids and plumes from the Mid-Atlantic Ridge:
F.R., eds., Tectonic and eustatic controls on sedimentary cycles: Tulsa, Marine Chemistry, v. 46, p. 217–235, doi: 10.1016/0304-4203(94)90079-
Oklahoma, Society for Sedimentary Geology, Concepts in Sedimentology 5.
and Paleontology, v. 4, p. 65–87. Murton, B.J., Redbourn, L.J., German, C.R., and Baker, E.T., 1999, Sources
Hein, J.R., Gibbs, A.E., Clague, D.A., and Torresan, M., 1996, Hydrothermal and fluxes of hydrothermal heat, chemicals and biology within a segment
mineralization along submarine rift zones, Hawaii: Marine Georesources of the Mid-Atlantic Ridge: Earth and Planetary Science Letters, v. 171,
and Geotechnology, v. 14, p. 177–203. p. 301–317, doi: 10.1016/S0012-821X(99)00157-0.
Hurtgen, M.T., Arthur, M.A., Suits, N.S., and Kaufman, A.J., 2002, The sulfur Nesbitt, H.W., and Young, G.M., 1982, Early Proterozoic climates and plate
isotopic composition of Neoproterozoic seawater sulfate: Implications for motions inferred from major element chemistry of lutites: Nature, v. 299,
a snowball Earth?: Earth and Planetary Science Letters, v. 203, p. 413– p. 715–717.
429, doi: 10.1016/S0012-821X(02)00804-X. Nyame, F.K., 1998, Mineralogy, geochemistry, and genesis of the Nsuta man-
Jacobsen, S.B., and Pimentiel-Klose, M.R., 1988, A Nd isotopic study of the ganese deposit, Ghana [Ph.D. dissertation]: Okayama, Japan, Okayama
Hamersley and Michipicoten banded iron formations: The source of REE University.
and Fe in Archean oceans: Earth and Planetary Science Letters, v. 87, Ohmoto, H., and Goldhaber, M.B., 1997, Sulfur and carbon isotopes, in Barnes,
p. 29–44, doi: 10.1016/0012-821X(88)90062-3. H.L., ed., Geochemistry of hydrothermal ore deposits: New York, John
Kampschulte, A., and Strauss, H., 2004, The sulfur isotopic evolution of Pha- Wiley and Sons, p. 517–611.
nerozoic seawater based on the analysis of structurally substituted sul- Okita, P.M., 1987, Geochemistry and mineralogy of the Molango manganese
fate in carbonates: Chemical Geology, v. 204, p. 255–286, doi: 10.1016/ orebody, Hidalgo State, Mexico [Ph.D. dissertation]: Cincinnati, Ohio,
j.chemgeo.2003.11.013. University of Cincinnati, 362 p.
Kirschvink, J.L., 1992, Late Proterozoic low-latitude global glaciation: The Okita, P.M., 1992, Manganese carbonate mineralization in the Molango Dis-
snowball Earth, in Schopf, J. W., and Klein, C., eds., The Proterozoic trict, Mexico: Economic Geology and the Bulletin of the Society of Eco-
biosphere: Cambridge, UK, Cambridge University Press, p. 51–58. nomic Geologists, v. 87, p. 1345–1366.
Klein, C., and Beukes, N.J., 1993, Sedimentology and geochemistry of the gla- Okita, P.M., and Shanks, W.C., 1992, Origin of stratiform sediment-hosted
ciogenic late Proterozoic Rapitan iron-formation in Canada: Economic manganese carbonate ore deposits: Examples from Molango Mex-
Geology and the Bulletin of the Society of Economic Geologists, v. 88, ico, and TaoJiang, China: Chemical Geology, v. 99, p. 139–164, doi:
p. 542–565. 10.1016/0009-2541(92)90036-5.
Klein, C., and Ladeira, E.A., 2004, Geochemistry and mineralogy of Neopro- Okita, P.M., Maynard, J.B., Spiker, E.C., and Force, E.R., 1988, Isotopic evi-
terozoic banded iron-formations and some selected manganese forma- dence for organic matter oxidation by manganese reduction in the forma-
tions from the Urucum district, Mato Grosso do Sul, Brazil: Economic tion of stratiform manganese carbonate ore: Geochimica et Cosmochi-
Geology and the Bulletin of the Society of Economic Geologists, v. 99, mica Acta, v. 52, p. 2679–2685, doi: 10.1016/0016-7037(88)90036-1.
p. 1233–1244. Potter, P.E., Maynard, J.B., and Depetris, P.J., 2005, Mud and mudstones. Ber-
Li, R., Chen, J., Zhang, S., Lei, J., Shen, Y., and Chen, X., 1999, Spatial and lin, Springer-Verlag, 297 p.
temporal variations in carbon and sulfur isotopic compositions of Sin- Poulsen, C.J., 2003, Absence of a runaway ice-albedo feedback in the
ian sedimentary rocks in the Yangtze platform: South China: Precambrian Neoproterozoic: Geology, v. 31, p. 473–476, doi: 10.1130/0091-
Research, v. 97, p. 59–75, doi: 10.1016/S0301-9268(99)00022-4. 7613(2003)031<0473:AOARIF>2.0.CO;2.
Li, Y-H., 2000, A Compendium of Geochemistry: Princeton University Press, Poulsen, C.J., Jacob, R., Pierrehumbert, R.T., and Huynh, T.T., 2002, Testing
475 p. paleogeographic controls on a Neoproterozoic snowball Earth: Geophysi-
Liu, T.-B., 1988, C-S-Fe correlation of shales hosting sedimentary manganese cal Research Letters, v. 29, p. 10–1 to 10–4.
deposits [Ph.D. dissertation]: Cincinnati, Ohio, University of Cincinnati, Raiswell, R., Buckley, F., Berner, R.A., and Anderson, T.F., 1988, Degree of
300 p. pyritization of iron as a paleoenvironmental indicator of bottom-water
Liu, T.-B., 1990, C-S relationships in shales hosting manganese ores from oxygenation: Journal of Sedimentary Petrology, v. 58, p. 812–819.
Mexico, China, and Newfoundland; implications for depositional envi- Rui, Z.Q., and Piper, J.D.A., 1997, Palaeomagnetic study of Neoproterozoic gla-
ronment and mineralization: Ore Geology Reviews, v. 5, p. 325–340, doi: cial rocks of the Yangzi Block: Palaeolatitude and configuration of South
10.1016/0169-1368(90)90037-N. China in the late Proterozoic Supercontinent: Precambrian Research,
Lottermoser, B.G., and Ashley, P.M., 2000, Geochemistry, petrology and ori- v. 85, p. 173–199, doi: 10.1016/S0301-9268(97)00031-4.
gin of Neoproterozoic ironstones in the eastern part of the Adelaide Geo- Schippers, A., and Jørgensen, B.B., 2001, Oxidation of pyrite and iron sulfide
syncline, South Australia: Precambrian Research, v. 101, p. 49–67, doi: by manganese dioxide in marine sediments: Geochimica et Cosmochi-
10.1016/S0301-9268(99)00098-4. mica Acta, v. 65, p. 915–922, doi: 10.1016/S0016-7037(00)00589-5.
Macouin, M., Besse, J., Ader, M., Gilder, S., Yang, Z., Sun, Z., and Agrinier, P., Shen, Y., 2002, C-isotope variations and paleoceanographic changes during
2004, Combined paleomagnetic and isotopic data from the Doushantuo the late Neoproterozoic on the Yangtze Platform, China: Precambrian
carbonates, south China: Implications for the “snowball Earth” hypothe- Research, v. 113, p. 121–133, doi: 10.1016/S0301-9268(01)00205-4.
sis: Earth and Planetary Science Letters, v. 224, p. 387–398, doi: 10.1016/ Sternbeck, J., and Sohlenius, G., 1997, Authigenic sulfide and carbonate min-
j.epsl.2004.05.015. eral formation in Holocene sediments of the Baltic Sea: Chemical Geol-
Maynard, J.B., 1991, Iron: Syngenetic deposition controlled by the evolving ogy, v. 135, p. 55–73, doi: 10.1016/S0009-2541(96)00104-0.
ocean-atmosphere system, in Force, E., Eidel, J.J., and Maynard, J.B., Strauss, H., 1997, The isotopic composition of sedimentary sulfur through time:
eds., 1991, Sedimentary and diagenetic mineral deposits: A basin analysis Palaeogeography, Palaeoclimatology, Palaeoecology, v. 132, p. 97–118,
approach to exploration: El Paso, Texas, Society of Economic Geologists, doi: 10.1016/S0031-0182(97)00067-9.
Reviews in Economic Geology, v. 5, p. 141–145. Studley, S.A., Ripley, E.M., Elswick, E.R., Dorais, M.J., Fong, J., Finkelstein,
Maynard, J.B., 1992, Chemistry of modern soils as a guide to interpreting Pre- D., and Pratt, L.M., 2002, Analysis of sulfides in whole rock matrices
cambrian paleosols: Journal of Geology, v. 100, p. 279–289. by elemental analyzer-continuous flow isotope ratio mass spectrom-
Maynard, J.B., 2004, Manganiferous sediments, rocks, and ores, in McKenzie, etry: Chemical Geology, v. 192, p. 141–148, doi: 10.1016/S0009-
F.T., ed., Sediments, diagenesis, and sedimentary rocks: Treatise on Geo- 2541(02)00162-6.
Tang, S.-Y., and Liu, T.-B., 1999, Origin of the early Sinian Minle manganese Wheat, C.G., Mottl, M.J., and Rudnicki, M., 2002, Trace element and REE
deposit, Hunan Province, China: Ore Geology Reviews, v. 15, p. 71–78, composition of a low-temperature ridge-flank hydrothermal spring: Geo-
doi: 10.1016/S0169-1368(99)00015-3. chimica et Cosmochimica Acta, v. 66, p. 3693–3705, doi: 10.1016/S0016-
Tsikos, H., 1999, Petrographic and geochemical constraints on the origin and 7037(02)00894-3.
post-depositional history of the Hotazel iron-manganese deposits, Kala- Yang, J.-D., Sun, W.-G., Wang, Z.-Z., Xue, Y.-S., and Tao, X.-C., 1999, Varia-
hari Manganese Field, South Africa [Ph.D. dissertation]: Grahamstown, tions in Sr and C isotopes and Ce anomalies in successions from China:
South Africa, Rhodes University, 217 p. Evidence for the oxygenation of Neoproterozoic seawater?: Precambrian
Urban, H., Stribny, B., and Lippolt, H.J., 1992, Iron and manganese deposits of Research, v. 93, p. 215–233, doi: 10.1016/S0301-9268(98)00092-8.
the Urucum district, Mato Grosso do Sul, Brazil: Economic Geology and Yeo, G.M., 1981, The Late Proterozoic Rapitan glaciation in the northern Cor-
the Bulletin of the Society of Economic Geologists, v. 87, p. 1375–1392. dillera: Geological Survey Canada, Paper 81-10, p. 25- 46.
Usui, A., and Someya, M., 1997, Distribution and composition of marine Xu, X., Huang, H., and Liu, B., 1990, Manganese deposits of the Proterozoic
hydrogenetic and hydrothermal manganese deposits in the northwest Datangpo Formation, South China: genesis and palaeogeography, in Par-
Pacific, in Nicholson, K., Hein, J.R., Bühn, B., and Dasgupta, S., eds., nell, J., Ye, L.-J., and Chen, C.-M., eds., Sediment hosted mineral depos-
Manganese mineralization: Geochemistry and mineralogy of terrestrial its: Boston, Blackwell Scientific, International Association of Sedimen-
and marine deposits: Geological Society [London] Special Publication tologists Special Publication 11, p. 39–50.
119, p. 177-198.
Wang, J., and Li, Z.-X., 2003, History of Neoproterozoic rift basins in South
China: Implications for Rodinia breakup: Precambrian Research, v. 122,
p. 141–158, doi: 10.1016/S0301-9268(02)00209-7.
Warning, B., and Brumsack, H.-J., 2000, Trace metal signatures of eastern
Mediterranean sapropels: Palaeogeography, Palaeoclimatology, Palaeo-
ecology, v. 158, p. 293–309, doi: 10.1016/S0031-0182(00)00055-9. MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA


Memoir 198
2006
An evaluation of diagenetic recycling as a source of iron

for banded iron formations
R. Raiswell*
School of Earth and Environment, University of Leeds, Leeds LS2 9JT, UK
ABSTRACT
REE and Nd isotope data indicate that most of the iron in banded iron formations
is derived from hydrothermal sources but do not exclude a significant contribution
from terrestrial sources, such as diagenetic recycling. A diagenetic model has been
used to estimate the recycling of iron into overlying seawater, due to microbial reduc-
tion and dissolution at depth in anoxic sediment pore waters, followed by diffusion
upward through a surface layer of sediment that contains oxygenated pore waters.
Rates of iron recycling increase with higher pore-water dissolved iron concentrations,
decreasing pH and temperature, and smaller thicknesses of the surface oxygenated
layer. Iron can be recycled at rates of 1000–5000 μg cm–2 yr–1 from Proterozoic (pO2
= PAL) pore waters with dissolved Fe2+ = 1–5 μg cm–3, pH 6.5 (and T < 65 °C), or pH
7.0 (and T < 40 °C), or pH 7.5 (and T < 20 °C), provided the thickness of the surface
oxygenated layer is less than 0.1 cm. Lower pO2 levels and more weakly oxygenated
surface layers do not significantly increase the maximum recycling rates but enable
these to be achieved at larger thicknesses of the surface layer, for all pH 6.5–7.5 and
temperatures from 10 to 65 °C. Rates of iron supplied by diagenetic recycling can be
substantially modified by the export efficiency (ε) from the source area and by the
ratio (Source Area)/(Sink Area), which can either disperse or concentrate the recycling
flux that is delivered to a sink area of banded iron formation. Banded iron formations
that require maximum iron delivery rates of 22500 μg cm–2 yr–1 can be produced only
by recycling rates of 5000 μg cm–2 yr–1 (and ε = 1) from a source area that is at least
four times larger than the area of banded iron formation. Modern basins have ratios
of shelf area (<200 m water depth) to deep basin area that commonly range from 0.2
to 4. Basins at either extreme have ratios of (Deep Basin Area)/(Shelf Area) or (Shelf
Area)/(Deep Basin Area) that exceed 4 and are potentially able to concentrate iron
either from a deep basin source area to banded iron formation on the shelf, or from a
shelf source area to a banded iron formation depositing in the deep basin. However,
these mass balance constraints require the existence of substantial areas of contempo-
raneous source sediments (or smaller areas of iron-enriched sediments) located either
on the shelf or in the deep basin.
Keywords: iron, diagenetic, recycling, sediment enrichment.
*[email protected]
Raiswell, R., 2006, An evaluation of diagenetic recycling as a source of iron for banded iron formations, in Kesler, S.E., and Ohmoto, H., eds., Evolution
of Early Earth’s Atmosphere, Hydrosphere, and Biosphere—Constraints from Ore Deposits: Geological Society of America Memoir 198, p. 223–238, doi:
10.1130/2006.1198(13). For permission to copy, contact [email protected]. ©2006 Geological Society of America. All rights reserved.
223
224 R. Raiswell
INTRODUCTION Barley et al. (1997), assuming that organic shales and sequence
boundaries represent periods of low to negligible deposition,
There is a widespread consensus that Superior-type iron respectively.
formations occurred in an ocean whose bottom waters con- This paper will be mainly concerned with using models of
tained significant concentrations of dissolved iron. Three major diagenetic mobilization to estimate iron fluxes and assess their
sources for the iron have generally been proposed (e.g., Holland, significance in relation to the maximum rates of iron delivery
1984; Morris, 1993): chemical erosion of the continents, diage- required to form the banded iron formations, but will not be
netic mobilization from sediments, and hydrothermal activity. directly concerned with the mechanisms of iron and silica trans-
Hydrothermal input to the oceans is currently the favored source port and precipitation, the nature of the depositional environ-
for the majority of the iron because of the strong similarities ment, or post-depositional alteration (see for example Trendall
in the rare earth element (REE) patterns of iron formations and Blockley, 1970; Ewers and Morris, 1981; Morris, 1993;
and seawater-dominated hydrothermal fluids (e.g., Morris and Krapez et al., 2003; Pickard et al., 2004). A brief consideration is
Horowitz, 1983; Klein and Beukes, 1989; Derry and Jacobsen, given to riverine fluxes of dissolved and particulate iron but other
1990; Beukes and Klein, 1992; Danielson et al., 1992; Bau and potential iron sources are considered beyond the present scope.
Moller, 1993). The Nd isotope studies also indicate a hydro-
thermal source for a large proportion of the iron (Jacobsen and PREVIOUS WORK
Pimental-Klose, 1988), although the isotope signatures indicate
perhaps as much as 50% may be derived from the continents Diagenetic mobilization has previously been invoked by
(Miller and O’Nions, 1985; Alibert and McCulloch, 1993). How- Borchert (1960), Holland (1973, 1984), and Drever (1974), but
ever, a hydrothermal source for the REE in seawater does not Holland (1973, 1984) provides the most mechanistic detail.
necessarily indicate that most Fe was hydrothermally sourced. Iron and manganese oxides undergo microbial reduction dur-
Holland (1984) has estimated the possible contributions ing diagenesis
from continental weathering and diagenetic mobilization and
concluded that continental weathering was unlikely to have CH2O + 4FeOOH + 8H+——-→ 4Fe 2+ + CO2 + 7H2O
made a significant contribution, because a relatively large river-
ine discharge is required and it would be difficult to achieve the CH2O + 2MnO2 + 3H+——-→ 2Mn2+ + HCO3- + 2H2O
large-scale separation of dissolved iron from the accompanying
sediment load (as is required to produce iron formations that releasing dissolved metals to the pore waters. Low concentrations
have a low terrigenous content). However, diagenetic mobiliza- of organic C give rise to suboxic diagenesis, where the effects of
tion was considered to be a potential source. It is the purpose sulfate reduction are diminished and dissolved iron accumulates
of this paper to estimate the possible iron contributions from in the pore waters rather than being precipitated as iron sulfides.
diagenetic mobilization, in the light of recent work on reactive Holland (1984) used data from Aller and Benninger (1981) to
iron supply to modern and ancient euxinic sediments (Wijsman show that rates of Mn release from pore waters ranged up to 850
et al., 2001; Anderson and Raiswell, 2004; Raiswell and Ander- μg cm–2 yr–1. He suggested that the fluxes of iron would be com-
son, 2005). parable and possibly larger, but that relatively little of either metal
Previous estimates of the iron supply to iron formations would reach the overlying water because of oxidation and precip-
have generally been based on a time frame supplied by the Dales itation as oxides in the uppermost oxygenated layer of sediment.
Gorge Member of the Brockman Iron Formation (Hamersley However, this limitation might be less important if bottom waters
Basin). The Dales Gorge Member contains alternations of iron- were depleted in O2, as would be the case if pO2 was less than
rich and silica-rich bands at three scales, termed macro-, meso-, the present atmospheric level (PAL). Holland (1984) also pointed
and microbands. The iron-rich sediments are believed to reflect out that the diagenetic mobilization of iron would be aided by a
relatively high levels of hydrothermal activity that are imposed sulfate-depleted ocean that minimized the precipitation of pore-
on the more silica-rich conditions resulting from periods of rel- water dissolved iron as iron sulfides, as a result of microbial sul-
atively low hydrothermal input (Morris, 1993). The mixing of fate reduction. There thus appear to be two main difficulties in
iron-rich deep waters with silica-rich surface waters occurs on relation to the supply of pore-water dissolved iron by diagenetic
variety of time scales, and the laterally extensive microbands mobilization, relating to the extent of removal by the formation
have been interpreted as annual varves (Trendall and Block- and precipitation of iron oxides in surface oxygenated layers of
ley, 1970; Ewers and Morris, 1981; Morris, 1993), indicating sediment and the competing effects of iron sulfide precipitation
sedimentation rates of 20–230 m Myr–1 that are also consistent due to sulfate reduction at depth in anoxic sediments.
with geochronological evidence (Trendall, 2000). These varves Recent studies of the Black Sea and the Cariaco Basin
provide a time control that allows the annual flux of iron to the (Canfield et al. 1996; Raiswell and Canfield 1998; Wijsman et
sediments to be estimated as 22.5 mg cm–2 yr–1 (Trendall and al. 2001; Lyons et al. 2003) have shown that their deep basin
Blockley, 1970). Faster rates of deposition (100–1000 m Myr–1) sediments, deposited under euxinic bottom waters, are enriched
have been estimated for Hamersley banded iron formations by in iron that is highly reactive toward sulfide (FeHR) relative to
An evaluation of diagenetic recycling as a source of iron for banded iron formations 225
oxic continental margin and deep-sea sediments. These studies to describe the effects of diffusion, advection, adsorption, and
suggested that highly reactive iron enrichment (recognized by a reaction on variations in solute concentration with time. Bou-
ratio of FeHR to total Fe or FeHR /FeT > 0.4; see Raiswell and Can- dreau and Scott (1978) have modified this equation to describe
field, 1998) resulted at least partly (and possibly entirely) from the flux of dissolved Mn2+ from a reduced pore water through an
the diagenetic mobilization of iron from basin margin sediments oxygenated layer of sediment to the sediment-water interface:
into overlying seawater, followed by lateral transport from the
∂C
basin margin to the deep basin (e.g., Canfield et al. 1996; Lyons
∂(ϕ C) ∂(ϕ Ds ∂x ) ∂(ϕ vC)
1997; Raiswell and Canfield 1998; Wijsman et al. 2001). Wijs- = − + ϕR (1)
man et al. (2001) also demonstrated that the extent of highly ∂t ∂x ∂x
reactive iron enrichment in Black Sea deep basin sediments
was within the range of iron release from oxic and dysoxic con- where Ds (cm2 sec–1) = the diffusion coefficient corrected for
tinental shelf sediments (150–450 μg cm–2 yr–1), as estimated tortuosity effects and C = the pore-water solute concentration
from a model of diagenetic iron recycling and in situ measure- (g cm–3), ϕ = porosity, t = time (sec–1), v = velocity of pore water
ments of iron fluxes. They proposed that highly reactive iron relative to the sediment-water interface (cm sec–1), x = distance
is first solubilized from iron oxides during anoxic diagenesis below the sediment-water interface (cm), and R = rate of oxida-
of the shelf sediments, and then released as dissolved Fe2+to tion in the oxygenated layer of sediment (g cm–3 sec–1). Equation
the overlying seawater. They also proposed that a significant (1) can be simplified by assuming that the advective velocity of
fraction of the dissolved Fe is reoxidized and redeposited into the porewater, the diffusion coefficient, and the porosity are inde-
the shelf sediments, but that some escapes and is transferred pendent of time. Boudreau and Scott (1978) also assume that the
to the deep basin where the dissolved Fe is precipitated as iron advective term is small relative to the diffusion and reaction
sulfide and deposited (so producing the observed enrichment). terms and that the reaction term can be expressed as:
However, Anderson and Raiswell (2004) suggested that the iron
oxides formed by the oxidation of Fe2+ after release to seawater R = –k1 (C – Cs) (2)
may not all be redeposited into the shelf sediments but may also
be transported to the deep basin sediments. where Cs is the saturation concentration of Fe (III) (g cm–3) with
Raiswell and Anderson (2005) have used a simple diage- respect to iron oxides and k1 is the first-order rate constant (sec–1)
netic model to show that rates of iron release from typical conti- for the oxidation of Fe(II) and precipitation of Fe(III) oxides.
nental margin sediments can range up to 1000 μg cm–2 yr–1 (for The saturation concentration of iron with respect to iron oxides
pore waters containing 1 μg cm–3 dissolved iron) but are cru- is assumed to be negligible, which is consistent with the solubili-
cially dependent on pore-water dissolved iron concentrations and ties of Fe(OH)3 (0.7 × 10−9 g cm–3), goethite (0.2 × 10−9 g cm–3),
the thickness of the surface oxygenated layer. Furthermore the and hematite (0.1 × 10−9 g cm–3) in seawater (de Baar and de
magnitude of the fluxes generated from clastic continental mar- Jong, 2001). Furthermore the concentration of Fe(III) in seawa-
gin sediments is significantly modified by the export efficiency ter is believed to be <0.5 × 10−9 g cm–3 in the absence of organic
(ε) of recycled iron from the source area, and by the ratio of the complexes although estimates are complicated by the formation
source sediment area to the basin sink area. Estimates of ε from of hydroxy complexes (Mackey et al., 2002). Concentrations
the Black Sea average at ~50% and the ratio (Source Area)/(Sink increase by approximately one order of magnitude for each unit
Area) is ~0.27/0.73 (Anderson and Raiswell, 2004), where the decrease in seawater pH (Waite, 2001), but values of Cs are still
source area is defined as those sediments at less than 200 m small enough to be ignored provided CP >> Cs (as will normally
water depth. Bottom waters at >200 m depth are sulfidic and iron be the case where pore waters are anoxic and the overlying sea-
released during diagenesis is immobilized by precipitation as sul- water is in equilibrium with pO2 values that are able to stabilize
fides. In the Black Sea, the ratio (Source Area)/(Sink Area) acts to iron oxides). For steady-state diagenesis and assuming that iron
disperse the flux of recycled iron, but other modern basins have is removed only by oxidation and precipitation of iron oxides
potential ratios of (Source Area)/(Sink Area) that can concentrate (see below), equation (1) then simplifies to:
recycled iron. The study by Raiswell and Anderson (2005) dem-
d 2C
onstrates that (1) the flux of iron recycled from clastic sediments Ds − k1 (C − Cs ) = 0 (3)
during diagenesis may be significant under favorable diagenetic dx 2
conditions, and (2) the ratio of the (Source Area)/(Sink Area) can This model (Fig. 1) envisages that transport processes main-
significantly enhance or diminish the rates of iron delivered to the tain a negligible concentration of Fe2+ above the sediment-water
sink area sediments. interface, below which there exists a thin zone of surface sedi-
ment where solid-phase iron oxides precipitate. Porewaters in the
BENTHIC FLUX MODEL surface zone are here termed oxygenated whether oxic (>2 ml
O2 liter–1), dysoxic (0.2–2 ml O2 liter–1), or suboxic (0–0.2 ml O2
A general one-dimensional diagenetic equation (Berner, liter–1), according to the classification of Wignall (1994). There is
1980; Boudreau, 1996) for a solute in pore water can be written a gradient in dissolved Fe2+ through this zone from the sediment-
226 R. Raiswell
d[FeII]
Seawater = − k[FeII][O 2 ][OH - ]2 (6)
dt
Oxygenated Layer
Sediment (thickness L cm)
Porosity ϕ
Millero et al. (1987) has the derived the following expression
Porewater for k:
Fe 2+(Cp)
log k = 21.56 − 1545/T − 3.29I0.5 + 1.52I
Figure 1. Schematic model for the benthic flux of recycled iron from
sediments to the overlying seawater. where T is temperature in degrees kelvin and I is the ionic
strength. Table 1 shows values of log k from 10 to 65 °C, assum-
ing I = 0.723 for Proterozoic seawater. Equation (6) can be modi-
fied to derive an apparent first-order rate constant k1,
water interface to the base of the iron oxide layer. Porewaters
below the surface oxygenated layer are reducing and maintain where
a uniform Fe2+ concentration with time. These assumptions pro- d[FeII]
= k1[FeII]
duce the boundary conditions that C = 0 at x = 0 and C = CP at x dt
= L. The benthic flux of Fe 2+ (Boudreau and Scott, 1978; Bou- and
dreau, 1999) is then given by:
ϕ (Ds k1 )0.5C p k1 = k [O2] [OH-]2. (7)
-3 -1
Flux(g cm sec )=
⎡⎛ k ⎞ 0.5
⎤ Values of k1 (sec–1) are given in Table 1 for the appropriate oxy-
sinh ⎢⎜ 1 ⎟ L⎥ (4)
⎢ ⎝ Ds ⎠ ⎥ gen concentrations (Benson and Krause, 1984) to achieve satu-
⎣ ⎦ ration at the specified temperature, and for varying pH (using
the apparent Kw for seawater from Millero, 2001). The results
where CP is the pore-water concentration of Fe2+ (g cm–3) and L in Table 1 show that k1 for shelf sediments increases by about
(cm) is the depth of the oxygenated layer. four orders of magnitude as T increases from 10 to 65 °C, and
Raiswell and Anderson (2005) have adapted this model to by two orders of magnitude for pH 6.5–7.5; furthermore k1 also
estimate the diagenetic mobilization of iron from clastic sedi- varies directly with O2 concentrations. Equation (4) also shows
ments. Following their approach, the infinite dilution diffusion that fluxes of Fe2+ from shelf sediments depend on pore-water
coefficient (Do) for Fe2+ is derived from Boudreau (1996) as Fe2+ concentrations (Cp) and the thickness of the oxygenated
Do (cm2 sec–1) = (3.31 + 0.15 T °C) 10−6. The corrections for layer (L) in addition to k1. The values of pH, temperature, and CP
tortuosity are derived from Ullman and Aller (1982), who give used in the models will be assumed to be those occurring in the
Ds = Do/ϕ F, where ϕ is the porosity (here assumed to be 0.85) pore waters below the surface oxygenated layer, following the
and the formation factor, F, is approximated as 1/ϕm and m = approach of Boudreau and Scott (1978). Some reasonable limits
2.5–3.0 for muddy sediments. Simplifying thus approximately are derived below for the temperature and composition of the
produces: overlying Proterozoic seawater.
Ds = Doϕ1.7 PALEO–MESOPROTEROZOIC SEAWATER

CHARACTERISTICS
and hence:
Most Superior-type iron formations occur over a time range
Ds = ϕ1.7 (3.31 + 0.15 T°C) 10−6. (5) that extends from ca. 2.7 to 1.8 Ga (Paleo–Mesoproterozoic)
although both older and younger examples are known (Isley,
The oxidation of Fe2+ can occur chemically or be biologi- 1995). The iron formations in the Hamersley Basin (from which
cally catalyzed. Chemical oxidation can be inhibited by the rates of iron delivery have been estimated; see earlier) have a
presence of humics and other natural organic species (Theis rather smaller age range (ca. 2.6–2.4 Ga). The values of seawater
and Singer, 1974) but is very rapid and is believed to predomi- temperature and pH over the Paleo–Mesoproterozoic have been
nate over biological oxidation (Nealson, 1997; Santschi et al., the subject of considerable debate and both these parameters are,
1990). The kinetics of Fe2+oxidation in seawater have been stud- directly or indirectly, dependent on atmospheric composition.
ied by many workers (e.g., Stumm and Morgan, 1980; Davison The models of diagenetic iron mobilization additionally require
and Seed, 1983; Roekens and van Grieken,1983; Millero et al., an estimate of pore-water Fe2+ concentrations (Cp). However, sul-
1987) and there is general agreement that the rate law has the fate reduction exerts an important influence on pore-water dis-
following form: solved iron profiles in modern marine sediments (e.g., Canfield
TABLE 1. VARIATION IN THE APPARENT FIRST ORDER RATE flux model. This influence arises because low sulfate concentra-
CONSTANT (k1) FOR FEII OXIDATION IN SEAWATER WITH
TEMPERATURE AND pH, ASSUMING SATURATION tions limit the extent of sulfate reduced and fixed as iron sulfides
WITH RESPECT TO THE ATMOSPHERE (the low solubility products of which exert an important control
Temp log k pKw O2 PM k1 in sec–1 at varying pH
on pore-water dissolved Fe2+ profiles). Hence the following dis-
(°C) pH 6.5 pH 7.0 pH 7.5
10 14.4 13.8 275 2.9 × 10–6 2.9 × 10–5 2.9 × 10–4 cussion is confined to establishing the range in Paleo–Mesopro-
terozoic seawater sulfate estimates. These data will subsequently
–5 –4
20 14.59 13.4 226 2.3 × 10 2.3 × 10 2.3 × 10–3
30 14.76 13.0 191 1.8 × 10–4 1.8 × 10–3 1.8 × 10–2
40 14.92 12.7 164 9.0 × 10–4 9.0 × 10–3 9.0 × 10–2 be used to select appropriate modern sediments from which pore-
65 15.3 12.0 31.7 1.1 × 10–2 1.1 × 10–1 1.1 water dissolved Fe2+ concentration ranges can be derived.
Note: Data for pKw from Millero (2001), log k from Millero et al. (1987) with
oxygen saturation concentrations from Benson and Krause (1984). The debate on the concentration levels of early seawater sul-
fate has mainly focused on the Archean (older than 2.5 Ga), and
evidence for sulfate levels is usually drawn from δ34S data on
and Raiswell, 1991), whereas Archean and Paleoproterozoic sea- sedimentary sulfides although other approaches can provide use-
water may have been depleted in sulfate relative to modern sea- ful constraints. For example, Walker and Brimblecombe (1985)
water (see later). Hence the evidence for seawater sulfate values have argued that sulfate concentrations up to ~1 mmol can be
over this time are also here briefly reviewed. There is a substantial formed by the oxidation of volcanic SO2 and H2S even in the
literature relevant to defining the characteristics of Archean and absence of atmospheric oxygen.
Proterozoic seawater but this review has the limited objective of The isotopic characteristics of sedimentary sulfides of
defining the most appropriate ranges in these characteristics and Archean age show relatively limited variations in δ34S, and only
is thus selective. small depletions in 34S (~0 ± 10‰). These isotopic compositions
are also only slightly depleted in 34S relative to Archean seawater
Paleo–Mesoproterozoic Seawater Sulfate Concentrations sulfate (probably +2‰ to +3‰; Ohmoto and Felder, 1987; Can-
field et al., 2000). These characteristics have been interpreted by
There are currently conflicting views as to early ocean sul- Ohmoto and Felder (1987) as arising from rapid sulfate reduction
fate concentrations, depending on whether the atmosphere is in a sulfate-rich ocean containing >1 mmol (Ohmoto and Felder,
regarded as oxygen-rich (pO2 0.5–1.0 PAL) or oxygen-depleted 1987) or ~10 mmol (Ohmoto et al., 1993) sulfate. It was argued
(pO2 < 0.01 PAL). These views on early Earth pO2 values are that rapid sulfate reduction (1) produces an instantaneous sulfide
summarized in Ohmoto (1997) and Holland (1999) and the refer- showing a small kinetic fractionation (0‰–5‰) in favor of 32S,
ences therein provide further detail. Prior to 2.3 Ga the Cloud- and (2) favors closed system behavior, which produces a cumu-
Walker-Kasting-Holland (CWKH) model proposes that oxygen lative sulfide product that is isotopically similar to the initial
was absent or only present at very low concentrations ranging sulfate (provided sulfate reduction is rapid enough to minimize
from 10−6 atm (Anbar and Holland, 1992) to 10–2.5 atm (Holland, sulfate resupply from overlying, sulfate-rich seawater). A warm
1984; Towe, 1996). Concentrations then rose rapidly to a pO2 of Archean ocean (30–50 °C) was invoked to produce rapid closed
~10−3 atm (Kasting, 1987) or 10–1.5 atm (Holland, 1984; Lenton, system sulfate reduction with small kinetic fractionations, and
2003) by 2.25–2.05 Ga. The alternative Dimroth-Kimberley- hence sulfides with δ34S values similar to those of contempora-
Ohmoto (DKO) model maintains that pO2 has remained at PAL neous sulfate.
± 0.5 since 4 Ga. The same isotopic characteristics of Archean sulfides have,
These models directly influence the debate on early seawa- however, been interpreted by Canfield and Teske (1996) and
ter sulfate because atmospheric O2 causes the oxidative weath- Habicht and Canfield (1996) as resulting from small kinetic
ering of sulfides to sulfate, which is transported by rivers into fractionations in favor of 32S that result from reduction in a sul-
the oceans. Numerous studies have demonstrated the effects of fate-depleted (<1 mmol) ocean. Canfield and Teske (1996) and
oxygen and ferric iron (whose stability also depends on pO2) on Canfield et al. (2000) have shown that sulfate reduction rate
pyrite oxidation, and empirical rate laws have been formulated exerts little effect on the kinetic fractionation factor and that high
by McKibben and Barnes (1986) and Williamson and Rimstidt sulfate reduction rates are associated with large δ34S depletions
(1994). Canfield et al. (2000) have used the rate law of William- (>20‰) in microbial mats and modern marine sediments (Habi-
son and Rimstidt (1994) to estimate that the oxidation of pyrite cht and Canfield, 1996, 2001). Furthermore, temperatures up to
to dissolved sulfate occurs at geologically significant rates even 40–45 °C produce significant fractionations (10%–20‰) with
with pO2 = 0.04 PAL. Thus low concentrations of seawater sul- Desulphovibrio desulphuricans (Kaplan and Rittenberg, 1964;
fate (compared to present-day) would indicate the existence of Chambers and Trudinger, 1979). Consistent with this, mixed
pO2 significantly lower than 0.04 PAL. microbial populations able to function at higher temperatures
It is not intended that this study should contribute to the (up to 85 °C) showed comparable fractionations of 13‰–28‰
atmospheric pO2 debate, as the diagenetic models to be used here (Canfield et al., 2000). A modeling study (Canfield et al., 2000)
do not directly require knowledge of seawater sulfate. However, also indicated that high sulfate reduction rates result in rapid
sulfate concentrations do potentially affect pore-water dissolved depletion of organic matter near the sediment surface and thus
Fe2+ concentrations, which are an important variable in the benthic do not produce closed system behavior. These arguments are
228 R. Raiswell
difficult to reconcile with the views of Ohmoto and Felder (1987) High seawater temperatures of 55–85 °C have also been
and Ohmoto et al. (1993) but there are still significant factors deduced by Knauth and Lowe (1978, 2003) from the δ18O values
to be evaluated. For example, it may be important that sulfate- of Archean cherts from the Swaziland Supergroup. The δ18O val-
reducing bacteria show interspecies variations in sulfur isotope ues of these cherts (+15‰ to + 22‰) are significantly lower than
fractionation (Detmers et al., 2001). However, these studies are their Phanerozoic counterparts (+18‰ to +34‰). This difference
all consistent in suggesting that Paleo–Mesoproterozoic sulfate has hitherto been attributed to formation during late diagenesis or
concentrations were lower than in the present ocean (28 mmol), to regional metamorphic, hydrothermal, or long-term resetting of
although the suggested values range from <200 μmol (Canfield the original values. However, Knauth and Lowe (2003) argued
et al., 2000) to 1–10 mmol (Ohmoto and Felder, 1987; Ohmoto that these explanations are inconsistent with the preservation of
et al., 1993). δ18O values through different metamorphic grades and the sys-
tematic δ18O differences between different types of chert, strati-
Paleo–Mesoproterozoic Seawater Temperature graphic units, and conglomerate clasts. Hence they attributed the
low δ18O values to seawater temperatures of 55–85 °C, rather
Stellar evolution models indicate that solar luminosity higher than have been inferred in the studies summarized above.
increased from ~70% of current values ca. 4.6 Ga to ~85% by Thus the model discussed here uses a range of 10–65 °C.
2.0 Ga (Gough, 1981; Newman and Rood, 1977). This weak
luminosity would be insufficient to prevent the oceans from Paleo–Mesoproterozoic Seawater pH
freezing solid, and enhanced concentrations of greenhouse
gases must therefore have been present to allow the existence of Previous studies (Holland, 1973, 1984; Walker, 1983;
a hydrological cycle (Sagan and Mullen, 1972). The most prob- Grotzinger and Kasting, 1993) have generally concluded that
able greenhouse gas candidates are assumed to be carbon diox- variations in seawater pH since the Archean have been relatively
ide, water vapor, and methane, and several studies (see below) small. These arguments, however, are dependent on assumptions
have concluded that the enhanced atmospheric concentrations of atmospheric pCO2, which are in turn constrained by the need
of carbon dioxide required to maintain a hydrological cycle do for greenhouse gases to maintain a surface temperature that per-
not indicate variations in ocean chemistry that are inconsis- mits the existence of a hydrological cycle (see earlier).
tent with the sedimentary rock record. However, a more direct Walker (1983) examined the dissolved equilibria that control
approach is to use evidence of temperature constraints from the existence of calcite and gypsum in relation to pCO2. The data
Phanerozoic glaciations to infer the climatic conditions dur- showed that higher pCO2 values produced only relatively small
ing the Paleoproterozoic Huronian glaciation, which occurred changes in dissolved carbonate, SO4 2–, and Ca2+, which were not
ca. 2.0–2.5 Ga (Kasting, 1987; Grotzinger and Kasting, 1993). inconsistent with the occurrence of gypsum and calcite in the
Thus sea-surface temperatures at the height of the most recent rock record. On this basis pCO2 values of 1 atm would need to be
glaciation are thought to be ~2 °C (Schneider and Londer, 1984) accompanied by a seawater pH of ~6.2. Grotzinger and Kasting
colder than at present (15 °C), and Kasting (1987) suggested (1993) have also examined the relationship between atmospheric
that 5 °C is probably a generous lower limit for seawater tem- pCO2 and seawater pH, using the approach of Holland (1984).
perature during the Paleoproterozoic. An upper limit to sea- This approach assumes an ocean saturated with respect to cal-
water temperature is also indicated by the absence of polar ice cite and siderite, with a dissolved Fe2+ concentration of 1–3 ppm
during the Mesozoic, when surface temperatures are believed and dissolved Ca 2+ concentrations of 3–9 ppm. The carbonate
to be around 25 °C or ~10 °C warmer than at present. However, equilibria then allow H+ to be expressed as a function of pCO2,
Antarctica did not become glaciated until the early Oligocene, producing a pH range of 7.1–7.4 for pCO2 = 0.03 atm, and pH
when global surface temperatures had declined substantially = 6.6–6.9 for pCO2 = 0.3 atm. Grotzinger and Kasting (1993)
from the Mesozoic. Thus Kasting (1987) and Grotzinger and argued that pCO2 = 0.3 atmosphere is a likely upper limit for the
Kasting (1993) place Paleoproterozoic surface seawater tem- Paleoproterozoic, which equates to a probable pH >6.5.
peratures at 5–20 °C. Seawater pH is clearly sensitive to atmospheric pCO2,
Higher estimates of 30–50 °C have been suggested by which is not well constrained by greenhouse gas requirements.
Ohmoto and Felder (1987) on the basis of their examination The main greenhouse gases are CO2, water vapor, and CH4
of sedimentary sulfides of Archean age (see earlier). Rela- and the same greenhouse conditions could be reached at lower
tively high Proterozoic temperatures (~50 °C) have also been pCO2 levels with higher concentrations of either water vapor
inferred by Schwartzmann et al. (1993) assuming that the first or CH4. Habicht et al., (2002) have recently pointed out that a
appearance of cyanobacteria and eukaryotes in the rock record low-sulfate ocean (<1 mmol) would result in most anaerobic
occurred at their upper temperature limit for viable growth mineralization of organic carbon occurring through methano-
(which thus corresponds to the existing surface temperature). genesis, which is favored when sulfate reduction is suppressed
Schwartzmann et al. (1993) stressed that this approach would (as in present-day lake sediments; Berner and Raiswell, 1983;
require a re-evaluation of the evidence for Paleoproterozoic gla- Davison et al., 1985). Habicht et al. (2002) found a significant
ciation, which is seemingly robust (see earlier). difference between the sulfur isotope fractionations above 200
μmol sulfate (22.6‰ ± 10.3‰) compared with those below 200 lower seawater sulfate concentrations (see above) and thus less
μmol (0.7‰ ± 5.2‰). These data suggest that the low kinetic sulfate reduction.
fractionations found in Archean sulfides are attributable to low The low Paleo–Mesoproterozoic seawater pH arising from
(200 μmol) concentrations of seawater sulfate. This new esti- higher atmospheric pCO2 values makes it unlikely that there
mate of early seawater sulfate is at least five times lower than could be further significant decreases with depth, for two rea-
hitherto suggested (see earlier) and thus permits an enhanced sons. First, higher atmospheric pCO2 produces higher concentra-
role for methanogenesis, and potentially increased fluxes of tions of dissolved carbonate species, thus providing more effec-
methane to the atmosphere. High atmospheric methane concen- tive buffering capacity against further decreases in pH (Gardner,
trations would imply that lower pCO2 levels would be needed to 1973). Second, depth changes in pH in modern sediments arise
maintain greenhouse conditions. Paleosol evidence (Rye et al., mainly through the effects of oxic respiration, iron reduction and
1995) also suggests a methane-rich atmosphere accompanied sulfate reduction (see earlier). Diagenetic models show that the
by a relatively low pCO2 (<0.04 atm), because iron lost from the largest pH decreases (to <7) result when sulfide accumulates in
tops of paleosols is precipitated at depth as silicate rather than solution rather than being precipitated as iron sulfide (Ben Yaa-
carbonate phases. However, the paleosol evidence has recently kov, 1973; Boudreau and Canfield, 1993. However, the effects of
been reassessed by Ohmoto et al. (2004), who concluded that sulfate reduction will be diminished if Paleo–Mesoproterozoic
the data were consistent with an atmosphere containing pCO2 of seawater sulfate was significantly lower than present-day. Less
0.02–0.06 atm and without significant concentrations of meth- sulfate reduction and less production of sulfide makes the accu-
ane. Overall the evidence on pCO2 limits indicates that seawa- mulation of sulfide in excess of reactive iron oxides less likely.
ter pH can be only loosely constrained to between 6.5 and 7.5. Overall, the above studies suggest that pore-water pH will fall
within essentially the same range as the pH of Paleo–Mesopro-
DIFFUSIVE FLUX MODEL: INPUTS AND RESULTS terozoic seawater (6.5–7.5, see above).
Porewater dissolved iron results from the reduction of iron
Model Inputs oxides (e.g., Canfield, 1989) and is closely followed by micro-
bial sulfate reduction in most continental margin sediments (e.g.,
The above discussion has placed some limits on the temper- Canfield and Raiswell, 1991),
ature, pH, and sulfate concentrations in Paleo–Mesoproterozoic
seawater, which are here used as a basis to constrain the vari- 2CH2O + SO42–⎯→ 2HCO3– + H2S
ables k1, Ds, and Cp in equation (4). The first-order rate constant
k1 is defined in terms of [OH-] and thus a temperature and pH with the dissolved sulfide being rapidly precipitated as iron sul-
dependence (see eq. 6) arises in relating [OH-] to the estimates fides. The rapid reaction between dissolved sulfide and dissolved
of pH, because the Kw for seawater is temperature-dependent Fe2+ (plus solid iron oxides) buffers dissolved sulfide at low con-
(see Table 1). The tortuosity-corrected diffusion coefficient Ds is centrations until the reactive iron oxides are exhausted. The pore
also temperature-dependent as described in equation (5). Hence waters are then dominated by dissolved sulfide. Thus a depth
the models used here will attempt to demonstrate the possible profile through modern continental margin sediments typically
effects of relatively high surface temperatures by considering shows a layer of iron-rich pore waters overlying a layer of dis-
a range of 10–65 °C for Paleo–Mesoproterozoic seawater (see solved sulfide-rich pore waters (Canfield and Raiswell, 1991).
above), and pore-water temperatures are assumed to fall within Raiswell and Anderson (2005) have conducted a literature survey
the same range. of pore-water Fe2+ concentrations in anoxic marine shelf sedi-
Literature estimates for the pH of seawater over this time ments, which shows that mean values of 1–5 μg cm–3 (Fig. 2A)
period ranged from 6.5 to 7.5 (see above). By contrast the pH of are obtained by averaging over depths of 10–30 cm. However,
present-day surface seawater is ~8.1 but the observed pH of pore a better analogue for a low-sulfate Paleo–Mesoproterozoic pore
waters in modern marine shelf sediments undergoing anoxic dia- water may be provided by pore waters from modern, non-acidic
genesis is usually between 7 and 8 (e.g., Boudreau and Canfield, lake sediments and Figure 2B shows the range in mean pore-
1988). The decrease in pH is due to the effects of oxic respiration water dissolved Fe2+ concentrations in sediments from lakes with
followed by the combined effects of iron reduction and sulfate sulfate concentrations mostly <200 μmol. Pore-water dissolved
reduction, which have been modeled by Ben Yaakov (1973), iron concentrations in lakes show considerable spatial and sea-
Gardner (1973), and Boudreau and Canfield (1993). The mod- sonal variations, but the range in mean dissolved iron concen-
els are consistent with field data indicating that pH varies only trations is a little larger (1–8 μg cm–3) than in modern marine
between ~7 and 8, with the lower values occurring where greater sediments, and is sustained over greater depths. Holland (1984)
proportions of sulfide accumulate in the pore waters (as opposed and Grotzinger and Kasting (1993) used values of 1–3 μg cm–3
to being precipitated as iron sulfide). An estimation of the pore- for Fe2+ in seawater that is saturated with respect to siderite at pH
water pH of Paleo–Mesoproterozoic clastic marine sediments 6.6–6.9 (pCO2 = 0.3 atm) and pH 7.1–7.4 (pCO2 = 0.03 atm). Sig-
thus needs to account for differences in both the initial pH of sea- nificantly higher dissolved iron concentrations would be difficult
water and the changes with depth as a possible consequence of to maintain against removal by siderite (see Ohmoto et al., 2004)
230 R. Raiswell
10 cation, manganese reduction, iron reduction, sulfate reduction,

and methanogenesis) and the precipitation of iron sulfides. The
8
A present model produces diffusive flux estimates that are in some
cases three times higher than the model outputs from Wijsman
Frequency
6 et al. (2002) and the calculated diffusive fluxes at the sites stud-
ied by McManus et al. (1997) and Elrod et al. (2004). However,
4
diffusive fluxes derived from equation (4) were consistent with
2
benthic flux measurements from Black Sea shelf sediments by
Friedl et al. (1998) and Friedrich et al. (2002) that average 920
0 μg cm–2 yr–1 (but see Raiswell and Anderson, 2005). The sim-
0 1 2 3 4 5 6 7 8 9 10 ple model used here does not reproduce the rapid redox cycling
μg Fe cm -3 found in more detailed models (e.g., Wang and Van Cappellen,
1996) and hence produces only semiquantitative estimates of dif-
8
fusive fluxes that are used here to examine iron recycling in a
7 B basinal context.
6 McManus et al. (1997) and Elrod et al. (2004) also report
Frequency
5 benthic flux measurements of iron from sites on the California

4 continental margin. The data are generally lower than those
3 reported for the Black Sea but show two significant features.
2 First, shallow-water bioirrigated sites (99 m depth) with well-
oxygenated bottom waters (101–185 μmol) have benthic flux
1
measurements in the range 2–22 μg cm–2 yr–1 that are on aver-
0
age 75 times higher than predicted by diffusion flux estimates
1 2 3 4 5 6 7 8 9 10
based on pore-water dissolved iron concentrations (Elrod et al.,
μg Fe cm -3
2004). Raiswell and Anderson (2005) suggest that bioirrigation
Figure 2. (A) Pore-water dissolved iron concentrations in marine sedi- decreases pore-water dissolved Fe by transporting Fe to the over-
ments, derived by averaging concentrations over depths of 10–30 cm lying waters (thus producing low diffusive fluxes but significant
(from Raiswell and Anderson, 2005). (B) Pore-water dissolved iron bioirrigation fluxes). Enhanced fluxes arising from bioirrigation
concentrations in fresh-water sediments, derived by averaging con- would not be possible for Paleo–Mesoproterozoic sediments in
centrations over depths of 10–50 cm. Data from Carignan and Lean
(1991), Emerson (1976), Feijtel et al. (1988), Matisoff et al. (1980), which benthic activity was absent.
Nembrini et al. (1983), Nriagu and Dell (1974), and Weiler (1973). Second, Elrod et al. (2004) show that deeper-water sites with
oxygen-depleted bottom waters (<100 μmol) have benthic fluxes
in the range 0.2–37 μg cm–2 yr–1 that are roughly in agreement
with estimated diffusive fluxes. Exceptions, however, occur for
although modern lake pore waters are able to maintain consider- the highest benthic flux measurements, which are significantly
able over-saturation with respect to siderite without precipitation smaller (by a factor of more than ten) than the estimated diffusive
occurring (e.g., Emerson, 1976). This argument together with fluxes. Elrod et al. (2004) suggest that dissolved Fe was reoxi-
the data in Figure 2 indicates that concentrations of Fe2+ of 1–5 dized in the sediment rather than diffusing into the benthic cham-
μg cm–3 represent a reasonable range for pore waters in Paleo– ber. Alternatively, Raiswell and Anderson (2005) suggest that
Mesoproterozoic pore waters whatever the sulfate concentrations iron was released from the sediment but reoxidized and precipi-
in seawater. Note also that the pH of the lake pore waters used tated from seawater in the benthic chamber without being mea-
in Figure 2B in no case falls below 7, and that the changes in pH sured. In the absence of the chamber such iron might have been
with depth tend to be small unless the overlying water has a pH transported away from the site and some fraction exported from
>7.5. I therefore conclude that pore-water characteristics can rea- the shelf (see later). Overall the benthic flux measurements show
sonably be assumed to approximate the ranges exhibited in the that iron fluxes may be significant but are clearly influenced by
overlying seawater for temperature and pH. processes that are currently poorly quantified. Thus the present
model estimates should be regarded as no more than semiquan-
Model Results titative.
Figure 3 shows the model estimates of benthic recycled
Diffusive fluxes estimated from the model described in equa- iron fluxes as a function of the oxygenated layer thickness (L =
tion (4) have been compared (Raiswell and Anderson, 2005) with 0.1–10 cm). The model predicts that fluxes continue to increase
the flux estimates derived from a more sophisticated diagenetic as L decreases below 0.1 cm. However, values of L <0.1 cm are
model (Wijsman et al., 2002) that takes into account all the main probably geologically unrealistic and also produce fluxes that
carbon mineralization reactions (aerobic respiration, denitrifi- require very rapid rates of iron deposition (see later). The three
graphs in Figure 3 have been drawn for pH values of 6.5, 7.0, (4) shows that using Cp = 5 μg cm–3 simply produces a five times
and 7.5, assuming bottom waters saturated with O2 at the speci- larger flux of recycled iron. Figure 3 shows that the diffusive flux
fied temperature (T = 10, 20, 30, 40, and 65 °C) of each plot. of recycled iron is significantly dependent on L and, to a lesser
The saturation levels of oxygen above 90 μmol (see Table 1) are extent, temperature and pH. The highest fluxes of recycled iron
classified as oxic, whereas a concentration of 31.7 μmol for T = are reached for L = 0.1 cm, but values of ~1000 μg cm–2 yr–1 are
65 °C is dysoxic (see earlier and Wignall, 1994). All three graphs only reached at pH 6.5 for T <65 °C, at pH 7 for T <40 °C, and at
use a concentration of Cp = 10−6 g cm–3 or 1 μg cm–3 but equation pH 7.5 for T <20 °C. The maximum diffusive flux represents the
case where there is diffusion from the pore waters into the over-
lying seawater with little or no oxidation in the surface layer. Fig-
ure 3 shows that the maximum diffusive fluxes are clearly most
1000 sensitive to variations in L for any given Cp, but that increasing
Flux (microgm cm-2 yr-1)
pH 6.5 temperatures play a subsidiary role in promoting faster oxidation,

which becomes more apparent at higher pH.
100 Figure 4 shows similar plots for the fluxes of recycled iron
o
40 C 30 C
o
at lower bottom water concentrations of dissolved O2, assuming
saturation with respect to atmospheric O2 (PAL = 0.1 and 0.01).
10 The O2 saturation levels for PAL = 0.1 are all dysoxic except
o o o
for the T = 65 °C case, where the level is suboxic (see earlier
65 C 20 C 10 C and Wignall, 1994). All O2 concentrations for PAL = 0.01 are
1 suboxic. There is a marginal increase in the maximum diffusive
0.1 1 10 fluxes (~30%–50%) compared to fluxes at pH 6.5 in Figure 3,
L (cm)
1000
1000 o pH 6.5
40 C
pH 7.0 PAL 0.1

100 o
100 30 C
o o o
20 C 10 C 10 C
10
10
o o
65 C 20 C
o o
40 C 30 C 1
1
0.1 1 10
0.1 1 10
L (cm) L (cm)
1000 1000
Flux microgm cm-2yr-1
pH 7.5
o
10 C
100 100
o
o 65 C
10 C
10 o 10 o
20 C 20 C
pH 6.5
o 40oC 30oC
30 C PAL 0.01
1 1
0.1 1 10 0.1 1 10
L (cm) L (cm)
Figure 3. Variations in the benthic flux of recycled iron with oxygen- Figure 4. Variations in the benthic flux of recycled iron with oxygenated
ated zone thickness (L cm) for temperatures of 10, 20, 30, 40, and 65 zone thickness (L cm) for temperatures of 10, 20, 30, 40, and 65 °C at pH
°C at pH 6.5, 7.0, and 7.5, assuming a dissolved iron concentration of 6.5, assuming a dissolved iron concentration of 1 μg cm–3 and saturation
1 μg cm–3 and saturation with the atmosphere at pO2 = PAL. with the atmosphere at pO2 = 0.1 PAL (top) and 0.01PAL (bottom).
232 R. Raiswell
but the principal difference is that fluxes of 1000 μg cm–2 yr–1 are receiving sink sediments, which are modified (Raiswell and
reached at higher values of L (approaching 1 cm.). The effects of Anderson, 2005) by the export efficiency (ε) of iron from the
increased oxidation in the pore waters at higher temperatures are source region and the relative sizes of the source-generating and
overcome by lower bottom water O2 concentrations. sink depositional areas (and by their spatial relationships). The
Jorgensen and Boudreau (2001) show that values of L export efficiency term arises because dissolved iron recycled to
decrease with increasing rates of organic C mineralization (and overlying seawater may be oxidized to iron oxides that are rede-
thus vary seasonally in many modern shelf sediments). Depths posited on the shelf and not exported to the area of banded iron
of oxygen penetration also increase with increasing bioirriga- formation. Raiswell and Anderson (2005) estimated ε to be at
tion. Nonetheless the thickness of the oxygenated zone is often least 30% and possibly 80%–90% for the Black Sea. In view of
only a few millimeters in shelf sediments. Proterozoic sediments this uncertainty it will be assumed here that ε = 1, which pro-
are likely to have relatively low rates of Corg mineralization but duces maximum iron delivery rates.
would be unaffected by bioirrigation as benthic fauna had not
then developed. Values of L of around 0.1 cm may thus be rea- SOURCE AND SINK AREA RELATIONSHIPS
sonable for Paleo–Mesoproterozoic sediments, and the model
then indicates that shelf sediments can recycle iron to the overly- The model estimates of recycled iron are clearly inadequate,
ing seawater at rates of 1000–5000 μg cm–2 yr–1 (for Cp = 1–5 μg on their own, to match the required value of ~22500 μg cm–2 yr–1
cm–3). These rates are achieved at pO2 = PAL for smaller values (based on the assumption that the microbands are annual varves).
of L when the pH lies toward the lower end of the probable range However, the relative sizes of the source generating area and the
for Paleo–Mesoproterozoic seawater pH. With pO2 levels below receiving sink area play a crucial role in enhancing or diminish-
PAL, near-maximum fluxes are reached even when the oxygen- ing the delivery rates of recycled iron. Two different cases may
ated layer is relatively thick, but the maximum values are similar be envisaged, depending on whether the sink area of banded iron
to the case where pO2 = PAL. formation is on the continental shelf (e.g., Morris, 1993) or in
Note that steady-state conditions require that the addition the deeper areas of the basin (Krapez et al., 2003; Pickard et al.,
of reactive iron by sedimentation to the shelf sediments at least 2004). The basic assumptions in both cases are as follows:
equals the loss by diagenetic recycling to the overlying seawa- 1. Atmospheric pO2 values are only loosely constrained but
ter. The mean reactive iron content (FeHR) of modern continen- need to be at least sufficient to stabilize iron oxides. Ther-
tal margin sediments is 0.83‰ ± 0.21% (Raiswell and Canfield, modynamic data indicate that iron oxides are stable at pO2
1998). Sediments with this composition that deposit lithogenous > 10−60 atm, but free O2 molecules disappear at pO2 <10−21
material at rates of at least 0.12 ± 0.03 g cm–2 yr–1 (equivalent atm (Ohmoto et al., 2004). The DKO model has pO2
to 0.3 cm yr–1 for ϕ = 0.85) are potentially capable of supply- >10−1 atm at the start of the Paleoproterozoic whereas the
ing reactive fluxes of 1000 μg cm–2 yr–1 provided all the depos- CWKH model variously estimates pO2 = 10−6 to 10–2.5 atm.
ited reactive iron is recycled. Higher reactive iron fluxes require The value of pO2 >10−6 atm (Fig. 5) used here is arbitrarily
proportionately faster sedimentation (fluxes of 5000 μg cm–2 chosen but is not in contradiction with either the DKO or
yr–1 require sedimentation rates of at least 1.5 cm yr–1) or sedi- the CWKH model and produces weakly suboxic seawater
ments with higher FeHR contents than modern continental margin in which iron oxides are stable.
sediments (see later). Thus the highest rates of diagenetic iron 2. A surface ocean layer in equilibrium with atmospheric pO2
recycling require relatively high shelf sedimentation rates and/or and a deeper Fe2+-bearing water layer that is less-oxygen-
iron-enriched sediments (see later) and will result in the loss of ated (but still weakly suboxic or almost anoxic) in which
significant proportions of reactive iron from the shelf sediment iron oxides are stable.
source. 3. Clastic shelf sediments deposited under steady-state con-
Wijsman et al. (2001) found that the Black Sea shelf sedi- ditions with sedimentation rates and compositions that
ments were depleted in reactive iron relative to mean continental are capable of maintaining rates of iron recycling of up to
margin sediments, but assessing the extent of depletion strictly 5000 μg cm–2 yr–1. Non-steady-state sedimentation effects
requires knowing the composition of the shelf sediments before may be important but are difficult to quantify in the pres-
reactive iron losses have occurred. This may be difficult to ent context and are here excluded. For example, physi-
achieve (see Anderson and Raiswell, 2004). Kump and Holland cal reworking of iron-enriched sediments on the Amazon
(1992) found that Precambrian shales generally exhibited Fe/Ti inner shelf (Aller et al., 1986) suppresses sulfate reduction
ratios that were close to the igneous rock trend and were thus and produces high concentrations of pore-water dissolved
neither Fe-enriched nor -depleted. However, some depletions iron (100 μg cm–3) that are periodically mixed into overly-
were observed relative to the igneous rock trend and these were ing seawater and oxidized to colloidal iron oxides that are
attributed to losses by weathering, transport, or diagenesis. This exported from the shelf.
approach would be worth further study. 4. Porewaters with pH and temperature characteristics that
Rates of iron sourced by diagenetic iron recycling do not, allow maximum iron recycling rates to be achieved. For
however, represent the rates at which iron accumulates in the pO2 = PAL this requires a pH of 6.5 and T of <65 °C, or
Shelf Deposition no more than 200 m) on an isolated platform or shelf on the mar-
Atmospheric pO2 > 10 -6 atm. gin of a continent (Morris and Horowitz, 1983; Morris, 1993),
and that the source area of clastic sediments is located on the
continental shelf or slope and/or in a deep basin area. Part of this
Seawater in equilibrium
with atmospheric pO 2
source area is envisaged to be in contact with the surface oxygen-
ated waters and part in contact with relatively deep Fe2+-bear-
Deposition of Fe
Clastic sediments
Suboxic/anoxic oxides by
ing waters that are suboxic/anoxic. Diagenetic iron is recycled as
recycle Fe2+ to
seawater where
water with Fe 2+ oxidation of particulate iron oxides from the source area either directly into
oxidation occurs. Fe
Fe2+ transported the deep suboxic/anoxic water or indirectly following downslope
on to shelf.
oxides transported Fe oxides reduced to transport (Fig. 5). These deep Fe2+-bearing waters are mixed on
to the deep basin. Fe2+ during diagenesis to the shelf area, where oxidation and deposition of iron oxides
occurs to form banded iron formation. Oxidation may occur inor-
Deep Basin Deposition ganically, microbially, and/or by photo-oxidation (Cairns-Smith,
Atmospheric pO 2 1978; Anbar and Holland, 1992; Konhauser et al., 2002).
> 10-6 atm. In this model, the deep ocean acts essentially as a connecting
reservoir and, assuming a steady state exists, the rates at which
Seawater in equilibrium diagenetically recycled iron is exported from the source area
with pO2 (Jsource × ε) must match the rates at which iron is deposited into
Clastic sediments
the banded iron formation. The rates of diagenetic iron recycling
Suboxic/anoxic water
recycle Fe 2+ to with Fe 2+ from the source area in contact with oxygenated bottom waters
seawater where are a maximum under the conditions defined by assumptions 3
oxidation occurs. Fe
oxides transported
Deposition of and 4. The rates of iron supply from a source area located under
Feoxides
to the deep basin. anoxic bottom waters are not modified by passage through a
surface oxygenated layer of sediment and equation (4) is thus
Figure 5. Schematic models of diagenetic recycling to banded iron
formations depositing in shelf and deeper basin environments. Heavy inappropriate. Under these circumstances the diffusive flux is
dark areas represent source sediments that recycle iron to overlying dependent only on the diffusion coefficient and the concentration
seawater and heavy light areas represent banded iron formations. gradient between the pore waters and the anoxic overlying sea-
water. A concentration difference of ~5 μg cm–3 between seawa-
ter and sediment pore waters <0.1 cm below the sediment-water
interface provides a flux of 5000 μg cm–2 yr–1.
pH 7.0 and T <40 °C, or pH 7.5 and T <20 °C. Lower
The flux of iron to the sink area (Jsink) is given by:
pO2 levels (to 0.01 PAL or less) allow the same rates to be
achieved at all pH values 6.5–7.5 and temperatures from
Jsink = Jsource × ε × (Source Area)/(Sink Area)
10 to 65 °C.
5. Transport of iron oxides can occur from the source areas to
and an iron supply of 5000 μg cm–2 yr–1 therefore requires that the
the depositional area of banded iron formation. Raiswell
product of ε × (Source Area)/(Sink Area) is at least 4 in order to
and Anderson (2005) point out that iron-enriched sedi-
achieve the delivery of 22,500 μg cm–2 yr–1 of iron to the banded
ments (FeHR/FeT > 0.4) commonly occur in modern (the
iron formation. Estimates of (Source Area)/(Sink Area) for the
Cariaco Basin, Baltic Sea, and Black Sea) and ancient
Hamersley Basin can be little more than conjecture, but values of
(e.g., the Jet Rock, the Kimmeridge Clay) euxinic sedi-
ε × (Source Area)/(Sink Area) as high as 5 have been estimated
ments, which suggests that intra-basinal transport is not
by Raiswell and Anderson (2005) for ancient and modern anoxic
unusual although the mechanisms may be difficult to iden-
sediments. However, the area of the Dales Gorge member is at
tify. In the Black Sea, currents at the depths of the suboxic
least 105 km2 (Morris, 1993) and the existence of substantially
zone and the sulfide interface have lateral velocities of
larger areas of clastic shelf sediment is thought to be unlikely
~0.5 cm s–1 (Buessler et al., 1991) and are able to transport
(Morris, 1993).
dissolved and particulate iron from the shelf to the deep
A rather different situation arises where the source area is
basin (Kempe et al., 1990; Wijsman et al., 2001). Trans-
located beneath dissolved iron-bearing water in the deep basin,
port from the shelf to the basin may also involve riverine
and the shelf area represents the sink area of banded iron forma-
delivery (see later) into ocean currents or turbidity cur-
tion. Table 2 shows the B/S ratios for some modern basins where
rents (see Krapez et al., 2003; Pickard et al., 2004), with
S (the shelf area) is defined as the area above 200 m depth (typi-
upwelling generally favored to transport dissolved iron on
cally the limit of surface water mixing) and B (the deeper basin
to the shelf (Ewers and Morris, 1981; Morris, 1993).
area) >200 m depth. The data in Table 2 do not imply that these
The first case (Fig. 5, top) envisages that banded iron forma- modern basins are capable either of exhibiting source-sink char-
tions are deposited below wave base (possibly at water depths of acteristics or of producing iron-enriched sediments. The data also
234 R. Raiswell
TABLE 2. RATIOS OF SHELF AREA (S, <200 m DEPTH) TO DEEP the microbands as varves) suggest that recycling is a potential
BASIN AREA (B, >200 m DEPTH) IN MODERN BASINS iron source provided the Hamersley Basin had values of S/B (for
Enclosed/semi-enclosed Basin area S/B B/S deep basin deposition) or B/S (for shelf deposition) of ~4. Both
basins (× 106 km2)
Gulf of California 0.15 0.89 1.1
of these extremes imply the existence of large areas of contempo-
Baltic Sea 0.39 99 0.01 raneous clastic sediment on the shelf or in the deep basin.
Caspian Sea 0.41 2.3 0.43
Black Sea 0.42 0.37 2.7
Red Sea 0.45 0.72 1.4 RIVERINE IRON FLUXES
Sea of Japan 1.01 0.30 3.4
East China Sea 1.20 4.2 0.23
Hudson Bay 1.23 13 0.08 Holland (1984) has suggested that riverine particulate
Gulf of Mexico
Sea of Okhotsk
1.29
1.39
0.92
0.87
1.1
2.7
and dissolved iron sources might also operate to augment the
Bering Sea 2.26 0.85 1.2 potential supply by diagenetic recycling. Thus Anderson and
Marginal Basins Raiswell (2004) have shown that the diagenetic mobilization
S. Australian Basin 1.53 0.22 4.6 of recycled iron was only responsible for the supply of 40‰
Argentine Basin 4.40 0.23 4.3
± 20% of the reactive iron enrichment in the deep-basin eux-
World Oceans 362 0.08 12
Note: After Raiswell and Anderson (2005). inic sediments of the Black Sea. The remaining enrichment was
attributed to enhanced reactivity of lithogenous iron, possibly
by the fractionation of the riverine particulate flux to produce
an iron oxide-enriched component.
reflect present tectonic, sea-level, and oceanographic conditions Poulton and Raiswell (2002) have shown that the ratio of
and are not necessarily representative of the Proterozoic. A basin highly reactive iron to total iron in riverine particulates (0.43 ±
source area will deliver iron to a shelf area of banded iron forma- 0.03) exceeds that found in modern continental margin and deep
tion at a rate given Jsource × ε × (Basin Source Area)/(Shelf Sink sea sediment (overall average 0.26 ± 0.08). It was suggested
Area), which would produce an iron delivery rate of 22,500 μg that highly reactive iron is at present preferentially trapped in
cm–2 yr–1 from a recycling flux of ~5000 μg cm–2 yr–1 for an area inner shore areas (e.g., estuaries, tidal flats) and a limited data
ratio of ~4. The B/S ratios in modern enclosed, semi-enclosed, set provided some support for this hypothesis. A significant pro-
and marginal basins typically range up to 4, which suggests that portion of the highly reactive iron in riverine particulates occurs
clastic sediments located in the deep basin area have the potential as iron oxide nanoparticulate spheres that adhere to clay miner-
to source most of the iron for banded iron formations. als (Poulton and Raiswell, 2005). Adhesion may be diminished
The second case (Fig. 5, bottom) is based on the deep marine on contact with saline estuarine waters owing to collapse of the
depositional environment inferred by Krapez et al. (2003) and double layer, and spheres that are detached and sedimented will
Pickard et al. (2004). Water depth is known only to be below cause the loss of iron oxides from riverine suspended sediments
wave base, but a turbiditic origin is deduced from the presence entering the marine environment. However, detachment with-
of density current rhythms indicating a depositional environment out sedimentation could produce an increased export of iron
that is distal from a shelf that supplied the clastics. It is useful to oxides to marine sediments. Such iron oxide-enriched sedi-
note that currents in the Black Sea disperse clastic sediment uni- ments might produce higher concentrations of Cp (pore-water
formly over an area comparable in size to the Hamersley Basin, dissolved iron) and corresponding increases in the fluxes of dia-
producing microlaminae that can also be traced for distances of genetically recycled iron (see eq. 4).
more than 1000 km over the basin floor (Lyons, 1991; Lyons and The long-distance transport of riverine suspended sedi-
Berner, 1992). Krapez et al. (2003) and Pickard et al. (2004) offer ment across the shelf to distal environments might also allow
no evidence for the degree of bottom water oxygenation of the sufficient fractionation to produce a component consisting pre-
iron formation depositional environment although the existence dominantly of iron oxide nanoparticulates. Evidence suggests
of iron oxides in the distal area is possible under bottom waters that riverine particulates can undergo long-range transport
that are almost anoxic. A diagenetic supply of iron in the form of into deep basin areas. For example rivers draining the active
iron oxide colloids generated by recycling might be transported margins of islands in the East Indies deliver fine particulates
by turbidity currents to a deep marine depositional environment. into the Equatorial Undercurrent, allowing long-range trans-
Typical S/B ratios for modern basins range up to 4 (although val- port across the Pacific (Gordon et al., 1997; Milliman et al.,
ues <1 are more common), which suggests that a shelf source 1999; Sholkovitz et al., 1999; Mackey et al. 2002). However,
area is also viable for the supply of recycled iron to banded iron this mechanism is problematic because of the large degrees of
formation in a deep basin environment. fractionation required to isolate iron oxides from detrital phases
Better geological constraints, and especially an improved (Holland, 1984).
understanding of the basin configuration and the depositional The flux estimates originating from recycling may also
environment of the banded iron formations, are required to reach be assisted by the riverine delivery of dissolved iron (see ear-
a definitive view on the contributions of iron from diagenetic lier). A rough estimate of the riverine flux into the Hamersley
recycling. However, the iron mass balance constraints (based on Basin can be made assuming that all silica in the microbands
was derived from riverine input, as is suggested by Ge/Si balance constraints indicate that a large area of contemporane-
ratios (Hamade et al., 2003). Garrels (1987) estimated that the ous clastic source sediments (or a smaller area of iron-enriched
microbands required an annual flux of 39 mg cm–2 yr–1, which sediments) would be needed to supply all the iron in banded
would need to be sustained over an area of ~105 km2, giving a iron formations.
total mass of silica of 39 × 1015 mg. The maximum silica con-
tent of rivers is typically around 10–30 ppm (Holland, 1972), ACKNOWLEDGMENTS
which therefore implies a discharge flux of at least 1–3 × 1015
liter yr–1, close to the estimate of 4 × 1015 liter yr–1 by Holland Lee Kump, Ariel Anbar, and Hiroshi Ohmoto supplied
(1984). This confirmation of Holland’s (1984) discharge esti- thoughtful and extremely valuable reviews.
mate supports his conclusion that a high riverine dissolved iron
concentration (~7 ppm) is needed to provide the iron in banded
REFERENCES CITED
iron formations. These dissolved iron concentrations require a
low atmospheric pO2 (<10−60 atm; Ohmoto et al., 2004), which Alibert, C., and McCulloch, M.T., 1993, Rare earth element and neodymium
might have existed only prior to 2.3 Ga (if atmospheric evolu- isotopic compositions of the banded iron-formations and associated
tion followed the CWKH model). Significant riverine dissolved shales from the Hamersley, Western Australia: Geochimica et Cosmochi-
mica Acta, v. 57, p. 187–204, doi: 10.1016/0016-7037(93)90478-F.
iron contributions to banded iron formations are thus unlikely. Aller, R.C., and Benninger, L.K., 1981, Spatial and temporal patterns of dis-
solved ammonium, manganese and silica fluxes from bottom sediments of
CONCLUSIONS Long Island Sound: Journal of Marine Research, v. 39, p. 295–314.
Aller, R.C., Mackin, J.E., and Cox, R.T., Jr., 1986, Diagenetics of Fe and S in
the Amazon inner shelf muds: Apparent dominance of Fe reduction and
A benthic recycling model has been used to estimate the implications for the genesis of ironstones: Continental Shelf Research,
rates at which dissolved iron can diffuse through a surface-oxy- v. 6, p. 269–289.
Anbar, A.D., and Holland, H.D., 1992, The photochemistry of manganese and
genated layer of sediment (where precipitation occurs as iron the origin of banded iron formations: Geochimica et Cosmochimica Acta,
oxides) and into the overlying seawater. The model shows that v. 56, p. 2595–2603, doi: 10.1016/0016-7037(92)90346-K.
rates of recycling are controlled mainly by the thickness of the Anderson, T.F., and Raiswell, R., 2004, Sources and mechanisms for the enrich-
ment of highly reactive iron in euxinic Black Sea sediments: American
surface-oxygenated layer (L) and the pore-water concentration Journal of Science, v. 304, p. 203–233.
of dissolved iron (Cp), and to a lesser extent by pH and tempera- Barley, M.E., Pickard, A.L., and Sylvester, P.J., 1997, Emplacement of a large
ture. Diffusive fluxes of up to 5000 μg cm–2 yr–1 of iron can be igneous province as a possible cause of banded iron formation 2.45 billion
years ago: Nature, v. 385, p. 55–58, doi: 10.1038/385055a0.
recycled to the overlying seawater when L < 0.1 cm and Cp = 5 Bau, M., and Moller, P., 1993, Rare earth element systematics of the chemi-
μg cm–3, for pH 6.5 (and T < 65 °C), or pH 7.0 (and T < 40 °C), cally-precipitated component in Early Precambrian iron formations
or pH 7.5 (and T < 20 °C). Lower atmospheric pO2 (0.1 and 0.01 and the evolution of the terrestrial atmosphere-hydrosphere-lithosphere
system: Geochimica et Cosmochimica Acta, v. 57, p. 2239–2249, doi:
PAL) has little effect on the maximum fluxes of iron recycling, 10.1016/0016-7037(93)90566-F.
although these are achieved at smaller thicknesses of the oxy- Benson, B.B., and Krause, D.K., Jr., 1984, The concentration and isotopic frac-
genated layer for all pH 6.5–7.5 and temperatures 10–65 °C. tionation of oxygen dissolved in freshwater and seawater in equilibrium
with the atmosphere: Limnology and Oceanography, v. 29, p. 620–632.
Recycled iron fluxes of 5000 μg cm–2 yr–1 are inadequate Ben-Yaakov, S., 1973, pH buffering of pore water of recent anoxic marine Sedi-
compared to the rates of 22,500 μg cm–2 yr–1 required for banded ments: Limnology and Oceanography, v. 18, p. 86–94.
iron formation. However, the flux of recycled iron derived Berner, R.A., 1980, Early diagenesis: A theoretical approach: Princeton, New
Jersey, Princeton University Press, 241 p.
from the source sediments may be modified substantially by Berner, R.A., and Raiswell, R., 1983, Burial of organic carbon and pyrite sulfur
two factors: the export efficiency (ε) from the source area and in sediments over Phanerozoic time: A new theory: Geochimica et Cosmo-
the ratio of (Source Area)/(Sink Area). The export efficiency chimica Acta, v. 47, p. 855–862, doi: 10.1016/0016-7037(83)90151-5.
Beukes, N.J., and Klein, C., 1992, Models for iron-formation deposition, in
term accounts for dissolved iron entering oxygenated overlying Schopf, J.W., and Klein, C., eds., The Proterozoic biosphere: A multidis-
waters that is oxidized to iron oxides and redeposited into the ciplinary study: New York, Cambridge University Press, p. 147–152.
source sediments, rather than being exported to banded iron Borchert, H., 1960, Genesis of marine sedimentary iron ores: Transactions of
the Institute of Mining and Metallurgy, v. 69, p. 261–277.
formations. Assuming ε = 1, then the ratio (Source Area)/(Sink Boudreau, B.P., 1996, Diagenetic models and their interpretation: Berlin,
Area) can act either to disperse the recycled flux over a rela- Springer-Verlag, 414 p.
tively large sink area or to focus the flux into a relatively small Boudreau, P., 1999, Metals and models: Diagenetic modeling in freshwater
lacustrine sediments: Journal of Paleolimnology, v. 22, p. 227–251, doi:
sink area. 10.1023/A:1008144029134.
Current depositional models for banded iron formation Boudreau, B.P., and Canfield, D.E., 1988, A provisional diagenetic model
suggest that deposition occurred either on a shelf or in a dis- for pH in anoxic porewaters: Application to the FOAM site: Journal of
Marine Research, v. 46, p. 429–455.
tal slope/deep basin environment. Modern basins have (Shelf Boudreau, B.P., and Canfield, D.E., 1993, A comparison of closed-system and
Area)/(Deep Basin Area) ratios that typically range from 0.2 open-system models for porewater pH and calcite saturation state: Geo-
to 4. Thus basins at either extreme would enable a source shelf chimica et Cosmochimica Acta, v. 57, p. 317–334, doi: 10.1016/0016-
7037(93)90434-X.
supply area to produce the required rates of iron delivery to a Boudreau, B.P., and Scott, M.R., 1978, A model for the diffusion-controlled
deep basin banded iron formation, or a deep basin source area growth of deep-sea manganese nodules: American Journal of Science,
to supply a banded iron formation on the shelf. The iron mass v. 278, p. 903–929.
236 R. Raiswell
Buessler, K.O., Livingston, H.D., and Casso, C., 1991, Mixing between oxic diagenetic pathways in the north-western Black Sea: Estuarine, Coastal
and anoxic waters in the Black Sea as traced by Chernobyl cesium iso- and Shelf Science, v. 54, p. 369–383, doi: 10.1006/ecss.2000.0653.
topes: Deep-Sea Research, v. 38, p. S725–S745. Gardner, L.R., 1973, Chemical models for sulfate reduction in closed anaerobic
Cairns-Smith, A.G., 1978, Precambrian solution photochemistry-inverse seg- marine environments: Geochimica et Cosmochimica Acta, v. 37, p. 53–
regation and banded iron formations: Nature, v. 276, p. 807–808, doi: 68, doi: 10.1016/0016-7037(73)90243-3.
10.1038/276807a0. Garrels, R.M., 1987, A model for the deposition of the micro-banded Precam-
Canfield, D.E., 1989, Reactive iron in marine sediments: Geochimica et Cosmo- brian iron-formations: American Journal of Science, v. 287, p. 81–106.
chimica Acta, v. 53, p. 619–632, doi: 10.1016/0016-7037(89)90005-7. Gordon, R.M., Coale, K.H., and Johnson, K.S., 1997, Iron distributions in
Canfield, D.E., and Raiswell, R., 1991, Carbonate precipitation and dissolu- the Equatorial Pacific: Implications for new production: Limnology and
tion: Its relevance to fossil preservation, in Allison, P.A., and Briggs, Oceanography, v. 42, p. 419–431.
D.E.G., eds., Taphonomy: Releasing the data locked in the fossil record: Gough, D.O., 1981, Solar interior structure and luminosity variations: Solar
New York, Plenum Press, p. 412–453. Physics, v. 74, p. 21–34, doi: 10.1007/BF00151270.
Canfield, D.E., and Teske, A., 1996, Late Proterozoic rise in atmospheric oxy- Grotzinger, J.P., and Kasting, J.F., 1993, New constraints on Precambrian ocean
gen concentration inferred from phylogenetic and sulphur-isotope stud- chemistry: Journal of Geology, v. 101, p. 235–243.
ies: Nature, v. 382, p. 127–132, doi: 10.1038/382127a0. Habicht, K.S., and Canfield, D.E., 1996, Sulphur isotope fractionation in mod-
Canfield, D.E., Lyons, T.W., and Raiswell, R., 1996, A model for iron deposi- ern microbial mats and the evolution of the sulphur cycle: Nature, v. 382,
tion to euxinic Black Sea sediments: American Journal of Science, v. 296, p. 342–343, doi: 10.1038/382342a0.
p. 818–834. Habicht, K.S., and Canfield, D.E., 2001, Isotope fractionation by sulfate-
Canfield, D.E., Habicht, K.S., and Thamdrup, B., 2000, The Archean sulfur reducing natural populations and the isotopic composition of sulfide
cycle and the early history of atmospheric O2: Science, v. 288, p. 658– in marine sediments: Geology, v. 29, p. 555–558, doi: 10.1130/0091-
661, doi: 10.1126/science.288.5466.658. 7613(2001)029<0555:IFBSRN>2.0.CO;2.
Carignan, R., and Lean, D.R.S., 1991, Regeneration of dissolved substances in Habicht, K.S., Gade, M., Thamdrup, B., Borg, P., and Canfield, D.E., 2002, Cal-
a seasonally anoxic lake: The relative importance of processes occurring ibration of sulfate levels in the Archean ocean: Science, v. 298, p. 2372–
in the water column and in the sediments: Limnology and Oceanography, 2374, doi: 10.1126/science.1078265.
v. 36, p. 683–707. Hamade, T., Konhauser, K.O., Raiswell, R., Goldsmith, S., and Morris,
Chambers, L.A., and Trudinger, P.A., 1979, Microbiological fractionation of R.C., 2003, Using Ge/Si ratios to decouple iron and silica fluxes in the
stable sulfur isotopes: A review and critique: Geomicrobiology Journal, Precambrian banded iron formations: Geology, v. 31, p. 35–38, doi:
v. 1, p. 249–293. 10.1130/0091-7613(2003)031<0035:UGSRTD>2.0.CO;2.
Danielson, A., Moller, P., and Dulski, P., 1992, The europium anomalies in Holland, H.D., 1972, The chemistry of the atmosphere and oceans: New York,
banded iron formations and the thermal history of the ocean crust: Chemi- Wiley- Interscience, 351 p.
cal Geology, v. 97, p. 89–100, doi: 10.1016/0009-2541(92)90137-T. Holland, H.D., 1973, The oceans: A possible source of iron in iron formations:
Davison, W., and Seed, G., 1983, The kinetics of the oxidation of ferrous iron Economic Geology and the Bulletin of the Society of Economic Geolo-
in synthetic and natural waters: Geochimica et Cosmochimica Acta, v. 47, gists, v. 68, p. 545–557.
p. 67–79, doi: 10.1016/0016-7037(83)90091-1. Holland, H.D., 1984, The chemical evolution of the atmosphere and oceans:
Davison, W., Lishman, J.P., and Hilton, J., 1985, Formation of pyrite in fresh- Princeton, New Jersey, Princeton University Press, 582 p.
water sediments: Implications for C/S ratios: Geochimica et Cosmochi- Holland, H.D., 1999, When did the Earth’s atmosphere become oxic? A reply:
mica Acta, v. 49, p. 1615–1620, doi: 10.1016/0016-7037(85)90266-2. Geochemical News. v. 100, p. 20–22.
de Baar, H.J.W., and de Jong, J.T.M., 2001, Distributions, sources and sinks Isley, A.E., 1995, Hydrothermal plumes and the delivery of iron to banded iron
of iron in seawater, in Turner, D.R., and Hunter, K.H., eds., The bio- formations: Journal of Geology, v. 103, p. 169–185.
geochemistry of iron in seawater: New York, John Wiley and Sons, p. Jacobsen, S.B., and Pimental-Klose, M.R., 1988, A Nd isotopic study of the
123–253. Hamersley and Michipicoten banded iron formations: The source of REE
Derry, L.A., and Jacobsen, S.B., 1990, The chemical evolution of Precam- and Fe in Archean oceans: Earth and Planetary Science Letters, v. 87,
brian seawater: Evidence from REEs in banded iron formations: Geo- p. 29–44, doi: 10.1016/0012-821X(88)90062-3.
chimica et Cosmochimica Acta, v. 54, p. 2965–2977, doi: 10.1016/0016- Jorgensen, B.B., and Boudreau, B.P., 2001, Diagenesis and sediment-water
7037(90)90114-Z. exchange, in Boudreau, B.P., and Jorgensen, B.B., eds., The benthic
Detmers, J., Bruchert, V., Habicht, K.S., and Kuever, J., 2001, Diversity of boundary layer: Oxford, UK, Oxford University Press, p. 211–244.
sulfur isotope fractionations by sulfate-reducing prokaryotes: Applied Kaplan, I.R., and Rittenberg, S.C., 1964, Microbiological fractionation of sul-
and Environmental Microbiology, v. 67, p. 888–894, doi: 10.1128/ phur isotopes: Journal of General Microbiology, v. 34, p. 195–212.
AEM.67.2.888-894.2001. Kasting, J.F., 1987, Theoretical constraints on oxygen and carbon dioxide con-
Drever, J.I., 1974, Geochemical model for the origin of Precambrian banded centrations in the Precambrian atmosphere: Precambrian Research, v. 34,
iron formations: Geological Society of America Bulletin, v. 85, p. 1099– p. 205–229, doi: 10.1016/0301-9268(87)90001-5.
1106, doi: 10.1130/0016-7606(1974)85<1099:GMFTOO>2.0.CO;2. Kempe, S., Liebezeit, G., Diercks, A.R., and Asper, V., 1990, Water balance in
Elrod, V.A., Berelson, W.M., Coale, K.H., and Johnson, K.S., 2004, The the Black Sea: Nature, v. 346, p. 419, doi: 10.1038/346419a0.
flux of iron from continental shelf sediments: A missing source for Klein, C., and Beukes, N.J., 1989, Geochemistry and sedimentology of a facies
global budgets: Geophysical Research Letters, v. 31, p. L12307, doi: transition from limestone to iron-formation deposition in the Early Pro-
10.129/2004GL020216, doi: 10.1029/2004GL020216. terozoic Transvaal Supergroup, South Africa: Economic Geology and the
Emerson, S., 1976, Early diagenesis in anaerobic lake sediments: Chemical Bulletin of the Society of Economic Geologists, v. 84, p. 1733–1774.
equilibria in interstitial waters: Geochimica et Cosmochimica Acta, v. 40, Knauth, L.P., and Lowe, D.R., 1978, Oxygen isotope geochemistry of cherts
p. 925–934, doi: 10.1016/0016-7037(76)90141-1. from the Onverwacht Group (3.4 billion years), Transvaal, South Africa
Ewers, W.C., and Morris, R.C., 1981, Studies of the Dales Gorge Member of with implications for secular variations in the isotopic compositions
the Brockman Iron Formation, Western Australia: Economic Geology and of cherts: Earth and Planetary Science Letters, v. 41, p. 209–222, doi:
the Bulletin of the Society of Economic Geologists, v. 76, p. 1929–1953. 10.1016/0012-821X(78)90011-0.
Feijtel, T.C., DeLaune, R.D., and Patrick, W.H., Jr., 1988, Biogeochemical Knauth, L.P., and Lowe, D.R., 2003, High Archean climatic temperatures
control on metal distribution and accumulations in Louisiana sediments: inferred from oxygen isotope geochemistry of cherts in the 3.5 Ga
Journal of Environmental Quality, v. 17, p. 88–94. Swaziland Supergroup, South Africa: Geological Society of America
Friedl, G., Dinkel, C., and Werhli, B., 1998, Benthic fluxes of nutrients in the Bulletin, v. 115, p. 566–580, doi: 10.1130/0016-7606(2003)115<0566:
northwestern Black Sea: Marine Chemistry, v. 62, p. 77–88, doi: 10.1016/ HACTIF>2.0.CO;2.
S0304-4203(98)00029-2. Konhauser, K.O., Hamade, T., Raiswell, R., Morris, R.C., Ferris, F.G.,
Friedrich, J., Dinkel, C., Friedl, G., Pimenov, N., Wijsman, J., Gomoiu, M.-T., Southam, G., and Canfield, D.E., 2002, Could bacteria have formed Pre-
Cociasu, A., Popa, L., and Wehrli, B., 2002, Benthic nutrient cycling and cambrian banded iron formations?: Geology, v. 30, p. 1079–1082, doi:
10.1130/0091-7613(2002)030<1079:CBHFTP>2.0.CO;2. Ohmoto, H., and Felder, R.P., 1987, Bacterial activity in the warmer, sul-
Krapez, B., Barley, M.E., and Pickard, A.L., 2003, Hydrothermal and resedi- phate-bearing Archean ocean: Nature, v. 328, p. 244–246, doi:
mented origins of the precursor sediments to banded iron formation: 10.1038/328244a0.
Sedimentological evidence from the Early Palaeoproterozoic Brockman Ohmoto, H., Kagegawa, T., and Lowe, D.R., 1993, 3.4 billion-year-old bio-
Supersequence of Western Australia: Sedimentology, v. 50, p. 979–1011, genic pyrites from Barberton, South Africa; sulfur isotope evidence: Sci-
doi: 10.1046/j.1365-3091.2003.00594.x. ence, v. 262, p. 555–557.
Kump, L.R., and Holland, H.D., 1992, Iron in Precambrian rocks: Implications Ohmoto, H., Watanabe, Y., and Kumazaawa, K., 2004, Evidence from mas-
for the global oxygen budget of the ancient earth: Geochimica et Cosmo- sive siderite beds for a CO2-rich atmosphere before 1.8 billion years ago:
chimica Acta, v. 56, p. 3217–3223, doi: 10.1016/0016-7037(92)90299- Nature, v. 429, p. 395–399, doi: 10.1038/nature02573.
X. Pickard, A.L., Barley, M.E., and Krapez, B., 2004, Deep-marine depositional
Lenton, T.M., 2003, The coupled evolution of life and atmospheric oxygen, in setting of banded iron formation: Sedimentological evidence from inter-
Rothschild, L., and Lister, A., eds., Evolution on the planet Earth: The bedded clastic sedimentary rocks in the early Palaeoproterozoic Dales
impact of the physical environment: London, Academic Press, p. 35–53. Gorge Member of Western Australia: Sedimentary Geology, v. 170,
Lyons, T.W., 1991, Upper Holocene sediments of the Black Sea: Summary of p. 37–62, doi: 10.1016/j.sedgeo.2004.06.007.
Leg 4 box cores (1988 Black Sea Oceanographic Expedition), in Izdar, Poulton, S.W., and Raiswell, R., 2002, The low temperature geochemical cycle
E., and Murray, J.W., eds., Black Sea oceanography: Dordrecht, Kluwer, of iron; from continental fluxes to marine sediment deposition: American
NATO ASI Series, ser. C, v. 351, p. 401–441. Journal of Science, v. 302, p. 774–805.
Lyons, T.W., 1997, Sulfur isotopic trends and pathways of iron sulfide forma- Poulton, S.W., and Raiswell, R., 2005, Chemical and physical characteristics of
tion in upper Holocene sediments of the anoxic Black Sea: Geochi- iron oxides in riverine and glacial meltwater sediments: Chemical Geol-
mica et Cosmochimica Acta, v. 61, p. 3367–3382, doi: 10.1016/S0016- ogy, v. 218, p. 203–221, doi: 10.1016/j.chemgeo.2005.01.007.
7037(97)00174-9. Raiswell, R., and Anderson, T.F., 2005, Reactive iron enrichment in sediments
Lyons, T.W., and Berner, R.A., 1992, Carbon-sulfur-iron systematics of the deposited beneath euxinic bottom waters: Constraints on supply by shelf
uppermost Holocene sediments of the anoxic Black Sea: Chemical Geol- recycling, in McDonald, I., Boyce, A.J., Butler, I.B., Herrington, R.J.,
ogy, v. 99, p. 1–27, doi: 10.1016/0009-2541(92)90028-4. and Polya, D.A., eds, Mineral deposits and Earth evolution: Geological
Lyons, T.W., Werne, J.P., Hollander, D.J., and Murray, R.W., 2003, Contrast- Society [London] Special Publication 248, p.179–194.
ing sulfur geochemistry and Fe/Al and Mo/Al ratios across the last oxic- Raiswell, R., and Canfield, D.E., 1998, Sources of iron for pyrite formation:
to-anoxic transition in the Cariaco Basin, Venezuela: Chemical Geology, American Journal of Science, v. 298, p. 219–245.
v. 195, p. 131–157, doi: 10.1016/S0009-2541(02)00392-3. Roekens, E.J., and van Grieken, R.E., 1983, Kinetics of iron oxidation in
Mackey, D.J., O’Sullivan, J.E., and Watson, R.J., 2002, Iron in the western seawater of various pH: Marine Chemistry, v. 13, p. 195–202, doi:
Pacific: A riverine or hydrothermal source for iron in the Equatorial 10.1016/0304-4203(83)90014-2.
Undercurrent?: Deep-Sea Research, v. 40, p. 877–893. Rye, R., Kuo, P.H., and Holland, H.D., 1995, Atmospheric carbon dioxide con-
Matisoff, G., Lindsay, A.H., Matis, S., and Soster, F.M., 1980, Trace metal centrations before 2.2 billion years ago: Nature, v. 378, p. 603–605, doi:
equilibria in Lake Erie sediments: Journal of Great Lakes Research, v. 6, 10.1038/378603a0.
p. 353–366. Sagan, C., and Mullen, G., 1972, Earth and Mars: Evolution of atmospheric and
McKibben, M.A., and Barnes, H.L., 1986, Oxidation of pyrite in low temperature surface temperature: Science, v. 177, p. 52–56.
acidic solutions: Rate laws and surface textures: Geochimica et Cosmochi- Santschi, P., Hohener, P., Benoit, G., and Bucholtz-ten-Brink, M., 1990, Chemi-
mica Acta, v. 50, p. 1509–1520, doi: 10.1016/0016-7037(86)90325-X. cal processes at the sediment-water interface: Marine Chemistry, v. 30,
McManus, J., Berelson, W.M., Coale, K.H., and Kilgore, T.E., 1997, Phos- p. 269–315, doi: 10.1016/0304-4203(90)90076-O.
phorus regeneration in continental margin sediments: Geochimica Schneider, S.H., and Londer, R., 1984, The coevolution of climate and life: San
et Cosmochimica Acta, v. 61, p. 2891–2907, doi: 10.1016/S0016- Francisco, Sierra Club Books, 543 p.
7037(97)00138-5. Schwartzmann, D., McMenamin, M., and Volk, T., 1993, Did surface tempera-
Miller, R.G., and O’Nions, R.K., 1985, Source of Precambrian chemical and tures constrain microbial evolution?: Bioscience, v. 43, p. 390–393.
clastic sediments: Nature, v. 314, p. 325–330, doi: 10.1038/314325a0. Sholkovitz, E.R., Elderfield, H.E., Szymczak, R., and Casey, K., 1999,
Millero, F.J., 2001. The physical chemistry of natural waters: New York, Wiley- Island weathering: River sources of rare earth elements to the Western
Interscience, 654 p. Pacific Ocean: Marine Chemistry, v. 68, p. 39–57, doi: 10.1016/S0304-
Millero, F.J., Sotolongo, S., and Izaguirre, M., 1987, The oxidation kinetics of 4203(99)00064-X.
Fe(II) in seawater: Geochimica et Cosmochimica Acta, v. 51, p. 793–891, Stumm, W., and Morgan, J.J., 1980, Kinetics and products of ferrous iron oxy-
doi: 10.1016/0016-7037(87)90093-7. genation in aqueous solutions: Environmental Science and Technology,
Milliman, J.D., Farnsworth, K.L., and Albertin, C.S., 1999, Flux and fate of v. 14, p. 561–568.
fluvial sediments leaving large islands in the East Indies: Journal of Sea Theis, T.L., and Singer, P.C., 1974, Complexation of iron (III) by organic matter
Research, v. 41, p. 97–107, doi: 10.1016/S1385-1101(98)00040-9. and its effect on iron (II) oxygenation: Environmental Science and Tech-
Morris, R.C., 1993, Genetic modeling for banded iron-formation of the Ham- nology, v. 8, p. 569–573, doi: 10.1021/es60091a008.
ersley Group, Pilbara craton, Western Australia: Precambrian Research, Towe, K., 1996, Environmental oxygen conditions during the origin and early
v. 60, p. 243–286, doi: 10.1016/0301-9268(93)90051-3. evolution of life: Advances in Space Research, v. 18, no. 12, p. 7–15, doi:
Morris, R.C., and Horowitz, R.C., 1983, The origin of the iron-formation- 10.1016/0273-1177(96)00022-1.
rich Hamersley Group of Western Australia—Deposition on a plat- Trendall, A., 2000, The significance of banded iron formation (BIF) in the Pre-
form: Precambrian Research, v. 21, p. 273–297, doi: 10.1016/0301- cambrian stratigraphic record: Geoscientist, v. 10, p. 4–7.
9268(83)90044-X. Trendall, A.F., and Blockley, J.G., 1970, The iron formations of the Precam-
Nealson, K.H., 1997, Sediment bacteria: Who’s there, what are they doing brian Hamersley Group, Western Australia: With special reference to the
and what’s new?: Annual Review of Earth and Planetary Sciences, v. 25, associated crocidolite: Western Australia Geological Survey Bulletin,
p. 403–434, doi: 10.1146/annurev.earth.25.1.403. v. 119, p. 336.
Nembrini, G.P., Capobianco, J.A., Veil, M., and Williams, A.F., 1983, A Moss- Ullman, W.J., and Aller, R.C., 1982, Diffusion coefficients in near shore marine
bauer and chemical study of the formation of vivianite in sediments sediments: Limnology and Oceanography, v. 27, p. 552–556.
of Lago Maggiore (Italy): Geochimica et Cosmochimica Acta, v. 47, Waite, T.D., 2001, Thermodynamics of the iron system in seawater, in Turner,
p. 1459–1464, doi: 10.1016/0016-7037(83)90304-6. D.R., and Hunter, K.H., eds., The biogeochemistry of iron in seawater:
Newman, M.J., and Rood, R.T., 1977, Implications of solar evolution for the New York, John Wiley and Sons, p. 292–342.
Earth’s early atmosphere: Science, v. 198, p. 1035–1067. Walker, J.C.G., 1983, Possible limits on the composition of the Archean ocean:
Nriagu, J.O., and Dell, C.J., 1974, Diagenetic formation of iron phosphates in Nature, v. 302, p. 518–520, doi: 10.1038/302518a0.
Recent lake sediments: American Mineralogist, v. 59, p. 934–946. Walker, J.C.G., and Brimblecombe, P., 1985, Iron and sulfur in the pre-biologic
Ohmoto, H., 1997, When did the Earth’s atmosphere become oxic?: Geochemi- ocean: Precambrian Research, v. 28, p. 205–222, doi: 10.1016/0301-
cal News, v. 93, p. 12–27. 9268(85)90031-2.
238 R. Raiswell
Wang, Y., and Van Cappellen, P., 1996, A multicomponent reactive transport Wijsman, J.W.M., Herman, P.M.J., Middelburg, J.J., and Soetaert, K., 2002, A
model of early diagenesis: Application to redox cycling in coastal marine model for early diagenetic processes in sediments of the continental shelf
sediments: Geochimica et Cosmochimica Acta, v. 60, p. 2993–3014, doi: of the Black Sea: Estuarine, Coastal and Shelf Science, v. 54, p. 403–421,
10.1016/0016-7037(96)00140-8. doi: 10.1006/ecss.2000.0655.
Weiler, R.R., 1973, The interstitial water composition in the sediments of the Williamson, M.A., and Rimstidt, D.J., 1994, The kinetics and electrochemical
Great Lakes. I. Western Lake Ontario: Limnology and Oceanography, rate-determining step of aqueous pyrite oxidation: Geochimica et Cosmo-
v. 18, p. 918–931. chimica Acta, v. 58, p. 5443–5454, doi: 10.1016/0016-7037(94)90241-0.
Wignall, P.B., 1994, Black shales: Oxford, UK, Clarendon Press, 127 p.
Wijsman, J.W.M., Middelburg, J.J., and Heip, C.H.R., 2001, Reactive iron in
Black Sea sediments: Implications for iron cycling: Marine Geology,
v. 172, p. 167–180, doi: 10.1016/S0025-3227(00)00122-5. MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA


Memoir 198
2006
Microbial mediation of iron mobilization and deposition in iron

formations since the early Precambrian
D. Ann Brown*
Department of Geological Sciences, University of Manitoba, Winnipeg, Manitoba R3T 2N2, Canada
ABSTRACT
It is suggested that sedimentary deposition of iron in sediments from the Archean

to the present day can be attributed largely to microbial mediation. Results from
laboratory experiments, using a microbial consortium enriched from the source of a
biofilm growing on a rock face in an underground research laboratory, are used to
advance a plausible explanation for the mobilization, precipitation, transport, and
deposition of iron. The consortium produces its own local environments independent
of the prevailing atmosphere. It is able to extract iron from minerals such as bio-
tite and magnetite, as well as from a chelated solution; this iron is then metabolized,
mainly through dissimulatory iron reduction, to provide cell energy, after which it
is immediately precipitated. All organisms require energy for growth and reproduc-
tion, but because iron redox reactions are inefficient a large amount of iron must be
processed, either directly through metabolism or indirectly due to the local microbial
redox microenvironment, with the result that vast quantities accumulate as waste.
This precipitate could be the main source of iron in sedimentary iron formations. Sid-
erite and ferrihydrite, the main precipitates, may occur in close juxtaposition within
a biofilm. The oxidation state of the iron precipitate is controlled by the nutrient sup-
ply, which in turn influences the metabolism of the biofilm organisms and hence their
redox. Subsequently, this iron, enmeshed within the biofilm, is either deposited in sed-
iments as fine layers or rolled by wave or current action into particles and granules,
which can form structures similar to those found in banded iron formations (BIFs).
Keywords: biofilm, microbial mediation, iron formations, siderite, ferrihydrite.
INTRODUCTION iron while investigating a biofilm forming underground in the

Canadian Shield. Evidence from experiments using the micro-
This paper describes field and laboratory evidence sug- bial consortium forming this biofilm is set out to show the sig-
gesting that microbial activity was important to the genesis of nificant microbial processes that the consortium can encom-
iron formations such as BIFs (banded iron formations). The pass, and a comparison is then made between field and labora-
work is the result of a long-standing dialogue between Gor- tory biofilm textures and some iron formation structures that
don A. Gross, a geologist with the Geological Survey of Can- have been supplied and described by G.A. Gross. Though in
ada, who has wide experience of iron formations, and me, a general I have used Gross’s (1996) nomenclature for stratiform
microbiologist who was introduced to microbial deposition of iron deposits, common practice is followed in the use of the
*Present address: A&A Research, Old Moorcocks, Rushlake Green, Heathfield, E. Sussex TN21 9PP, UK.
Brown, D.A., 2006, Microbial mediation of iron mobilization and deposition in iron formations since the early Precambrian, in Kesler, S.E., and Ohmoto,
H., eds., Evolution of Early Earth’s Atmosphere, Hydrosphere, and Biosphere—Constraints from Ore Deposits: Geological Society of America Memoir 198,
p. 239–256, doi: 10.1130/2006.1198(14). For permission to copy, contact [email protected]. ©2006 Geological Society of America. All rights reserved.
239
240 D.A. Brown
term BIF for fine-grained iron minerals closely interbanded The role of sulfur was not investigated because no iron sul-
with quartz, chert, and/or carbonate. fur compounds were found in the laboratory studies or in the
Serious consideration has been given for many years to underground biofilm. The status of silica was also not explored,
the processes required to produce the enormous concentrations although it is a principal component of many iron formations, in
of iron in BIF sediments (Gross, 1988; Brown et al., 1995). part because silica cannot be metabolized by single-cell organ-
Past studies of the genesis of sedimentary rocks have gener- isms known as prokaryotes (Archaea and Bacteria), the only
ally emphasized abiotic physical and chemical processes, and fauna extant at the time of the early BIF deposits, and because
iron in sediments was widely believed to be supplied primar- some authors (Valley et al., 2002; Knauth et al., 2004) consider
ily by inorganic chemical precipitation from hydrothermal that early Archean temperatures on Earth were sufficiently low
waters (Gross, 1983). However, today there is a much better for marine silica saturation.
understanding of the influence of biological processes on the The results suggest that both the dissolution of iron from gra-
geochemistry of the atmosphere and hydrosphere, as well as nitic crystalline rocks and its subsequent reprecipitation through
on sedimentation processes in general. Better data are now the action of natural microbial consortia may offer a plausible
also available on the sources of the iron incorporated into and satisfactory alternative explanation for the mobilization,
sediments, and on the direct microbial reduction of iron as transport, precipitation, and deposition of iron in at least some
opposed to its oxidation. BIF sediments reaching back in age to the Archean.
The aspects of microbial processes that have been inves-
tigated in this study are summarized schematically in Figure 1, History of Geomicrobiological Research
and are based on laboratory investigations using microscopic
organisms (henceforth called microbes or bacteria) from the Evidence now seems to indicate that microbes emerged
biofilm consortium. These include the primary source of the almost as soon as water could form on the surface of Earth. A
iron; the leaching of this iron into solution where it is utilized brief survey shows that the association of microbes with iron
through redox reactions by bacteria as their energy source; the deposition has been recognized for nearly two centuries, ever
consequent precipitation of this iron in either the ferrous or since Ehrenberg (1836) suggested that bacteria were capable of
ferric state depending upon the local microenvironment; and precipitating bog-iron as hydrated iron oxides. Later, when Win-
the possible diagenesis of these precipitates during their depo- ogradsky (1888) was investigating microbes in the natural envi-
sition in iron formations. Various particles, granules, layers, ronment, he discovered that bacteria could obtain energy through
and other structures in BIFs are then compared morphologi- oxidation of inorganic compounds such as ferrous iron, and was
cally with textures found in the field and from biofilm pro- able to show that the bacterium Leptothrix requires this dissolved
cesses in the laboratory. iron for growth. An extensive literature review by Harder (1919)
Figure 1. Schematic diagram showing

a summary of the various microbial
aspects of iron conversion investigated
in the laboratory. The dissolution of
iron sources (direct mineral/microbe
contact not shown) for the provision
of cell energy through microbial me-
tabolism results in the precipitation of
either ferrous or ferric minerals. These
precipitates can become incorporated in
sediments where they may be altered by
diagenesis and ultimately concentrated
in iron formations.
Microbial mediation of iron mobilization and deposition in iron formations 241
further noted the importance of bacteria for the deposition of sed- microbial communities (Baross and Deming, 1983). The mass
imentary iron. More recently, it has been recognized that bacteria flux of the deeply circulating hydrothermal systems, manifested
can grow anaerobically through the coupling of carbon oxidation in part by these vents, is enormous. According to a recent calcu-
to the sedimentary reduction of manganese or iron oxides (Neal- lation by Harris et al. (2004), the flux from sea mounts equals
son and Myers, 1990). that from mid-ocean ridges, indicating that it could well have
The slow accumulation of knowledge on microbial com- been one of the major sources of iron in marine waters, whether
munities, particularly those in oligotrophic (nutrient-poor) envi- from biologic or abiotic processes. Furthermore, researchers in
ronments, is relevant to the investigation. It was initially thought the Ocean Drilling Program not only have found an incredibly
that microorganisms could only exist at the surface of Earth, but rich diversity of microbes living 1000 m below the seafloor, but
Waksman (1916) demonstrated that the populations and their also have reported bacteria and remnants of biofilms in waters
physiological activities extend throughout the soil profile. Bas- from an artesian borehole in basalt of the mid-ocean Juan da Fuca
tin (1926) discovered that bacteria were present in deep subsur- Ridge (Cowen et al., 2003). These discoveries suggest that pro-
face waters of oil field brines, at depths known as the profound karyotic microbes can thrive in seemingly hostile, hot, and anaer-
groundwater zone, and ZoBell (1952) reported the presence of obic environments similar to those thought to have been common
bacteria in seawater at the bottom of the Philippine Trench, more in shallow Archean marine waters (Baross and Hofmann, 1985).
than 10 km beneath the sea surface. Morita and ZoBell (1955) If these microbes were able to exist in such environments, it is
reported viable bacteria from 8 m below the seafloor of Pacific highly likely that they would have been involved in the mobiliza-
pelagic sediments, and suggested that they had been incorpo- tion and deposition of iron.
rated contemporaneously with the million-year-old sediment.
ZoBell (1943) proposed that, when nutrients were limited, bac- Sources and Deposition Sites for Iron in BIFs
teria formed biofilms that preferentially grew as slime-encased
microbes on rock surfaces (sessile) rather than as free-swimming Though iron formations are known throughout the geo-
(planktonic) organisms. The importance of such biofilm consor- logical record, from 3.8 Ga (Isua, West Greenland) to contem-
tia has been gradually recognized, but progress was slow until it porary deposition in the Red Sea basins (El Shazly, 1990), the
was realized in the 1970s there was a universal interdependence major iron deposits were formed between the Late Archean
amongst all the organisms found within the surface-associated and the middle Proterozoic (Gross, 1983; Isley, 1995). Many
biofilms. The recent investigation of crystalline rocks for the BIFs are extensive (Baur et al., 1985) and comprise alternately
burial of radioactive waste has shown that the groundwater of the banded iron- and silica-rich beds, microbands, and lamina-
subsurface zone harbors numerous and diverse microbial populations, where magnetite, hematite, and/or siderite are the main
tions (Fredrickson and Onstott, 1996; Amy and Haldeman, 1997; primary iron minerals. Nisbet (2000) discussed the evidence
Pedersen, 1997; Brown and Sherriff, 1999). for Archean life (possibly as early as 3.7 Ga) and the environ-
The morphologies of filamentous and of domed microfos- ments/ecological sites where it may have flourished. These
sils, first discovered in the early Proterozoic Gunflint Chert, Can- largely coincide with those outlined by Gross (1996, Fig. 3.2)
ada (1.9 Ga) by Tyler and Barghoorn in 1954, was thought at that as sites of iron formation deposition. Both authors consider
time to have cyanobacterial (blue-green algae) origins, and thus these habitats to be similar to those formed through tecto-
possibly to indicate oxygenic photosynthesis. This formation has nism today, even though independent cratonic masses may
a marked similarity to other, more iron-rich cherts such as BIFs, have been smaller and less long lived. The ecological sites are
and several modern authors (Gross, 1972; Hofmann et al., 1991; (1) shallow back-arc waters and shores with stromatolite and
Schopf, 1992) have established that autochthonous microbial microbial mat communities with oxygenic photosynthesis at
microfossils and stromatolites were indeed incorporated in Late surface and anoxygenic photosynthesis and methanogenesis
Archean BIFs. There are now numerous reports of the presence below; (2) deeper waters and commonly oligotrophic open
of microbes during deposition of silica-, carbonate-, and iron-rich ocean, with photosynthetic plankton subject to blooms when
sedimentary rocks from modern to Precambrian ages (Braga and nutrient supply is above normal; and (3) andesitic island arcs,
Martin, 2000; Sumner, 2000; Bonny and Jones, 2003). Even the komatiitic shield volcanoes, and mid-ocean ridges, each driv-
presence of hydrothermal vent microbial communities during the ing a hydrothermal circulation supplying nutrients to chemot-
Middle Archean has been suggested by Rasmussen (2000) in a rophic, thermophilic bacteria, together with some photosyn-
report of microfossils in a deep-sea volcanogenic massive sulfide thetic bacteria. Gross (1995, 1996) associates Lake Superior
deposit from the Pilbara Craton in Australia (3.2 Ga). type iron formations with continental shelf and slope; Algoma
Oceanographic research over the last 60 years, especially type with volcanic arcs and spreading ridges; and his Rapitan
the discovery of deep-sea thermal vents, has further expanded type with passive margin grabens. The action of the micro-
our knowledge of primitive life. These vents not only create a bial consortium that was used for the laboratory experiments
wide range of local environmental conditions but also discharge implies an early independence of bacteria from an obligatory
large quantities of iron and manganese into the ocean, which are photosynthetic basis and, to a certain extent, also from the gen-
then available to provide the primary energy source for extensive eral atmospheric composition.
242 D.A. Brown
The iron that is ultimately deposited in iron formations is work shows how complex the route to the final isotope value of a
derived from terrestrial cratonic and volcanic rocks, as well as deposit can be, as it is affected by microbially direct and indirect
from oceanic basalts and from crystallizing magmas. Terrestrial precipitation, by residual material from the leached rock-form-
and submarine groundwaters and hydrothermal fluids acquire ing minerals, as well as their solution and resultant precipitation,
iron from these sources through both biotic and abiotic reactions and then further by diagenesis during transport and sedimenta-
with iron-rich oxides and silicates. These waters can transport tion. Though Brantley et al. (2001) have shown that fractionation
the iron in ionic or complex organic molecular solution or, as through microbial iron reduction can provide a useful signature,
is shown, in microbially reprecipitated particles. Because iron, both they and Icopini et al. (2004) have warned of the need to
even in the reduced state, is only marginally soluble in water, the fully understand the mechanisms involved before firm conclu-
question has been how such vast quantities of iron found in iron sions can be reached. Therefore, the laboratory and field work
formations came to accumulate in the depositional basins. Sev- on the microbial dissolution of iron from rock-forming miner-
eral investigators, for instance Holland (1973), consider that the als and its concurrent reprecipitation in a biofilm environment
periodic upwelling from an ocean reservoir into a restricted basin suggests that the iron in BIFs could have been multi-sourced—
would have provided sufficient amounts of iron to produce exten- some from terrestrial weathering and some from marine hydro-
sive BIFs. Some (Garrels et al., 1973) suggest that the ocean acts thermal venting.
as a steady-state system with iron, from whatever source, and The question of whether mediation by microbes could
silica fluctuating only within narrow limits. Others (Lepp, 1987) account for the amount of iron deposited in BIFs has been inves-
think the laterization of cratonic rocks and fluvial transport are tigated by Konhauser et al. (2002). Their calculations of the num-
sufficient to provide all the necessary iron, and the ocean only ber of organisms and of the nutrients required for their growth
acts to homogenize the iron and trace elements. Konhauser et show that direct metabolic oxidation of iron has the potential to
al. (2002) in their study of marine systems suggested that the generate most, if not all, of the ferric iron in BIFs, even when
major oceanic source of the iron in BIFs was hydrothermal, pos- no account is taken of any indirect biomineralization that might
sibly supplemented by continental drainage. Homogenization occur. They conclude that microbes could clearly have been an
of the precipitated minerals in the oceanic environment spread important contributor to the deposition of iron, and hence micro-
them uniformly over large depositional basins. In their model bial mediation must be considered a viable and demonstrable
the richer hematite/magnetite bands were formed during major mechanism for BIF development in the Precambrian.
hydrothermal pulses and dependent blooms of iron-precipitat-
ing planktonic bacteria, whereas the silica-rich sequences sug- MICROBIAL COMMUNITIES
gest quieter periods when hydrothermal fluids did not reach the
basin, or did so only in small amounts. Hamade et al. (2003) have Microbial Metabolism
shown that silica and iron sources are decoupled. They suggested
that the silica comes from continental landmass weathering and Appreciation of the influence of various microorganisms on
the iron sources are marine hydrothermal. sedimentary geochemistry has developed slowly. The microbes
On the basis of element content and concentration, Gross involved are predominantly Archaea, which tend to inhabit
(1993) suggested that almost all of the iron in iron formations of extreme environments, and Bacteria, which prefer more moder-
all ages was volcano-hydrothermal in origin. Beard et al. (2003) ate conditions and include the main chemolithotrophic organisms
summarized the flux and isotope composition of the sources (those that obtain their energy from inorganic sources). Both of
delivering iron to the ocean under present-day levels of atmo- these are single-celled prokaryotic microbes that lack a mem-
spheric oxygen: riverine particulate, atmospheric particulate, brane-bound nucleus. The imperative that drives their activities
and hydrothermal loads accounting for 72%, 19%, and 6.5% by is the need to acquire sufficient energy for individual growth
weight respectively. The narrow range of δ56Fe values for ter- and reproduction. The rate at which this occurs depends on the
restrial igneous rocks, clastic sediments, and airborne materials amount of energy that can be released through redox-coupled
implies that chemical weathering, sedimentary transport, and dia- microbial metabolism, but the reactions that do occur are those
genesis play only a minor role in iron isotope variation. At mid- that are thermodynamically possible within a specific local envi-
ocean ridges hot hydrothermal fluids (>300 °C) shift the isotope ronment. It is noteworthy, however, that in some instances the
signature, again in a narrow peak, to a lower value. However, the microbes can adjust their environment to better suit their activi-
wider range of values in Archean iron formations (attributed by ties (see niche construction, Laland et al., 2004).
Beard et al. [2003] to Johnson et al., 2003) implies a mixing of All organisms, whether single or multicellular, have the
terrestrial and marine hydrothermal sources. same basic metabolic pathways, such as those for the biosyn-
Our leaching experiments show that microbial action can thesis of proteins or DNA, and including the energy-producing
supply terrestrial-source iron to riverine transport. However, pathways like the tricarboxylic acid (or Krebs) cycle. Today
from this source it is scattered throughout marine clastic sedi- most organisms, including all eukaryotes (those that contain a
ments, thus considerably lowering the percentage of riverine par- defined nucleus), ultimately obtain their energy from the sun,
ticulate flux to iron formation deposition sites. In addition, this either directly through photosynthesis, or indirectly by using the
oxygen so produced as their electron acceptor. Although there is determined by the velocity and the nutrient content of the local
little agreement on the composition of the early terrestrial atmo- water. Seed organisms adhere to a surface, then multiply and
sphere, I concur with Bekker et al. (2004) that it is unlikely to grow to form new colonies while excreting an EPS envelope
have contained much molecular oxygen before ca. 2.45 Ga. Our that provides both a stable environment for the microbes and a
present-day atmosphere has been produced through the oxygenic medium within which chemical exchange can take place. The
photosynthesis of water, initially by prokaryotic cyanobacteria, negative charges of the EPS and the cell walls of the microbes
but mainly today by eukaryotic green plants. Because the phylo- also “trap” abiotic material, notably positively charged metal
genic record suggests that cyanobacteria were one of the last lin- ions such as iron, which precipitate and form considerable min-
eages to diverge from the bacterial tree, it was probably not until eral concentrations (McLean et al., 1996). As biofilms grow
after 2.3 Ga that molecular oxygen would actually have begun to mass transfer from bulk liquid to the cells becomes more dif-
accumulate (Blank, 2004). ficult. The emphasis on a molecule’s utilization shifts from its
In the likely absence of available oxygen in the Archean, intrinsic metabolic return to its ability to diffuse into the bio-
microbes needed to develop alternative electron acceptors. The film (Riding, 2002; Battin et al., 2003). As was observed, for
chemiosmotic theory (Gottschalk, 1986) describes how chemi- instance, in the field at the site of the original biofilm, both
cal energy, in a variety of forms, can be changed into an elec- the internal runnels that facilitated nutrient transport and the
trochemical potential across a membrane to produce a proton all-enveloping slime provided local and very different micro-
motive force (pmf) that is utilized to synthesize ATP (used by environments organized by the microbes themselves. Thus the
all living organisms to transfer cell energy). In principle, there- consortium works together and is able to survive many harsh
fore, any electron potential between electron donor and accep- and extreme external conditions where individual planktonic
tor can be used to pump protons to the exterior of the cell and organisms are not able to exist.
thus induce a pmf. Hydrogen carriers are produced, usually by A modern study (Battin et al., 2003) confirms that microbial
enzymatic oxidation of an organic substrate that interacts with biofilms can change ecosystem processes to enhance their envi-
a membrane-bound electron transport chain, which channels the ronment. These changes within the restricted system of a biofilm
electrons to an oxidant while at the same time pumping protons increase competition between similarly acting species of microbes
to the exterior. It should be noted that most of the reactants and (in-groups) but enhance cooperation between those of dissimilar
products in the sedimentary environment are oxidants or reduc- function (out-groups), thus building a more efficient consortium
tants that can be utilized by organisms that follow the chemios- (Queller, 2004). These biofilm-forming consortia contain a wide
motic model (Nealson, 1997). Bacteria are thus able to use an variety of different microbes, but which group of these organ-
extraordinary diversity of ways to obtain energy from either the isms is active at any particular moment is determined by what
oxidation or the reduction of inorganic compounds and transition nutrients are available and also by the organisms’ ability to uti-
elements. So long as the necessary enzymatic machinery is avail- lize any accessible energy-producing redox reactions. It is this
able, it makes little difference to the organism what that source activity that controls the local microenvironment. Furthermore,
of energy is, be it nitrate, manganese, uranium, iron, sulfate, or because a consortium contains a variety of species, it is able to
carbon dioxide; these follow a sequence of decreasing redox adapt rapidly to any change in nutrients or the environment. Thus
potential and thermodynamic feasibility (Stumm and Morgan, at any one time some organisms will be dormant and will become
1981). It is this ability that enables bacteria to collectively fill vir- active only when the environment becomes suitable. Pertinent to
tually every ecological niche found on or under Earth’s surface, some of our experiments, where oxygen is initially present it is
as long as water is present (Brown and Sherriff, 1999). In the first consumed by aerobic organisms, and only when the oxygen
natural environment, however, iron is the most important source is exhausted do the anaerobic microbes become dominant (Neal-
of energy because, although the energy it provides is limited, it son, 1997). Metabolic cascades can occur when the waste product
is the most abundant and available element that undergoes such from one organism can be used by another as its substrate, thus
redox reactions. allowing further energy to be extracted (Nee, 2004). This forms a
stepwise degradation of complex organic compounds to simpler
Biofilms ones, and the final product is generally carbon dioxide, which
is lost to the atmosphere. Residual organic carbon is therefore
Although indigenous Bacteria and Archaea can be pres- found only in sediments where such degradation is incomplete,
ent as free-swimming planktonic organisms, in oligotrophic and thus is unlikely to be found in BIFs, although there may be a
groundwaters they flourish best as microbial consortia that signature left from accumulated sediment and minerals.
congregate by forming a biofilm at water-rock interfaces. Bio-
films are complex communities of independent, but interact- Biomats
ing, organisms. They consist of a layer of slime comprising
extracellular polymeric substances (EPS) that are usually nega- Biofilms are common today on rock and sediment surfaces
tively charged polysaccharides excreted by, and encompassing, in terrestrial lakes and streams, as well as on the walls of subsur-
the microbes (Costerton et al., 1994). Biofilm development is face groundwater channels and pores. In bodies of calm water
244 D.A. Brown
they can form free-floating gelatinous clouds several tens of cen- dynamic systems affected by tidal and seasonal fluctuations that
timeters across. However, both marine and fresh-water biofilms influence nutritional gradients and thus effectively balance sedi-
can also become an integral part of planar, soft organic layers mentation (whether mechanical or biochemical) and intermittent
and biomats that carpet lake beds and the seafloor over a wide lithification to produce millimeter-scale laminations, similar to
area, trapping sediment as the seasonal supply of nutrients varies. those described above in Baja California algal mats. Again, sedi-
Their resilience to erosion appears to be due in part to filamentous ment trapping smothers the organic surface. This is then recolo-
bacteria that stabilize the mat, making it resistant to high-energy nized by new organisms that utilize the organic remains of the
environments and allowing the mat to acquire a significant thick- previous microbes as their carbon source, effectively removing
ness. Biomats of this type have been recognized in South African any organic carbon and leaving only a thin layer of sediment
Archean siliciclastic rocks dated at 2.9 Ga, although it is uncer- as their memorial. An example of this in strongly microlami-
tain what group of bacteria this would have been (Noffke et. al., nated stromatolites from a Lake Superior type iron formation is
2003). Clastic carbonaceous particles of mats, eroded from a described later.
shallow-water, euphotic zone, have also been found by Tice and On aging, the mats entrap added minerals and become thicker
Lowe (2004) in similar South African rocks but dated to 3.4 Ga. and heavier so that they either break up or slough off excess frag-
In modern settings, the distribution of microbial species in ments. These fragments, and the grains of iron and silica minerals
these mats is mainly determined by the amount of light and oxy- they contain, can form the primary granules for sedimentation,
gen available (Stolz, 2000), with the surface dominated by photo- whether they are deposited in situ or are transported, possibly
trophic microbes (capable of photosynthesis) such as filamentous fluvially, farther afield.
cyanobacteria and algal eukaryotes. But in anaerobic environ-
ments, even today, consortia can form without phototrophs, as Interactions between Microbes and Minerals
occurs in the catotelm (body) of a bog, where anaerobes are able
to cycle electrons through both reduction and oxidation redox Microbial metabolic reactions have an extensive influ-
reactions (Brown, 1994). In the anaerobic Archean environment ence on the geochemistry of natural groundwaters by cycling
similar microbial consortia would also have been present, but iron between oxidized and reduced states (Freeze and Cherry,
they probably were composed of different populations. 1979; Lovley, 1995). This work confirms the ability of bacterial
In Baja California thick algal mats are formed by several fila- biofilms to accumulate metal ions from their environment and
mentous organisms that contribute to the primary organic deposit to precipitate them in particulate form. This immobilization of
(Stolz and Margulis, 1984). Microbes and their EPS within the metal ions (Douglas and Beveridge, 1998) and their subsequent
mats, as well as potentially precipitating iron and/or manganese transport into aquatic sediments represents a means of concen-
minerals, trap a layer of fine, inorganic, siliceous sediment that trating metals in the environment and incorporating them into
smothers the microbes during flooding. This is then recolonized mineral deposits that are important in shaping our planet.
and anaerobically reworked so no evidence of the original organ- Two common rock-forming minerals that contain ferrous
isms survives, although a recognizable microlayer may be left. iron—magnetite and ilmenite—are significant in controlling
The textures of some laminae in BIFs, rather than being the heterogeneous redox reactions at low temperatures (White and
result of settling in quiet water from a bloom of iron-precipitating Peterson, 1996). The solid-state mass balance requires that one
microbes as suggested by Konhauser et al. (2002), are strongly Fe(II) atom be removed from magnetite for every two Fe(II)
reminiscent of such sequential biomat growth. atoms oxidized to Fe(III) in the structure to form a new min-
Microbial mats currently forming in Icelandic hot springs eral, hematite. On “weathering” at surface, whether biologic or
show that their ability to bind iron and/or silica helps to pre- not, both electrons and ferrous iron are released to the aqueous
vent the degradation of the organisms by heterotrophic (organic environment, while the surface of the primary mineral is oxi-
nutrient source) bacteria (Konhauser and Ferris, 1996). The dized, possibly in a fashion similar to that of the chemiosmotic
encrusted cells are then incorporated into a matrix of amorphous model. Much of the iron liberated by weathering is deposited
silica prior to their deposition in the sedimentary record. This as solid ferric oxyhydroxides such as ferrihydrite, which are
suggests that contemporaneous biomineralization, even under a characteristic components of young iron oxide accumulations
different atmosphere, may have structures similar to those of the (Schwertmann, 1988). Therefore we investigated the microbial
geologic past. leaching of magnetite and also of biotite (to represent the mafic
Sometimes only small areas of mat are stable, and these silicates) in our experiments, although other researchers have
can build up into stromatolites, which provide some of the more investigated hornblende, common in rocks more mafic than
convincing Precambrian field evidence for microbial mediation granite (Brantley et al. 2001).
in iron deposition. They have been found in rocks as old as the In contrast to the oxic zone at the surface, in the subsur-
Steep Rock Lake deposit in Canada (3.0 Ga; Blackburn et al., face zone below a few meters the groundwater is reducing, a
1991). Stromatolites are small, local, domed or columnar forms state that is almost entirely due to the microbial reduction of
of microbial mats, with similar patterns of growth, that com- iron (Schwertmann and Taylor, 1989; Lovley, 1995). At one
monly occur in shallow seas or pools (Reid et al., 2000). They are time it was thought that both oxidation and reduction reactions
in groundwater were entirely abiotic chemical processes, or, especially around water runnels, and ferrihydrite formed on
if bacteria were involved, that it was only through their abil- much of the outer, aerobic surface, while gypsum was precipi-
ity to alter the redox and indirectly influence the environment tated on the outside of the internal water runnels. This precipi-
(Halvorson and Starkey, 1927). However, specific enzymatic tation led us to suggest that iron-rich minerals from the granite
bacterial reduction reactions have now been clearly demon- were being used in energy-producing microbial redox reac-
strated (Lovley, 1991; Bridge and Johnson, 1998), whereas tions (Brown et al., 1994, 1999). In addition, because both sid-
purely abiotic extraction of iron from rock minerals into surface erite and ferrihydrite were precipitated in such close proximity,
and subsurface groundwater has not been found to be signifi- it was also suggested that different biofilm microbes within a
cant because there appears to be insufficient energy available consortium were able to generate specific microenvironments
for its dissolution. Thus, without the “catalysis” of microbial with a very different redox. Significantly, the overall reaction
intervention, the rate of iron accumulation in major deposits when using the WATEQ4F modeling program (Ball and Nord-
would be extremely slow if not geologically improbable. strom, 1991) showed, from chemical considerations alone, that
siderite was undersaturated and therefore should not have been
LABORATORY EXPERIMENTS formed, suggesting that microbial mediation within the biofilm
was necessary for its precipitation (Brown et al., 1994).
At the University of Manitoba, together with B.L. Sher- This unexpected juxtaposed formation of different iron
riff, I investigated a biofilm-forming consortium that was first minerals in the underground led us to investigate the apparent
discovered growing where mining water from an iron-poor self-organization of aerobic and anaerobic environments pro-
source, typical of the Canadian Shield, flowed over a fresh rock duced by the biofilm microbes. A series of laboratory experi-
face in the Underground Research Laboratory (URL) exca- ments was carried out, utilizing a consortium enriched from
vated by Atomic Energy of Canada Ltd. in the Lac du Bonnet the surface water used for mining at the URL, to elucidate the
granitic batholith in Manitoba, Canada (Brown et al., 1989). various microbial aspects outlined in Figure 1. A more detailed
The biofilm formed following an oil spill on the 420 m level, flow chart of the dissolution of iron from primary sources,
which provided just sufficient carbon for cell mass, but not for through various processes before it is deposited as secondary
cell energy. The biofilm was ~10 mm thick and covered over a minerals in sediments, is shown in Figure 2. The boxes in the
square meter in extent (Brown et al., 1994). Iron minerals, both latter chart indicate the processes that occur during the trans-
in the oxidized and reduced state, were deposited in close jux- port of iron, and the bold type the microbial reactions that were
taposition within the biofilm; siderite (identified by Mössbauer investigated in the laboratory and that are discussed in detail in
spectroscopy) was precipitated against the anaerobic rock face the following sections.
Figure 2. Flow sheet of the various reac-

tions involving iron from the primary min-
eral to its sedimentation. Bold type—mi-
crobially mediated reactions investigated
in the laboratory. Heavy boxes—process-
es investigated.
246 D.A. Brown
URL Biofilm The microenvironments produced by the microbes are quite

specific and extremely local. Within three days of inoculation
The consortium contains a variety of bacteria with dif- with the consortium, the pH of the medium in stationary labora-
fering morphologies as well as a protozoan (eukaryote), which tory flasks increased from 7.0 to 9.0, and the Eh was reduced
would not have been present in the Archean (Brown et al., to around −300 mV, even though the bungs in these flasks were
1998a). The majority are Gram-negative bacteria (that would permeable to air. Thus, although the flasks were in the normal
include Geobacter metallireducens), and only a few Gram- present-day atmosphere, microbial metabolism was sufficient not
positive bacteria are present. Phospholipid fatty acid analysis only to use up all the oxygen in the flask, but also to maintain an
suggests the presence of Shewanella sp., as well as Desulfovi- anaerobic environment. However, when the iron was present in
brio sp. There are also indications of the presence of Pseudo- solid mineral form and the flasks were rocked, the metabolism
monas and Actinomycetes. Transmission electron microscopy was slower and the interchange with the atmosphere was much
(TEM) shows the distinctive morphology of Gram-negative greater, so that an anaerobic environment was not achieved.
bacteria and confirms the presence of a curved flagellated rod
similar to Desulfovibrio sp. (Brown et al., 1998b). Energy
Not only does the consortium appear to be very stable,
as it has been maintained in the laboratory for nearly 10 yr, The main source of bacterial energy was found to be dis-
but this type of consortium also appears to be widespread. simulatory iron reduction. This occurs when the iron is not incor-
Three seemingly similar consortia were enriched from the porated as a nutrient into the cell, but instead is utilized as an
URL area. A sample taken from a stream in the English Weald, electron acceptor in a redox reaction to generate energy. The iron
a historic siderite ironstone mining area, was also found to is reduced from the ferric to the ferrous state with the precipita-
contain a consortium with comparable reactions (personal tion of siderite, while at the same time the major electron donor,
observation). citrate, is ultimately oxidized to carbon dioxide (Brown et al.,
Several methods were tried for growing biofilms by run- 1999). Both citrate and glucose can be fermented, producing an
ning the culture medium over rock surfaces, but it was difficult Eh of around −150 mV, and under the experimental conditions
both to maintain sterility and to ensure the biofilm remained citrate was broken down more rapidly than glucose (Table 1).
moist. It tended to dry out, cracking partially into thin gelati- When iron was added to the glucose it initially inhibited the fer-
nous platelets that morphologically resemble a texture noted mentation, but after 40 days the amount of glucose metabolized
in some BIFs. All the experiments were therefore conducted in was similar whether iron was present or not. However, when both
flasks standing on the laboratory bench, occasionally swirled citrate and iron were present in the medium, there was a rapid
to aid mixing. Only when a solid substrate, such as mineral enzymatic reduction of the iron and utilization of citrate, virtually
chips, was used were the flasks put on an orbital rocker to all being consumed within ten days; notably this did not occur
ensure that all parts of the culture were in contact with the when glucose was the carbon source. There was a considerable
solid. In this latter case the precipitate enmeshed in EPS precipitation of siderite at an Eh below −300 mV, confirming
formed small discrete flocs due to the motion; this becomes that this lower redox is necessary for its formation (Garrels and
important when discussing ovoid granule textures in BIFs. Christ, 1990), and also that, here, citrate is the source of the car-
A bioreactor was maintained to produce inocula for the bonate anion rather than carbon dioxide (although it may have
experiments. The reactor contained granite chips as a source of been originally synthesized from carbon dioxide through the tri-
trace nutrients, in a medium of 5.0 g ferric ammonium citrate, carboxylic acid cycle). Using citrate as the carbon source in labo-
0.5 g K 2HPO4, 0.2 g MgSO4.7H2O, and 0.01 g CaCl2.2H2O ratory cultures proved doubly efficient as it provided the electron
per liter at an initial pH of 7.0 (Brown and Hamon, 1994). In
laboratory cultures the ferric ammonium citrate used in the
bioreactor was replaced with ammonium citrate and ferrous
chloride to allow variation in the relative concentrations of TABLE 1. INCUBATION EXPERIMENTS IN TRIPLICATE OVER 40 DAYS
iron and carbon. A 10% inoculum was used throughout, and Days Citrate + iron Citrate Glucose + iron Glucose
all experiments were carried out at least in triplicate at room (% citrate) (% citrate) (% glucose) (% glucose)
0 100 100 100 100
temperature (23 ºC) with sterile (not inoculated) controls. 5 3 67 99 80
Dissolved iron in the URL groundwater has a concentra- 10 2 66 98 80
15 1 44 92 60
tion of less than 1.0 mg L –1 (Gascoyne et al., 1989), but in the 20 – 35 33 54
surrounding granite the iron-bearing minerals biotite, magne- 25 – 29 30 44
30 – 12 19 33
tite, ilmenite, and hematite make up 2% of the rock mass, and 35 – 9 18 29
therefore would seem to be the most probable source of the iron. 40 – 4 15 22
Note: Citrate and glucose media are used both with and without iron.
Silica was also found in the biofilm, but was not investigated Maintained at 23 °C on the bench, with 50 mL of medium in 250 mL flasks
because silica cannot be directly metabolized by bacteria; pre- and with 10% inoculum. Shown as percentage remaining after a number of
days incubation.
sumably it was adsorbed indirectly onto the EPS slime.
donor and also chelated the iron, which by itself is poorly soluble
under these pH/Eh conditions.
Although the concentration of sulfate in the groundwater
at the URL was considerable (48 mg L–1) no iron sulfides were
detected, indicating that the reduction of iron was thermodynam-
ically preferred to that of sulfate.
Dissolution
Direct contact between cells and the mineral involved was

once thought to be essential for microbial reduction of iron (Lov-
ley, 1995), but it has now been shown that iron reducing microbes
such as Shewanella sp. can overcome this problem by releasing
soluble quinones into the water (Childers et al., 2002). These act
as electron shuttles by transferring electrons from the cell sur-
face to insoluble metal oxides that can be located at a significant
distance from the cell. The first organism, G. metallireducens,
Figure 3. The amount of iron in accumulated precipitate resulting from
reported by Caccavo et al. (1994) to be capable of completely microbial leaching over time of 5 g each of biotite (ò), and magnetite
oxidizing organic compounds to carbon dioxide by utilizing oxi- (¢) in 50 mL of medium, from an average of 3 flasks for each mea-
dized iron or manganese as electron acceptors uses two methods surement; iron in solution at each measurement (p).
to extract iron. When grown on relatively insoluble oxides this
organism specifically expresses flagella for mobility to reach the
mineral and pili for attachment to it, whereas when grown on
soluble iron citrate these appendages are not produced, suggest- idly, but gradually tailed off as the available iron became dis-
ing that the organism is able to detect whether soluble electron tanced from the broken cleavage edges. In contrast, the rate of
acceptors are potentially available (Childers et al., 2002). This precipitation from the magnetite substrate, although quite low,
confirms that in some cases direct contact with insoluble miner- remained steady while the amount of precipitate increased
als may indeed be necessary. slowly over time. A very approximate deposition rate of 20 m of
Mineral dissolution preferentially occurs at surface disloca- hematite per 1 m.y., calculated from the more reliable magnetite
tion sites. That biogeochemical reactions also occur at these sites experiment, compares well with the lower rate (20–225 m per 1
can be seen from scanning electron micrograph images, where m.y.) suggested from some sequences at Hamersley, Australia
more bacteria cluster on the mineral surface at fracture edges (Trendall, 2000).
than on the crystal and cleavage faces (Brown et al., 1997). This So far as is known, iron must be in solution, if only fleet-
implies that the former sites are preferred for the extraction of ingly, for it to be metabolized by bacteria (Nealson and Little,
nutrients, presumably because of the presence of a greater con- 1997). Therefore, there has to be a rapid flux of iron from the min-
centration of unsatisfied bonds, or possibly, in the case of a sheet eral, through a dilute solute phase, to the microbial sites where
silicate such as biotite, more ready access to the layers of iron metabolism takes place; the iron is then rapidly precipitated and
atoms in the lattice. removed from solution. Presumably the small pool of dissolved
Because the consortium formed in waters draining a granitic iron, which remains at a constant low concentration, is sufficient
batholith, chips of the iron-containing minerals biotite and mag- for active metabolism to occur so long as this pool is maintained
netite were used as the source of iron in the medium for the dis- by the chemical equilibrium. As microbes precipitate the iron,
solution experiments. Although the amount of iron in solution more primary iron must be drawn into solution to preserve this
was extremely low (~1.0 mg L–1), a week after adding the inocu- equilibrium, suggesting that the microbial process drives the dis-
lum the rocked flasks showed a significant amount of orange fer- solution of iron from its host minerals. Even when the pH and Eh
ric precipitate enmeshed within the biofilm slime. These results were adjusted to the same conditions as the cultures and killed
show that the consortium is able to grow on solid iron-contain- organisms were present, no precipitation occurred until viable,
ing mineral fragments, suggesting that organisms similar to She- metabolizing microbes from the consortium were added (Brown
wanella and G. metallireducens must be present. et al., 1999).
In the leaching experiments, after eight weeks the average
amount of iron precipitate in the medium from biotite was 54.7 Metabolism
mg L–1, and that from magnetite 6.8 mg L–1 (Fig. 3). During this
period, precipitation from the biotite dissolution was initially In the natural oxygenated environment little iron is found
quite low while the consortium organisms became conditioned in solution (Beard et al., 2003), except in the extremely acidic
to extracting iron from a solid substrate. It then increased rap- and aerobic environments that are produced at massive sulfide
248 D.A. Brown
exposures or in man-made acid mine drainage. Although the iron took place in solution while the iron was still chelated
simple ionizable ferrous salts are moderately soluble, ferrous (nuclear magnetic resonance data from B.L. Sherriff, 1997, per-
iron is much less soluble when bound in complex minerals or sonal commun.). Alternatively, if the ratio was equal to or less
organic compounds: a phenomenon exhibited in the reducing than 1:5, this smaller amount of citrate was rapidly consumed by
groundwaters of the URL. Also, in the laboratory experiments, the consortium, removing the chelation. In this latter case the iron
unless chelated, concentrations of dissolved iron greater than was precipitated in the oxic state as very fine grained orange fer-
~1.0 mg L–1 were rare. In the natural environment microbes are rihydrite, which rapidly darkened to brown as it was reduced, but
always present, and if iron is available it will be used as an energy here the average percentage of ferrous iron was lower, and the
source and be precipitated. Thus it is unlikely that dissolved iron protein tended to be less. The iron remained reduced so long as
in Archean waters was greater than is found today under similar there were sufficient carbon metabolites present for the anaerobic
anaerobic conditions. environment to be maintained.
Iron is thought to have been one of the earliest, as well as Thus the reduction of ferric iron in the laboratory can occur
one of the most important, external electron acceptors to have through two different mechanisms, depending on the initial
been utilized by microbes (Vargas et al., 1998). Depending upon ratio of iron to carbon as citrate (Table 2B). Two populations
the local redox conditions, bacteria can obtain energy from either are found when the percentage of ferrous iron is plotted against
the oxidation or the reduction of iron, although this change of protein produced: one, where the precipitate is greenish, fer-
state is less efficient for iron than for manganese (see chemios- rous iron averages 56% and protein 16 mg mL–1, and the other,
motic model). Iron is almost universally and abundantly present where the precipitate is initially orange, ferrous iron averages
in rocks and so it is the most likely element to have been uti- 33% and protein 11 mg mL–1. Sterile controls showed no sign of
lized. Whereas in today’s oxic world iron reduction is predomi- any precipitation. The ability of the consortium to reduce iron is
nant because much of the iron in soils and minerals is in the oxi- apparently constitutional (the enzyme system is always present
dized state, in the anaerobic environment of the Archean readily and does not need to be induced by the presence of substrate).
available iron would have been in the ferrous state, and the main After 24 serial transfers without iron in the medium, the rate of
reaction would have been its oxidation. This oxidation does not reduction when iron was reintroduced was the same as that of
require the presence of free oxygen but can be driven instead by the original incubations.
light; anoxygenic phototrophs are bacteria that only have pho- This set of experiments shows that it is the relative amount
tosystem I and can use ferrous iron as an electron donor. This of iron to carbon that determines the state of the iron precipi-
anaerobic photosynthesis implies that oxygen-independent bio- tate. When all the other parameters are kept the same, the larger
logical iron oxidation would have been possible before the evo- relative concentration of carbon in the medium produces greater
lution of two-step oxygenic photosynthesis (Widdel et al., 1993;
Ehrenreich and Widdel, 1994), thus making it possible for oxi-
dized iron to be laid down under an anoxic environment. Tice and
TABLE 2A. DIFFERENT RATIOS OF IRON (AS FeCl2)
Lowe (2004) concluded from the shallow-water setting (euphotic TO CARBON (AS CITRATE) IN THE MEDIA
zone) and carbon isotopic composition of the Buck Reef Chert, 15 mmol carbon 20 mmol carbon 25 mmol carbon
2.5 mmol iron
South Africa, that photosynthetic, probably anoxygenic microbes Fe:C ratio 1:6 1:8 1:10
were active in the 3.4 Ga ocean. However, because the modern Fe(II) % 50 60 57
Protein (mg mL–1) 11 18 20
URL consortium is unable to photosynthesize, oxygen has to be Munsell standard 5Y 6/4 5Y 5/2 5Y 5/4
available before it can form ferric iron. 5.0 mmol iron
To study the metabolism of the consortium we varied the Fe:C ratio 1:3 1:4 1:5
Fe(II) % 25 34 40
relative concentrations of both iron and carbon in the medium Protein (mg mL–1) 8 11 17
(Brown et al., 1999). Ammonium citrate was used to provide Munsell standard 5YR 5/6 5YR 4/4 10YR 4/2
both carbon and nitrogen for cell growth, and ferrous chloride 7.5 mmol iron
Fe:C ratio 1:2 1:2.7 1:3.3
(mostly oxidized during sterilization) supplied iron for cell Fe(II) % 26 38 33
energy. The relative efficiency of the incubations is shown, first Protein (mg mL–1) 6 14 12
Munsell standard 5YR 6/6 5YR 3/4 5YR 2/2
by the amount of protein produced as a measure of growth and Note: This table shows the average results, after eight days incubation in
second by the percentage of ferrous iron precipitated. As well, five separate experiments, of the ferrous iron and protein produced, as well
we used the Munsell standard card to define the color of the as the Munsell standard color of the precipitates.
precipitate because this indicates the balance between the iron

valency states within the precipitate (Table 2A). When the ratio
of iron to citrate (as molar concentrations) was greater than 1:5, TABLE 2B. AVERAGES OF THE RESULTS IN TABLE 2A
the amount of ferrous iron produced was over 50% of the total Fe(II) % Protein (mg mL–1) Munsell standard
iron, and consisted of siderite, ferrous hydroxide, and vivianite Fe:C ratio
1:6 to 1:10 56 16 5Y
(where the phosphate came from the medium) and the color of 1:2 to 1:5 33 11 5 to 10YR
the precipitate was an olive green. In these cases the reduction of Note: This table shows there are two separate populations.
metabolic activity, hence more rapid iron reduction, so that the iron
is directly and rapidly precipitated as siderite. With relatively less
carbon, the citrate is first metabolized, thereby removing the chela-
tion, and the iron is precipitated initially as ferrihydrite. This is later
reduced, possibly to ferrous hydroxide, which is much less stable
than siderite and so is easily reoxidized by ambient oxygen in the
air. Thus in this natural consortium only a small difference in the
ratio of available nutrients results in completely different iron min-
erals being precipitated, either a carbonate or an oxide.
The ability of the consortium to define its own redox depend-
ing upon the nutrients available has a distinct bearing on the type
of iron supplied to sediments. Furthermore, the state of this iron
could indicate the changeable environment to which such biofilm
consortia are commonly subjected and contribute to the differing
laminae found in BIFs. Whether iron is laid down in iron forma-
tions as carbonate or as oxide minerals appears to depend upon
the local microbial environment at that particular moment rather
than on the influence of the general prevailing atmosphere.
Precipitation
The consortium cultures regularly produced both siderite

Figure 4. Transmission electron micrograph of an unstained thin sec-
and ferrihydrite, which altered to hematite through the removal tion of biofilm bacteria showing iron directly precipitated by metabolic
of water, and although magnetite was not identified, its formation activity. One cell has lysed where the iron precipitate has become too
under similar laboratory conditions has been reported (Lovley et dense for metabolism to occur. Other cells are surrounded by a parac-
al., 1987). Thus three major iron minerals can all be precipitated rystalline proteinaceous S-layer on which iron has been metabolically
through microbial activity, and both siderite and ferrihydrite have precipitated instead of directly on the cell wall. Scale bar, 0.25 μm.
been seen to do so in the field within the URL biofilm. Further-
more, even though the consortium did not synthesize magnetite,
when utilized as a substrate magnetite was rapidly oxidized to
hematite. This clearly demonstrates the consortium’s ability, (Schultze-Lam et al., 1992). Iron is also precipitated indirectly
when oxygen is available, to extract energy from iron through in the EPS and on cell walls owing to their negative charge
either reduction or oxidation reactions. Mössbauer spectros- (Fig. 5).
copy showed that 11% of the magnetite was altered to hematite The spatial distributions of these two types of precipitate
within three weeks of incubation, but interestingly, maghemite, have distinct fractal dimensions. Using 12 transmission electron
the usual intermediate in this transformation, was not identified photomicrographs, we measured the density of the iron precipi-
(Brown et al., 1997). Because the reaction is rapid, FeO is prob- tate for the biofilm matrix and for the cell wall (Fig. 6). Although
ably released directly from the lattice surface and oxidized to the clusters do overlap, the means are significantly different; the
hematite, and the lack of maghemite (γFe2O3) suggests that once fractal dimension of the iron mineral spatial distribution is lower
released it is immediately altered to the αFe2O3 form. Scanning for the cell wall environs than that for the more general EPS
electron micrographs of the surface of the oxidized magnetite matrix. This supports the likelihood that they are formed by two
showed hexagonal plates, suggesting the formation of crystals different mechanisms, that is, either by direct metabolic activity
(Brown et al., 1997) similar to those that have been reported in or indirectly through the microbially induced environment pro-
BIFs (LaBerge et al., 1987). duced within the biofilm itself.
Biomineralization occurs as the consortium microbes extract In these experiments the precipitation of iron-bearing min-
energy from the iron, precipitating it as siderite, ferrihydrite, erals, by whatever means, results in its entrapment within the
and/or magnetite. However, precipitation directly onto the cell biofilm. In some of the cultures aggressive protozoa actively
wall clogs the active metabolic sites and causes death. The cells and preferentially ingested the iron-coated bacteria. When these
then lyse and leave only a ring (i.e., sphere) of iron-rich minerals bacteria had all been consumed, the protozoa encysted, thus effi-
around an empty cell (Fig. 4), but many cells are able to protect ciently concentrating and stabilizing the iron in even larger par-
themselves from such a death by excreting a paracrystalline pro- ticles within the biofilm. Furthermore, because protozoa would
teinaceous envelope, or S-layer. This layer surrounds the organ- not have been extant in the Archean, cultures that did not contain
ism and prevents the iron from coming into direct contact with them formed iron microcrystallites of around 8 μm in diameter
the cell wall, thus allowing normal metabolic activity to continue that were also tightly enmeshed within the EPS (Brown et al.,
250 D.A. Brown
Figure 6. Measurements of fractal dimension, D, from 12 TEM prints

showing the different clusters formed by iron precipitates, eight data
Figure 5. Transmission electron micrograph of an unstained thin sec- points for mineralization at the cell wall and four for within the biofilm
tion of the biofilm showing the iron indirectly precipitated throughout matrix. The boxes are one standard deviation about the mean.
the EPS matrix due to its negative charge. The S-layer, with metaboli-
cally precipitated iron around the central microbe, has been distorted
during preparation of the section. Scale bar, 0.25 μm.
the ferrous precipitate to be much more stable than the ferric.
The ferrous compounds remained reduced except when oxygen
was present in the headspace, whereas the ferric compounds,
1998a). Thus, regardless of the probable differences between under the same experimental conditions, were not reduced but
consortia, the iron would be securely entrapped under Archean lost water and were altered to hematite under all headspace gases
conditions, ensuring that it could be transported as particles and (Brown et al. 1998b). This suggests that ferrous iron microbi-
deposited during sedimentation. In all the cases tested, almost all ally precipitated in sediments as siderite would be more likely
the iron involved in the microbial mediation was precipitated, to remain in the reduced state, whereas iron deposited as ferrihy-
with only a minimal amount remaining in solution. drite would tend to form hematite. Notably, the control experi-
Finally, it should be noted that because of the influence of ments using chemically precipitated ferric iron formed goethite
pre-glacial marine waters, the URL groundwater composition is as well as hematite. Whether ferrihydrite or goethite is formed
quite saline (245 mg L–1 of chloride; Gascoyne et al., 1989) indi- in the natural environment is under considerable discussion (van
cating that consortia flourishing in fractures underground could der Zee et al., 2003), but this was the only time that goethite was
also have been active in Archean seas. identified in any of the experiments.
Diagenesis Summary Interpretation of Laboratory Experiments
The post-deposition low-temperature diagenesis of the The type of microbial consortium isolated from the URL
microbially precipitated iron minerals in different oxidation states biofilm is nowadays widespread under both anaerobic and aero-
was also investigated. Experiments at 80 °C for 12 weeks under bic conditions in the natural environment, indicating that micro-
different headspace gases (air, nitrogen, and hydrogen) showed bial mediation is likely to have an important influence on the
geologic deposition of iron, and has probably changed little since to oxidation than is siderite, and thus perhaps is the origin of
the Archean. The various processes performed by the consor- the iron oxides found in BIFs. Ultimately, however, because all
tium were entirely dependent on microbial activity, given that these metabolic reactions are inefficient, a large quantity of iron
even under simulated microbial conditions no activity could be has to be metabolized to produce an adequate supply of energy
detected without the presence of viable organisms. The consor- for bacterial growth, resulting in vast accumulations of iron as
tium can utilize iron either from general solution or by close con- a waste product.
tact with solid minerals, and the very low concentration of iron
commonly reported in groundwaters demonstrates its utilization IRON FORMATION STRUCTURES
by such microbial activity.
In the natural environment, the microbial imperative is to If biofilms mobilized iron from a primary source to its
acquire energy, which drives all the necessary steps for iron to sedimentation site, some structure or texture indicative of this
be leached, mobilized, and reprecipitated. This precipitate then microbial mediation ought to be visible in suitable formations.
provides particulate minerals for transportation to their ultimate In fact many rock structures in iron formations either resemble
depositional sites. Under present global conditions, the consor- microbially related features observed in the laboratory or can
tium microbes are able to manipulate the biofilm microenviron- be attributed to biological activities. Some of these features that
ment independently of the surrounding atmosphere. In the exper- have been identified in examples of iron formation deposits,
iments the dominant and constitutive energy-producing redox from the Archean to the present day, are listed in Table 3.
reaction is dissimulatory iron reduction, which has persisted The observed textures that result from the processes dis-
possibly since it evolved under the anoxic atmosphere of the cussed earlier are strong indicators of the depth of water in
Archean. Where oxygen is available, both ferric oxides, mainly which the deposition occurs and of whether conditions are gen-
as ferrihydrite, and ferrous carbonate can be precipitated in close erally quiet or are influenced by current and wave action. They
juxtaposition. Under suitable conditions this same consortium may be roughly divided into three: (1) very fine particles of iron
can also oxidize magnetite to hematite. In the laboratory the minerals originating as precipitates either from within a biofilm
ratio of carbon (for cell growth) to iron (for energy) determines or from a bloom (and in addition, silica particles, though these
the metabolism and thus which mineral is precipitated. Sider- are not discussed here); (2) considerably larger granules of iron
ite is preferentially formed when the concentration of carbon is minerals and/or jaspilite formed from the breakup of biofilms
greater, and because it is also resistant to oxidation under low and biomats or from repeated oolitic growth of biofilm and
temperature diagenesis, it is more likely to be retained in iron enmeshed precipitate; and (3) layering at various scales due to
sediments. Ferrihydrite is formed when carbon is limited and the successive sedimentation of particles and granules of different
environment is less reducing, and although it can be microbially composition, including the gradual accumulation of biofilms
reduced, possibly to ferrous hydroxide, it is more susceptible and biomats.
TABLE 3. COMPOSITION AND GEOMETRY OF GRANULAR MINERALS AND OTHER TEXTURES IN IRON FORMATION DEPOSITS
Description Fig. Name Type Age
(Ga)
Microbanded magnetite with a little silica: relict microovoids of magnetite Moose Mountain, Canada Algoma L. Archean 2.7–2.6
Siderite microovoids and spheroids and pyrite/pyrrhotite in thick siderite beds 8 Wawa, Canada Algoma L. Archean 2.7
Magnetite, hematite, and silica in microovoids accumulating on bedding planes in 9 Timagami, Canada Algoma L. Archean 2.7–2.6
magnetite/jasper lithofacies
Magnetite microovoids in magnetite and siliceous magnetite beds Anshan, China ? L. Archean 2.6
Magnetite and hematite microovoids in magnetite, hematite and siliceous Krivoy-Rog, Ukraine Lake Superior E. Proterozoic 2.4–2.2
magnetite microbands in magnetite/jasper lithofacies
Hematite and magnetite particles in siliceous oolite and microovoids; hematite in 7 Sokoman, Canada Lake Superior E. Proterozoic 2.2–1.8
jasper ovoids and larger siliceous granules; hematite micro-particles in jasper
and fine-grained magnetite and hematite as rims on siliceous ovoids and oolites
Magnetite (?) microovoids in massive magnetite beds Bayanobo, China ? E. to M. Proterozoic 1.8–1
Hematite (?) ovoids and oolites in massive hematite beds and siliceous granules Rapitan, Canada Rapitan L. Proterozoic 0.6
Magnetite (?) microovoids in magnetite micro- and mesobands Austin Brook, Canada ? Ordovician
Ovoids and spheroids of Fe and Mn oxides in microbands and beds Woodstock, Canada Algoma Silurian
Microovoids of Fe and Mn oxides in mud Red Sea Recent
Note: The global formations range in age from the Archean to Recent. Data provided by G.A. Gross, Geological Survey of Canada (2004).
252 D.A. Brown
Fine Particles together with the presence of dark, possibly organic material, in
coarse siderite layers at Wawa, Canada (Fig. 8) suggests a series
The sizes and shapes of the fine particles of iron com- of dehydrated biomats that dried in the intertidal zone. Morpho-
pounds and their aggregates that formed in the laboratory logically this strongly resembles the platelets (~0.5–10 mm) of
experiments are comparable to those forming naturally today partially dehydrated biofilm grown on granite plaques in the
in biofilms (field observation), and also to those in the least laboratory.
metamorphosed iron formations, which are, in many cases, Although plausible abiotic origins have been suggested for
reported as spheroids (Table 3). Examination of particles in the small spheroids and microlaminae, cyclical microbial precip-
photomicrographs from the consortium (Brown et al., 1998a) itations, the growth of biofilms and biomats, and near sea-surface
suggests that although the smallest particles on microbe walls blooms seem to be the most likely processes to produce these
may be less than 5 nm in average diameter, they are commonly textures, at least on the finer scale.
between 5 and 50 nm (Fig. 4), and spheroidal particles within
the matrix of the biofilm are 50–100 nm. Encrusted microbes
leave iron aggregates 750–1000 nm in diameter, but these
together with other precipitates within the biofilm comprise
either irregular clumps of spheroids and shapeless particles,
350–450 nm in the longest dimension, or irregular networks of
rods comprising three to five spheroids and averaging 250 nm
in length by 40 nm in diameter. These particle sizes are com-
parable to those reported for ferrihydrite in springs (2–5 nm;
Carlson and Schwertmann, 1981) or iron hydroxide in micro-
bial mats (100 nm; Konhauser and Ferris, 1996).
Encysted protozoa, appropriate only to post–middle Protero-
zoic deposits, form much larger secondary particles 2.5–3.5 μm
in diameter, whereas secondary (post-microbial action) hexago-
nal hematite crystals are up to 20 μm in diameter and aggregates
of crystallites may be 8 μm in width. Particles from iron for-
mations themselves are reportedly larger (LaBerge et al., 1987;
Robbins et al., 1987), but these authors also report submicron
hematite “dust” disseminated in jaspilite. They regard their larger
grains of spheroidal or crystalline hematite (1–30 μm) as diage-
netic-metamorphic recrystallization of the submicron hematite.
Larger Granules
In some deposits thicker (10–300 mm) layers of larger sphe-

roidal to platy granules (0.5–10 mm in diameter) of hematite,
siderite, and/or jaspilite interrupt zones of microlaminae. The
granules are encountered in depressions on the bedding surface,
and show evidence of current action such as imbrication and
graded bedding. The granules, especially those of jaspilite, also
exhibit repeated growth with outer layers enclosing one or more
earlier-formed granules, as shown in specimens from Sokoman,
Canada (Fig. 7). This is a common texture developed by bio-
films in high energy environments that are intermittently quies-
cent. These granules resemble the small, spherical flocs of iron-
mineral-filled biofilm formed during the orbitally rocked flask
experiments described earlier.
All biomats, gelatinous biofilms, and derived flocs must
dewater at some stage during sedimentation. Cracks, regarded as Figure 7. A polished, sectioned hand specimen exhibiting ovoids and
syneresis because the outer layers are unbroken, can be observed irregular granules in a local jasper lithofacies from a Lake Superior
type, Proterozoic (2.2–1.8 Ga) iron formation from Sokoman Lake,
within jaspilite granules in examples from Sokoman. Fossil bio- Labrador, Canada. Some granules show successive layers of biofilm
mats are difficult to recognize. The irregular form (viewed from overgrowth and probable internal cracking due to syneresis. Supplied
above) of more spheroidal (viewed in cross section) granules, by G.A. Gross, Geological Survey of Canada.
Figure 9. A polished, sectioned hand specimen exhibiting laminae of

magnetite, hematite, and silica in magnetite-jasper lithofacies from an
Algoma type, Archean (2.7–2.6 Ga) iron formation from the Tima-
gami area, Ontario, Canada. Iron-rich laminae probably represent the
accumulation in quiescent conditions of both direct and indirect fine
Figure 8. A polished, sectioned hand specimen exhibiting granules and microbial precipitation. Supplied by G.A. Gross, Geological Survey
broken layering in siderite lithofacies from an Algoma type, Archean of Canada.
(2.7 Ga) iron formation from Wawa, Ontario, Canada. These probably
represent successive, partially dried, microbial mats containing pri-
mary microbially precipitated siderite, metamorphically recrystallized.
Supplied by G.A. Gross, Geological Survey of Canada.
Because the experiments suggest that a minor change in sub-
strate supply (the iron to carbon ratio) readily changes the state
Layering of the iron in microbial precipitates, microlayering would seem
to be likely in the natural state where environmental conditions
Multiple, sequential laminae and banding are common, can change rapidly. Seasonality may also cause variations in the
though not ubiquitous, structures observed in iron formations. supply of iron and sediment or, alternatively, variations could
Although they range widely in thickness from 10−4 m to possibly be associated with periods of volcanic activity. Konhauser et al.
10 m, only the thinner layers are pertinent to this study. Quiet- (2002) have suggested that annual marine blooms of iron-utiliz-
water sedimentation of the finer particles leads to millimeter- ing bacteria are richer when the supply of iron is greater.
scale microlaminae, but some current action is suggested by mild A specimen from the Lake Superior Biwabik iron formation,
scour-and-fill structures and grains of earlier lithified bands in an United States (1.9 Ga; Fig. 10), shows small stromatolites coeval
example from Timagami, Canada (Fig. 9). I suggest that, before with the sediment. A group of these, strongly microlaminated
diagenesis and metamorphism, these laminae would comprise with jaspilite layers, are separated from each other by much less
the finer particles described earlier. structured, approximately contemporaneous, zones of siliceous
254 D.A. Brown
CONCLUSIONS
It has been difficult to envisage how iron, although ubiq-

uitous in rocks but to a large extent insoluble in the natural
environment, could have been mobilized and transported from
primary sources, both terrestrial and submarine, and then
redeposited in iron formations. Although abiologic reactions
are possible, thermodynamic considerations and deposition
rates make it unlikely that they could have been wholly respon-
sible for the amount of iron amassed. The laboratory and field
work reported here suggests that the direct and indirect effects
of microbial metabolism may have played a significant role in
producing the massive accumulations of iron, in geologically
credible amounts and time scales necessary for the deposition
of stratified sediments, particularly those of early BIFs. Primi-
tive microbial life is thought to have become established very
early in the history of Earth, possibly soon after 4.0 Ga. In the
anaerobic environment extant at that time, sources of energy
for microbial metabolism were limited, since oxygenic photo-
synthesis had yet to evolve. However, iron is widespread, and
redox reactions between the ferrous and ferric states are able
to provide a limited but sufficient amount of energy for cell
growth. Calculations confirm that many billions of microbes
could, by utilizing iron as their energy source, discharge into
the environment the many tons of waste iron that is recognized
today in formidable iron deposits, such as those of BIFs.
ACKNOWLEDGMENTS
I would like to thank my colleagues, especially G.A. Gross, for

the many discussions and their support in this research; B.L. Sher-
Figure 10. Polished, sectioned hand specimen of stromatolites from
the Lake Superior Biwabik iron formation, Minnesota, United States. riff, who was instrumental in obtaining my funding and produced
The complete right-hand stromatolite is 55 mm tall. The specimen the NMR analyses; and the Guelph Regional Scanning Transmis-
is oriented stratigraphically up toward the top of the page. Zones of sion Electron Microscope Facility of the National Sciences and
10–15 pairs that are more iron-rich than adjacent zones stretch across Engineering Research Council of Canada (NSERC) for supplying
several stromatolites in the complete hand specimen, probably indicat- the TEM figures. I would also like to thank Anton Brown, S. Kesler,
ing periods of general increase in the supply of iron.
H. Ohmoto, K. Towe, and an anonymous reviewer, who have done
much work to improve this manuscript.
sediment enclosing magnetite grains. If individual pairs of the REFERENCES SITED

iron-rich and iron-poor jaspilite laminae are regarded as annual
cycles, then they imply an accumulation rate of ~1000 m per 1 Amy, P.S., and Haldeman, D.L., eds., 1997, The microbiology of the terrestrial
deep subsurface: New York, CRC Lewis Publishers, 344 p.
m.y.; this is barely an order of magnitude faster than the aver- Ball, J.W., and Nordstrom, K.D., 1991, User’s manual for calculating specia-
age rate (20–225 m per 1 m.y.) suggested over much larger spa- tion of major, trace, and redox elements in natural waters: U.S. Geological
tial and time scales for BIFs in the Hamersley Group (Trendall, Survey Open-File Report 91-183, 189 p.
Baross, J.A., and Deming, J.W., 1983, Growth of ‘black smoker’ bacte-
2000; Isley, 1995). More importantly, there are several zones of ria at temperatures of at least 250°C: Nature, v. 303, p. 423–426, doi:
especially iron-rich, adjacent microlaminae that stretch at the 10.1038/303423a0.
same stratigraphic level across all the eight stromatolites in this Baross, J.A., and Hofmann, S.E., 1985, Submarine hydrothermal vents and
associated gradient environments as sites for the origin and evolution of
specimen. These groups of 10–15 lamina pairs suggest a possible life: Dordrecht, Kluwer Academic Publishers, Origins of Life and Evolu-
decadal climatic control on the iron source in the sedimentary tion of the Biosphere, v. 15, p. 327–345, doi: 10.1007/BF01808177.
basin. However, in this example, although magnetite content in Bastin, E.S., 1926, The presence of sulphate-reducing bacteria in oil-field
waters: Science, v. 63, p. 21–24.
the intervening sediment is weakly correlated with the groups of Battin, T.J., Kaplan, L.A., Newbold, J.D., and Hansen, C.M.E., 2003, Contribu-
iron-rich stromatolitic microlaminae, it is not correlated with the tions of microbial biofilms to ecosystem processes in stream mesocosms:
individual microlaminae. Nature, v. 426, p. 439–442, doi: 10.1038/nature02152.
Baur, M.E., Hayes, J.M., Studley, S.A., and Walter, M.R., 1985, Millimeter- Carlson, L., and Schwertmann, U., 1981, Natural ferrihydrites in surface depos-
scale variation of stable isotope abundances in carbonates from banded its from Finland and their association with silica: Geochimica et Cosmo-
iron-formations in the Hamersley group of Western Australia: Economic chimica Acta, v. 45, p. 421–429, doi: 10.1016/0016-7037(81)90250-7.
Geology and the Bulletin of the Society of Economic Geologists, v. 80, Childers, S.E., Ciufo, S., and Lovley, D.R., 2002, Geobacter metallireducens
p. 270–282. accesses insoluble Fe(III) oxide by chemotaxis: Nature, v. 416, p. 767–
Beard, B.L., Johnson, C.K., Von Damm, K.L., and Poulson, R.L., 2003, 769, doi: 10.1038/416767a.
Iron isotope constraints on Fe cycling and mass balance in oxygen- Costerton, J.W., Lewandowski, Z., DeBeer, D., Caldwell, D., Korber, D., and
ated Earth oceans: Geology, v. 31, p. 629–632, doi: 10.1130/0091- James, G., 1994, Biofilms, the customized microniche: Journal of Bacte-
7613(2003)031<0629:IICOFC>2.0.CO;2. riology, v. 176, p. 2137–2142.
Bekker, A., Holland, H.D., Wang, P.-L., Rumble, D., III, Stein, H.J., Hannah, Cowen, J.P., Giovannoni, S.J., Kenig, F., Johnson, H.P., Butterfield, D., Rappe,
J.L., Coetzee, L.L., and Beukes, N.J., 2004, Dating the rise of atmospheric M.S., Hutnak, M., and Lam, P., 2003, Fluids from aging ocean crust that
oxygen: Nature, v. 427, p. 117–120, doi: 10.1038/nature02260. support microbial life: Science, v. 299, p. 120–123, doi: 10.1126/sci-
Blackburn, C.E., Johns, G.W., Ayer, J., and Davis, D.W., 1991, Wabigoon Sub- ence.1075653.
province, in Thurston, P.C., Williams, H.R., Sutcliffe, R.H., and Stott, Douglas, S., and Beveridge, T.J., 1998, Mineral formation by bacteria in natural
G.M., eds., Geology of Ontario: Ontario Geological Survey Special Vol- microbial communities: FEMS Microbiology Ecology, v. 26, p. 79–88,
ume 4, p. 303 −381. doi: 10.1016/S0168-6496(98)00027-0.
Blank, C.E., 2004, Evolutionary timing of the origins of mesophilic sulphate Ehrenberg, D.C.G., 1836, Vorläufige Mittheilungen über das wirkliche Vor-
reduction and oxygenic photosynthesis: A phylogenomic dating approach: kommen fossiler Infusorien und ihre grosse Verbreitung: Annalen der
Geobiology, v. 2, p. 1–20, doi: 10.1111/j.1472-4677.2004.00020.x. Physik und Chemie, v. 38, p. 213–227.
Bonny, S., and Jones, B., 2003, Microbes and mineral precipitation, Miette Ehrenreich, A., and Widdel, F., 1994, Anaerobic oxidation of ferrous iron by
Hot Springs, Jasper National Park, Alberta, Canada: Canadian Journal of purple bacteria, a new type of phototrophic metabolism: Applied and
Earth Sciences, v. 40, p. 1483–1500, doi: 10.1139/e03-060. Environmental Microbiology, v. 60, p. 4517–4526.
Braga, J.C., and Martin, J.M., 2000, Subaqueous siliciclastic stromatolites: A El Shazly, E.M., 1990, Red Sea deposits, in Chauvel, J.J., Yuqi, C., El Shazly,
case history from late Miocene beach deposits in the Sorbas Basin of SE E.M., eds., Ancient banded iron formations: Athens, Theophrastus Publi-
Spain, in Riding, R.E., and Awramik, S.M., eds., Microbial sediments: cations, p. 157–222.
Berlin, Springer-Verlag, p. 226–232. Fredrickson, J.K., and Onstott, R.C., 1996, Microbes deep inside the earth: Sci-
Brantley, S.L., Liemann, L., and Bullen, T.D., 2001, Fractionation of Fe iso- entific American, October, p. 68–73.
topes by soil microbes and organic acids: Geology, v. 29, p. 535–538, doi: Freeze, R.A., and Cherry, J.A., 1979, Groundwater: Englewood Cliffs, New
10.1130/0091-7613(2001)029<0535:FOFIBS>2.0.CO;2. Jersey, Prentice-Hall, 524 p.
Bridge, T.A.M., and Johnson, D.B., 1998, Reduction of soluble iron and reduc- Garrels, R.M., and Christ, C.L., 1990, Solutions minerals and equilibria: Bos-
tive dissolution of ferric iron-containing minerals by moderately thermo- ton, Jones and Bartlett, 400 p.
philic iron-oxidizing bacteria: Applied and Environmental Microbiology, Garrels, R.M., Perry, E.A., and Mackenzie, F.T., 1973, Genesis of Precambrian
v. 64, p. 2181–2186. iron-formations and the development of atmospheric oxygen: Economic
Brown, A., Soonawala, N.A., Everitt, R.A., and Kamineni, D.C., 1989, Geol- Geology and the Bulletin of the Society of Economic Geologists, v. 68,
ogy and geophysics of the Underground Research Laboratory site, Lac du p. 1173–1179.
Bonnet Batholith, Manitoba: Canadian Journal of Earth Sciences, v. 26, Gascoyne, M., Ross, J.D., Watson, R.L., and Kamineni, D.C., 1989, Soluble
p. 404–425. salts in a Canadian Shield granite as contributors to groundwater salinity,
Brown, D.A., 1995, Carbon cycling in peat and the implications for the reha- in Miles, D.L., ed., Proceedings of the 6th International Symposium on
bilitation of bogs, in Cox, M., Straker, V., and Taylor, D., eds., Wetlands: Water-Rock Interaction, Malvern, U.K., 1989, p. 247–249.
Archaeology and nature conservation: London, HMSO, p. 99–107. Gottschalk, G., 1986, Bacterial metabolism (2nd edition): New York, Springer-
Brown, D.A., and Hamon, C., 1994, Initial investigation of groundwater micro- Verlag, 359 p.
biology at AECL’s Underground Research Laboratory: Pinawa, Atomic Gross, G.A., 1972, Primary features in cherty iron-formations: Sedimentary
Energy of Canada Technical Record, TR-608, 50 p. Geology, v. 7, p. 241–261, doi: 10.1016/0037-0738(72)90024-3.
Brown, D.A., and Sherriff, B.L., 1999, Evaluation of the effect of microbial Gross, G.A., 1983, Tectonic systems and the deposition of iron-formation: Precam-
subsurface ecosystems on spent nuclear fuel repositories: Environmental brian Research, v. 20, p. 171–187, doi: 10.1016/0301-9268(83)90072-4.
Technology, v. 20, p. 469–477. Gross, G.A., 1988, A comparison of metalliferous sediments, Precambrian to
Brown, D.A., Kamineni, D.C., Sawicki, J.A., and Beveridge, T.J., 1994, Min- Recent: Krystalinikum, v. 19, p. 59–74.
erals associated with biofilms occurring on exposed rock in a granitic Gross, G.A., 1993, Element distribution patterns as metallogenic indicators in
Underground Research Laboratory: Applied and Environmental Micro- siliceous metalliferous sediments: Resource Geology Special Issue 17,
biology, v. 60, p. 3182–3191. p. 96–107.
Brown, D.A., Gross, G.A., and Sawicki, J.A., 1995, A review of the microbial Gross, G.A., 1995, The distribution of rare earth elements in iron-formations:
geochemistry of banded iron-formations: Canadian Mineralogist, v. 33, Global Tectonics and Metallogeny, v. 5, p. 63–67.
p. 1321–1333. Gross, G.A., 1996, Stratiform iron, in Eckstrand, O.R, Sinclair, W.D., and
Brown, D.A., Sherriff, B.L., and Sawicki, J.A., 1997, Microbial transforma- Thorpe, R.I., eds., Geology of Canadian mineral deposit types: Geologi-
tion of magnetite to hematite: Geochimica et Cosmochimica Acta, v. 61, cal Survey of Canada, Geology of Canada, no. 8, p. 41–54.
p. 3341–3348, doi: 10.1016/S0016-7037(97)00159-2. Halvorson, H.O., and Starkey, R.L., 1927, Studies on the transformation of iron
Brown, D.A., Beveridge, T.J., Keevil, C.W., and Sherriff, B.L., 1998a, Evalu- in nature. I: Theoretical considerations: Journal of Physical Chemistry,
ation of microscopic techniques to observe iron precipitation in a natural v. 31, p. 626\7–631, doi: 10.1021/j150274a016.
microbial biofilm: FEMS Microbiology Ecology, v. 26, p. 297–310, doi: Hamade, T., Konhauser, K.O., Raiswell, R., Goldsmith, S., and Morris, R.C.,
10.1016/S0168-6496(98)00044-0. 2003, Using Ge/Si ratios to decouple iron and silica fluxes in Precambrian
Brown, D.A., Sawicki, J.A., and Sherriff, B.L., 1998b, Alteration of micro- banded iron formations: Geology, v. 31, p. 35–38, doi: 10.1130/0091-
bially precipitated iron oxides and hydroxides: American Mineralogist, 7613(2003)031<0035:UGSRTD>2.0.CO;2.
v. 83, p. 1419–1425. Harder, E.C., 1919, Iron-depositing bacteria and their geologic relations: U.S.
Brown, D.A., Sherriff, B.L., Sawicki, J.A., and Sparling, R., 1999, Precipi- Geological Survey Professional Paper 113, 89 p.
tation of iron minerals by a natural microbial consortium: Geochi- Harris, R.J., Fisher, A.T., and Chapman, D.S., 2004, Fluid flow through sea-
mica et Cosmochimica Acta, v. 63, p. 2163–2169, doi: 10.1016/S0016- mounts and implications for global mass fluxes: Geology, v. 32, p. 725–
7037(99)00188-X. 728, doi: 10.1130/G20387.1.
Caccavo, F., Lonergan, D.J., Lovely, D.R., Davis, M., Stolz, J.F., and McIn- Hofmann, H.J., Sage, R.P., and Berdusco, E.N., 1991, Archean stromato-
erney, M.J., 1994, Geobacter sulfurreducens sp. nov., a hydrogen- and lites in Michipicoten group siderite ore at Wawa, Ontario: Economic
acetate-oxidizing dissimilatory metal-reducing microorganism: Applied Geology and the Bulletin of the Society of Economic Geologists, v. 86,
and Environmental Microbiology, v. 60, p. 3752–3759. p. 1023–1030.
256 D.A. Brown
Holland, H.D., 1973, The oceans: A possible source of iron in iron-formations: Riding, R., 2002, Biofilm architecture of Phanerozoic cryptic carbon-
Economic Geology and the Bulletin of the Society of Economic Geolo- ate marine veneers: Geology, v. 30, p. 31–34, doi: 10.1130/0091-
gists, v. 68, p. 1169–1172. 7613(2002)030<0031:BAOPCC>2.0.CO;2.
Icopini, G.A., Anbar, A.D., Ruebush, S.S., Tien, M., and Brantley, S.L., 2004, Robbins, E.I., LaBerge, G.L., and Schmidt, R.G., 1987, A model for the bio-
Iron isotope fractionation during microbial reduction of iron: the impor- logical precipitation of Precambrian iron formations—B: Morphologi-
tance of adsorption: Geology, v. 32, p. 205–208, doi: 10.1130/G20184.1. cal evidence and modern analogs, in Appel, P.W.U., and LaBerge, G.L.,
Isley, A.E., 1995, Hydrothermal plumes and the delivery of iron to banded iron eds., Precambrian iron-formations: Athens, Theophrastus Publications,
formation: Journal of Geology, v. 103, p. 169–185. p. 92–139.
Knauth, L., Paul, L., and Lowe, D.R., 2004, High Archean climatic temperature Schopf, J.W., 1992, The oldest fossils and what they mean, in Schopf, J.W., ed.,
inferred from oxygen isotope geochemistry of cherts in the 3.5 Ga Swazi- Major events in the history of life: Boston, Jones and Bartlett, p. 29–61.
land Supergroup, South Africa: Nature, v. 42, p. 566–580. Schultze-Lam, S., Harauz, G., and Beveridge, T.J., 1992, Participation of a cya-
Konhauser, K.O., and Ferris, F.G., 1996, Diversity of iron and silica pre- nobacterial S-layer in fine-grain mineral formation: Journal of Bacteriol-
cipitation by microbial mats in hydrothermal waters, Iceland: Implica- ogy, v. 174, p. 7971–7981.
tions for Precambrian iron formations: Geology, v. 24, p. 323–326, doi: Schwertmann, U., 1988, Occurrence and formation of iron oxides in various
10.1130/0091-7613(1996)024<0323:DOIASP>2.3.CO;2. pedoenvironments, in Stucki, J.W., Goodman, B.A., and Schwertmann,
Konhauser, K.O., Hamade, T., Raiswell, R., Morris, R.C., Ferris, F.G., U., eds, Iron in soils and clay minerals: Dordrecht, D. Reidel Publishing
Southam, G., and Canfield, D.E., 2002, Could bacteria have formed the Co., p. 267–308.
Precambrian banded iron formations?: Geology, v. 30, p. 1079–1082, doi: Schwertmann, U., and Taylor, R.M., 1989, Iron oxides, in Boris, J., and Weed,
10.1130/0091-7613(2002)030<1079:CBHFTP>2.0.CO;2. S.B., eds., Minerals in soil environments: Madison, Wisconsin, Soil Sci-
LaBerge, G.L., Robbins, E.I., and Han, T.-M., 1987, A model for the biological ence Society of America, p. 379–438.
precipitation of Precambrian iron-formations—A: Geological evidence, Stolz, J.F., 2000, Structure of microbial mats and biofilms, in Riding, R.E.,
in Appel, P.W.U., and LaBerge, G.L., eds., Precambrian iron-formations: and Awramik, S.M., eds., Microbial sediments: Berlin, Springer-Verlag,
Athens, Theophrastus Publications, p. 69–95. p. 1–8.
Laland, K.N., Odling-Smee, J., and Feldman, M.W., 2004, Causing a commo- Stolz, J.F., and Margulis, L., 1984, The stratified microbial community at
tion: Nature, v. 429, p. 609, doi: 10.1038/429609a. Laguna Figueroa, Baja California, Mexico: A possible model for PrePha-
Lepp, H., 1987, Chemistry and origin of Precambrian iron-formations, in nerozoic laminated microbial communities preserved in chert: Origins
Appel, P.W.U., and LaBerge, G.L., eds., Precambrian iron-formations: of Life and Evolution of the Biosphere, v. 14, p. 671–679, doi: 10.1007/
Athens, Theophrastus Publications, p. 3–30. BF00933720.
Lovley, D.R., 1991, Dissimilatory Fe(III) and Mn(IV) reduction: Microbiologi- Stumm, W., and Morgan, J.J., 1981, Aquatic chemistry: New York, John Wiley
cal Reviews, v. 55, p. 259–287. and Sons, 742 p.
Lovley, D.R., 1995, Microbial reduction of iron, manganese and other minerals: Sumner, D.Y., 2000, Microbial vs. environmental influences on the morphology
Advances in Agronomy, v. 54, p. 176–217. of late Archean fenestrate microbialites, in Riding, R.E., and Awramik,
Lovley, D.R., Stolz, J.F., Nord, G.L., and Phillips, E.J.P., 1987, Anaerobic pro- S.M., eds., Microbial sediments: Berlin, Springer-Verlag, p. 307–314.
duction of magnetite by a dissimilatory iron-reducing microorganism: Tice, M.M., and Lowe, D.R., 2004, Photosynthetic microbial mats in the 3,414-
Nature, v. 330, p. 252–254, doi: 10.1038/330252a0. Myr-old ocean: Nature, v. 431, p. 549–552, doi: 10.1038/nature02888.
McLean, R.J.C., Fortin, D., and Brown, D.A., 1996, Microbial metal-binding Trendall, A., 2000, The significance of banded iron formation (BIF) in the Pre-
mechanisms and their relation to nuclear waste disposal: Canadian Jour- cambrian stratigraphic record: Geoscientist, v. 10, p. 4–7.
nal of Microbiology, v. 42, p. 392–400. Tyler, S.A., and Barghoorn, E.S., 1954, Occurrence of structurally preserved
Morita, R.Y., and ZoBell, C.E., 1955, Occurrence of bacteria in pelagic sedi- plants in Precambrian rocks of the Canadian Shield: Science, v. 119,
ments collected during the Mid-Pacific expedition: Deep-Sea Research, p. 606–608.
v. 3, p. 66–73. Valley, J.W., Peck, W.H., King, E.M., and Wilde, S.A., 2002, A cool early Earth:
Nealson, K.H., 1997, Sediment bacteria: who’s there, what are they doing, Geology, v. 30, p. 351–354, doi: 10.1130/0091-7613(2002)030<0351:
and what’s new?: Annual Review of Earth and Planetary Sciences, v. 25, ACEE>2.0.CO;2.
p. 403–434, doi: 10.1146/annurev.earth.25.1.403. van der Zee, C., Roberts, D.R., Rancourt, D.G., and Slomp, C.P., 2003, Nano-
Nealson, K.H., and Myers, C.R., 1990, Iron reduction by bacteria: A potential goethite is the dominant reactive oxyhydroxide phase in lake and marine
role in the genesis of banded iron formations: American Journal of Sci- sediments: Geology, v. 31, p. 993–996, doi: 10.1130/G19924.1.
ence, v. 290-A, p. 35–45. Vargas, M., Kashefi, K., Blunt-Harris, E.L., and Lovley, D.R., 1998, Micro-
Nealson, K.H., and Little, B., 1997, Breathing manganese and iron: Solid-state biological evidence for Fe(III) reduction on early earth: Nature, v. 395,
respiration: Advances in Microbiology, v. 45, p. 213–239. p. 65–67, doi: 10.1038/25720.
Nee, S., 2004, More than meets the eye: Nature, v. 429, p. 804–805, doi: Waksman, S.A., 1916, Bacterial numbers in soils, at different depths, and in
10.1038/429804a. different seasons of the year: Soil Science, v. 2, p. 331–343.
Nisbet, E., 2000, The realms of Archaean life: Nature, v. 405, p. 625–626, doi: White, A.F., and Peterson, M.L., 1996, Reduction of aqueous transition metal
10.1038/35015187. species on the surfaces of Fe(II)-containing oxides: Geochimica et Cosmo-
Noffke, N., Hazen, R., and Nhleko, N., 2003, Earth’s earliest microbial mats in chimica Acta, v. 60, p. 3799–3814, doi: 10.1016/0016-7037(96)00213-X.
siliciclastic marine environment (2.9 Ga Mozaan Group, South Africa): Widdel, F., Schnell, S., Heising, S., Ehrenreich, A., Assmus, B., and Schink,
Geology, v. 31, p. 673–676, doi: 10.1130/G19704.1. B., 1993, Ferrous iron oxidation by anoxygenic phototrophic bacteria:
Pedersen, K., 1997, Microbial life in deep granite rock: FEMS Microbiology Nature, v. 362, p. 834–836, doi: 10.1038/362834a0.
Reviews, v. 20, p. 399-414, doi: 10.1016/S0168-6445(97)00022-3. Winogradsky, S., 1888, Ueber Eisenbacterien: Botanische Zeitschrift, v. 46,
Queller, D.C., 2004, Kinship is relative: Nature, v. 430, p. 975–976, doi: p. 261–270.
10.1038/430975a. ZoBell, C.E., 1943, The effect of solid surfaces upon bacterial activity: Journal
Rasmussen, B., 2000, Filamentous microfossils in a 3,235-million-year-old of Bacteriology, v . 46, p. 39–56.
volcanogenic massive sulphide deposit: Nature, v. 405, p. 676–679, doi: ZoBell, C.E., 1952, Bacterial life at the bottom of the Philippine Trench: Sci-
10.1038/35015063. ence, v. 115, p. 507–508.
Reid, R.P., Visscher, P.T., Decho, A.W., Stolz, J.F., Bebout, B.M., Dupraz, C.,
Macintyre, I.G., Paerl, H.W., Pinckney, J.L., Prufert-Bebout, L., Steppe,
T.F., and DesMarais, D.J., 2000, The role of microbes in accretion, lami-
nation and early lithification of modern marine stromatolites: Nature,
v. 406, p. 989–992, doi: 10.1038/35023158. MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA


Memoir 198
2006
Oxygen isotope composition of hematite and genesis

of high-grade BIF-hosted iron ores
J. Gutzmer*
Paleoproterozoic Mineralization Research Group, Department of Geology, University of Johannesburg,
P.O. Box 524, Auckland Park 2006, South Africa
J. Mukhopadhyay
Paleoproterozoic Mineralization Research Group, Department of Geology, University of Johannesburg, P.O. Box 524, Auckland
Park 2006, South Africa, and Department of Geology, Presidency College, 86/1 College Street, Calcutta, India, 700073
N.J. Beukes
Paleoproterozoic Mineralization Research Group, Department of Geology, University of Johannesburg,
P.O. Box 524, Auckland Park 2006, South Africa
A. Pack
Department of Earth and Planetary Sciences, University of New Mexico, Albuquerque, New Mexico 87131-0001, USA,
and CRPG/CNRS Centre de Recherches Petrographiques et Geochimiques, 15 Rue Notre Dame des Pauvres,
54501 Vandoeuvre-les-Nancy, France
K. Hayashi
Department of Mineralogy, Petrology and Economic Geology, Graduate School of Science, Tohoku University,
Aoba-ku, Sendai 980-8578 Japan
Z.D. Sharp
Department of Earth and Planetary Sciences, University of New Mexico, Albuquerque, New Mexico 87131-0001, USA
ABSTRACT
High-grade BIF-hosted iron ore deposits are widely believed to have formed by
epigenetic residual enrichment of hematite at the expense of other constituents, most
notably chert. Processes responsible for the enrichment to high-grade iron ores are,
however, only poorly understood and a range of metallogenetic models have been
proposed. Field relationships have been used to distinguish three major groups of
BIF-hosted high-grade iron ore deposits, namely deposits of ancient supergene,
hydrothermal, and supergene-modified hydrothermal origin. Iron ores from all three
deposit types are essentially composed of hematite; among different morphological
types of hematite, microcystalline platy hematite and martite predominate.
In this contribution, the oxygen isotope geochemistry of ore-forming hematite
and martite from several high-grade iron ore deposits is examined, in an attempt to
differentiate deposits of hydrothermal origin from those formed in ancient supergene
environments. The δ18O composition of martite and microplaty hematite from depos-
*[email protected]
Gutzmer, J., Mukhopadhyay, J., Beukes, N.J., Pack, A., Hayashi, K., and Sharp, Z.D., 2006, Oxygen isotope composition of hematite and genesis of high-grade
BIF-hosted iron ores, in Kesler, S.E., and Ohmoto, H., eds., Evolution of Early Earth’s Atmosphere, Hydrosphere, and Biosphere—Constraints from Ore Deposits:
Geological Society of America Memoir 198, p. 257–268, doi: 10.1130/2006.1198(15). For permission to copy, contact [email protected]. ©2006 Geological
Society of America. All rights reserved.
257
258 J. Gutzmer et al.
its presumed to be of hydrothermal origin range from +0.9‰ to –7.3‰. Microcrystal-

line platy hematite from high-grade ores of ancient supergene origin, in contrast, has
δ18O values ranging between +2.0‰ and –3.9‰. The latter range overlaps with the
range that is defined by hematite and magnetite from weakly metamorphosed Arche-
an–Paleoproterozoic BIF (+5‰ to −4‰). The results obtained for ancient supergene
deposits developed along the 2.2 Ga Gamagara-Mapedi unconformity strengthen the
argument that the Paleoproterozoic atmosphere-hydrosphere-lithosphere system was
very similar to that of modern Earth. The marked shift to negative δ18O values dis-
played by hematite and martite from hydrothermal iron ore deposits, on the other
hand, provides support for the suggestion that aqueous fluids of shallow crustal origin
were responsible for the hydrothermal enrichment of banded iron formations to high-
grade iron ores.
Keywords: oxygen isotope geochemistry, hematite, martite, iron ore deposits, BIF.
INTRODUCTION From field geological evidence, Beukes et al. (2002a) rec-

ognized three major groups of BIF-hosted high-grade iron ore
Most world-class high-grade iron ore deposits (Fig. 1) are deposits, namely deposits of ancient supergene, hydrothermal
the product of epigenetic enrichment of Precambrian banded and supergene-modified hydrothermal origin. Type examples for
iron formation (BIF). Despite the economic importance of this ancient supergene hematite iron ore deposits are those hosted by
deposit type, processes responsible for the enrichment of BIF the ca. 2.45 Ga iron formations of the Asbestos Hills Subgroup of
(containing 30–40 wt% Fe) to high-grade iron ores with 60–67 the Transvaal Supergroup in Griqualand West, South Africa (Fig.
wt% Fe are poorly understood. Different models, ranging from 1). High-grade iron ores that constitute these deposits are thought
syngenetic and diagenetic (King, 1989) to hydrothermal (Powell to have formed as products of deep lateritic weathering (Gutzmer
et al., 1999; Barley et al., 1999; Taylor et al., 2001) and ancient and Beukes, 1998) immediately below the ca. 2.2 Ga Gamag-
supergene (Morris, 1980, 1985; van Schalkwyk and Beukes, ara-Mapedi unconformity. The high-grade ore grades from the
1986) or modern supergene (MacLeod, 1966), have been sug- unconformity down into oxidized but not markedly enriched iron
gested. Uncertainties about the origin of high-grade hematite formation. The largest and economically most important exam-
deposits are largely due to the monomineralic composition of the ples for ancient supergene enriched iron ores are the Sishen and
ores. They are almost exclusively composed of hematite, a min- Beeshoek deposits (Fig. 2A). The ores are typically overlain by
eral with a wide stability field and simple chemical composition red beds that contain conglomerates composed of hematite ore
that reveals little about its origin. Furthermore, later deformation and oxidized BIF pebbles (Fig. 2A).
and geologically recent deep chemical weathering overprint and In contrast, high-grade iron ores of hydrothermal origin, of
obscure many of the primary ore characteristics. which hard hematite ores of Thabazimbi, South Africa (Fig. 2B),
Figure 1. Location map illustrating the

Krivoj Rog distribution of important BIF-hosted
high-grade hematite ore deposits. The
Noamundi
diameter of the circle shows relative
Dalli-Rajhara size of the iron ore resource hosted by
Mt Nimba Donimalai the deposit/district. The Rooinekke and
Zeekoebaart deposits, which are close
Carajas to the Sishen-Beeshoek deposits in the
QF Northern Cape Province of South Af-
Urucum rica, are not displayed on this map be-
Hamersley
District cause they are currently of no economic
significance. Abbreviations: QF—Iron
Sishen-Beeshoek Thabazimbi Quadrangle, Minas Gerais, Brazil.
A Sishen/Beeshoek deposits C
N4E Mine Section 600N (Carajas)
V
V
V V V V V
V V
V V V V V
Al Al Al V
Al V V
V V
Al V V V
Al
Al V V V
V
V
V V
V V V 200 m
~50 to 200 m
V
V V f V
V
V V V
V V
V V V V
Surface laterite and canga Goethitic ore

V V Lava Friable ore
Friable iron-formation
Unaltered iron-formation
Hard ore
Carbonate-hematite rock
~200 to 500 m V Weathered lava
Manganore Iron
Gamagara Formation
Unconformity
] Formation
D
Al
Red and cream shale
Transgressive quartzite
Al-rich shale
Laminated ore
Cherty iron formation ] W Noamundi West Mine E
Wolhaarkop breccia
Conglomeratic ore
Dissolution surface
Campbellrand dolomite
100 m
200 m
Laterite & Canga Goethite-cemented

hard ore
Blue dust ore Weathered BIF
Friable flaky ore BIF
B Hard ore Shale
Donkerpoort West, Thabazimbi
Cretaceous-Tertiary African Land Surface
Present day surface Figure 2. (A) Schematic diagram illus-

trating the setting of ancient supergene
hematite ores in karstic solution col-
lapse structures at the Sishen and Bee-
shoek deposits, South Africa. (B–D)
Schematic cross sections of high-grade
hematite iron ore bodies: (B) Donker-
poort West, Thabazimbi; (C) N4E de-
posit, Carajas, Brazil (source CVRD,
Diabase Carajas); (D) Noamundi West deposit,
Hematite ore India. All cross sections modified from
Beukes et al. (2002a).
Carbonate ore
BIF
Shale
0m Malmani dolomite
100 m
and Mount Tom Price, Mount Whaleback, Paraburdoo, and New- mal origin that were investigated during this study. In contrast,
man in the Hamersley District of Western Australia (Fig. 1) are high-grade iron ores of the Sishen and Rooinekke deposits, both
type examples, are not associated with any erosional unconfor- presumably of ancient supergene origin, are predominantly com-
mity. They are typically located at the base of the hosting iron- posed of microplaty hematite, with only minor amounts of mar-
formation succession, from which they grade up into unmineral- tite. Martite from all the studied deposits contains tiny remnants
ized iron formation. In many of these deposits structural controls of magnetite and kenomagnetite. Trace amounts of cryptocrystal-
on ore distribution, such as faults, sills, or dykes, have been rec- line goethite are present in samples from the Noamundi, Mount
ognized (Taylor et al., 2001; Beukes et al., 2002a). Tom Price and Paraburdoo deposits (Fig. 1).
Supergene-modified hydrothermal iron ores, such as the Patchy hematite (Fig. 3B) and specular hematite (Figs. 3A,
giant deposits of the Carajas District in Brazil (Fig. 2C) or the 3D) are much less abundant than microcrystalline platy hematite
Noamundi deposit in India (Fig. 2D), are characterized by the and martite. Patchy hematite, characterized by its anhedral shape
presence of friable hematite ores derived from geologically and grain sizes not exceeding 50 μm, is thought to have formed
recent supergene enrichment of carbonate-metasomatized iron by recrystallization of microplaty hematite or martite. Specular
formation (Beukes et al., 2002a). Typically, these friable ores are hematite is distinctly coarser grained than the other three mor-
closely associated with hard, non-friable high-grade hematite ore phological types of hematite, with grain sizes that may exceed
bodies of hydrothermal origin. 1 mm (Fig. 3D). Crystals are often developed as thin blades that
Here, we report the results of an investigation into the oxy- may be arranged in radial sheaves or massive aggregates. The
gen isotope geochemistry of hematite from several high-grade formation of specular hematite always postdates ore formation;
iron ore deposits, in an attempt to differentiate deposits of hydro- it is restricted to the infill of vugs or veinlets that crosscut pre-
thermal origin from those formed in ancient supergene envi- existing iron ore.
ronments. Samples from eight deposits, hosted by banded iron On the basis of mesoscopic appearance (porosity and rem-
formations of Archean to Paleoproterozoic age, were studied. nant sedimentary lamination) and petrographic features (pres-
Two of these deposits (Sishen, Rooinekke) are considered to be ence and abundance of morphological types of hematite), hard
of ancient supergene origin, the other six (Thabazimbi, Mount iron ores from all studied deposits have been grouped into mas-
Tom Price, Paraburdoo, Noamundi, Zeekoebaart, and Carajas) sive martite, massive martite-microplaty hematite, laminated
are thought to be of hydrothermal origin (Beukes et al., 2002a). microplaty hematite, and conglomeratic ore.
The apparent lack of a significant metamorphic overprint on all Massive martite ore is defined by massively intergrown
of these deposits was an important selection criterion, to avoid martite with minor intergranular porosity filled by microplaty
the effects of metamorphic recrystallization and resetting of oxy- hematite. This ore type was identified at Carajas, Thabazimbi,
gen isotope compositions as described by Hoefs et al. (1982). Noamundi, and Paraburdoo. It appears to have originated as
We restricted our investigation to hard ores that we expect to large masses of magnetite that were subsequently transformed
be either exclusively hydrothermal or supergene in origin, and into martite. Massive martite–microplaty hematite ore is com-
avoided friable ores that are likely to have a supergene-modified posed of approximately equal amounts of microplaty hematite
hydrothermal origin. and martite with conspicuous porosity. It was encountered at
Mount Tom Price, Thabazimbi, and Carajas. In contrast to the
PETROGRAPHY AND CLASSIFICATION OF THE ORES more massive ores, laminated microplaty hematite ore often
displays microbanding defined by the regular intercalation of
The high-grade iron ores from the studied deposits are millimeter-thin laminae of more densely and less densely inter-
essentially composed of hematite, with only trace amounts of grown microplaty hematite, with open pore space between the
magnetite, kenomagnetite, goethite, quartz, carbonate (calcite, hematite platelets. This ore type constitutes a large part of the
dolomite), and apatite. Four morphological types of hematite ore reserves at Sishen, Rooinekke, and Noamundi. Conglom-
are present in the ores, namely microcrystalline platy hematite eratic ore is restricted to the Sishen and Beeshoek deposits in
(Figs. 3A, 3C), martite (Figs. 3B, 3C), patchy hematite (Fig. South Africa (van Schalkwyk and Beukes, 1986). Lenses of
3B), and specular hematite (Figs. 3A, 3D). Microcystalline platy conglomeratic ore are developed along the Gamagara-Mapedi
(microplaty) hematite and martite predominate in all deposits. unconformity and are composed of hematite pebbles derived
The microplaty hematite occurs as subhedral to euhedral plates, from the underlying laminated microplaty hematite ores. The
typically less than 25 μm in diameter (Fig. 3C). Martite is closely hematite ore pebbles are hosted by a matrix of microplaty
associated with the microplaty hematite in the form of euhedral to hematite and some specularite.
subhedral magnetite octahedra (<100 μm in size) that are partly
or completely replaced by a porous scaffold of hematite laths ANALYTICAL METHODS
(Fig. 3C) (Ramdohr, 1980). In all of the deposits thought to be of
hydrothermal origin, martite is much more abundant than magne- Samples of different ore and deposit types were taken either
tite in the iron-formation protolith. Indeed, martite predominates from drill core or deep open-pit exposures, far below the zone of
in all hard iron ores from the six deposits of presumed hydrother- modern weathering. Oxygen isotope analyses were conducted at
Oxygen isotope composition of hematite and genesis of high-grade BIF-hosted iron ores 261
A B
1 mm 50 μm
C D
50 μm 150 μm
Figure 3. Reflected light photomicrographs illustrating different morphological types of hematite. (A) Laminated microcrystalline platy hema-
tite ore from the Sishen deposit. Note the apparent dissolution collapse brecciation. The breccia clasts are composed of microcrystalline platy
hematite and minor hematite; interstices are filled by a porous mass of specularite with trace amounts of quartz (dark gray). (B) Massive martite
ore from the Noamundi West deposit (India). Note the characteristic cross-hatch extinction pattern of martite that occurs as poorly defined grains
intimately intergrown with patchy hematite (crossed nicols, oil immersion). (C) Martite grain with well-preserved octahedral shape, intergrown
with microplaty hematite. Massive martite-microcrystalline platy hematite ore, Noamundi West deposit (note different shades of gray, which
are due to anisotropy of hematite under crossed nicols, oil immersion). (D) Coarse crystalline specularite sheaves in vein crosscutting laminated
microcrystalline platy hematite ore from the Sishen deposit (South Africa).
the Department of Earth and Planetary Sciences, University of overnight in order to remove impurities and moisture from the
New Mexico, USA, and at the Department of Mineralogy, Petrol- surfaces. Samples were heated and reacted in a BrF5 atmosphere
ogy and Economic Geology, Tohoku University, Japan. using a CO2 laser with IR emission at 10.6 μm. Excess BrF5 was
At the University of New Mexico, oxygen isotope analy- frozen on a cold trap cooled with liquid nitrogen. Traces of F2 were
ses were conducted by the modified fluorination technique as eliminated by reaction with hot KCl. Chlorine gas was removed
described by Sharp (1990). Monomineralic hematite grains, from the analyte O2 gas by freezing in a cold trap. Sample oxygen
between 1 mg and 2 mg in weight, were loaded along with stan- gas was collected on two successive 13X molecular sieve cold
dards in a Ni-sample holder and placed in the reaction cham- fingers and released into the dual inlet system of a Finnigan-MAT
ber of the laser fluorination line. The chamber was covered with Delta Plus XL mass spectrometer. All samples are reported rela-
an IR-transparent BaF2 window. Samples were pre-fluorinated tive to standard mean ocean water (VSMOW), defined such that
NBS-28 quartz has a δ18O value of 9.6‰. Calibrated in-house TABLE 1. COMPILATION OF G18OVSMOW SIGNATURES OF
standards were routinely run with each set of analyses. Repro- ORE-FORMING HEMATITE FROM SELECTED HIGH-GRADE
BANDED IRON FORMATION–HOSTED IRON ORE DEPOSITS
ducibility was found to be better than 0.15‰ (1σ).
Sample no. Description G18OVSMOW Lab
At Tohoku University, samples were analyzed both in situ
Paraburdoo (Hamersley District, Australia)
on polished chips and by complete evaporation of 1–2 mg ore PB-1-7 Martite –5.0 1
chips that were previously separated from a larger specimen. PB-1-8 Martite –3.7 1
Oxygen was extracted using a 0–17 Watt JLC-200 CO2-laser PB-1-9 Martite –4.7 1
(λ = 10.6 μm) and BrF5 by a fluorination method similar to PB-1-10 Martite –6.4 1
Mount Tom Price (Hamersley District, Australia)
the technique used at the University of New Mexico. The opti-
MTP-1-1 Martite + Microplaty hematite –4.4 1
cal system to introduce the laser beam into the stainless steel MTP-1-2 Martite + Microplaty hematite –4.3 1
reaction chamber through a BaF2 window is coaxial with a MTP-1-3 Martite + Microplaty hematite –3.8 1
reflected microscope, which enables accurate in situ laser abla- MTP-1-4 Martite + Microplaty hematite –5.1 1
tion. The diameter of the focused laser beam is ~100 μm; how- MTP-1-5 Martite + Microplaty hematite –5.0 1
ever, the area of thermal halo affected by laser ablation is ~200 MTP-1-6 Martite + Microplaty hematite –4.6 1
MTP-2-11 Martite + Microplaty hematite –4.7 1
μm. Therefore, oxygen extracted by CO2-laser ablation from Noamundi (India)
a single spot is generated from the area of ~200 μm. Oxygen IO-20-1 Microplaty hematite –2.2 1
was converted to CO2 quantitatively by the reaction with a IO-20-2 Microplaty hematite –2.6 1
heated diamond. The 18O/16O ratios are measured by a Finni- IO-20-3 Microplaty hematite –1.3 1
gan-MAT252 mass spectrometer. The accuracy of in situ CO2- IO-20-5 Microplaty hematite –2.4 1
IO-10-1 Martite –3.9 1
laser fluorination oxygen isotope analyses at Tohoku University
IO-16C-7 Martite –6.7 1
was examined by the repeated analyses of working standards IO-16C-8 Martite –5.7 1
(Hayashi et al., 2001). The precision of measured δ18O of this IO-16C-10 Martite –4.6 1
study is ± 0.3‰ (1σ) per mil. IO-16C-11 Martite –2.6 1
All data are expressed in per mil (‰) relative to VSMOW. IO-4A-5 Martite –4.8 1
IO-4A-6 Martite –5.0 1
A total of 87 samples of hematite and martite from a total of
IO-4A-7 Martite –4.4 1
eight deposits were analyzed (Table 1). Oxygen isotope analy- IO-4A-9 Martite –5.8 1
ses of specular hematite were obtained only from the Thaba- N4E Mine, Carajas (Brazil)
zimbi and Sishen deposits, South Africa (Table 1). A summary BH529-120-7 Martite –4.0 1
of the data is presented in Table 2. BH529-120-6 Martite –5.4 1
The high-grade iron ores studied here typically contain BH529-120-5 Martite –6.0 1
BH529-120-4 Martite –4.4 1
only trace amounts of gangue minerals, including quartz, car- BH529-90-3 Martite –5.7 1
bonate, and apatite. These trace minerals generally occur finely BH529-90.0-2 Martite –7.4 2
disseminated and are too fine-grained to permit isotopic analy- BH529-90-1 Martite –4.4 1
ses. The occurrence of megaquartz is restricted to rare veins that BH-384-12 Martite –6.5 1
are either associated with paragenetically late specularite or in BH384-160 Martite –6.7 2
BH541-177.5 Martite –7.0 2
an uncertain paragenetic relationship to ore-forming event(s). Thabazimbi (RSA)
Coarser-grained, sparitic carbonates (calcite, dolomite) associ- BH-607-4 Martite –4.1 1
ated with high-grade iron ores at Thabazimbi and Carajas show BH-607-3 Martite –3.4 1
textural evidence suggesting that they are coeval with micro- BH-607-2 Martite –4.7 1
platy hematite and therefore part of the ore-forming event. BH-607-1 Martite + Microplaty hematite –1.5 1
BH2099 Martite –2.7 1
Detailed fluid inclusion studies and stable isotope analyses car-
BH-1706-89.5 Martite –6.0 2
ried out on sparitic carbonates from Thabazimbi, which are not DON-383-159 Specularite +3.4 2
presented here, confirm the cogenetic hydrothermal origin of DON-396-215 Martite –5.6 2
these minerals (Gutzmer et al., 2001a; Netshiozwi, 2002). DKP442-358 Martite +0.9 2
KW 309-89.92 Martite –3.4 2
BOB#3 Specularite –1.4 2
RESULTS
BE#12 Martite –0.3 2
BW#3 Martite + Microplaty hematite –1.9 2
The δ18O composition of martite-textured hematite (referred VDB#2 Specularite –2.4 2
to below as martite) from the different deposits ranges from KF-83-9 Martite + Microplaty hematite –3.7 2
+0.9‰ to −7.3‰, with an average of −4.8‰ (Tables 1, 2). Com- (continued)
pared to massive martite ores, ores composed of martite and
microplaty hematite are enriched in 18O, with δ18O ranging from
+2.0‰ to −5.1‰, with an average of −1.6‰ (Table 2). Still more
enriched in 18O is specularitic hematite hosted in veins that cross-
TABLE 1. COMPILATION OF G18OVSMOW SIGNATURES OF TABLE 2 SUMMARY OF OXYGEN ISOTOPE COMPOSITION

ORE-FORMING HEMATITE FROM SELECTED HIGH-GRADE OF HEMATITE FROM SELECTED HIGH-GRADE
BANDED IRON FORMATION–HOSTED IRON ORE DEPOSITS (continued) BANDED IRON FORMATION–HOSTED IRON ORE DEPOSITS
Deposit/district Textural type N Range Average
Sample no. Description G18OVSMOW Lab
Carajas (Brazil) Martite 10 –4.0 to –7.3 –5.6
Zeekoebaart (RSA)
Mt Tom Price (Hamersley) Martite + Microplaty 7 –3.8 to –5.1 –4.9
Baart-1 Martite –6.1 2 Paraburdoo (Hamersley) Martite 4 –3.7 to –6.4 –4.9
Rooinekke (RSA) Noamundi (India) Martite 9 –2.6 to –6.7 –4.3
Rooinekke#1a Microplaty hematite –2.3 2 Microplaty 4 –1.3 to –2.6 –2.1
Rooinekke#1b Microplaty hematite –1.2 2 Thabazimbi (RSA) Martite 9 +0.9 to –6.0 –3.9
Rooinekke#2 Microplaty hematite –1.5 2 Martite + Microplaty 2 –1.5 to –1.9 –1.7
Sishen (RSA) Specularite 3 +3.4 to –2.4 +0.1
WP-CL-1 Microplaty hematite (conglomerate clast) –2.3 1 Zeekoebaart (RSA) Martite 1 –6.1 –6.1
Sishen (RSA) Microplaty 28 +2.0 to –3.9 –1.5
WP-CL-3 Microplaty hematite (conglomerate clast) –3.5 1
Specularite 5 +3.0 to 0.5 +1.9
Rooinekke (RSA) Microplaty 3 –1.2 to –2.3 –1.7
WP-MT-3 Microplaty hematite (conglomerate matrix) +0.2 1
WP-MT-2 Microplaty hematite (conglomerate matrix) –1.9 1
WP-MT-1 Microplaty hematite (conglomerate matrix) +0.2 1 values ranging between +0.9‰ and –7.3‰ (Fig. 4A). δ18O val-
SHV-3-8 Specularite +2.0 1 ues below –4‰ were observed in all six deposits considered to
SHV-3-7 Specularite +2.7 1 be of hydrothermal origin, but are conspicuously absent from the
SHV-PT-3 Specularite +1.0 1 two deposits deemed to be of ancient supergene origin. Results
SHV2-PT-0 Specularite +3.0 1
obtained in this study are in excellent agreement with similar
SHV-3-9 Microplaty hematite –3.9 1
1975-175-5 Microplaty hematite –1.1 1
data reported by Powell et al. (1999) for hematite in microplaty
1975-189-1 Microplaty hematite –2.3 1 hematite ore from Mount Tom Price (–1‰ to –4‰, n = 4) and by
1975-189-2 Microplaty hematite –2.9 1 Bird et al. (1993) for hard hematite ore (−5.0‰, n = 1) and soft
1975-199-3 Microplaty hematite –2.7 1 hematite ore (−5.0‰ to –6.5‰, n = 4) from the N4E deposit in
the Carajas District.
1975-238-12 Microplaty hematite –2.5 1 DISCUSSION
1772#3 Microplaty hematite (conglomerate clast) +2.0 2
1772#5b Specularite +0.5 2 Introduction
1772#5c Microplaty hematite –0.1 2
(coating around chert pebble)
1772#7 Microplaty hematite (conglomerate clast) +0.9 2 There is growing consensus that high-grade iron ore bodies
ADR#15 Microplaty hematite (conglomerate clast) –2.1 2 hosted by banded iron formations form by epigenetic residual
4552#9b Microplaty hematite (conglomerate clast) –0.5 2 enrichment of iron at the expense of other constituents (Barley
4552#9d Microplaty hematite (conglomerate clast) +0.4 2 et al., 1999; Taylor et al., 2001; Beukes et al., 2002a). Residual
4225#9a Microplaty hematite –0.7 2
(coating around chert pebble) enrichment of iron is attributed to effective leaching of SiO2 and
4225#9e Microplaty hematite (conglomerate clast) –0.8 2 other constituents present in the protolith BIF by either meteoric
3226#3 Microplaty hematite –0.9 2 or hydrothermal fluids. The BIF protolith is an intimate and fine-
3226#4b Microplaty hematite +1.3 2 grained mixture of quartz/chert and iron oxides (hematite and
3226#12 Microplaty hematite (conglomerate clast) –2.3 2
magnetite), as well as Fe2+-rich carbonates (ankerite, siderite) and
Note: All data are expressed in per mil relative to Vienna standard mean
ocean water (VSMOW). silicates (e.g., greenalite, grunerite). Ferrous iron hosted in mag-
Data source/Lab: 1—Dept. of Mineralogy, Tohoku University, Japan; 2—Dept. netite, carbonates, or silicates is oxidized to form hematite, i.e.,
of Earth Sciences, University of New Mexico, USA.
ferric iron oxide, during the enrichment process. The predomi-
nance of martite in samples from some high-grade hematite ore
deposits suggests that the oxidation of carbonates and silicates to
cut the high-grade ore; the latter values range from +3.4‰ to hematite proceeds via magnetite as an intermediate product.
−2.4‰, with an average of +0.4‰ (Table 2). Little is known about the fluids involved in the enrichment
There is a very obvious difference in the results obtained for process. For the ancient supergene deposits of Sishen and Roo-
hard high-grade hematite ores from deposits that have been clas- inekke, descending meteoric water may have driven the altera-
sified by Beukes et al. (2002a) as ancient supergene in origin, i.e., tion process. Hydrothermal fluids that might have been respon-
Sishen and Rooinekke, and those presumed to be of hydrother- sible for the formation of high-grade BIF-hosted iron ore have
mal origin. Microcrystalline platy hematite from high-grade ores been characterized only for Mount Tom Price (Hagemann et al.,
from Sishen and Rooinekke has δ18O values ranging between 1999; Taylor et al., 2001) and Thabazimbi (Netshiozwi, 2002).
+2.0‰ and –3.9‰ (Fig. 4A). Compared to this, samples from Hagemann et al. (1999) identified three compositionally differ-
deposits of hydrothermal origin are depleted in 18O, with δ18O ent aqueous fluid inclusion types in megaquartz-hematite veins
A δ 18O VSMOW (in ‰)

-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4
Thabazimbi N=11
Supergene Hydrothermal
Hamersley N=11
Carajas N=10
Noamundi N=13
Zeekoebaart N=1
Rooinekke N=3
Sishen N=28
error bar
N=number of samples
δ 18O VSMOW (in ‰)

B -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5
Age (Ga)
3.8 N=number of samples Hematite

Magnetite + hematite
3.4
3.0 Weld Range N=10
N=60
Dales Gorge Mb.,
2.6
Brockman IF N=4
Kuruman N=8
2.2 N=4 (Perry & Bonnichsen, 1966)
Biwabik
N=10 (Perry et al., 1973)
1.8
Hydrothermally
Diagenetic to low-grade metamorphic BIF
altered BIF
Figure 4. (A) Oxygen isotope composition of hematite from high-grade iron ores in selected deposits. Division of deposits studied into those
of presumably hydrothermal origin and those of presumably ancient supergene origin is based on field geological relationships as described by
Beukes et al. (2002a). Values obtained for specularitic hematite are not displayed. (B) Oxygen isotope composition of iron oxides (hematite,
magnetite) in low-grade metamorphic BIF of Archean-Paleoproterozoic age. Data sources: Weld Range (Perry and Ahmad, 1983); Dales Gorge
Member (Becker and Clayton, 1976); Kuruman Iron Formation (Perry and Ahmad, 1983); Biwabik Iron Formation (Perry and Bonnichsen,
1966; Perry et al., 1973). Data for Mount Tom Price and Paraburdoo are combined and displayed as Hamersley.
at the Mount Tom Price deposit, including a low-salinity NaCl- ing fluid. A detailed investigation by Netshiozwi (2002) revealed
dominated fluid type (salinity [s] = 0.7–3.3 wt% NaClequivalent, the presence of two fluid end members, a high-salinity brine (s
homogenization temperature TH = 138–190 °C), a moderately = 27 wt% NaClequivalent) and a fluid of moderate to low salinity
saline NaCl-dominated fluid (s = 12.3–14.1 wt% NaClequivalent, TH (s = 7–10 wt% NaClequivalent), with homogenization temperatures
= 130–209 °C), and a CalCl2-rich brine (s = 18–25 wt% CaCl2). ranging from 120 ºC to 180 ºC for both fluids. The results from
However, megaquartz veins studied by Hagemann et al. (1999) Thabazimbi compare well with those obtained by Hagemann et
are spatially but not necessarily genetically associated with the al. (1999) at Mount Tom Price, suggesting that hydrothermal
formation of the high-grade iron ore. Textural evidence available enrichment of iron-formation protolith to high-grade hematite
for sparitic dolomite and calcite intimately associated with high- iron ore is the result of interaction of iron formation with water-
grade hematite ore at Thabazimbi, on the other hand, suggests rich fluid mixtures at moderate temperatures (Barley et al., 1999;
that these carbonates formed cogenetic with ore-forming micro- Hagemann et al., 1999; Netshiozwi, 2002).
platy hematite (Gutzmer et al., 2001a). The study of inclusions It is likely that the oxygen isotope composition of fine-
within these carbonates should thus characterize the ore-form- grained ore-forming hematite/ martite represents a mixture
between the oxygen isotopic composition of iron oxides that that had undergone intensive hydrothermal alteration. Similarly,
were present in the BIF protolith and oxygen introduced by the the detailed petrographic descriptions by Gole (1980) of Weld
ore-forming fluid. Evaluating the significance of the δ18O values Range BIF suggest widespread hydrothermal alteration as indi-
of ore-forming hematite thus requires knowledge of the oxygen cated by the presence of, for example, chalcopyrite and arseno-
isotopic composition of iron oxide in the protolith and of the pyrite. It thus appears reasonable to attribute the presence of iron
aqueous fluid responsible for enrichment, as well as the tempera- oxides with unusually low δ18O values to hydrothermal alteration,
ture of ore-formation and temperature-dependent oxygen isotope and to exclude them from the expected range of isotopic compo-
fractionation between water and iron oxide minerals. sitions of magnetite and hematite in weakly metamorphosed BIF
of Paleoproterozoic and Archean age.
Oxygen Isotope Geochemistry of Iron Oxides in BIF
Protolith Temperature-Dependent Isotope Fractionation
Hoefs (1992) provided a comprehensive compilation of δ18O The temperature-dependent isotope fractionation between
values of iron oxides (magnetite, hematite) in banded iron forma- iron oxides and water, especially at low temperature, has been at
tions. Data are available for only four weakly metamorphosed the center of much debate (see review by Yapp, 2001). However,
iron formations of Archean and Paleoproterozoic age, namely for our purpose it is important to note that similar, though not
the ca. 3.0 Ga Weld Range Iron Formation of the Yilgarn Cra- identical, temperature-dependent oxygen isotope fractionation
ton (Perry and Ahmad, 1983), the 2.45 Ga Dales Gorge Member curves for the hematite-water system were obtained experimen-
of the Brockman Iron Formation (Becker and Clayton, 1976), tally by Yapp (1990) and Bao and Koch (1999), and that these
the 2.45 Ga Kuruman Iron Formation of the Transvaal Super- curves correspond reasonably well to the curve constructed by
group (Perry and Ahmad, 1983), and the 1.9 Ga Biwabik Iron Clayton and Epstein (1961) from natural assemblages. In this
Formation of the Superior Craton (Perry and Bonnichsen, 1966; contribution, we have chosen the fractionation curve experimen-
Perry et al., 1973). Oxygen isotopic compositions of fine-grained tally determined by Yapp (1990) in the temperature range 20 ºC
magnetite and hematite of these iron formations range from +5‰ to 120 ºC and extrapolated it to 300 ºC (Fig. 5). The resultant
to −8‰ (Fig. 4B). This range, together with equivalent data for curve shows an important crossover point at ~100 ºC. At tem-
coexisting quartz, is thought to represent a disequilibrium array peratures below this point, hematite precipitating in isotopic
indicative of open-system fluid-rock interaction and recrystalli- equilibrium from an aqueous solution should be enriched, and
zation of iron oxide minerals during diagenesis, and low-grade above this point it should be depleted in 18O, relative to the parent
metamorphism (Gregory, 1986; Hoefs, 1992). solution (Fig. 5).
We critically evaluated the original data sets available.
From mineralogical and petrographical descriptions of the mate- Hydrothermal Hematite Formation
rial analyzed, and original descriptions of the geological setting
of the deposits that were sampled, we were able to define the δ18O values as low as –7.3‰ observed for hematite ores of
range of δ18O values of magnetite and hematite in very weakly hydrothermal origin (Table 1) suggest a depletion of at least 3‰
metamorphosed BIF to −4‰ and +5‰, a smaller range than that in δ18O compared to iron oxides in the BIF protolith (Fig. 4B).
reported by Hoefs (1992). Most of the data points in Figure 4B Fluid inclusion constraints available for the Thabazimbi (Netsh-
fall within the latter range, with the notable exception of four data iozwi, 2002) and Mount Tom Price deposits (Hagemann et al.,
points from the Weld Range (Perry and Ahmad, 1983) and three 1999; Taylor et al., 2001) suggest temperatures of ore formation
data points from the Biwabik Iron Formation (from the study in the range of 120 ºC to 209 ºC. No reliable pressure correction
of Perry et al., 1973). In their contribution, Perry et al. (1973) can be applied to these values, but because of the apparent lack of
state explicitly that iron oxides unusually depleted in 18O in the any significant regional metamorphic overprint recorded by the
Biwabik Iron Formation are restricted to a part of the succession host rock strata, it is unlikely that it would exceed 10 ºC.
20 A 1000lnα=(1.63x106/T2) - 12.3
Fractionation (1000 lnα)
B 1000lnα=(0.413x106/T2) - 2.56
15
Figure 5. Temperature-dependent oxygen isotope frac-
10 tionation curves for the system hematite-water. (A) was
determined experimentally by Yapp (1990) in the tem-
5 A
B
perature interval 20 °C to 120 °C; (B) was suggested
100 150 200 250 300 by Clayton and Epstein (1961) on the basis of natural
0
50 T (oC) assemblages in the temperature interval 25 °C to 120
-5 °C. Both curves are extrapolated to 300 °C.
-10 experimental interval extrapolated interval

If fluid inclusion homogenization temperatures reflect the tion and ore formation in the ancient supergene Rooinekke and
temperature of ore formation at the Thabazimbi and Mount Tom Sishen deposits obviously did not result in a significant shift in
Price deposits, estimates of the maximum 18O/16O ratio of the oxygen isotopic compositions of iron oxides. This is illustrated
ore-forming fluid is possible from the most 18O-depleted oxygen by the complete overlap between the range of δ18O values mea-
isotope compositions in hematite of the two deposits (−5.1‰ for sured for ore-forming hematite (+2.0‰ and –3.9‰, Fig. 4A) and
Mount Tom Price, −6.0‰ for Thabazimbi; Table 1). These val- the reference range for iron oxides from weakly metamorphosed
ues are regarded as the most fluid-dominated, because they are BIF (Fig. 4B). This apparent lack of isotopic resetting is despite
farthest removed from the δ18O values reported for iron oxide indications that isotopic equilibrium is usually attained between
in the BIF protolith. However, they are still likely to constitute a meteoric water and iron oxide precipitates in a range of modern
mixture of fluid-derived oxygen and oxygen sourced from pro- and ancient surface environments (Yapp, 2001).
tore iron oxides; i.e., calculations can be used only to derive a Recent investigations suggest that ancient supergene iron
maximum value for δ18O of the ore-forming fluid. The assess- ores along the Paleoproterozoic Gamagara-Mapedi unconfor-
ment is based on two further assumptions: (1) there is equilib- mity formed at low latitudes (Evans et al., 2002) and that lateritic
rium partitioning of oxygen isotopes, and (2) the temperature- weathering conditions prevailed (Gutzmer and Beukes, 1998).
dependent oxygen isotope fractionation curve of Yapp (1990) for Consequently, ore formation is attributed to residual enrichment
the hematite-water system remains valid at temperatures exceed- under lateritic weathering conditions. Assuming that equilibrium
ing 120 ºC (Fig. 5). isotopic fractionation was attained between meteoric water and
Calculations suggest that the ore-forming fluid at Thaba- neoformed iron oxides, the δ18O values of Paleoproterozoic mete-
zimbi had a maximum δ18O composition of −4.4‰ at 120 ºC, or oric water may be constrained (−3‰ to −8‰) from the isotopic
−1.7‰ at 180 ºC. At Mount Tom Price the value was −2.4‰ at composition of ore-forming hematite. This range corresponds
140 ºC, and +0.2‰ at 200 ºC. Because these values are estimates very closely to the isotopic composition of modern precipitation
of the maximum possible δ18O values of the ore-forming fluid, in low latitudes, tropical and subtropical environments (see com-
we conclude that the actual hydrothermal fluid must have been pilation by Fricke and O’Neil, 1999).
depleted in 18O relative to VSMOW. Such depletion is typical for Although only a very crude first-order approximation, this
shallow crustal fluids, i.e., fluids of meteoric origin that have not result suggests that the oxygen isotopic composition of Paleo-
interacted extensively with silicate rocks (Hoefs, 1997). This is proterozoic meteoric water was similar to modern values, and
in excellent agreement with the results of a fluid chemistry study possibly controlled by very similar processes. It is also in agree-
at Mount Tom Price indicating that the ore-forming fluids were ment with the conclusions of Godderies and Veizer (2000) and
similar in composition to those of modern basinal brines (Hage- Gutzmer et al. (2001b) that Paleoproterozoic ocean water had
mann et al., 1999). an oxygen isotopic composition very similar to modern ocean
It is also interesting to speculate on possible reasons for the water.
observation that martite-textured hematite is more depleted in
18
O than microplaty hematite from the same deposit (Table 2). CONCLUSIONS
This apparent difference could be explained by inherent differ-
ences in temperature-dependent isotope fractionation between The oxygen isotope geochemistry of ore-forming hema-
the systems magnetite-H2O and hematite-H2O. Limited informa- tite may be regarded as the first geochemical tool to differenti-
tion available on oxygen isotope fractionation in the magnetite- ate between high-grade BIF-hosted iron ore deposits of ancient
water system (Chacko et al., 2001) suggests that magnetite should supergene or hydrothermal origin. Compared to the oxygen iso-
be somewhat depleted in 18O compared to cogenetic hematite. tope composition of magnetite and hematite in the BIF protolith,
Unfortunately, we have no means to assess the possible effects hydrothermal hematite ores are depleted in 18O, a shift that is not
of martitization on the oxygen isotopic composition nor can we evident in hematite ores of ancient supergene origin. The oxygen
ensure that martite and microplaty hematite are formed cogeneti- isotope composition thus has the potential to become a tool to
cally. From petrographic evidence we do, however, assume that categorize high-grade iron ore deposits of unknown origin, espe-
paragenetically older magnetite may have formed at a higher cially in geologically complex or poorly exposed areas.
temperature, followed by martitization and microplaty hematite The 18O depletion of ore-forming hematite in deposits of
deposition. Martite, formed at the expense of magnetite, may hydrothermal origin provides support for the hypothesis that flu-
thus have inherited its 18O-depleted oxygen isotope composition ids of shallow crustal origin, such as brines and meteoric water,
from its magnetite precursor. are responsible for enrichment of banded iron formations to high-
grade iron ores. Results obtained for ancient supergene deposits
Ancient Supergene Ore Formation developed along the Paleoproterozoic Gamagara-Mapedi uncon-
formity in South Africa, in contrast, strengthen the argument that
Bird et al. (1993) proposed that prolonged chemical weath- the Paleoproterozoic atmosphere-hydrosphere-lithosphere sys-
ering should lead to resetting of the pre-existing isotopic compo- tem at that time was very similar to modern Earth (Beukes et al.,
sition of iron oxide minerals contained in BIF. However, oxida- 2002b). A more comprehensive evaluation of the oxygen isotope
geochemistry is at present hampered by the lack of constraints meteoric water and surface temperature: Its use in investigating terrestrial
climate change over geologic time: Earth and Planetary Science Letters,
regarding the temperature of ore formation and the nature of the v. 170, p. 181–196, doi: 10.1016/S0012-821X(99)00105-3.
fluids involved in the enrichment process. Godderies, Y., and Veizer, J., 2000, Tectonic control of chemical and isotopic
composition of ancient oceans: The impact of continental growth: Ameri-
can Journal of Science, v. 300, p. 434–461.
ACKNOWLEDGMENTS Gole, M.J., 1980, Mineralogy and petrology of very-low-metamorphic grade
Archaean banded iron-formations, Weld Range, Western Australia: Amer-
The authors acknowledge the support of geologists of ican Mineralogist, v. 65, p. 8–25.
Gregory, R.T., 1986, Oxygen isotope systematics of quartz-magnetite pairs
Kumba Resources of South Africa, CVRD of Brazil, Hamersley from Precambrian iron formations: Evidence for fluid-rock interaction
Iron of Australia, and TISCO of India for permitting access and during diagenesis and metamorphism, in Walther, J.V., and Wood, B.J.,
providing guidance during the visits to deposits. The Indo-South eds., Fluid-rock interactions during metamorphism: Heidelberg, Springer-
Verlag, Advances in Physical Geochemistry, v. 5, p. 132–153.
African Inter-Government Science and Technology Coopera- Gutzmer, J., and Beukes, N.J., 1998, Earliest laterites and possible evidence for
tion Programme, project no. 5/1/12/15, supported this research terrestrial vegetation in the Early Proterozoic: Geology, v. 26, p. 263–266,
project. Mukhopadhyay acknowledges a postdoctoral fellowship doi: 10.1130/0091-7613(1998)026<0263:ELAPEF>2.3.CO;2.
Gutzmer, J., Beukes, N.J., Netshiozwi, S., and Szabo, A., 2001a, Genesis of
grant from the National Research Foundation of South Africa. world-class, high-grade iron ore deposits—The South African experience,
Pack gratefully acknowledges funding by the Deutscher Akade- in Piestrzynski, A., ed., Mineral deposits at the beginning of the 21st cen-
mischer Austauschdienst, and support by Viorel Atudorei and Toti tury: Proceedings of the 6th Biennial Joint SGA-SEG meeting, Krakow,
26-29 August 2001: Lisse, A.A. Balkema, p. 1079–1082.
Larson during analytical work at the University of New Mexico. Gutzmer, J., Pack, A., Lueders, V., Wilkinson, J.J., Beukes, N.J., and Van
Gutzmer, Mukhopadhyay, and Beukes are grateful to Carlos Niekerk, H.S., 2001b, Formation of hydrothermal jasper and andradite
Rosiere for numerous discussions regarding the origin of high- during low-temperature hydrothermal seafloor metamorphism, Ongeluk
Formation, Northern Cape Province: Contributions to Mineralogy and
grade iron ores. Many thanks also to Simon Netshiozwi for the Petrology, v. 142, p. 27–42.
preparation of some of the hematite samples from Thabazimbi Hagemann, S.G., Barley, M.E., Folkert, S.E., Yardley, W.D., and Banks, D.A.,
and Carajas. Finally, we gratefully acknowledge Steve Kesler 1999, A hydrothermal origin for the giant BIF-hosted Tom Price iron ore
deposit, in Stanley et al. Rankin, A.H., Bodnar, R.J., Naden, J., Yard-
and two anonymous reviewers for their constructive criticism, ley, B.W.D., Criddle, A.J., Hagni, R.D., Gize, A.P., Pasava, J., Fleet, A.J.,
which helped to improve this manuscript. Seltmann, R., Halls, C., Stemprok, M., Williamson, B., Herrington, R.J.,
Hill, R.E.T., Prichard, H.M., Wall, F., eds., Mineral deposits: Processes to
processing: Proceedings of the 5th Biennial SGA Meeting and the 10th
REFERENCES CITED Quadrennial IAGOD Symposium, London, 22–25 August 1999: Rotter-
dam, A.A. Balkema, v. 1, p. 41–44.
Bao, H., and Koch, P.L., 1999, Oxygen isotope fractionation in ferric oxide- Hayashi, K., Maruyama, T., and Satoh, H., 2001, Precipitation of gold in a low-
water systems: Low temperature synthesis: Geochimica et Cosmochimica sulfidation epithermal gold deposit: Insight from a submillimeter-scale
Acta, v. 63, p. 599–613, doi: 10.1016/S0016-7037(99)00005-8. oxygen isotope analysis of vein quartz: Economic Geology and the Bul-
Barley, M.E., Pickard, A.L., Hagemann, S.G., and Folkert, S.L., 1999, Hydro- letin of the Society of Economic Geologists, v. 96, p. 211–216.
thermal origin for the 2 billion year old Mount Tom Price giant iron ore Hoefs, J., 1992, The stable isotope composition of sedimentary iron oxides with
deposit, Hamersley Province, Western Australia: Mineralium Deposita, special reference to banded iron formations, in Clauer, N., and Chau-
v. 34, p. 784–789, doi: 10.1007/s001260050238. duri, S., eds., Isotopic signatures and sedimentary records: Heidelberg,
Becker, R.H., and Clayton, R.N., 1976, Oxygen isotope study of a Precam- Springer-Verlag, Lecture Notes in Earth Sciences, v. 43, p. 199–213.
brian banded iron-formation, Hamersley Range, Western Australia: Geo- Hoefs, J., 1997, Stable isotope geochemistry (4th edition): Berlin, Springer-Ver-
chimica et Cosmochimica Acta, v. 40, p. 1153–1165, doi: 10.1016/0016- lag, 201 p.
7037(76)90151-4. Hoefs, J., Mueller, G., and Schuster, A.K., 1982, Polymetamorphic relations
Beukes, N.J., Gutzmer, J., and Mukhopadhyay, J., 2002a, The geology and gen- in iron ores from the Iron Quadrangle, Brazil: The correlation of oxygen
esis of high-grade hematite iron ore deposits, in Conference Proceedings, isotope variations with deformation history: Contribution to Mineralogy
Iron Ore 2002, 9-11 September 2002, Perth: Carlton, Australia, Austral- and Petrology, v. 79, p. 241–251, doi: 10.1007/BF00371515.
asian Institute of Mining and Metallurgy, p. 25–30. King, H.F., 1989, The rocks speak: Australasian Institute of Mining and Metal-
Beukes, N.J., Dorland, H., Gutzmer, J., Nedachi, M., and Ohmoto, H., 2002b, lurgy Monograph 15, p. 1–316.
Tropical laterites, life on land, and the history of atmospheric oxygen: MacLeod, W.N., 1966, The geology and iron deposits of the Hamersley Range
Geology, v. 30, p. 491–494. area, Western Australia: Perth, Geological Survey of Western Australia
Bird, M.I., Longstaffe, F.J., Fyfe, W.S., Kronberg, B.I., and Kishida, A., 1993, Bulletin, v. 117, p. 1–170.
An oxygen-isotopic study of weathering in the Eastern Amazon Basin, Morris, R.C., 1980, A textural and mineralogical study of the relationship of
Brazil, in Swart, P.K., Lohman, K.C., McKenzie, J., and Savin, S., eds., iron ore to banded iron-formation in the Hamersley iron province of West-
Climate change in continental isotopic record: American Geophysical ern Australia: Economic Geology and the Bulletin of the Society of Eco-
Society Monograph 78, p. 295–307. nomic Geologists, v. 75, p. 184–209.
Chacko, T., Cole, D.R., and Horita, J., 2001, Equilibrium oxygen, hydrogen Morris, R.C., 1985, Genesis of iron ore in banded iron-formation by super-
and carbon isotope fractionation factors applicable to geologic systems, gene and supergene-metamorphic processes—A conceptual model, in
in Valley, J.W., and Cole, D.R., eds., Stable isotope geochemistry: Wash- Wolff, K.H., ed., Regional studies and specific deposits: Amsterdam,
ington, D.C., Mineralogical Society of America, Reviews in Mineralogy Elsevier, Handbook of Strata-bound and Stratiform Ore Deposits, v. 13,
and Geochemistry, v. 43, p. 1–81. p. 72–235.
Clayton, R.N., and Epstein, S., 1961, The use of oxygen isotopes in high-tem- Netshiozwi, S.T., 2002, Origin of high-grade hematite ores at Thabazimbi
perature geological thermometry: Journal of Geology, v. 69, p. 447–452. Mine, Limpopo Province, South Africa [M.Sc. thesis]: Johannesburg,
Evans, D.A.D., Beukes, N.J., Kirschvink, J.L., 2002, Paleomagnetism of a lat- Rand Afrikaans University, 135 p.
eritic paleo-weathering horizon and overlying Paleoproterozoic redbeds Perry, E.C., and Ahmad, S.N., 1983, Oxygen isotope geochemistry of Pro-
from South Africa: Implications for the Kaapvaal apparent polar wander terozoic chemical sediments, in Medaris, L.G., Byers, C.W., Mickelson,
path and a confirmation of atmospheric oxygen enrichment. Journal of D.M., and Shanks, W.C., eds., Proterozoic geology: Selected papers from
Geophysical Research-Solid Earth, v. 107, p. 2326–2333. an international Proterozoic symposium: Geological Society of America
Fricke, H.C., and O’Neil, J.R., 1999, The correlation between 18O/16O ratios of Memoir 161, p. 253–263.
Perry, E.C., and Bonnichsen, B., 1966, Quartz and magnetite oxygen-18-oxy- ince, Western Australia: Economic Geology and the Bulletin of the Soci-
gen-16 fractionation in metamorphosed Biwabik iron formation: Science, ety of Economic Geologists, v. 96, p. 837–878.
v. 153, p. 528–529. van Schalkwyk, J.F., and Beukes, N.J., 1986, The Sishen iron ore deposit, in
Perry, E.C., Tan, C., and Morey, G.B., 1973, Geology and stable isotope geo- Anhaeusser, C.R., and Maske, S.S., eds., Mineral deposits of Southern
chemistry of the Biwabik iron formation, Northern Minnesota: Economic Africa: Johannesburg, Geological Society of South Africa, v. 1, p. 157–
Geology and the Bulletin of the Society of Economic Geologists, v. 68, 182.
p. 1110–1125. Yapp, C.J., 1990, Oxygen isotopes in iron (III) oxides. I: Mineral-water fraction-
Powell, McA., Oliver, N.S., Li, Z.X., Martin, D.McB., and Ronaszecki, J., 1999, ation factors: Chemical Geology, v. 85, p. 329–335, doi: 10.1016/0009-
Synorogenic hydrothermal origin for giant Hamersley iron oxide ore bod- 2541(90)90010-5.
ies: Geology, v. 27, p. 175–178, doi: 10.1130/0091-7613(1999)027<0175: Yapp, C.J., 2001, Rusty relics of Earth History: Iron (III) oxides, isotopes and
SHOFGH>2.3.CO;2. surficial environments: Annual Review of Earth and Planetary Sciences,
Ramdohr, P., 1980, The ore minerals and their intergrowths. Oxford, UK, Per- v. 29, p. 165–199, doi: 10.1146/annurev.earth.29.1.165.
gamon Press, 2 v., 1207 p.
Sharp, Z.D., 1990, A laser-based microanalytical method for the in-situ-
determination of oxygen isotope ratios of silicates and oxides: Geochi-
mica et Cosmochimica Acta, v. 54, p. 1353–1357, doi: 10.1016/0016-
7037(90)90160-M.
Taylor, D., Dalstra, H.J., Harding, A.E., Broadbent, G.C., and Barley, M.E.,
2001, Genesis of high-grade hematite orebodies of the Hamersley Prov- MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA


Memoir 198
2006
Rare earth elements in Precambrian banded iron formations:

Secular changes of Ce and Eu anomalies and evolution of
atmospheric oxygen
Yasuhiro Kato
Department of Geosystem Engineering, University of Tokyo, Bunkyo-Ku, Tokyo 113-8656, Japan
Kosei E. Yamaguchi
Yokosuka, Kanagawa 237-0061, Japan
Hiroshi Ohmoto
The Pennsylvania State University, University Park, Pennsylvania 16802, USA
ABSTRACT
Rare earth element (REE) analyses of Precambrian banded iron formations

(BIFs) show that distinct negative Ce anomalies, although rather weak or moder-
ate (Ce/Ce* = 0.5–0.9), are commonly present in Algoma-type BIFs of the Early and
Middle Archean, and even in the 3.8–3.7 Ga Isua iron formation (IF). This indicates
that the seawater columns from which the BIFs precipitated were not entirely anoxic
and that Ce oxidation mechanisms already existed in the 3.8–3.7 Ga oceans. The pres-
ence of pronounced negative Ce anomalies (Ce/Ce* = 0.1–0.5) in Late Archean (2.9–
2.7 Ga) Algoma-type BIFs suggests that strongly oxygenated oceanic conditions like
today emerged by 2.9–2.7 Ga. This suggestion is consistent with geologic evidence that
small but widespread Mn deposits formed during Late Archean time. The Hamersley
and Transvaal IFs (2.7–2.4 Ga in age) have less noticeable Ce anomalies (Ce/Ce* =
0.7–1.0). These BIFs were deposited on an evolving rift in a land-locked ocean that
became anoxic due to intense hydrothermal activity. The 2.2–2.1 Ga Superior-type
IFs exhibit distinct negative Ce anomalies (Ce/Ce* = 0.2–0.7), but the ca. 1.9 Ga IFs
and the ca. 0.7 Ga IFs have less distinct Ce anomalies. These variations in Ce/Ce* val-
ues of the post-2.7 Ga BIFs may reflect the episodicity in global mantle plume activity
that created locally anoxic basins.
The Archean Algoma-type BIFs have distinctly positive but variable Eu anoma-
lies (Eu/Eu* = 0.8–7). Their strong positive Eu anomalies suggest large contributions
of hydrothermal fluids to the seawater involved in BIF precipitation. The large varia-
tion in Eu anomalies in Algoma-type BIFs reflects the large variation in mixing ratios
of hydrothermal fluids and ambient seawaters at various depositional sites. However,
the post-2.7 Ga Superior-type BIFs exhibit much lower and constant Eu/Eu* values
(0.7–1.9). This implies that the REE chemistry of the basin water that hosted volu-
Kato, Y., Yamaguchi, K.E., and Ohmoto, H., 2006, Rare earth elements in Precambrian banded iron formations: Secular changes of Ce and Eu anomalies and evo-
lution of atmospheric oxygen, in Kesler, S.E., and Ohmoto, H., eds., Evolution of Early Earth’s Atmosphere, Hydrosphere, and Biosphere—Constraints from Ore
Deposits: Geological Society of America Memoir 198, p. 269–289, doi: 10.1130/2006.1198(16). For permission to copy, contact [email protected]. ©2006
Geological Society of America. All rights reserved.
269
270 Y. Kato et al.
minous Superior-type BIFs was influenced by a riverine influx from the surrounding
continent that grew rapidly due to global mantle plume activity, besides the intense
hydrothermal influx.
Keywords: banded iron formation, Precambrian, rare earth element, Ce anomaly, Eu

anomaly, redox condition, oxygen, geochemistry.
INTRODUCTION of remarkable similarities in geologic settings and REE geo-

chemistry between the BIFs and modern MOR hydrothermal
Banded iron formations (BIFs) are chemical sedimentary Fe-Mn sediments. The recognition of horizontal shortening
rocks deposited in marine environments mostly during the by thrust duplication of BIF successions, along with new geo-
Precambrian era (James, 1954). They occur mostly in cratons chronological data, has led many researchers to propose a new
with 3.8–1.9 Ga sedimentary sequences, and are classified into geologic and stratigraphic framework for the Archean green-
two types based on their formation age, size, and associated stone belts in the plate tectonic regime (de Wit et al., 1992;
lithologies (Gross, 1965). Algoma-type BIFs mainly formed Kusky and Kidd, 1992; Krapez, 1993; Barley, 1993, Kato et
during the Archean and developed in volcanic sequences on al., 1998; Komiya et al., 1999). In this framework, depositional
a relatively small scale (lateral extents <10 km and thickness sites of Algoma-type BIFs and terrigenous clastic sediments
<100 m), whereas Superior-type BIFs mostly formed during were completely different; the BIFs are interpreted to have
the Late Archean to Paleoproterozoic period and are associated been deposited in deep-sea pelagic environments, far away
with clastic sedimentary rocks as well as volcanic rocks. Supe- from continental margins, and to have been placed in their
rior-type BIFs, the dominant iron resource in the world, are present position as accretionary prism complexes by obduc-
typically very large; the largest one exceeds 105 km2 in aerial tion/subduction.
distribution (Trendall and Blockley, 1970). In addition to these Although opinions greatly vary on the depositional envi-
Precambrian iron formations (IFs), many IFs were deposited ronments and the depositional sites relative to hydrothermal
during the Neoproterozoic (ca. 0.7 Ga) and Phanerozoic peri- discharge areas for BIFs, the consensus among researchers is
ods on a small scale (Ohmoto et al., this volume). that the geochemical features of BIFs reflect those of the water
A currently popular theory (e.g., Jacobsen and Pimentel- column from which BIFs precipitated. Among these geochemi-
Klose, 1988) postulates that the ultimate sources of iron and cal features, REE chemistry of BIFs has been well documented.
silica for both types of BIFs are submarine hydrothermal ema- Several researchers have attempted to reconstruct the chemical
nations, mostly from mid-ocean ridges (MORs). This theory evolution of the Precambrian oceans from the REE features of
was developed primarily from rare earth element (REE) data BIFs (Fryer, 1977; Derry and Jacobsen, 1990; Danielson et al.,
on BIFs, which show the predominance of mantle-like Nd iso- 1992; Bau and Möller, 1993; Kato et al., 1998).
topes and the ubiquitous existence of positive Eu anomalies Redox conditions of Phanerozoic seawater have been
(e.g., Jacobsen and Pimentel-Klose, 1988; Bau and Dulski, inferred from Ce anomalies in marine sedimentary rocks (e.g.,
1996; Kato et al., 1998). Recently, it has been recognized that MacLeod and Irving, 1996; Kato et al., 2002b). Unlike other
submarine volcanic rocks are much more abundantly associated REEs, in the presence of dissolved oxygen Ce3+ in seawater is
with Superior-type BIFs than previously thought (e.g., Barley et removed as CeO2 or Ce(OH)4, especially with Mn hydroxides
al., 1997). Hence, global magmatism (probably plume-related (Elderfield et al., 1981). Because of this oxidative removal of
mafic volcanism) could have played an important role in BIF Ce, modern oxygenated deep-sea waters are depleted in Ce and
deposition (e.g., Isley and Abbott, 1999; Kato, 2003). show very large negative Ce anomalies. These large negative Ce
However, depositional environments for BIFs are still con- anomalies are also seen in the hydrothermal Fe-Mn sediments
troversial. Many researchers have regarded BIFs as chemical on ridge flanks (Bender et al., 1971; Ruhlin and Owen, 1986).
precipitates in shallow-sea environments along continental Kato et al. (1998, 2002a) have suggested that some Algoma-
shelves or near island arcs, because BIFs are often associated type BIFs are ancient counterparts of modern ridge-flank Fe-
with terrigenous clastic rocks (e.g., Jacobsen and Pimentel- Mn sediments. Europium anomalies of seawaters are mainly
Klose, 1988; Manikyamba et al., 1993). According to this determined by the relative contributions of hydrothermal and
scenario, Fe2+-rich deep-sea water from a MOR hydrothermal riverine influxes (Elderfield, 1988). Therefore, we could moni-
source upwelled to the continental margin, and mixed with tor the outline of ocean chemistry evolution controlled by these
oxygenated surface water to produce deposition of BIFs that two major influxes by using the secular variation of Eu anoma-
are conformable with terrigenous clastic sediments. lies preserved in BIFs. In this paper, we examine published and
In contrast, Kato et al. (1998, 2002a) suggested that some unpublished data on REEs of Precambrian BIFs, especially the
Algoma-type BIFs deposited on the flanks of MORs where Ce and Eu anomalies, in order to constrain the redox condi-
hydrothermal Fe-oxyhydroxide particulates precipitated in tions and relative importance of hydrothermal versus riverine
situ. This suggestion was made because of the recognition (continental) influxes in the Precambrian oceans.
Rare earth elements in Precambrian banded iron formations 271
REES AS AN INDICATOR FOR PALEOCEAN of Cretaceous age (MacLeod and Irving, 1996). Murray et al.
CHEMISTRY (1992b) and Kato et al. (2002b) have shown that strong negative
Ce anomalies in Phanerozoic pelagic cherts were not severely
Total concentrations and relative abundances of REEs in affected by diagenesis, but they were predominantly influenced
marine sediments are very useful geochemical indicators to by the chemistry of the immediately overlying seawater. These
determine the origins and depositional environments of marine results confirm the validity of the Ce anomaly in bulk sediments
sediments and to estimate various geologic processes during as a paleoredox indicator of the oceans. Similarly, Eu anomalies
sedimentation (e.g., Murray et al., 1992a; Kato and Nakamura, in sedimentary rocks are likely to reflect Eu anomalies of ancient
2003). REEs typically exist in a trivalent oxidation state, but Ce ocean waters, and the magnitude of the Eu anomaly may be used
and Eu are unique in that they can also exist in tetravalent and to evaluate the relative contributions of two major sources of
divalent oxidation states, respectively, giving rise to their anoma- REEs (riverine and hydrothermal) in the ancient seawater (Elder-
lous behaviors. Both are very important elements providing some field, 1988). Mitra et al. (1994) have shown that REEs, including
constraints on paleocean chemistry. The Ce anomaly in ancient Eu, behave conservatively during mixing of hydrothermal fluids
submarine sediments is one of the most useful indicators to mon- and ambient seawater at venting sites.
itor oceanic redox conditions. The Ce anomaly of seawater cor- It is generally believed that REEs of bulk-rock samples are
responds well to varying redox conditions (de Baar et al., 1988; virtually unaffected by post-depositional processes, including
German and Elderfield, 1989; Sholkovitz and Schneider, 1991) diagenesis, metamorphism, and weathering, unless the fluid/rock
owing to rapid reduction or oxidation reactions between Ce3+ and ratios were very large and the dissolution/reprecipitation of REE-
Ce4+ in response to changes in the redox conditions (Sholkovitz bearing minerals (as well as major minerals) occurred extensively
et al., 1992). For example, in the Cariaco Trench, the overlying (e.g., Taylor and McLennan, 1985; McLennan and Taylor, 1991).
oxygenated waters exhibit striking negative Ce anomalies, but in For example, many investigations (Grauch, 1989, and refer-
the oxic/anoxic transition zone Ce anomalies disappear, and in ences therein) have recognized that REEs are basically immobile
the anoxic bottom water no or positive Ce anomalies are found during high-grade metamorphism under relatively fluid-absent
(de Baar et al., 1988). conditions, whereas several studies (e.g., Nyström, 1984) have
However, there is no linear relationship between the mag- reported REE mobility under low-grade metamorphic conditions
nitudes of Ce anomalies and the contents of dissolved oxygen in where metamorphic fluids were prevalent. In some instances
seawater. This is because REEs, and thus Ce anomalies, are also where carbonate and/or fluoride complexes of REEs could have
influenced by ocean circulation and biogeochemical cycling of been important in the metamorphic fluids, REEs were apparently
nutrients and other trace metals. The presence of either negative remobilized even during high-grade (e.g., the granulite facies)
or positive Ce anomalies, however, suggests that the redox-related metamorphism (e.g., Pan and Fleet, 1996). However, carbonate
Ce fractionation took place in the oceans, although qualitatively. veins and fluorine minerals (fluorapatite and fluorite) contain-
Hence, if ancient sediments have mimicked the Ce anomaly of ing high REE concentrations are rare in BIFs, and there is no
contemporaneous seawater, we can get useful information on the evidence for REE mobility during the metamorphism of BIFs.
paleoredox condition. Earlier studies have related Ce anomalies Because the “distribution coefficients” of REEs between BIFs
of fossil apatites and carbonates in marine sediments to anoxic and metamorphic fluids must have been very large (~105-106),
events in the Phanerozoic, assuming that these minerals record as in the case of diagenesis mentioned above, the REE concen-
the Ce anomalies of contemporaneous seawater (e.g., Wright et trations and patterns of bulk BIF samples would not have been
al., 1987; Liu et al., 1988). However, it was reported subsequently significantly affected by metamorphic processes, except when
that the Ce anomalies in fossil minerals change as a result of dia- the fluid/rock ratios were extremely high (e.g., Elderfield and
genetic alteration, and thus these fossil materials do not retain Sholkovitz, 1987; Michard, 1989).
their marine Ce anomalies at the time of sedimentation (German The mobility and redistribution of REEs by weathering have
and Elderfield, 1990; Sholkovitz and Schneider, 1991). Because been reported for lateritic soil profiles (e.g., Braun et al., 1998)
the pore waters in reducing nearshore sediments have high REE and soil formations associated with granitic rocks (e.g., Nesbitt,
concentrations (Elderfield and Sholkovitz, 1987; Sholkovitz et 1979; Aubert et al., 2001). Intensive and/or extensive weathering
al., 1989) and because these apatite-rich fossils can take up large has resulted in a complete disappearance of original rock texture,
amounts of REEs, Ce anomalies in small fossil minerals may dissolution of the primary REE-bearing minerals (e.g., apatite,
be easily altered during sediment diagenesis; thus, they cannot sphene, and monazite), as well as feldspars and silicates, in the
be used as a paleoredox proxy. On the other hand, it is unlikely upper soil sections, and reprecipitation of secondary phosphate
that REE signatures of bulk sediments are significantly affected minerals (e.g., cerianite, florencite, and rhabdophane) in the
during diagenesis, because the amounts of REEs in bulk sedi- lower soil sections (Braun et al., 1998 and references therein).
ments, such as BIFs, are greater than those in diagenetic fluids However, the REE data selected for this study were obtained on
by five or six orders of magnitude. In fact, a good correlation of BIF samples that have well-preserved primary banding, indicat-
Ce anomalies with other independent indicators of oceanic paleo- ing that these samples were not subjected to severe weathering.
redox conditions has been reported for bulk limestone samples Therefore, it would be reasonable to conclude that the selected
272 Y. Kato et al.
BIF samples have retained their original REE chemistry, which tive Pr/Pr* values, whereas true positive Ce anomalies are repre-
reflected the chemistry of waters involved in BIF precipitation. sented by positive Ce/Ce* and negative Pr/Pr* values (see Fig-
ures 2 and 4). Slightly positive Gd anomalies have been reported
Ce AND Eu ANOMALIES IN MODERN OCEAN for modern seawaters (de Baar et al., 1985; Alibo and Nozaki,
WATERS AND SUBMARINE HYDROTHERMAL Fe-Mn 1999) and ancient sediments (Bau and Dulski, 1996). However,
SEDIMENTS anomalous abundances of Gd are not common; Gd anomalies are
generally absent in most BIFs. Thus, in this paper Eu anomalies
REE data of modern seawaters, marine sediments, and are calculated as EuN/[(SmN)(GdN)]1/2, not by using Tb instead of
post-Archean sedimentary rocks are generally normalized to Gd as suggested by Bau and Dulski (1996).
those of post-Archean average Australian shale (PAAS; Tay- Chondrite-normalized REE patterns of modern seawaters
lor and McLennan, 1985) or North American Shale Composite (deep-sea and surface), black smoker fluids, hydrothermal Fe-
(NASC; Goldstein and Jacobsen, 1988). The data indicate that Mn sediments on the ridge flank of the East Pacific Rise (EPR),
the REE concentration ratios of post-Archean shales are remark- hydrothermal Fe-Mn umbers in the Phanerozoic accretionary
ably similar, suggesting the similar chemical compositions for complexes, and hydrogenetic Mn crusts are shown in Figure
the continental crust after the Archean era. Similarly, in order to 1. The Ce/Ce*-Pr/Pr* diagrams for them are given in Figure
understand the differences in behaviors of REEs in the Archean 2. It is clearly shown that seawaters plot in the region of true
oceans, REE data on Archean marine sedimentary rocks should negative Ce anomalies. Modern oxygenated deep-sea waters
be normalized to the average Archean shale, which represents (>2000 m) exhibit very large negative Ce anomalies (Ce/Ce*
the average Archean upper continental crust. However, there is = 0.03–0.25) and moderate negative Eu anomalies (Eu/Eu* =
no consensus among researchers on the average Archean crustal 0.63–0.72) together with positive La and Gd anomalies (Fig.
composition, because there are large variations in the REE ratios 1A). Moderate HREE (heavy REE) enrichment is also recog-
among rocks of the same age and there may be an evolution- nized. The positive La and Gd anomalies may be caused by
ary trend in the REE ratios of Archean shales (e.g., Taylor and unusually high stability of La and Gd of the lanthanide series in
McLennan, 1985; Condie, 1993). Therefore, it is not meaning- seawater (e.g., de Baar et al., 1985). Surface seawaters (<200
ful to select REE data from a single shale for normalization of m) have less striking negative Ce anomalies (Ce/Ce* = 0.33–
various-aged BIFs. In this study, the REE abundances of mod- 0.43) and HREE enrichment, but similar Eu anomalies (Eu/Eu*
ern ocean waters, sediments, and ancient BIFs are normalized to = 0.68–0.73), compared to deep-sea waters. The REE patterns
those of the Leedey chondrite obtained by Masuda et al. (1973) of high-temperature black smoker fluids differ greatly from
and Masuda (1975). Cerium and europium anomalies are defined those of seawater, and are characterized by large positive Eu
quantitatively as Ce/Ce* = CeN/[(LaN)(PrN)]1/2 and Eu/Eu* = EuN/ anomalies (Eu/Eu* = 7.2–15.1), no Ce anomaly, and remark-
[(SmN)(GdN)]1/2, respectively, where Ce* and Eu* are the hypo- able LREE (light REE) enrichment (Fig. 1B). In addition, black
thetical concentrations that strictly trivalent Ce and Eu would smoker fluids have higher REE concentrations than seawaters
have (Taylor and McLennan, 1985). The subscript “N” indicates by two or three orders of magnitude.
chondrite-normalized values. When Ce and Eu are depleted, Ce/ Hydrogenetic Mn crusts (except for one sample) have con-
Ce* and Eu/Eu* values are below unity and vice versa. Another siderably positive Ce anomalies (Ce/Ce* = 1.3–3.4; Fig. 1E),
method of calculation, [Ce/Ce* = 2CeN/(LaN+PrN), Eu/Eu* = which are caused by the oxidative scavenging of Ce and its pref-
2EuN/(SmN+GdN)], is often used for seawaters and marine sedi- erential removal from seawater by Mn hydroxides (Elderfield et
ments (e.g., de Baar et al., 1985), but differences between the al., 1981). The hydrogenetic crusts plot in the region of Ce/Ce*>1
values of these two types of calculations are negligible for a wide and Pr/Pr*<1 in the Ce/Ce*-Pr/Pr* diagram (Fig. 2). The magni-
range of anomalies (Kato et al., 2002a). tude of positive Ce anomalies is probably related to precipitation
Evaluation of the presence or absence of a Ce anomaly is rates; a very low growth rate results in a pronounced positive
often complicated because of the anomalous behavior of La Ce anomaly in hydrogenetic Mn crusts (Kuhn et al., 1998). The
(Bau and Dulski, 1996). Lanthanum enrichment (i.e., positive Eu/Eu* values fall in a narrow range (from 0.67 to 0.78), which
La anomalies), together with the well-known Ce depletion, has is the same as (or slightly higher than) that of seawater.
been identified in modern seawaters (e.g., Zhang et al., 1994; Submarine hydrothermal Fe-Mn sediments, which typically
Alibo and Nozaki, 1999). Positive La anomalies have also been occur on flanks of MORs, generally mimic the striking depletion
reported for hydrothermal Fe-Mn sediments on oceanic crust of Ce and the less striking depletion of Eu in deep-ocean waters
(Barrett and Jarvis, 1988) and some BIFs (Barrett et al., 1988; because of scavenging of REEs from the deep-ocean waters
Bau and Dulski, 1996). Consequently, anomalous La enrichment (Fig. 1C). German et al. (1990) have suggested that hydrother-
can create false negative Ce anomalies in some cases. In order to mal plume particles (the precursor of Fe-Mn sediments) scav-
distinguish true negative Ce anomalies from false ones, plots of enge REEs with the proportions of 0.1% hydrothermal fluid and
Ce/Ce* against Pr/Pr* (Pr/Pr* = PrN/[(CeN)(NdN)]1/2), which were 99.9% seawater. Furthermore, most of the uptake of REEs from
suggested by Bau and Dulski (1996), are often used: true nega- seawater occurs after the hydrothermal precipitates have settled
tive Ce anomalies are represented by negative Ce/Ce* and posi- on the seafloor, because the REE concentrations of plume par-
10-4 10 -2
A B
10-5 deep-sea (>2000 m) 10 -3
10-6 10 -4
surface (<200 m)
seawater
MOR hydrothermal fluids
10-7 10 -5
C D
chondrite-normalized
1000 1000
100 100
10 10
EPR Fe-Mn sediments Phanerozoic Fe-Mn umbers
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1000 E
100
10
hydrogenetic Mn crusts
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Figure 1. (A) Chondrite-normalized REE patterns of modern seawater. (B) MOR hydrothermal fluids. (C) EPR (East Pacific Rise) hydrothermal
Fe-Mn sediments. (D) Phanerozoic hydrothermal Fe-Mn umbers. (E) Hydrogenetic Mn crusts. The data sources are summarized in Table 1.
ticles are found to be one order of magnitude lower than those of est ΣREE tends to have the Eu/Eu* value closest to the seawater
sediments. That is, the REEs in hydrothermal Fe-Mn sediments value (Ruhlin and Owen, 1986; Olivarez and Owen, 1991). This
are dominated by seawater components. Consequently, both the is an important factor to keep in mind when we interpret BIF
Ce/Ce* (0.12–0.26) and Pr/Pr* values of Fe-Mn sediments agree REE data.
well with those of some deep-sea waters (Fig. 2). On the other It is well recognized that Phanerozoic Fe-Mn umbers have
hand, the Eu/Eu* values of Fe-Mn sediments (0.75–0.93) are less similar Ce and Eu features of the MOR hydrothermal Fe-Mn
negative compared to deep-sea waters (0.63–0.72), because of a sediments (Fig. 1D), as suggested by Ravizza et al. (1999). The
contribution of submarine hydrothermal fluids with strong posi- umbers that deposited in the oceans but subsequently accreted
tive Eu anomalies. Even a quite trivial contribution from hydro- onto land have wide ranges of Ce/Ce* (0.03–0.24) and Pr/Pr*
thermal fluids can be detected as Eu anomalies of hydrothermal values, which correspond to the ranges of modern deep-sea
Fe-Mn sediments, because the hydrothermal fluids have striking waters (Fig. 2). Moreover, the umbers have Eu/Eu* values rang-
positive Eu anomalies and much higher Eu contents than ocean ing from 0.60 to 0.90, which are also similar to those of sea-
water. In general, however, the subsequent overprinting of REEs waters. These REE features of modern and Phanerozoic subma-
from the overlying deep-ocean water with a negative Eu anom- rine hydrothermal Fe-Mn sediments are excellent geochemical
aly decreases the Eu anomaly values of the hydrothermal Fe-Mn proxies of the Ce and Eu anomalies of oxygenated deep-ocean
sediments to the seawater value, with increasing concentrations water. Moderate enrichment of HREE is recognized in modern
of REEs; hence the hydrothermal Fe-Mn sediment with the high- deep-sea waters (Fig. 1A). However, HREE enrichment is absent
274 Y. Kato et al.
0.5
0.4
3
0.3
0.2
Ce/Ce*
2 0.1
0
0 1 2 3 4 5 6
deep-sea water (>2000 m)

1 Figure 2. Ce/Ce* versus Pr/Pr* relationships
surface seawater (<200 m)
for modern seawater, EPR Fe-Mn sediments,
hydrogenetic Mn crusts
Phanerozoic Fe-Mn umbers, and hydrogenet-
ic Mn crusts. The region of true negative Ce
anomalies is shown as a shaded area.
0
0 1 2 3 4 5 6
0.5
surface seawater
0.4 EPR Fe-Mn sediments
Phanerozoic Fe-Mn umbers
0.3
0.2
deep-sea water
0.1
0
0 1 2 3 4 5 6
Pr/Pr*
in modern and Phanerozoic MOR sediments. This suggests a samples analyzed for REEs by Morinaga (1997) have suffered
continuing preferential uptake of LREEs, compared to HREEs, greenschist-facies metamorphism, as indicated by the mineral
by hydrothermal Fe-Mn sediments; this has been experimentally assemblage of chlorite, epidote, albite, and quartz (Komiya et al.,
substantiated by Byrne and Kim (1990) and Koeppenkastrop and 1999). The reconstructed lithostratigraphy comprises a simple
De Carlo (1992). succession of greenstones with low-K tholeiitic characteristics,
BIFs, and turbidites in ascending stratigraphic position (Komiya
BRIEF DESCRIPTIONS OF SELECTED BIFS et al., 1999); this sequence is very similar to the oceanic plate
stratigraphy of Phanerozoic age. The Isua IF consists of black,
The BIFs selected for this study, excepting the Braemar IF, white, gray, green, and red layers that vary in thickness from
are chemical sediments free from detritus in order to obtain infor- several millimeters to several centimeters. Black and gray lay-
mation about the compositions of contemporaneous seawaters; ers are dominant and are composed mainly of magnetite, quartz,
data on shaly BIFs and ferruginous shales were excluded. Data amphibole, and chlorite. The red layers contain an unidentified
sources for the selected BIFs are given in Table 1. Reliable data red-colored mineral. Hematite is absent in the Isua IF.
on all REEs (14 elements except for Pm) are necessary in order to
discuss Ce and Eu anomalies in detail. However, complete analy- Marble Bar, Australia
ses of BIFs using the inductively coupled plasma–mass spectrom-
etry (ICP-MS) method are much fewer than those determined by The Marble Bar jasper (ferruginous chert) is a member of the
the ICP or neutron activation analyses. Therefore, where there are Towers Formation of the Warrawoona Group in the Marble Bar
no ICP-MS data, we use the ICP, neutron activation, and isotope greenstone belt, Pilbara, Western Australia (Hickman, 1983; Van
dilution data: the Marra Mamba, Brockman, Hotazel, Väyrylän- Kranendonk et al., 2001). Their depositional age is between 3463
kylä, Gunflint, Biwabik, Sokoman, and Rapitan (Braemar) IFs. and 3454 Ma (Thorpe et al., 1992; McNaughton et al., 1993).
These rocks have undergone relatively low-grade metamorphism
Isua, Greenland of prehnite-pumpellyite to greenschist facies. The Marble Bar jas-
per/chert formation is composed of red, brown, light green, gray,
The 3.8–3.7 Ga Isua IF, the oldest in the world, is from black, and white layers that vary from 1 mm to 2 cm in thick-
the Isua supracrustal belt, southern West Greenland. The BIF ness. Black, gray, and light green cherts are mainly composed of
microcrystalline quartz, siderite, dolomite, and black aggregates layers, the association with mid-ocean ridge basalt–like green-
(probably organic compounds). The brown chert layers mostly stones) and geochemical signatures of the Marble Bar cherts are
consist of quartz and goethite, whereas the red chert contains essentially identical to those of typical Algoma-type BIFs (Kato
quartz, hematite, and siderite, and the white chert is composed and Nakamura, 2003).
solely of quartz. Goethite is probably a recent weathering prod-
uct. However, there is no systematic difference between REE Cleaverville, Australia
features of samples with goethite and those without goethite,
suggesting that REEs are virtually unaffected by goethite forma- The Middle Archean Cleaverville Formation in the West Pil-
tion. The average Fe2O3* content of the ferruginous cherts is 10.2 bara Granite-Greenstone Terrane, Western Australia, is estimated
wt%, which is significantly lower than in common BIFs. How- to be between 3.2 and 3.0 Ga in age (Kiyokawa and Taira, 1998).
ever, field occurrences (e.g., alternation of iron-rich and cherty The Formation comprises weakly metamorphosed (greenschist
276 Y. Kato et al.
facies) greenstone, BIF, mudstone, sandstone, and minor amounts Mn formations, and carbonaceous phyllites. These lithological
of conglomerate. Geochemical and petrological characteristics of variations suggest that there was a great variation in depositional
the underlying greenstones suggest they are Fe-rich low-K tho- environments for the Bababudan Group.
leiites that formed on a mid-ocean ridge (Ohta et al., 1996). The The IF samples of the Bababudan belt, collected from the
Cleaverville IF consists of alternating Fe-rich and cherty layers core library at Chickmagalur, are cherty, consisting of magne-
that are several millimeters to several centimeters thick (Kato et tite, quartz, grunelite, Mg riebeckite, hornblende, and actino-
al., 1998). The Fe-rich layers are composed mainly of hematite, lite (Arora et al., 1995). The cherty IF from the Kushtagi belt is
goethite, and quartz. Chert consists dominantly of quartz with entirely composed of alternating bands of hematite and quartz
minor amounts of goethite, hematite, and/or kaolinite, and illite. that were separately analyzed (Khan et al., 1996). The IF col-
lected from the northern Holenarsipur belt consists of two, very
Sargur, India thin (usually 1–2 mm thick) alternating lithologies, containing
Fe-rich and cherty layers (Kato et al., 2002a). The Fe-rich layers
The other Middle Archean BIF is the Sargur IF in the Sargur consist dominantly of magnetite and quartz with minor hematite
greenstone belt, South India. The Sargur and equivalent green- and goethite. The cherty layers are solely composed of quartz
stone belts occur as small remnants in the Peninsular Gneiss and with or without small inclusions of magnetite. Cherty IF, ferrugi-
have been metamorphosed to the upper amphibolite and gran- nous chert, and chert of the Sandur belt, collected from borehole
ulite facies. These belts consist of ultramafic to mafic igneous cores and recent railway and road cuttings, comprise hematite,
rocks, fuchsite-bearing quartzites, pelites, calc-silicate rocks, and magnetite, and quartz with lesser amounts of siderite and ankerite
BIFs. Their depositional age is constrained between 3.13 and (Manikyamba et al., 1993).
2.96 Ga (Nutman et al., 1992; Chadwick et al., 2000). The Sar-
gur IF is composed of alternating Fe-rich and cherty thin layers, Hamersley, Australia
approximately 1–2 mm in thickness (Kato et al., 1996). The Fe-
rich layers consist of magnetite and minor amphibole, and the The Hamersley Group, in the Hamersley Basin, Western
cherty layers consist of quartz. The associated amphibolites have Australia, includes three BIF groups (Marra Mamba IF, Brock-
mid-ocean ridge basalt–type geochemical signatures (Kato et al., man IF, and Boolgeeda IF, in ascending stratigraphic position).
1996). Together they provide a significant proportion of the world’s iron
ore supply. The depositional ages of these IFs are well determined
Bababudan, Kushtagi, Northern Holenarsipur, and by U-Pb sensitive high-resolution ion microprobe (SHRIMP) zir-
Sandur, India con dating (Arndt et al., 1991; Barley et al., 1997; Trendall et al.,
1998), although the upper limit of the Boolgeeda IF is not tightly
Late Archean Algoma-type BIFs widely occur in the Dharwar constrained because of a lack of age data for the overlying Turee
Supergroup of the Dharwar Craton, South India (Radhakrishna Creek Group. The Marra Mamba IF and Brockman IF deposited
and Naqvi, 1986). The depositional age of the Dharwar Super- between 2684 and 2561 Ma and between 2561 and 2449 Ma,
group is broadly bracketed between 2.9 Ga and 2.7 Ga (Nutman respectively. The Woongarra Rhyolite underlying the Boolgeeda
et al., 1996; Trendall et al., 1997a, 1997b; Chadwick et al., 2000). IF has been precisely dated at 2449 Ma (Barley et al., 1997);
The deficiency of precise geochronological data and the absence the apparent conformity between them shows that the Boolgeeda
of good exposure spanning the entire succession preclude geolog- IF formed soon after 2.45 Ga (Martin et al., 2000). The Marra
ical correlation between the Dharwar greenstone schist belts. The Mamba and Brockman IFs deposited in a clastic-starved environ-
low-grade-metamorphosed (up to amphibolite facies) greenstone ment and are dominated by either chemical sediments or extru-
belts contain three beds of BIFs interbedded with different types sive volcanic rocks. The Boolgeeda IF is somewhat different
of sedimentary and volcanic rocks (Naqvi et al., 1988). The thick- from the lower IFs (Marra Mamba and Brockman) and is over-
est and most extensive BIF deposition occurred in the Bababudan lain by turbidites, shallow-marine carbonate rocks, and fluvial
Group of the lower Dharwar Supergroup. The Bababudan Group and marine siliciclastic rocks of the Turee Creek Group (Mar-
is extensively exposed in many greenstone belts, including the tin et al., 2000). The Hamersley Group has suffered low-grade
Bababudan, Kudremukh, Kushtagi, Chitradurga, Shimoga, and metamorphism varying from the prehnite-pumpellyite facies to
Sandur schist belts. Among these, data sets on BIFs from four greenschist facies (Smith et al., 1982). The analyzed drill core
schist belts were selected for the present study: the Bababudan samples of the lower Marra Mamba IF comprise magnetite-rich
belt (Arora et al., 1995), Kushtagi belt (Khan et al., 1996), San- IF, minnesotaite-chert with subordinate stilpnomelane, and chert-
dur belt (Manikyamba et al., 1993), and the northern extension carbonate (Alibert and McCulloch, 1993). Sulfides are ubiqui-
of the Holenarsipur belt (Kato et al., 2002a). Some schist belts tous as disseminated grains, nodules, and millimeter-thick beds.
(Bababudan, Kudremukh, Kushtagi, northern Holenarsipur) are The upper Marra Mamba IF is characterized by the ubiquitous
predominated by oxide facies BIFs and lack stromatolitic carbon- presence of organic carbon and sulfides, thin laminations, and the
ate rocks and Mn formations, whereas other belts (Chitradurga, presence of both magnetite and relict hematite. The Brockman
Shimoga, Sandur) are associated with stromatolitic carbonates, IF samples are mainly magnetite-rich mesobands and riebeckite-
chert. The Boolgeeda IF is composed mostly of magnetite-rich IF the northern shore of Lake Superior, and the Biwabik IF of the
and ferruginous chert with millimeter-thick hematite- and chert- Mesabi Range is its western equivalent. Stille and Clauer (1986)
rich layers (Kato and Kimura, 1999). obtained a whole rock Sm-Nd isochron age of 2.08 ± 0.25 Ga
from argillites in the Gunflint Formation, but its depositional age
Kuruman and Penge, South Africa are generally estimated to be ca. 1.9 Ga (Jacobsen and Pimentel-
Klose, 1988). The Sokoman IF is exposed throughout the Labra-
The Kuruman and Penge IFs occur in the Griqualand West dor Trough and known as one of the world’s largest iron deposits.
sub-basin and the Eastern Transvaal sub-basin, respectively, in Fryer (1972) reported a whole rock Rb-Sr isochron age of 1.87 ±
South Africa where the Transvaal Supergroup is extensively 0.05 Ga for a series of slate samples from the Sokoman IF.
exposed. The Penge IF is correlative with and lithologically
similar to the Kuruman IF (Bau and Dulski, 1996). According to Rapitan, Canada, and Braemar, Australia
U-Pb ages for zircons from volcanic ash beds, the Kuruman IF
is well dated between 2521 and 2432 Ma (Trendall et al., 1990; Two data sets of Neoproterozoic IFs were included in this
Sumner and Bowring, 1996). Most of the Transvaal Supergroup study, the Rapitan IF in Canada (Fryer, 1977; Klein and Beukes,
was subject to zeolite-facies metamorphism, although metamor- 1993) and the Braemar IF in South Australia (Lottermoser and
phic grades locally reached hornblende-hornfels facies adjacent Ashley, 2000). The Rapitan IF probably deposited between 755
to the Bushveld Complex. The Kuruman and Penge IFs analyzed and 730 Ma (see Klein and Beukes, 1993). Selected samples
for REEs by Bau and Dulski (1996) are drill-core samples and from the Rapitan IF are banded, nodular, or arenose (Al2O3 < 1.2
consist of alternating layers of siderite, hematite, and magnetite wt%). Their mineralogy is very simple, consisting of hematite
as well as chert. cemented by microquartz. The Braemar IF occurs in the glacio-
marine sequences of the Sturtian glaciation (750–700 Ma). The
Hotazel, South Africa Braemar ironstone facies consists of lenticular laminated and
diamictic ironstones. The mineralogy of the laminated ironstones
The Hotazel Formation represents the youngest episode of and matrix of diamictic ironstones is simple: magnetite, hema-
BIF deposition in the Transvaal Supergroup and overlies andes- tite, and quartz with minor muscovite, chlorite, biotite, carbonate,
itic lavas of the Ongeluk Formation, which are dated at 2222 Ma apatite, plagioclase, and tourmaline. It should be noted that Al2O3
(Cornell et al., 1996). Four distinct BIF units are interbedded contents in the Braemar IF are exceptionally high (2.7–4.0 wt%),
with three Mn beds in this formation. The lower Mn bed is exten- reflecting the abundance of aluminous minerals such as musco-
sively mined as the world’s largest Mn resource (Kalahari Mn vite, chlorite, biotite, and plagioclase.
field). The Hotazel IF comprises laminae of chert, iron oxides
(magnetite, hematite), iron silicates (greenalite, minnesotaite, Ce AND Eu ANOMALIES IN BIFS
stilpnomelane, riebeckite, iron-rich mica), carbonates (calcite,
ankerite, siderite), and pyrite (Tsikos and Moore, 1997). There REE characteristics of the selected BIFs are summarized in
is a stratigraphic change in mineralogy. Oxide-dominated facies Table 2, and chondrite-normalized REE patterns and Ce/Ce*-Pr/
characterize the lower parts of the Hotazel IF, whereas silicate- Pr* diagrams for them are given in Figures 3 and 4, respectively.
and carbonate-dominated facies increase in upper sections. Histograms of Ce/Ce* and Eu/Eu* values in the BIFs together
with modern seawater, EPR Fe-Mn sediments, and Phanerozoic
Väyrylänkylä, Finland umbers are also given in Figures 5 and 6, respectively. The REE
indices, such as ∑REE, Ce/Ce*, and Eu/Eu* values, of Early
The Väyrylänkylä IF (Pääkkö and Iso Vuorijärvi IFs) depos- Archean IFs (the Isua and Marble Bar IFs) are very similar with
ited in the Kainuan schist belt, Salmijärvi Basin, Finland. The each other (Table 2, Figs. 3–6), in spite of their differences in
whole rock Pb-Pb isochron age of the Pääkkö IF is 2080 ± 45 Ma mineralogy. Both BIFs are characterized by distinct positive Eu
(Sakko and Laajoki, 1975). The IF samples are composed mainly anomalies and weak but indisputable negative Ce anomalies (Ce/
of quartz and magnetite (Laajoki, 1975; Danielson et al., 1992). Ce* = 0.6–0.9). Middle Archean IFs (the Cleaverville and Sargur
The Väyrylänkylä IF is characterized by small dimensions, low IFs) have distinct positive Eu anomalies, but their magnitudes are
Fe grade (Fetot ~26 wt%), and exceptionally high P content (~1.2 not as large as those of the Early Archean IFs (Fig. 6). Negative
wt%), and hence are uneconomic (Laajoki, 1975). Manganifer- Ce anomalies are generally weak to moderate (Ce/Ce* = 0.5–0.9;
ous rocks locally occur in the formation. Fig. 5C) in the Cleaverville IF, whereas the Ce anomaly values of
the Sargur IF are almost close to one (no Ce anomaly; Fig. 5D).
Gunflint, Biwabik, and Sokoman, Canada Generally, positive Eu anomalies in the Late Archean IFs
are remarkable and are similar to those of the Early Archean IFs
Data on the Gunflint, Biwabik, and Sokoman IFs for the (Fig. 6). Eu anomalies of the Kushtagi and northern Holenarsi-
present paper are from Fryer (1977), Derry and Jacobsen (1990), pur IFs become less striking with increasing ∑REE. This gradual
and Danielson et al. (1992). The Gunflint IF is exposed along change in Eu anomaly with increasing ∑REE is also observed in
the Marble Bar (3.5 Ga) and Cleaverville (3.2–3.0 Ga) IFs (Kato out the possible presence of slightly negative Ce anomalies in
et al., 1998; Kato and Nakamura, 2003). Negative Ce anomalies the Isua IF. However, whether or not negative Ce anomalies truly
are distinguished in many Late Archean BIF samples, with the exist in the 3.8–3.7 Ga IF has been ambiguous, because their REE
Ce/Ce* values less than 0.6 (Fig. 5). Furthermore, some portions data were determined by instrumental neutron activation analy-
of samples exhibit strong negative Ce anomalies, like modern sis and hence were not complete. Especially, the absence of Pr
seawaters, with the Ce/Ce* values less than 0.3. The 2.9–2.7 Ga analyses has made it difficult to determine whether the apparent
Sandur IF shows a sizable variation in the REE patterns (Fig. negative Ce anomalies were simple reflections of the presence of
3H). Although La enrichment is recognizable in some samples, positive La anomalies. However, data obtained by the ICP-MS
true negative Ce anomalies are striking and close to those of method (Morinaga, 1997) clearly demonstrate the true existence
modern oxygenated seawaters. The Ce anomalies vary from of negative Ce anomalies in the 3.8–3.7 Ga IF. This suggests that
extremely positive (Ce/Ce* = 5.1 at maximum) to greatly nega- some sort of Ce oxidation mechanism already existed in the 3.8–
tive (Ce/Ce* = 0.02 at minimum; Figs. 4H, 5H). The Sandur IF 3.7 Ga ocean and that the seawater column from which the Isua
shows the greatest variation of Eu/Eu* from 0.5 to 7.1, although IF precipitated was not entirely anoxic. The Ce/Ce* values range
positive Eu anomalies are predominant (Fig. 6H). from ~0.5 (moderately negative) to ~1.0 (no anomaly) in BIFs of
The 2.7–2.4 Ga Hamersley IFs (the Marra Mamba, Brock- ca. 3.8–3.5 Ga in age, suggesting the development of relatively
man, and Boolgeeda IFs in ascending stratigraphic position) show oxygenated oceanic conditions where aqueous Fe2+ precipitated
small or no Ce anomalies (Ce/Ce* = 0.72–1.0; Fig. 5), which con- as Fe-oxyhydroxides.
trasts with the strong negative Ce anomalies in the Late Archean Stronger negative Ce anomalies are observed in many BIFs
Indian BIFs. The Eu anomalies in the Hamersley BIFs vary from of 2.9–2.7 Ga age (Late Archean), suggesting that the same oxy-
negative to positive (Eu/Eu* = 0.84–1.38; dominated by positive genated conditions as those of today existed in the 2.9–2.7 Ga
anomalies; Figs. 6I, 6J, and 6L). Eu/Eu* values decrease from oceans. Even the average Ce/Ce* values of the Late Archean
the Marra Mamba to the Boolgeeda IFs. Generally, the Kuruman Indian IFs, excepting the Sandur IF, are distinctly negative (0.50
and Penge IFs show concave REE patterns that are similar to for the Holenarsipur IF) or moderately negative (~0.75 for the
the Hamersley IFs (Fig. 3K). The Ce/Ce* values fall in a narrow Bababudan and Kushtagi IFs) (Table 2), unequivocally demon-
range from 0.76 to 0.96, and there is no significant difference strating the widespread presence of negative Ce anomalies in
between the Hamersley and Kuruman/Penge IFs (Fig. 5K). Late Archean seawaters.
True and significant negative Ce anomalies are well recog- The above suggestion is consistent with the presence of
nized in the 2.2–2.1 Ga IFs (the Hotazel and Väyrylänkylä IFs; small but widespread Mn deposits that formed at the same time
Figs. 4, 5). The Hotazel IF has distinctly low Eu/Eu* values from as the BIFs in the Dharwar Craton. The geochemical behavior
0.50 to 0.93, whereas the Väyrylänkylä IF has the Eu/Eu* values of dissolved Ce is similar to that of dissolved Mn, because the
narrowly ranging from 0.86 to 1.11. Negative Ce anomalies are free-energy changes for Ce4+ and Mn4+ reduction are quite similar
generally absent or moderate in the ca. 1.9 Ga Gunflint, Biwawik (Elderfield, 1988). In modern oceans, oxidation of Ce3+ to Ce4+-
and Sokoman IFs (Figs. 4O, 5O), and their Eu/Eu* values are rel- oxide mostly takes place in the upper water column (Alibo and
atively constant (Fig. 6O; 0.89 ± 0.15). The ca. 0.7 Ga IFs exhibit Nozaki, 1999). Even without additional removal of Ce from the
nearly flat or subtly LREE-enriched patterns (Fig. 3P). The Brae- deep water, negative Ce anomalies would be enhanced with time
mar IF, which has higher Al2O3 contents compared to other IFs, because of the preferential regeneration of REE3+ except Ce in
has greater REE concentrations (chondrite-normalized La val- the oxygenated deep-sea water (Alibo and Nozaki, 1999). Hence,
ues = 14.3–40.2) compared to the other BIF samples discussed the contemporaneous Mn deposits in the Dharwar Craton may
above, clearly reflecting contributions of REEs in clastic sedi- have formed in shallow environments associated with the oxida-
ments. The Ce anomalies of the Braemar ironstones are absent tive removal of Ce, although positive Ce anomalies have not yet
or very weak (Ce/Ce* = ~0.92), whereas those of the Rapitan IF been reported for these Mn deposits (Manikyamba and Naqvi,
are moderately negative (Ce/Ce* = ~0.66). The contamination 1997). The Mn formations, which occur with some BIFs (e.g.,
of detrital materials may have obscured negative Ce anomalies the Sandur and Chitradurga IFs), are characterized by an associa-
in the Braemar IF. The magnitudes of negative Eu anomalies in tion with stromatolitic carbonate rocks, suggesting their precipi-
the Braemar IF are very similar to those of the Rapitan IF, with tation in shallow environments. Other Indian BIFs (the Bababu-
remarkably constant Eu/Eu* values (0.75 ± 0.08; Fig. 6P). dan, Kushtagi, and northern Holenarsipur IFs), which lack Mn
formations and stromatolitic carbonate rocks, were most likely
DISCUSSION deposited in deep seas (Kato et al. 2002a). Thus, it is very likely
that fully oxygenated seawaters like those today existed in both
The Redox History of the Precambrian Oceans deep and shallow environments at 2.9–2.7 Ga.
In addition to negative Ce anomalies, huge positive Ce
Figure 7, which shows the secular variation of Ce anomaly anomalies exist in some parts of the Sandur IF, suggesting that
in BIFs, gives us new insight into the oxygen scenario for the redox-related Ce fractionation took place in the Late Archean
Precambrian Earth’s surface. Dymek and Klein (1988) pointed ocean. As the mixing of hydrothermal Fe-Mn precipitates with
280 Y. Kato et al.

100 100
A Isua B Marble Bar
(3.8-3.7 Ga) (3.5 Ga)
10 10
1 1
0.1 0.1
100 100
Cleaverville
(3.2-3.0 Ga)
D Sargur
(3.1-3.0 Ga)
10
10
Chondrite-normalized
1
1
0.1 C
0.1
100 100
E Bababudan F Kushtagi
(2.9-2.7 Ga) (2.9-2.7 Ga)
10 10
1 1
0.1 0.1
100 100
Holenarsipur Sandur
(2.9-2.7 Ga) (2.9-2.7 Ga)
10
10
1
1
G 0.1 H
0.1
La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er Tm Yb Lu La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er Tm Yb Lu
Figure 3 (on this and following page). Chondrite-normalized REE patterns of BIFs.
hydrogenetic Mn crusts occurs in the modern ocean system (Kuhn of ∑REE and negative Ce anomalies in the lower soil sections
et al., 1998), it seems likely that the Sandur IF precursors were (e.g., Koppi et al., 1996; Braun et al., 1998). Detailed investiga-
formed by the mixing of hydrothermal Fe-Mn (with negative Ce tions on the Sandur IF are necessary to determine the origins of
anomalies) and hydrogenetic Mn (with positive Ce anomalies) these Ce anomalies.
precipitates in various proportions. The Sandur IF samples were The Hamersley and Transvaal BIFs, which formed during
collected from borehole cores and recent railway and road cut- the period from 2.7 Ga to 2.4 Ga, have weak negative Ce anoma-
tings (Manikyamba et al., 1993), and hence weathering effects lies, varying narrowly from 0.7 to 1.0 in their Ce/Ce* ratios. The
on these samples were probably insignificant. Nevertheless, the magnitudes of their negative Ce anomalies are clearly smaller
fact that the average Ce/Ce* value of the Sandur IF samples is than those of Early and Middle Archean BIFs. This suggests that
close to one (Table 2) raises the possibility that these positive Ce the water columns from which these BIFs precipitated were less
anomalies, as well as negative Ce anomalies, are products of Ce oxygenated than those for the pre-2.7 Ga BIFs. A plausible expla-
fractionation by modern weathering. It is well known that posi- nation is that the depositional environments for the Hamersley
tive and negative Ce anomalies are found in some (not all) mod- and Transvaal BIFs were related to evolving rifts (Alibert and
ern soil profiles: typically, depletions of ∑REE and positive Ce McCulloch, 1993; Ohmoto, 1993). Their depositional settings
anomalies are found in the upper soil sections, and enrichments are quite different from those for the pre-2.7 Ga Algoma-type
100 100
I Marra Mamba J Brockman

(2.7-2.6 Ga) (2.5 Ga)
10 10
1 1
0.1 0.1
100 100
K Kuruman & Penge L Boolgeeda
(2.5-2.4 Ga) (2.4 Ga)
10 10
1 1
Chondrite-normalized
0.1 0.1
100 100
Hotazel
(2.2 Ga)
10 10
1 1
Väyrylänkylä
M N (2.1 Ga)
0.1 0.1
100 100
Gunflint, Biwabik &
Sokoman (1.9 Ga)
10 10
1 1
Rapitan & Braemar

O P (0.7 Ga)
0.1 0.1
La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er Tm Yb Lu La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er Tm Yb Lu
Figure 3 (continued).
BIFs that were mostly deposited in the open-ocean environments anomalies. The Ce/Ce* fluctuation, which shows diminishing
(Kato et al., 1998). A Fe2+-rich brine pool may have accumulated negative Ce anomalies at 2.7–2.5 Ga and ca. 1.9 Ga, may have
in land-locked, anoxic basins much like the modern Red Sea been related to the fluctuation in the global mantle plume activity
brine pools (Ohmoto, 1993; Ohmoto et al., this volume). In this (Isley and Abott, 1999). The intensive global mantle plume activ-
scenario, the geochemical signatures preserved in the post-2.7 ity at 2.7–2.5 Ga and ca. 1.9 Ga created many rift-related basins
Ga BIFs are indicators only of local environments. Globally, the (e.g., the Hamersley Basin and the Lake Superior region) and
atmospheric and oceanic redox conditions may have been oxy- supplied great amounts of Fe2+ to the closed basins by subma-
genated persistently since at least 2.9–2.7 Ga. rine hydrothermal fluids. The hydrothermal Fe2+ reacted with O2
The Hotazel and Väyrylänkylä IFs with distinct nega- in the water column to precipitate large amounts of iron oxides
tive Ce anomalies (Ce/Ce* = 0.2–0.7) deposited at 2.2–2.1 Ga. (i.e., Superior-type BIFs) and to create anoxic water bodies. Fur-
Although their negative anomalies are not as strong as those thermore, the observed Ce/Ce* fluctuation in Late Archean to
of the Late Archean Indian BIFs, the close association of these Paleoproterozoic BIFs may represent the redox conditions only
IFs with Mn deposits suggests that oxygenated conditions pre- of local seas, rather than of the global oceans. Such explanations
vailed in the Hotazel and Väyrylänkylä oceans. The ca. 1.9 Ga can also explain why the ca. 0.7 Ga Rapitan and Braemar IFs
Gunflint, Biwabik, and Sokoman IFs exhibit less significant Ce exhibit only minor negative Ce anomalies.
1.2 1.2
1.8
Isua Marble Bar Cleaverville
(3.8-3.7 Ga) (3.5 Ga) 1.6 (3.2-3.0 Ga)
1.0 1.0
1.4
0.8 0.8 1.2
1.0
0.6 0.6 0.8
0.4
A 0.4
B 0.6
0.4
C
0.8 1.0 1.2 1.4 0.8 1.0 1.2 1.4 0.8 1.0 1.2 1.4
1.2 1.2 1.2
Sargur Bababudan Kushtagi
(3.1-3.0 Ga) (2.9-2.7 Ga) (2.9-2.7 Ga)
1.0 1.0 1.0
0.8 0.8 0.8
0.6 0.6 0.6
0.4
D 0.4
E 0.4
F
0.8 1.0 1.2 1.4 0.8 1.0 1.2 1.4 0.8 1.0 1.2 1.4
1.2 6 1.2
Holenarsipur Sandur Marra Mamba
1.0 (2.9-2.7 Ga) 5 (2.9-2.7 Ga) (2.7-2.6 Ga)
1.0
0.8 4
0.6 3 0.8
0.4 2
Ce/Ce*
0.6
0.2 1
0
G 0
H 0.4
I
0.5 1.0 1.5 2.0 2.5 3.0 0 1 2 3 4 5 6 0.8 1.0 1.2 1.4
1.2 1.2 1.2
Brockman Kuruman & Penge Boolgeeda
(2.5 Ga) (2.5-2.4 Ga) (2.4 Ga)
1.0 1.0 1.0
0.8 0.8 0.8
0.6 0.6 0.6
0.4
J 0.4
K 0.4
L
0.8 1.0 1.2 1.4 0.8 1.0 1.2 1.4 0.8 1.0 1.2 1.4
1.2 1.2 1.2
Hotazel Väyrylänkylä Gunflint, Biwabik &
(2.2 Ga) 1.0 (2.1 Ga) Sokoman (1.9 Ga)
1.0 1.0
0.8
0.8 0.6 0.8
0.4
0.6 0.6
0.2
0.4
M 0
N 0.4
O
0.8 1.0 1.2 1.4 0.5 1.0 1.5 2.0 2.5 3.0 0.8 1.0 1.2 1.4
1.2
Rapitan & Braemar
1.0
(0.7 Ga)
Pr/Pr* (Nd/Nd*)
0.8
Figure 4. Ce/Ce* versus Pr/Pr* (Nd/Nd*) relationships for BIFs. Nd/
Nd* values are used instead of Pr/Pr* values in the Marra Mamba,
0.6
Brockman, Väyrylänkylä, Lake Superior, and Rapitan/Braemar IFs be-
0.4
P cause of the lack of Pr data for these IFs. The region of true negative
Ce anomalies is shown as a shaded area.
0.8 1.0 1.2 1.4 1.6 1.8
30
Isua A 20 Marra Mamba I 6
(3.8-3.7 Ga) (2.7-2.6 Ga) 4
10 2
0 0
20 Marble Bar B 12
Brockman J
(3.5 Ga) 8 (2.5 Ga)
10
4
0 40 0 15
Cleaverville C 30 Kuruman & Penge K 10
(3.2-3.0 Ga) (2.5-2.4 Ga)
20
5
10
0 20 0
12
2.0
8
Sargur
(3.1-3.0 Ga)
D 15 Boolgeeda
(2.4 Ga)
L
10
4 5
0 0 5
Bababudan
(2.9-2.7 Ga)
E 6
Hotazel M 4
4 (2.2 Ga) 3
2
2
1
0 0
16 4
12 Kushtagi
(2.9-2.7 Ga)
F 3 Väyrylänkylä
(2.1 Ga)
8 2
4 1
N
0 8 0 5
Holenarsipur
(2.9-2.7 Ga)
6 Gunflint, Biwabik & O 4
Sokoman (1.9 Ga) 3
4
2
G 2
1
0 0
0 0.5 1.0 1.5
10
6
Rapitan & Braemar P
8 Sandur 4 (0.7 Ga)
6 (2.9-2.7 Ga)
2
4
2
0
H 0
deep-sea water
50
40
0 1 2 3 4 5 6 EPR Fe-Mn sediments
30
Phanerozoic Fe-Mn umbers
20
Ce/Ce*
surface seawater
Q 10
0
0 0.5 1.0 1.5
Figure 5. Histograms of Ce/Ce* values in BIFs. For comparison, Ce/Ce* values of modern seawater, EPR Fe-Mn sediments, and Phanerozoic
Fe-Mn umbers are presented (Q). True negative Ce anomalies, evaluated from the Ce/Ce*-Pr/Pr* diagrams, are shaded in these BIF histograms.
12 7
10 6
Isua 5 Marra Mamba
8 (3.8-3.7 Ga) (2.7-2.6 Ga)
4
6
3
A I
4 2
2 1
0 0
10 14
8 12
Marble Bar Brockman 10
(3.5 Ga) 6 (2.5 Ga) 8
4 6
B 2
0
J 4
2
0
25 12
20 Cleaverville 10 Kuruman & Penge
(3.2-3.0 Ga) 8 (2.5-2.4 Ga)
15
6
10
C K
4
5 2
0 0
8 20
Sargur 6 Boolgeeda 15
(3.1-3.0 Ga) (2.4 Ga)
4 10
D 2
0
L 5
0
10 5
8 Bababudan 4 Hotazel
6 (2.9-2.7 Ga) 3 (2.2 Ga)
4 2
2
0
E 1
0
M
6 6
Kushtagi 5 Väyrylänkylä 5
4 4
(2.9-2.7 Ga) (2.1 Ga)
3 3
F N
2 2
1 1
0 0
7 7
6 Holenarsipur 6 Gunflint, Biwabik &
5 5
4 (2.9-2.7 Ga) 4 Sokoman (1.9 Ga)
3 3
2
1 G 2
1 O
0 0
8
H
5
Sandur 4 Rapitan & Braemar 6
50
(2.9-2.7 Ga) 3 (0.7 Ga) 4
P
2 40
2
1
30
0 70 0
0 1 2 3 4 5 6 7 8
60 20 seawater
50 EPR Fe-Mn sediments
40 10 Phanerozoic Fe-Mn umbers
30
Eu/Eu* 20
10
0
0 0.2 0.4 0.6 0.8 1.0 Q
0
0 1 2 3 4 5 6 7 8
Figure 6. Histograms of Eu/Eu* values in BIFs. The Eu/Eu* values of modern seawater, EPR Fe-Mn sediments, and Phanerozoic Fe-Mn umbers
are also presented (Q).
5.1
4.6
3.8
2 19
true negative Ce anomaly

false negative Ce anomaly
1.5 positive Ce anomaly
Ce/Ce*
10 12
1 21
4
9
11
16
0.5 1 2 15
3 13
5-8 18
17
14
20
0
4 3 2 1 0
Age (Ga)
Figure 7. Secular variation of Ce/Ce* values in BIFs. The Ce/Ce* values of modern seawater, MOR hydrothermal fluids, EPR Fe-Mn sediments,
Phanerozoic Fe-Mn umbers, and hydrogenetic Mn crusts are also given. Numbers correspond to the stratigraphic units in Table 1.
Secular Variation in Hydrothermal versus Riverine Abbott, 1999; Kato, 2003). The global mantle plume activ-
Influxes into the Precambrian Oceans ity could also have caused the episodic growth of continen-
tal crust and the formation of rift basins (e.g., Condie, 2000),
The secular variation of Eu anomalies in BIFs is summa- which created the conditions to increase the riverine influx into
rized in Figure 8. The figure shows that the pre-2.7 Ga Algoma- the land-locked basins and surpass the hydrothermal influx to
type BIFs have significantly positive Eu anomalies, although lower the Eu/Eu* values in the BIF-precipitating water column.
quite variable. Their remarkably positive Eu anomalies sug- The relatively constant Eu/Eu* values in the post-2.7 Ga BIFs
gest that the water column from which the BIFs precipitated may further suggest that the riverine flux from the surrounding
had strong positive Eu anomalies, probably caused by larger continent into the closed basins was nearly constant during the
contributions of hydrothermal fluids. The large variation in the deposition of BIFs. Within the Hamersley Basin, a gradual and
Eu anomalies was probably caused by a large variation in the slight decrease in the Eu/Eu* values from the 2.7–2.6 Ga Marra
mixing ratios of hydrothermal fluids and seawaters at various Mamba IF (~1.10), to the 2.5 Ga Brockman IF (~1.07), and
depositional sites in deep-sea pelagic environments (Kato et then to the 2.4 Ga Boolgeeda IF (~0.91) is recognizable (Figs.
al., 1998). 6I, 6J, and 6L). This may mean the hydrothermal component
Figure 8 shows a drastic change in Eu/Eu* values of BIFs decreased and the riverine (continental) component increased
ca. 2.7 Ga. The Superior-type BIFs and the Rapitan-type IFs, comparatively during the period between 2.7 and 2.4 Ga.
which are all younger than 2.7 Ga, exhibit much lower Eu/Eu*
values than the pre-2.7 Ga Algoma-type BIFs. Furthermore, CONCLUSIONS
their Eu/Eu* range is quite narrow and constant. The signifi-
cantly lower Eu/Eu* values probably resulted because the BIF- 1. Negative Ce anomalies in the 3.8–3.7 Ga Isua BIFs sug-
precipitating water columns were dominated by the riverine gest that the water column from which the BIFs pre-
influx with lower Eu/Eu* values. That is, the riverine influx cipitated was not entirely anoxic and that a Ce oxidation
overwhelmed the hydrothermal influx into the land-locked mechanism already existed in the 3.8–3.7 Ga oceans.
basins where the post-2.7 Ga Superior-type BIFs formed. The 2. The presence of strong negative Ce anomalies in the
global mantle plume activity was probably responsible for the 2.9–2.7 Ga BIFs indicates that strongly oxygenated
widespread hydrothermal activity that formed the voluminous oceanic conditions, much like today, already existed at
Superior-type BIFs at 2.7–2.4 Ga and ca. 1.9 Ga (Isley and 2.9–2.7 Ga.
286 Y. Kato et al.
21
10
5
Figure 8. Secular variation of Eu/Eu*
values in BIFs. The Eu/Eu* values of
8
modern seawater, MOR hydrothermal
1
2 fluids, EPR Fe-Mn sediments, Phanerozoic
5
Fe-Mn umbers, and hydrogenetic Mn crusts
Eu/Eu*
6
3 are given. Numbers correspond to the
7 stratigraphic units in Table 1. Box and bar
4
represent an average Eu/Eu value and ±1σ,
9 11 respectively.
1 14
10 12 15
0.9 18
16
0.8 13 17 19
0.7
0.6 20
0.5
4 3 2 1 0
Age (Ga)
3. The 2.7–2.4 Ga IFs and the ca. 1.9 Ga IFs have slightly
Yoshimichi Kajiwara, Shigenori Maruyama, Tsuyoshi Komiya,
negative Ce anomalies. These BIFs were probably depos-
Michael Bau, Yumiko Watanabe, and the late Yoshiyuki Nozaki
ited in evolving rifts where less oxygenated water con-
are deeply appreciated. We also thank Stephen E. Kesler, Richard
ditions were created by intense hydrothermal activity in
W. Murray, and two anonymous reviewers for helpful comments
closed basins.
that significantly improved the manuscript.
4. The presence of strong positive Eu anomalies in the pre-
2.7 Ga Algoma-type BIFs suggests large contributions of
REFERENCES CITED
hydrothermal fluids into the seawater column. The large
variation of Eu anomalies in the Algoma-type BIFs reflects Alibert, C., and McCulloch, M.T., 1993, Rare earth element and neodymium
a large variation in the mixing ratios of hydrothermal flu- isotopic compositions of the banded iron-formations and associated
ids and seawaters at various depositional sites. shales from Hamersley, western Australia: Geochimica et Cosmochimica
Acta, v. 57, p. 187–204, doi: 10.1016/0016-7037(93)90478-F.
5. The post-2.7 Ga BIFs exhibit much lower and constant Alibo, D.S., and Nozaki, Y., 1999, Rare earth elements in seawater: Particle
Eu/Eu* values than the pre-2.7 Ga BIFs, suggesting that association, shale-normalization, and Ce oxidation: Geochimica et Cos-
water in land-locked basins where the post-2.7 Ga BIFs mochimica Acta, v. 63, p. 363–372, doi: 10.1016/S0016-7037(98)00279-8.
Arndt, N.T., Nelson, D.R., Compston, W., Trendall, A.F., and Thorne, A.M.,
accumulated was strongly influenced by a riverine influx 1991, The age of the Fortescue Group, Hamersley Basin, Western Austra-
compared to hydrothermal influx. lia, from ion microprobe zircon U-Pb results: Australian Journal of Earth
Sciences, v. 38, p. 261–281.
Arora, M., Govil, P.K., Charan, S.N., Uday Raj, B., Balaram, V., Manikyamba,
ACKNOWLEDGMENTS C., Chatterjee, A.K., and Naqvi, S.M., 1995, Geochemistry and origin of
Archean banded iron-formation from the Bababudan schist belt, India:
Economic Geology and the Bulletin of the Society of Economic Geolo-
This work was supported by the Japanese Ministry of Edu- gists, v. 90, p. 2040–2057.
cation, Culture, Sports, Science and Technology through Grant- Aubert, D., Stille, P., and Probst, A., 2001, REE fractionation during granite
weathering and removal by waters and suspended loads: Sr and Nd isoto-
in Aid Nos. 07238105, 06041038, 08041102, 06740415, and pic evidence: Geochimica et Cosmochimica Acta, v. 65, p. 387–406, doi:
07740426 to Kato and 07041081 to Ohmoto, and by funds from 10.1016/S0016-7037(00)00546-9.
cooperative programs (No. 40 in 1994, No. 28 in 1995, and No. Barley, M.E., 1993, Volcanic, sedimentary and tectonostratigraphic environ-
ments of the ~3.46 Ga Warrawoona Megasequence: A review: Precam-
32 in 1997) to Kato provided by the Ocean Research Institute, brian Research, v. 60, p. 47–67, doi: 10.1016/0301-9268(93)90044-3.
University of Tokyo. Ohmoto is also supported by NASA Astro- Barley, M.E., Pickard, A.L., and Sylvester, P.J., 1997, Emplacement of a large
biology Institute (NCC2–1057; CAN#NNA04CC06A), NASA igneous province as a possible cause of banded iron formation 2.45 billion
years ago: Nature, v. 385, p. 55–58, doi: 10.1038/385055a0.
Exobiology Program (NAG5–9089), and National Science Foun- Barrett, T.J., and Jarvis, I., 1988, Rare-earth element geochemistry of metallifer-
dation (EAR-9706279, EAR-0229556). Yamaguchi acknowl- ous sediments from DSDP Leg 92: The East Pacific Rise transect: Chemi-
edges financial support from IFREE and NASA for his stay at the cal Geology, v. 67, p. 243–259, doi: 10.1016/0009-2541(88)90131-3.
Barrett, T.J., Fralick, P.W., and Jarvis, I., 1988, Rare-earth-element geochem-
University of Wisconsin, Madison. Stimulating discussion with istry of some Archean iron formations north of Lake Superior, Ontario:
Canadian Journal of Earth Sciences, v. 25, p. 570–580. Elderfield, H., Hawkesworth, C.J., Greaves, M.J., and Calvert, S.E., 1981, Rare
Bau, M., and Dulski, P., 1996, Distribution of yttrium and rare-earth elements earth element geochemistry of oceanic ferromanganese nodules and asso-
in the Penge and Kuruman iron-formations, Transvaal Supergroup, ciated sediments: Geochimica et Cosmochimica Acta, v. 45, p. 513–528,
South Africa: Precambrian Research, v. 79, p. 37–55, doi: 10.1016/0301- doi: 10.1016/0016-7037(81)90184-8.
9268(95)00087-9. Fryer, B.J., 1972, Age determinations in the Circum-Ungava Geosyncline and
Bau, M., and Dulski, P., 1999, Comparing yttrium and rare earths in hydro- the evolution of Precambrian banded iron-formations: Canadian Journal
thermal fluids from the Mid-Atlantic Ridge: Implications for Y and REE of Earth Sciences, v. 9, p. 652–663.
behavior during near-vent mixing and for the Y/Ho ratio of Proterozoic Fryer, B.J., 1977, Rare earth evidence in iron-formations for changing Precam-
seawater: Chemical Geology, v. 155, p. 77–90, doi: 10.1016/S0009- brian oxidation states: Geochimica et Cosmochimica Acta, v. 41, p. 361–
2541(98)00142-9. 367, doi: 10.1016/0016-7037(77)90263-0.
Bau, M., and Möller, P., 1993, Rare earth element systematics of the chemically Fujinaga, K., Kato, Y., Kiminami, K., Miura, K., and Nakamura, K., 1999, Geo-
precipitated component in early Precambrian iron formations and the evo- chemistry of red shale from the Mugi Formation in the Upper Cretaceous
lution of the terrestrial atmosphere-hydrosphere-lithosphere system: Geo- Shimanto Belt, Shikoku, Japan: Memoirs of Geological Society of Japan,
chimica et Cosmochimica Acta, v. 57, p. 2239–2249, doi: 10.1016/0016- no. 52, p. 205–216 (in Japanese with English abstract).
7037(93)90566-F. German, C.R., and Elderfield, H., 1989, Rare earth elements in Saanich Inlet,
Bau, M., Koschinsky, A., Dulski, P., and Hein, J.R., 1996, Comparison of British Columbia, a seasonally anoxic basin: Geochimica et Cosmochi-
the partitioning behaviors of yttrium, rare earth elements, and titanium mica Acta, v. 53, p. 2561–2571, doi: 10.1016/0016-7037(89)90128-2.
between hydrogenetic marine ferromanganese crusts and seawater: Geo- German, C.R., and Elderfield, H., 1990, Application of the Ce anomaly as
chimica et Cosmochimica Acta, v. 60, p. 1709–1725, doi: 10.1016/0016- a paleoredox indicator: The ground rules: Paleoceanography, v. 5,
7037(96)00063-4. p. 823–833.
Bender, M., Broecker, W., Gornitz, V., Middle, U., Kay, R., Sun, S., and Bis- German, C.R., Klinkhammer, G.P., Edmond, J.M., Mitra, A., and Elderfield,
caye, P., 1971, Geochemistry of three cores from the East Pacific Rise: H., 1990, Hydrothermal scavenging of rare-earth elements in the ocean:
Earth and Planetary Science Letters, v. 12, p. 425–433, doi: 10.1016/0012- Nature, v. 345, p. 516–518, doi: 10.1038/345516a0.
821X(71)90028-8. Goldstein, S.J., and Jacobsen, S.B., 1988, Rare earth elements in river waters:
Braun, J.J., Viers, J., Dupré, B., Polve, M., Ndam, J., and Muller, J.P., 1998, Earth and Planetary Science Letters, v. 89, p. 35–47, doi: 10.1016/0012-
Solid/liquid REE fractionation in the lateritic system of Goyoum, East 821X(88)90031-3.
Cameroon: The implication for the present dynamics of the soil covers Grauch, R.I., 1989, Rare earth elements in metamorphic rocks, in Lipin, B.R.,
of the humid tropical regions: Geochimica et Cosmochimica Acta, v. 62, and McKay, G.A., eds., Geochemistry and mineralogy of rare earth ele-
p. 273–299, doi: 10.1016/S0016-7037(97)00344-X. ments: Reviews in Mineralogy, v. 21, p. 147–167.
Byrne, R.H., and Kim, K., 1990, Rare earth element scavenging in sea- Gross, G.A., 1965, Geology of iron deposits in Canada: I. General geology
water: Geochimica et Cosmochimica Acta, v. 54, p. 2645–2656, doi: and evaluation of iron deposits: Geological Survey of Canada Economic
10.1016/0016-7037(90)90002-3. Geology Report 22, 181 p.
Chadwick, B., Vasudev, V.N., and Hedge, G.V., 2000, The Dharwar craton, Hickman, A.H., 1983, Geology of the Pilbara Block and its environs: Geologi-
southern India, interpreted as the result of Late Archaean oblique con- cal Survey of Western Australia Bulletin, v. 127, 268 p.
vergence: Precambrian Research, v. 99, p. 91–111, doi: 10.1016/S0301- Isley, A.E., and Abbott, D.H., 1999, Plume-related mafic volcanism and the
9268(99)00055-8. deposition of banded iron formation: Journal of Geophysical Research,
Condie, K.C., 1993, Chemical composition and evolution of the upper conti- v. 104, p. 15461–15477, doi: 10.1029/1999JB900066.
nental crust: Contrasting results from surface samples and shales: Chemi- Jacobsen, S.B., and Pimentel-Klose, M.R., 1988, A Nd isotopic study of the
cal Geology, v. 104, p. 1–37, doi: 10.1016/0009-2541(93)90140-E. Hamersley and Michipicoten banded iron formations: The source of REE
Condie, K.C., 2000, Episodic continental growth models: Afterthoughts and and Fe in Archean oceans: Earth and Planetary Science Letters, v. 87,
extensions: Tectonophysics, v. 322, p. 153–162, doi: 10.1016/S0040- p. 29–44, doi: 10.1016/0012-821X(88)90062-3.
1951(00)00061-5. James, H.L., 1954, Sedimentary facies of iron-formations: Economic Geol-
Cornell, D.H., Schutte, S.S., and Eglington, B.L., 1996, The Ongeluk basaltic ogy and the Bulletin of the Society of Economic Geologists, v. 49,
andesite formation in Griqualand West, South Africa: Submarine altera- p. 235–293.
tion in a 2222 Ma Proterozoic sea: Precambrian Research, v. 79, p. 101– Kato, Y., 2003, Banded iron formations: Their genesis and implications for
123, doi: 10.1016/0301-9268(95)00090-9. Precambrian Earth: Geochimica et Cosmochimica Acta, v. 67, no. 18S,
Danielson, A., Möller, P., and Dulski, P., 1992, The europium anomalies in p. A204.
banded iron formations and the thermal history of the oceanic crust: Chem- Kato, Y., and Kimura, S., 1999, Bulk-rock chemical compositions of the Bool-
ical Geology, v. 97, p. 89–100, doi: 10.1016/0009-2541(92)90137-T. geeda Iron Formation of the Hamersley Group, Western Australia: Its
de Baar, H.J.W., Bacon, M.P., and Brewer, P.G., 1985, Rare earth elements origin and implication from a very low concentration of minor elements:
in the Pacific and Atlantic Oceans: Geochimica et Cosmochimica Acta, Shigen-Chishitsu, v. 49, p. 175–189.
v. 49, p. 1943–1959, doi: 10.1016/0016-7037(85)90089-4. Kato, Y., and Nakamura, K., 2003, Origin and global tectonic significance of
de Baar, H.J.W., German, C.R., Elderfield, H., and van Gaans, P., 1988, Rare Early Archean cherts from the Marble Bar greenstone belt, Pilbara Cra-
earth element distributions in anoxic waters of the Cariaco Trench: Geo- ton, Western Australia: Precambrian Research, v. 125, p. 191–243, doi:
chimica et Cosmochimica Acta, v. 52, p. 1203–1219, doi: 10.1016/0016- 10.1016/S0301-9268(03)00043-3.
7037(88)90275-X. Kato, Y., Kawakami, T., Kano, T., Kunugiza, K., and Swamy, N.S., 1996,
Derry, L.A., and Jacobsen, S.B., 1990, The chemical evolution of Precam- Rare-earth element geochemistry of banded iron formations and
brian seawater: Evidence from REEs in banded iron formations: Geo- associated amphibolite from the Sargur belts, south India: Journal of
chimica et Cosmochimica Acta, v. 54, p. 2965–2977, doi: 10.1016/0016- Southeast Asian Earth Sciences, v. 14, p. 161–164, doi: 10.1016/S0743-
7037(90)90114-Z. 9547(96)00054-2.
de Wit, M.J., Roering, C., Hart, R.J., Armstrong, R.A., de Ronde, C.E.J., Green, Kato, Y., Ohta, I., Tsunematsu, T., Watanabe, Y., Isozaki, Y., Maruyama, S., and
R.W.E., Tredoux, M., Peberdy, E., and Hart, R.A., 1992, Formation of an Imai, N., 1998, Rare earth element variations in mid-Archean banded iron
Archaean continent: Nature, v. 357, p. 553–562, doi: 10.1038/357553a0. formations: Implications for the chemistry of ocean and continent and
Dymek, R.F., and Klein, C., 1988, Chemistry, petrology and origin of banded plate tectonics: Geochimica et Cosmochimica Acta, v. 62, p. 3475–3497,
iron-formation lithologies from the 3800 Ma Isua Supracrustal Belt: West doi: 10.1016/S0016-7037(98)00253-1.
Greenland: Precambrian Research, v. 39, p. 247–302, doi: 10.1016/0301- Kato, Y., Kano, T., and Kunugiza, K., 2002a, Negative Ce anomaly in the Indian
9268(88)90022-8. banded iron formations: Evidence for the emergence of oxygenated deep-
Elderfield, H., 1988, The oceanic chemistry of the rare-earth elements: Royal sea at 2.9–2.7 Ga: Resource Geology, v. 52, p. 101–110.
Society of London Philosophical Transactions, ser. A, v. 325, p. 105–126. Kato, Y., Nakao, K., and Isozaki, Y., 2002b, Geochemistry of Late Permian to
Elderfield, H., and Sholkovitz, E.R., 1987, Rare earth elements in the pore Early Triassic pelagic cherts from southwest Japan: Implications for an
waters of reducing nearshore sediments: Earth and Planetary Science Let- oceanic redox change: Chemical Geology, v. 182, p. 15–34, doi: 10.1016/
ters, v. 82, p. 280–288, doi: 10.1016/0012-821X(87)90202-0. S0009-2541(01)00273-X.
288 Y. Kato et al.
Khan, R.M.K., Das Sharma, S., Patil, D.J., and Naqvi, S.M., 1996, Trace, rare- McNaughton, N.J., Compston, W., and Barley, M.E., 1993, Constraints on
earth element, and oxygen isotopic systematics for the genesis of banded the age of the Warrawoona Group, eastern Pilbara Block, Western
iron-formations: Evidence from Kushtagi schist belt, Archaean Dharwar Australia: Precambrian Research, v. 60, p. 69–98, doi: 10.1016/0301-
Craton, India: Geochimica et Cosmochimica Acta, v. 60, p. 3285–3294, 9268(93)90045-4.
doi: 10.1016/0016-7037(96)00172-X. Michard, A., 1989, Rare earth element systematics in hydrothermal fluids: Geo-
Kiminami, K., Miyashita, S., and Kawabata, K., 1994, Ridge collision and in chimica et Cosmochimica Acta, v. 53, p. 745–750, doi: 10.1016/0016-
situ greenstones in accretionary complexes: An example from the Late 7037(89)90017-3.
Cretaceous Ryukyu Islands and southwest Japan margin: The Island Arc, Mills, R.A., and Elderfield, H., 1995, Rare earth element geochemistry of
v. 3, p. 103–111. hydrothermal deposits from the active TAG Mound, 26°N Mid-Atlan-
Kiyokawa, S., and Taira, A., 1998, The Cleaverville Group in the West Pilbara tic Ridge: Geochimica et Cosmochimica Acta, v. 59, p. 3511–3524, doi:
Coastal Granitoid-Greenstone Terrain of Western Australia: An example 10.1016/0016-7037(95)00224-N.
of a Mid-Archaean immature oceanic island-arc succession: Precambrian Mitra, A., Elderfield, H., and Greaves, M.J., 1994, Rare earth elements in subma-
Research, v. 88, p. 109–142, doi: 10.1016/S0301-9268(97)00066-1. rine hydrothermal fluids and plumes from the Mid-Atlantic Ridge: Marine
Klein, C., and Beukes, N.J., 1993, Sedimentology and geochemistry of the gla- Chemistry, v. 46, p. 217–235, doi: 10.1016/0304-4203(94)90079-5.
ciogenic Late Proterozoic Rapitan iron-formation in Canada: Economic Morinaga, C., 1997, Geochemical study of the oldest sedimentary rocks from
Geology and the Bulletin of the Society of Economic Geologists, v. 88, the Isua supracrustal belt [M.S. thesis]: Yamaguchi, Japan, Yamaguchi
p. 542–565. University, 160 p. (in Japanese with English abstract).
Koeppenkastrop, D., and De Carlo, E.H., 1992, Sorption of rare-earth ele- Murray, R.W., Buchholtz ten Brink, M.R., Gerlach, D.C., Russ, G.P., III, and
ments from seawater onto synthetic mineral particles: An experimental Jones, D.L., 1992a, Interoceanic variation in the rare earth, major, and
approach: Chemical Geology, v. 95, p. 251–263, doi: 10.1016/0009- trace element depositional chemistry of chert: Perspectives gained from
2541(92)90015-W. the DSDP and ODP record: Geochimica et Cosmochimica Acta, v. 56,
Komiya, T., Maruyama, S., Masuda, T., Nohda, S., Hayashi, M., and Oka- p. 1897–1913, doi: 10.1016/0016-7037(92)90319-E.
moto, K., 1999, Plate tectonics at 3.8–3.7 Ga: Field evidence from the Murray, R.W., Buchholtz ten Brink, M.R., Gerlach, D.C., Russ, G.P., III, and
Isua accretionary complex, southern West Greenland: Journal of Geology, Jones, D.L., 1992b, Rare-earth, major, and trace element composition
v. 107, p. 515–554, doi: 10.1086/314371. of Monterey and DSDP chert and associated host sediment: Assess-
Koppi, A.J., Edis, R., Field, D.J., Geering, H.R., Klessa, D.A., and Cockayne, ing the influence of chemical fractionation during diagenesis: Geochi-
D.J.H., 1996, Rare earth element trends and cerium-uranium-manga- mica et Cosmochimica Acta, v. 56, p. 2657–2671, doi: 10.1016/0016-
nese associations in weathered rock from Koongarra, Northern Territory, 7037(92)90351-I.
Australia: Geochimica et Cosmochimica Acta, v. 60, p. 1695–1707, doi: Naqvi, S.M., Sawkar, R.H., Subba Rao, D.V., Govil, P.K., and Gnaneswar
10.1016/0016-7037(96)00047-6. Rao, T., 1988, Geology, geochemistry and tectonic setting of Archaean
Krapez, B., 1993, Sequence stratigraphy of the Archean supracrustal belts of greywackes from Karnataka nucleus, India: Precambrian Research, v. 39,
the Pilbara Block, Western Australia: Precambrian Research, v. 60, p. 1– p. 193–216, doi: 10.1016/0301-9268(88)90042-3.
45, doi: 10.1016/0301-9268(93)90043-2. Nesbitt, H.W., 1979, Mobility and fractionation of rare earth elements dur-
Kuhn, T., Bau, M., Blum, N., and Halbach, P., 1998, Origin of negative Ce ing weathering of a granodiorite: Nature, v. 279, p. 206–210, doi:
anomalies in mixed hydrothermal-hydrogenetic Fe-Mn crusts from the 10.1038/279206a0.
Central Indian Ridge: Earth and Planetary Science Letters, v. 163, p. 207– Nutman, A.P., Chadwick, B., Ramakrishnan, M., and Viswanatha, M.N.,
220, doi: 10.1016/S0012-821X(98)00188-5. 1992, SHRIMP U-Pb ages of detrital zircon in Sargur supracrustal rocks
Kusky, T.M., and Kidd, W.S.F., 1992, Remnants of an Archean oceanic plain western Karnataka, southern India: Journal of Geological Society of
teau, Belingwe greenstone belt, Zimbabwe: Geology, v. 20, p. 43–46, doi: India, v. 39, p. 367–374.
10.1130/0091-7613(1992)020<0043:ROAAOP>2.3.CO;2. Nutman, A.P., Chadwick, B., Krishna Rao, B., and Vasudev, V.N., 1996,
Laajoki, K., 1975, Rare-earth elements in Precambrian iron formations in SHRIMP U/Pb zircon ages of acid volcanic rocks in the Chitradurga and
Väyrylänkylä, South Puolanka area, Finland: Bulletin of Geological Soci- Sandur Groups, and granites adjacent to the Sandur schist belt, Karnataka:
ety of Finland, v. 47, p. 93–107. Journal of Geological Society of India, v. 47, p. 153–164.
Liu, Y.-G., Miah, M.R.U., and Schmitt, R.A., 1988, Cerium: A chemical tracer Nutman, A.P., Bennett, V.C., Friend, C.R.L., and Rosing, M.T., 1997, ~3710
for paleo-oceanic redox conditions: Geochimica et Cosmochimica Acta, and ≥3790 Ma volcanic sequences in the Isua (Greenland) supracrustal
v. 52, p. 1361–1371, doi: 10.1016/0016-7037(88)90207-4. belt; structural and Nd isotope implications: Chemical Geology, v. 141,
Lottermoser, B.G., and Ashley, P.M., 2000, Geochemistry, petrology and ori- p. 271–287, doi: 10.1016/S0009-2541(97)00084-3.
gin of Neoproterozoic ironstones in the eastern part of the Adelaide Geo- Nyström, J.O., 1984, Rare earth element mobility in vesicular lava during low-
syncline, South Australia: Precambrian Research, v. 101, p. 49–67, doi: grade metamorphism: Contributions to Mineralogy and Petrology, v. 88,
10.1016/S0301-9268(99)00098-4. p. 328–331, doi: 10.1007/BF00376757.
MacLeod, K.G., and Irving, A.J., 1996, Correlation of cerium anomalies with Ohmoto, H., 1993, The banded iron formation in the Hamersley Basin, Austra-
indicators of paleoenvironment: Journal of Sedimentary Research, v. 66, lia: Products of the oxygen-rich Archean atmosphere?: Geological Soci-
p. 948–955. ety of America Abstracts with Programs, v. 25, no. 6, p. A-89.
Manikyamba, C., and Naqvi, S.M., 1997, Mineralogy and geochemistry of Ohmoto, H., Watanabe, Y., Yamaguchi, K.E., Naraoka, H., Haruna, M.,
Archean greenstone belt-hosted Mn formations and deposits of the Dhar- Kakegawa, T., Hayashi, K., and Kato, Y., 2006, this volume, Chemi-
war Craton: redox potential of proto-oceans: Geological Society [Lon- cal and biological evolution of early Earth: Constraints from banded
don] Special Publication 119, p. 91–103. iron formations, in Kesler, S.E., and Ohmoto, H., eds., Evolution of
Manikyamba, C., Balaram, V., and Naqvi, S.M., 1993, Geochemical signatures Early Earth’s Atmosphere, Hydrosphere, and Biosphere—Constraints
of polygenetic origin of a banded iron formation (BIF) of the Archaean from Ore Deposits: Geological Society of America Memoir 198, doi:
Sandur greenstone belt (schist belt) Karnataka nucleus, India: Precam- 10.1130/2006.1198(17).
brian Research, v. 61, p. 137–164, doi: 10.1016/0301-9268(93)90061-6. Ohta, H., Maruyama, S., Takahashi, E., Watanabe, Y., and Kato, Y., 1996,
Martin, D.McB., Powell, C.McA., and George, A.D., 2000, Stratigraphic archi- Field occurrence, geochemistry and petrogenesis of the Archean Mid-
tecture and evolution of the early Paleoproterozoic McGrath Trough, Oceanic Ridge Basalts (AMORBs) of the Cleaverville area, Pilbara
Western Australia: Precambrian Research, v. 99, p. 33–64, doi: 10.1016/ Craton, Western Australia: Lithos, v. 37, p. 199–221, doi: 10.1016/0024-
S0301-9268(99)00053-4. 4937(95)00037-2.
Masuda, A., 1975, Abundances of monoisotopic REE, consistent with the Olivarez, A.M., and Owen, R.M., 1991, The europium anomaly of seawa-
Leedey chondrite values: Geochemical Journal, v. 9, p. 183–184. ter: Implications for fluvial versus hydrothermal REE inputs to the
Masuda, A., Nakamura, N., and Tanaka, T., 1973, Fine structures of mutually oceans: Chemical Geology, v. 92, p. 317–328, doi: 10.1016/0009-
normalized rare-earth patterns of chondrites: Geochimica et Cosmochi- 2541(91)90076-4.
mica Acta, v. 37, p. 239–248, doi: 10.1016/0016-7037(73)90131-2. Pan, Y., and Fleet, M.E., 1996, Rare earth element mobility during prograde
McLennan, S.M., and Taylor, S.R., 1991, Sedimentary rocks and crustal evo- granulite facies metamorphism: Significance of fluorine: Contribu-
lution: Tectonic setting and secular trends: Journal of Geology, v. 99, tions to Mineralogy and Petrology, v. 123, p. 251–262, doi: 10.1007/
p. 1–21. s004100050154.
Radhakrishna, B.P., and Naqvi, S.M., 1986, Precambrian continental crust of Trendall, A.F., Compston, W., Williams, I.S., Armstrong, R.A., Arndt, N.T.,
India and its evolution: Journal of Geology, v. 94, p. 145–166. McNaughton, N.J., Nelson, D.R., Barley, M.E., Beukes, N.J., de Laeter,
Ravizza, G., Sherrell, R.M., Field, M.P., and Pickett, E.A., 1999, Geochem- J.R., Retief, E.A., and Thorne, A.M., 1990, Precise zircon U-Pb chrono-
istry of the Margi umbers, Cyprus, and the Os isotope composition of logical comparison of the volcano-sedimentary sequences of the Kaapvaal
Cretaceous seawater: Geology, v. 27, p. 971–974, doi: 10.1130/0091- and Pilbara cratons between about 3.1 and 2.4 Ga, in Glover, J.E., and
7613(1999)027<0971:GOTMUC>2.3.CO;2. Ho, S.E., eds., International Archean Symposium, 3rd, Extended Abstract
Ruhlin, D.E., and Owen, R.M., 1986, The rare earth element geochemistry of Volume: Perth, Department of Geology, University of Western Australia,
hydrothermal sediments from the East Pacific Rise: Examination of a sea- p. 81–83.
water scavenging mechanism: Geochimica et Cosmochimica Acta, v. 50, Trendall, A.F., de Laeter, J.R., Nelson, D.R., and Bhaskar Rao, Y.J., 1997a,
p. 393–400, doi: 10.1016/0016-7037(86)90192-4. Further zircon U-Pb age data for the Daginkatte Formation, Dharwar
Sakko, M., and Laajoki, K., 1975, Whole rock Pb-Pb isochron age for the Supergroup, Karnataka Craton: Journal of Geological Society of India,
Pääkkö iron formation in Väyrylänkylä, South Puolanka Area, Finland: v. 50, p. 25–30.
Bulletin of Geological Society of Finland, v. 47, p. 113–117. Trendall, A.F., de Laeter, J.R., Nelson, D.R., and Mukhopadhway, D., 1997b, A
Sholkovitz, E.R., and Schneider, D.L., 1991, Cerium redox cycles and rare precise zircon U-Pb age for the base of the BIF of the Mulaingiri Forma-
earth elements in the Sargasso Sea: Geochimica et Cosmochimica Acta, tion (Bababudan Group, Dharwar Supergroup) of the Karnataka Craton:
v. 55, p. 2737–2743, doi: 10.1016/0016-7037(91)90440-G. Journal of Geological Society of India, v. 50, p. 161–170.
Sholkovitz, E.R., Piepgras, D.J., and Jacobsen, S.B., 1989, The pore water Trendall, A.F., Nelson, D.R., de Laeter, J.R., and Hassler, S.W., 1998, Precise
chemistry of rare earth elements in Buzzards Bay sediments: Geochi- zircon U-Pb ages from the Marra Mamba Iron Formation and Wittenoom
mica et Cosmochimica Acta, v. 53, p. 2847–2856, doi: 10.1016/0016- Formation, Hamersley Group, Western Australia: Australian Journal of
7037(89)90162-2. Earth Sciences, v. 45, p. 137–142.
Sholkovitz, E.R., Shaw, T.J., and Schneider, D.L., 1992, The geochemistry of Tsikos, H., and Moore, J.M., 1997, Petrography and geochemistry of the Paleo-
rare earth elements in the seasonally anoxic water column and porewaters proterozoic Hotazel iron-formation, Kalahari manganese field, South
of Chesapeake Bay: Geochimica et Cosmochimica Acta, v. 56, p. 3389– Africa: Implications for Precambrian manganese metallogenesis: Eco-
3402, doi: 10.1016/0016-7037(92)90386-W. nomic Geology and the Bulletin of the Society of Economic Geologists,
Smith, R.E., Perdrix, J.L., and Parks, T.C., 1982, Burial metamorphism in v. 92, p. 87–97.
the Hamersley Basin, Western Australia: Journal of Petrology, v. 23, Van Kranendonk, M.J., Hickman, A.H., Williams, I.R., and Nijman, W., 2001,
p. 75–102. Archaean geology of the East Pilbara Granite-Greenstone Terrane, West-
Stille, P., and Clauer, N., 1986, Sm-Nd isochron-age and provenance of the ern Australia—A field guide: Geological Survey of Western Australia,
argillites of the Gunflint Iron Formation in Ontario, Canada: Geochi- Record 2001/9, 134 p.
mica et Cosmochimica Acta, v. 50, p. 1141–1146, doi: 10.1016/0016- Wright, J., Schrader, H., and Holser, W.T., 1987, Paleoredox variations in
7037(86)90395-9. ancient oceans recorded by rare earth elements in fossil apatite: Geo-
Sumner, D.Y., and Bowring, S.A., 1996, U-Pb geochronologic constraints chimica et Cosmochimica Acta, v. 51, p. 631–644, doi: 10.1016/0016-
on deposition of the Campbellrand Subgroup, Transvaal Supergroup, 7037(87)90075-5.
South Africa: Precambrian Research, v. 79, p. 25–35, doi: 10.1016/0301- Zhang, J., and Nozaki, Y., 1996, Rare earth elements and yttrium in seawater:
9268(95)00086-0. ICP-MS determinations in the East Caroline, Coral Sea, and South Fiji
Taylor, S.R., and McLennan, S.M., 1985, The continental crust: Its composition basins of the western South Pacific Ocean: Geochimica et Cosmochimica
and evolution: Oxford, Blackwell, 312 p. Acta, v. 60, p. 4631–4644, doi: 10.1016/S0016-7037(96)00276-1.
Thorpe, R.I., Hickman, A.H., Davis, D.W., Mortensen, J.K., and Trendall, A.F., Zhang, J., Amakawa, H., and Nozaki, Y., 1994, The comparative behaviors of
1992, U-Pb zircon geochronology of Archean felsic units in the Marble yttrium and lanthanides in the seawater of the North Pacific: Geophysical
Bar region, Pilbara Craton, Western Australia: Precambrian Research, Research Letters, v. 21, p. 2677–2680, doi: 10.1029/94GL02404.
v. 56, p. 169–189, doi: 10.1016/0301-9268(92)90100-3.
Trendall, A.F., and Blockley, J.G., 1970, The iron formations of the Precam-
brian Hamersley Group, Western Australia, with special reference to the
associated crocidolite: Geological Survey of Western Australia Bulletin,
v. 199, 366 p. MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA


Memoir 198
2006
Chemical and biological evolution of early Earth:

Constraints from banded iron formations
Hiroshi Ohmoto
Yumiko Watanabe
Pennsylvania State University, University Park, Pennsylvania 16802, USA
Kosei E. Yamaguchi
Yokosuka, Kanagawa 237-0061, Japan
Hiroshi Naraoka
Department of Chemistry, Okayama University, Okayama, Okayama 700-8530, Japan
Makoto Haruna
Uchida Yoko Co., Ltd., Chuo-ku, Tokyo 104-8282, Japan
Takeshi Kakegawa
Department of Earth Science, Tohoku University, Sendai, Miyagi 980-8578, Japan
Ken-ichiro Hayashi
Department of Earth Evolution Sciences, University of Tsukuba, Tsukuba, Ibaraki 305-8572, Japan
Yasuhiro Kato
Department of Geosystem Engineering, University of Tokyo, Bunkyo-ku, Tokyo 113-8656, Japan
ABSTRACT
Geological and geochemical characteristics of banded iron formations (BIFs) sug-

gest that they formed by mixing locally (or regionally) discharged submarine hydro-
thermal fluids with local seawater, rather than by upwelling deep ocean water. Sub-
marine hydrothermal fluids typically evolved from local seawater by acquiring heat,
metals, and sulfur during deep circulation through a variety of rocks (e.g., volcanics,
evaporites) in greenstone terranes that developed under a variety of tectonic settings.
In general, when the fluids were heated above ~350 °C, they may have produced Cu-
and Zn-rich volcanogenic massive sulfide deposits (VMSDs), whereas those heated
less than ~200 °C were generally poor in H2S and heavy metals, except Fe, and may
have subsequently produced BIFs.
Depending on the salinity contrast between discharging hydrothermal fluids
(evolved seawater) and local seawater, hydrothermal fluids may (1) mix rapidly with
local seawater to form smoker-type BIFs or (2) create a metal- and silica-rich brine
Ohmoto, H., Watanabe, Y., Yamaguchi, K.E., Naraoka, H., Haruna, M., Kakegawa, T., Hayashi, K., and Kato, Y., 2006, Chemical and biological evolution of early
Earth: Constraints from banded iron formations, in Kesler, S.E., and Ohmoto, H., eds., Evolution of Early Earth’s Atmosphere, Hydrosphere, and Biosphere—
Constraints from Ore Deposits: Geological Society of America Memoir 198, p. 291–331, doi: 10.1130/2006.1198(17). For permission to copy, contact editing@
geosociety.org. ©2006 Geological Society of America. All rights reserved.
291
292 H. Ohmoto et al.
pool, mix slowly with the overlying water body, and form brine pool-type BIFs. BIFs
associated with VMSDs and volcanic rocks generally belong to smoker-type BIFs;
many formed at seawater depths >2.5 km. Large BIFs, including the 2.6–2.4 Ga BIFs
in the Hamersley Basin, Australia, the 2.5 Ga Kuruman IF in South Africa, and the
1.87 Ga BIFs in the Lake Superior region, United States-Canada, belong to brine
pool-type BIFs. The Hamersley Basin and possibly other large BIF-hosting basins
were probably land-locked seas (like the Black Sea) where river waters diluted the
surface water zone and the underlying water bodies were anoxic.
During the accumulation of a BIF sequence, the dominant Fe mineralogy fre-
quently changed from ferric (hydr)oxides (oxide BIFs) to siderite (carbonate BIFs)
and to pyrite (sulfide BIFs). Such changes were probably caused by changes in the
relative amounts of dissolved O2 (DO), ∑CO32–, and ∑S2– in local seawater. From the
Fe2+-O2 mass balance calculations for the formation of iron oxides in smoker-type
BIFs, and the relationship between the atmospheric pO2 and oceanic O2 depth profile,
we conclude that the atmosphere and oceans have been fully oxygenated since ca. 3.8
Ga, except in local anoxic basins. Thermodynamic analyses of the formational condi-
tions of siderite and analyses of the carbon isotopic composition of siderite associated
with major BIFs suggest that the pre–1.8 Ga atmosphere was CO2-rich (pCO2 >100
PAL) and CH4-poor (pCH4 ≈10 ppm); therefore, CO2, rather than CH4, was the major
greenhouse gas throughout geologic history.
After a decline of hydrothermal fluid flux, BIF-hosting basins generally became
euxinic (H2S-rich) because of the increased activity of sulfate-reducing bacteria (SRB)
and SO42–-rich seawater, and thereby accumulated organic carbon-rich and pyrite-
rich black shales (sulfide-type BIFs). The SO42– contents and SRB activity in the
oceans have been essentially the same since ca. 3.8 Ga. The Archean oceans were most
likely poor in both Fe2+ and silica, much like modern oceans. Our study also suggests
that diverse communities of organisms, including cyanobacteria, SRB, methanogens,
methanotrophs, and eukaryotes, evolved very early in Earth’s history, probably by
the time the oldest BIFs (ca. 3.8 Ga) formed.
BIFs have been found in rocks of all geologic age. Therefore, they cannot be indi-
cators of an anoxic atmosphere and/or anoxic oceans as suggested by many previous
researchers. Instead, BIFs indicate that the atmosphere and ocean chemistry have
been regulated at present compositions (except pCO2 ) through geologic history by
interactions with the biosphere. The general trend of declining size and abundance of
BIFs with geologic time reflects the cooling history of Earth’s interior.
Keywords: banded iron formation, atmospheric evolution, oceanic evolution, bio-

logical evolution, Archean.
INTRODUCTION iron ores are those that developed from BIFs by the leaching of
silica and conversion of primary/diagenetic iron minerals (hema-
“Banded iron formation” (BIF) is a lithological term for a tite, magnetite, and siderite) to “secondary” hematite.
chemical sediment consisting of a thinly layered or laminated Since Cloud published his model in 1968, theories for the
rock composed of alternating layers (bands) of chert and primar- chemical and biological evolutions of both early Earth and BIF
ily iron-rich minerals (hematite, magnetite, siderite, pyrite, and formation have influenced each other and continually evolved
various Fe-rich silicates); its silica content (as SiO2) ranges from as new geologic concepts (e.g., plate tectonic and mantle
40 to 50 wt% whereas its iron content (as Fe) typically ranges dynamics, submarine hydrothermal processes, and the evolu-
from 15 to 35 wt% (James, 1954). Large BIF deposits are most tionary history of microbes) have been developed and new
abundant in sedimentary sequences older than 1.8 Ga and range data have become available. Here, we review how BIF mod-
from a few centimeters to several hundred meters in thickness els have evolved over the past five decades, and present a new
and from <~1 km2 to >100,000 km2 in aerial extent. model (synthesis) to explain the latest (unpublished) observa-
The term “iron ore” is typically reserved for a rock contain- tions as well as those in literature. In the first section (Back-
ing more than ~35 wt% Fe. Economically, the most important ground) of this paper, we review the characteristics of BIFs and
Chemical and biological evolution of early Earth 293
the evolution of BIF models. In the second section, we review (mesobands) to a ~1 m to ~30 m scale (macrobands) (Trendall
general characteristics of submarine hydrothermal systems, and Blockley, 1970).
because they are essential in building a new BIF model. In the Contrary to a popular notion that BIFs are restricted to rocks
third section, we present our interpretations of the geological of Precambrian age and öölitic ironstones to those of Phanerozoic
and hydrological processes for the formation of major BIFs, age, BIFs are also common in Phanerozoic sequences (Tables
including, but not restricted to ca. 2.7 Ga BIFs in the Abitibi 1, 2) and öölitic iron ores are not uncommon in Precambrian
district, Canada; ca. 2.5 Ga BIFs in the Hamersley district, sequences (Kimberley, 1989b). For example, some BIFs, such
Australia, and in the Kuruman district, South Africa; and ca. as the ca. 1.9 Ga Sokoman BIFs in Canada, contain both banded
1.9 Ga BIFs in the Lake Superior district, United States-Can- and öölitic types of iron formations (Klein and Fink, 1976). A
ada. In the fourth section, we present our models for the pre- large öölitic iron formation occurs in the 2.35 Ga Timeball Hill
cipitation mechanisms of iron oxides, siderite, pyrite, and chert Formation in South Africa (Beukes, 1983; Bekker et al., 2004).
in BIFs and for the conversions of hematite to magnetite and In this paper, we will focus on (bio)chemically precipitated iron
magnetite to hematite. In the final section, we relate BIFs to the formations and exclude from Tables 1 and 2 deposits comprised
evolutionary histories of atmospheric chemistry (pO2, pCO2 , solely of öölitic iron ores.
and pCH4), oceanic chemistry (Fe2+, silica, O2, H2S, SO42–, and
pH), constituents of the biosphere, and thermal dynamics of Oxide-, Carbonate-, Sulfide-, and Silicate-type BIFs
the mantle. James (1954) proposed a BIF classification system based
on the type of dominant Fe-rich mineral present: oxide (hema-
BACKGROUND tite and/or magnetite), carbonate (siderite, ankerite, and/or dolo-
mite), sulfide (pyrite ± pyrrhotite), or silicate (greenalite, min-
Over 1300 books, papers, and abstracts have been pub- nesotaite, and/or stilpnomelane). This classification scheme
lished on BIFs. Comprehensive reviews on the geology, min- is useful because it demonstrates that BIFs formed in several
eralogy, geochemistry, and models of depositional mechanisms distinct chemical environments and/or through several distinct
have been assembled in several recent books written or edited (bio)chemical reactions.
by Mel’nik (1982), Trendall and Morris (1983), Appel and Jaspers are low-grade oxide-type BIFs. They are composed
LaBerge (1987), Young and Taylor (1989), Farrell (1990), and of thin bands (typically 20 μm to 1 mm thick) of iron oxides
Eriksson et al. (2004). Other valuable review papers include, (hematite ± magnetite ± siderite ± pyrite) and thicker bands of
but are not restricted to, Dimroth (1976), Cole and Klein chert (~1 mm to ~1 cm); their bulk-rock Fe contents are typi-
(1981), Kimberley (1989a, 1989b), Gross (1991), and Klein cally ~1–5%, but Fe contents of the iron-rich layers frequently
and Beukes (1992). exceed 15%. The iron formations that are closely associated with
volcanogenic massive sulfide deposits (VMSDs) are sometimes
Classifications of Iron Formations called “exhalites” or “ferruginous cherts.” They should also be
included in oxide-type BIFs, because they have essentially the
Banded Iron Formations versus Öölitic Ironstones same mineralogical, chemical, and physical characteristics as
Although opinions vary as to whether BIFs in specific those not associated with VMSDs.
regions accumulated in open seas, semi-closed seas, lagoons, Most researchers have neglected the importance of sulfide-
or lakes (e.g., Eugster and I-Ming Chou, 1973), all researchers type BIFs in constraining the environments and processes for the
agree that BIFs accumulated in large water bodies (i.e., subaque- formation of oxide- and carbonate-type BIFs and also the chemi-
ous). Thus, BIFs contrast with laterites (ferric oxide-rich soils) cal evolution of the atmosphere and oceans. The sulfide-type
that also contain >15 wt% Fe but formed subaerially by in situ BIFs in James’s (1954) classification refer to pyrite- and organic
rock weathering. Öölitic ironstones, in which goethite and/or C-rich black shales with high Fe contents (often >15 wt%) that
hematite primarily occur as 1–5 mm öölites and/or pisolites, frequently occur underneath, interbedded with, and/or overlying
may have formed mostly (but not always) by the accumulation oxide- or carbonate-type BIFs of similar extent. A well-known
and reworking of detrital grains of laterites in shallow water example is the Mount McRae Shale that underlies the ca. 2.5 Ga
(Young, 1989). Brockman IF. Pyrite contents in sulfide-type BIFs are variable
In contrast, most iron minerals in BIFs (except some Fe- and may be exceeded by siderite contents in some formations.
rich silicates) precipitated by (bio)chemical processes utilizing James’s (1954) sulfide-type BIFs are distinguished from
the dissolved Fe (mostly Fe2+) in overlying water bodies. Fe-pre- VMSDs of pyrite ± pyrrhotite ± chalcopyrite ± sphalerite ±
cipitation was episodic and resulted in alternating Fe-rich and galena that underlie some oxide-type BIFs. Massive sulfide ore
silica-rich bands. The thickness of an individual iron-rich layer bodies are defined as rocks with sulfide contents more than
(microband) is typically from ~50 μm to 2 mm, whereas that of 50% in volume. Sulfides in the black shales, which typically
an individual chert layer ranges more widely, from ~50 μm to ~1 range from 1 to 10 wt% in S content, most likely formed by
cm. Because of the variation in thicknesses of Fe-rich and Si-rich utilizing the H2S generated from local seawater SO42– by sul-
microbands, Fe/Si ratios also vary from a ~1 mm to ~10 cm scale fate-reducing bacteria (SRB), whereas the sulfides in massive
TABLE 1. PARTIAL LIST OF VOLCANIC-ASSOCIATED OXIDE/CARBONATE-TYPE IRON FORMATIONS

Associated
massive sulfide Banded iron
Age Footwall type ores formation type References
Shanks and Bischoff (1980);
Atlantis II Deep, Red Sea 28,000 yr basalt Cu, Zn oxide Pottorf and Barnes (1983)
Fukazawa mine, Japan 12 Ma rhyolite Cu, Zn, Pb oxide Kalogeropoulos and Scott (1983)
Jalisco, Mexico Tertiary tuff oxide Zantop (1981)
Iimori, Japan Jurassic basalt Cu oxide Kanehira and Tatsumi (1970)
Okuki, Japan Jurassic basalt Cu oxide Imai (1978)
Windy Craggy, British Columbia Triassic basalt Cu, Zn oxide,carb. Peter and Scott (1998)
Wolverine, Yukon Territory Mississippian rhyolite Zn, Cu carb., oxide Piercey et al. (2001)
Nhuelbuta-Queule Mountain, central
Chile Mississippian basalt Cu oxide Collao et al. (1990)
Avnik, Turkey Ordovician rhyolite oxide Helvaci (1984)
Lahn-Dill, Germany Devonian spilitic tuff oxide,carb. Quade (1976)
Altai, western Siberia and eastern
Kazakhstan Devonian rhyolitic tuff oxide Kalugin (1973)
van Staal and Williams (1984);
Bathurst, New Brunswick Ordovician rhyolite Cu, Zn, Pb carb., oxide Peter and Goodfellow (1996)
Fernández and Moro (1998);
Iberian pyrite belt, Spain/Portugal Ordovician rhyolite Cu, Zn, Pb oxide Soriano and Marti (1999)
Trondheim, Norway Ordovician basalt Cu oxide Grenne and Vokes (1990)
Mount Windsor subprovince, Duhig et al. (1992);
Queensland, Australia Cambro-Ord. rhyolite Cu, Zn, Pb oxide Davidson et al. (2001)
Mahuilque-Relún area, central Chile Paleozoic mica schist Cu oxide Collao et al. (1990)
West Georgia Piedmont, South L. Proterozoic/
Carolina E. Paleozoic rhyolite Cu, Zn oxide Abrams and McConnell (1984)
Pilot Knob, Missouri 1.4–1.5 Ga tuffs oxide Rucker et al. (2003)
Starra, Mount Isa, Australia ca. 1.65 Ga rhyolite oxide Davidson (1992)
Haydon and McConachy (1987);
Broken Hill, Australia ca. 1.65 Ga rhyolite Zn, Pb oxide Page and Laing (1992)
Prescott-Jerome, Arizona ca. 1.8 Ga rhyolite Cu, Zn oxide, carb. DeWitt (1979)
Skellette, Sweden Proterozoic basalt Cu, Zn, Pb oxide Parák (1991)
basalt Morey and Southwick (1995);
Cuyuna North range, Minnesota 1.87 Ga (+rhyolite) oxide, carb. Schneider et al. (2002)
Iron-River Crystal City, Michigan 1.87 Ga (+rhyolite) oxide,carb. Schneider et al. (2002)
Menominee, Michigan 1.87 Ga (+rhyolite) oxide Schneider et al. (2002)
Bergslagen, Sweden 1.9 Ga rhyolite Zn, Pb oxide, carb. Allen et al. (1996)
Boolgeeda IF, Hamersley, Australia 2.42 Ga rhyolite oxide, carb. Trendall and Blockley (1970)
Weeli Wolli IF, Hamersley, Australia 2.45 Ga basalt oxide, carb. Trendall and Blockley (1970)
Olary Block, Australia Proterozoic rhyolite Cu oxide, carb. Lottermoser and Ashley (1996)
Aggeneys, South Africa Proterozoic rhyolite Cu, Zn, Pb carb., oxide Hoffmann (1994)
Otjihase-Matchless Belt, Namibia Proterozoic basalt Cu oxide Adamson and Teichmann (1986)
Murmac Bay, Newfoundland ca. 2.7 Ga basalt/quartzite oxide Hartlaub et al. (2004)
Costa et al. (1983); Liaghat and
Matagami, Quebec ca. 2.7 Ga rhyolite Cu, Zn, Pb oxide, carb. MacLean (1992)
Hemlo, Ontario ca. 2.7 Ga rhyolite barite oxide Lin (2001); Davis and Lin (2003)
Wawa, Ontario ca. 2.7 Ga rhyolite carb. Morton and Nabel (1984)
Sturgeon Lake, Ontario ca. 2.7 Ga rhyolite Cu, Zn, Pb oxide Shegelski (1987); Koopman (1993)
Manitouwage, Ontario ca. 2.7 Ga rhyolite Cu, Zn, Pb oxide Zalenski and Peterson (1995)
Temagami, Ontario ca. 2.7 Ga rhyolite oxide Bowing (1989)
Yellowknife Greenstone Belt ca. 2.9 Ga rhyolite (stockwork) oxide Ootes and Lentz (2002)
Isua, Greenland ca. 3.8 Ga mafic volcanics Cu oxide, carb. Appel (1987)
TABLE 2. PARTIAL LIST OF SEDIMENT-ASSOCIATED OXIDE/CARBONATE-TYPE IRON FORMATIONS

Associated
massive sulfide Banded iron
Deposit/District Age Footwall Type ores formation type References
Afar Rift, Ethiopia 0.2 Ma congl./carb. oxide Bonatti et al. (1972)
Granada, Spain Permian-Triassic carbonate oxide,carb Torres-Ruiz, J. (1983)
Lower Russell (1975); Clifford et al.
Tynah, Ireland Carboniferous limestone Zn, Pb oxide (1986)
Himalaya front, Carboniferous/
Kashmir–Bhutan Permian limestone oxide O'Rourke (1962)
L. Proterozoic/
Urucum, Brazil E. Paleozoic sandstone/congl. oxide Dorr and Van (1973)
Damara, Namibia ca. 0.75 Ga carbonate/tuff oxide Beukes (1973)
Rapitan, NWT, and
Yukon, Canada ca. 0.75 Ga shale/sandstone oxide Klein and Beukes (1993)
Grenville Province, Quebec ca. 1.2 Ga limestone Zn, Pb oxide Gauthier et al. (1987)
Sokoman IF, Labrador Trough 1.85 Ga quartzite/shale oxide, carb Klein and Fink (1976)
Morey and Southwick (1993);
Mesabi IF, Minnesota 1.87 Ga quartzite oxide, carb Schneider et al. (2002)
Morey and Southwick (1993);
Gunflint, Ontario 1.87 Ga quartzite oxide, carb Schneider et al. (2002)
Gamsberg, South Africa Proterozoic clastic sed. Zn, Pb oxide Rozendaal (1986)
Hotazel, Kalahar, South Africa ca. 2.05 Ga carbonate oxide, carb Tsikos et al. (2003)
Beukes (1973); Klein and Beukes
Kuruman, South Africa 2.5 Ga carbonate oxide, carb (1992)
Brockman IF, Hamersley, Trendall and Blockley (1970);
Australia ca. 2.5 Ga shale oxide, carb Ewers and Morris (1981)
Marra Mamba IF, Hamersley,
Australia ca. 2.6 Ga shale oxide, carb Trendall and Blockley (1970)
Boquira, Brazil ca. 2.7 Ga clastic sed. Zn, Pb oxide, carb Beeson (1990)
Gnaneshwar Rao and Naqvi
Chitrandurga schist belt, India >2.6 Ga quartzite/carb. oxide, carb (1995)
sulfide ores, which typically exceed 50 wt% S, most likely uti- than hematite, is stable. Implications of this assumption, although
lized the H2S in locally discharged submarine hydrothermal not explicitly stated, are that the transformation of goethite to
fluids (Ohmoto, 1996). hematite occurred when BIFs were buried more than ~2 km in a
Hydrothermal, diagenetic, and metamorphic processes sediment column where the temperature exceeded the transition
have modified most BIFs. Thus, important questions among temperature (~80 °C) of goethite to hematite (Langmuir, 1997),
researchers concern their primary mineralogy, especially that and that the conversion of hematite to magnetite occurred at even
of hematite (Fe2O3), magnetite (Fe3O4), siderite (FeCO3), and greater depths and higher temperatures. However, this assump-
pyrite (FeS2), because these minerals have been used to con- tion lacks the support of textural evidence (e.g., hematite pseudo-
strain the chemistry of the atmosphere and oceans. Most morphs after goethite). Some hematite crystals may have directly
researchers agree that the majority of hematite in BIFs formed crystallized from hydrothermal solutions at T ∼> ~80 °C. Most sid-
prior to magnetite when they occur in the same microbands. erite and pyrite crystals in BIFs formed within the water column
This is because the hematite in BIFs that were subjected only of anoxic basins (i.e., syngenetic minerals), but some also formed
to zeolite-facies metamorphism is typically much finer grained during the early diagenesis of sediments (i.e., diagenetic miner-
(often <1 μm) than the magnetite (typically >10 μm) (Ahn and als) (Ohmoto et al., 2004).
Buseck, 1990), and textures showing the replacement of hema- Silicate-type BIFs probably developed from tuff-rich sedi-
tite crystals by magnetite crystals are common (e.g., Trendall ments that incorporated abundant Fe-rich carbonates and silica
and Blockley, 1970; LaBerge et al., 1987). during sediment accumulation. Subsequent diagenetic/metamor-
Most researchers have concluded that the primary fer- phic reactions between the Fe-rich carbonate, silica, and silicates,
ric minerals in BIFs were amorphous and/or crystalline ferric such as
hydroxides, such as goethite (FeOOH), which were transformed
to hematite during the early diagenesis of BIFs. This conclusion 3FeCO3 (siderite) + 2SiO2 + 2H2O ⇒ Fe3Si2O5(OH)4 (greenalite) + 3CO2
was based on an assumption that the nucleation of primary min-
erals in BIFs occurred at a temperature where goethite, rather were probably responsible for the formation of iron-rich silicates
(e.g., French, 1973; Klein, 1983). For this reason, most research- James’s Model (1954)
ers have not used Fe-silicates to constrain the depositional chem- By the time James (1954) proposed his famous BIF model
ical environments of BIFs. (Fig. 1A), most geologists had agreed that the dominant form
of Fe in BIF-forming waters was ferrous iron (Fe2+), rather than
Algoma- versus Lake Superior–type BIFs ferric iron (Fe3+), because the concentration of Fe3+ in water that
Gross (1965) has recognized that most Archean (>2.5 Ga) is in equilibrium with hematite (Fe2O3), or its possible precursor
BIFs, such as those in the Abitibi Greenstone Belt of Canada, are goethite (FeOOH), is much too low (<<10−6 m or <<60 ppb at
closely associated with submarine volcanic rocks and are typi- pH >~3) to form vast Fe-rich sediments. It was also known that
cally small (<100 m thick and <100 km2 in area); Gross (1965) to transport a sufficient amount of Fe2+ at pH >3, the water must
termed these BIFs Algoma-type. In contrast, some (not all) BIFs be free of molecular O2, and some oxidation mechanisms must
from ca. 2.5 to ca. 1.8 Ga in age, such as those in the Lake Supe- exist in depositional environments to precipitate Fe2+ as ferric
rior district of North America and the Hamersley Basin in Austra- (hydr)oxides.
lia, appear to have formed in large sedimentary basins with minor James (1954) proposed that the Fe2+ was supplied by con-
volcanism, and are often more than 100 m × 1000 km2 in size; tinental rock weathering under an anoxic atmosphere and trans-
Gross (1965) termed this BIF group as Lake Superior–type. ported by rivers to chemically stratified basins. As an Fe2+-rich
water mass sank, the Fe2+ was successively sequestered as (1)
Previous Models for the Origin of BIFs ferric (hydr)oxides by reactions with free O2 molecules that were
generated by cyanobacteria in the surface layer, (2) Fe-rich car-
Numerous researchers have presented many different mod- bonates by reactions with CO2 (and HCO3–) in intermediate water
els for the origin of BIFs. Here we review only models that have depths, and (3) Fe-sulfides by reactions with the H2S (and HS-)
significantly influenced geologists’ perception of the connections generated by SRB in the bottom water (Fig. 1A).
between BIFs and the evolution of the atmosphere, hydrosphere, An implication of James’s (1954) model is that cyanobacte-
and biosphere. ria and SRB evolved by ca. 3.8 Ga (Isua BIFs). His model also
A sea level
Oxide I.F.
Fe(OH)3 completely or semi-closed
deep basin
Fe2O3 Carbonate I.F.
Fe3O4 circulation
FeCO3 barrier
Sulfide I.F.
FeS
FeS2
B
decreasing organic productivity maximum organic productivity
Figure 1. (A) James’s (1954) model
evaporation for the depositional environment of
BIFs. (B) Drever’s (1974) model for the
sea level formation of BIFs.
surface water
transport of O2 fO2 > ~10-60, Fe2+ = ~0 ppm, SO42- = ~ 15 ppm
thermocline
primary hematite zone
upwelling deep water
primary fO2 = ~10-70
(Hematite may be replaced diagenetically
by magnetite, derrous silicates or siderite.)
siderite Fe2+ = ~10 ppm
zone SO42- = ~ 0 ppm
pyrite HS- = ~ 0 ppm
zone SiO2 = ~ 120 ppm
increasing carbon and/or reduced (locally

iron minerals in sediment developed only)
implies that a BIF bed may change its dominant Fe mineralogy ing ocean water. However, the Drever-Holland model proposed
from ferric oxide to siderite to pyrite along a strike, although no a different mechanism to transport Fe from land to the oceans.
documented example of this exists. Typically, changes from one Whereas Cloud suggested fluvial transportation of Fe2+, Drever
BIF type to another occur in the stratigraphic sequence of a sedi- and Holland suggested that ferric (hydr)oxide minerals were
mentary basin, indicating that the environments for BIF forma- transported to deep oceans as detrital minerals, which were sub-
tion changed (e.g., from oxic to euxinic) over time, rather than sequently dissolved in pore fluids as Fe2+ in organic-rich sedi-
spatially, during the history of a sedimentary basin. ments to generate Fe2+-rich global oceans.
James (1954, 1983) did not explicitly relate BIFs to atmo- Holland (1973) and Drever (1974) noted that hematite (or
spheric and biological evolution. In fact, James (1983) lists goethite) is stable at pO2 >~10−60 atm and concluded that the gen-
ca. 375 Ma BIFs (medium size) in the Altai region of Siberia- eration of a large amount of ferric (hydr)oxide minerals during
Kazakhstan (Table 1) as examples that BIFs are not restricted to soil formation could take place even under an anoxic atmosphere.
Precambrian time. However, such an argument will negate red beds, where hematite
crystals are characteristic minerals, as important evidence for an
Cloud’s Model (1968) oxygenated atmosphere.
Cloud (1968) suggested that the Fe2+ brought in by rivers A more serious problem in relating a very low pO2 value to a
was not totally removed as Fe-rich minerals near the coasts, but real world scenario is that a free-O2 molecule is not present, even
that instead, most of the Fe2+ was carried to the open oceans in a volume as large as the entire atmosphere (V = 4 × 1024 cc),
to create Fe2+-rich and anoxic oceans. According to his model, when the pO2 is below ~10−43 atm because 1 mol of O2 ( = 6 ×1023
Fe2+ in the deep ocean beneath the photic zone was periodically molecules; Avogadro’s number) occupies 2.24 × 104 cc volume
moved into the photic zone by overturning (instigated by cli- at pO2 = 1 atm and T = 25 °C. For a 1L volume of water, a free
matic change) to form oxide-type BIFs. He suggested that BIF O2 molecule disappears at pO2 <~10−23 atm when equilibrium is
size was dictated by O2 availability and that large Superior-type established between the water and air (Ohmoto et al., 2004). At
BIFs were evidence for increased O2 production between ca. 2.5 pO2 <10−43 (or <10−23) atm, pO2 is merely a “virtual value” that is
and 1.8 Ga. By ca. 1.8 Ga, almost all of the Fe2+ in the oceans was calculated from a real pH2 value (i.e., pH2 > 10−23 atm) by assum-
removed by O2, causing the ocean waters to become Fe-poor, ing equilibrium in the reaction: H2 + 1/2O2 = H2O (Ohmoto et al.,
BIFs to disappear, the atmosphere to become oxic, and red beds 2004). In the absence of real O2 molecules, ferric (hydr)oxides
to appear. Subsequently, Cloud (1973, 1978) acknowledged the cannot form unless other mechanisms (e.g., photochemical reac-
existence of BIFs much younger than ca. 1.8 Ga, including those tions) exist in the formational environments.
of Phanerozoic age, but suggested that these formed by different,
unspecified mechanisms. Jacobsen-Holland-Klein-Beukes model (ca. 1990)
In the 1970s and 1980s, researchers began to recognize
Trendall and Blockley’s Model (1970) the importance of seawater-rock interactions in the formation
The most detailed and extensive BIF investigation was of submarine ore deposits (e.g., VMSDs) and in the global
made by Trendall and Blockley (1970) on ca. 2.5 Ga BIFs in geochemical cycles of Na, Mg, Sr, S, and many additional
the Hamersley Basin, Australia. Their recognition of a basin- elements (e.g., Ohmoto et al., 1970; Sleep, 1978; Ohmoto and
wide correlation of individual micro-, meso-, and macrobands, Skinner, 1983; Holland, 1984). Researchers (e.g., Jacobsen and
and their suggestion that a pair of Fe-rich and silica-rich micro- Pimentel-Klose, 1988; Derry and Jacobsen, 1990; Alibert and
bands represents an annual precipitation (Fe precipitation during McCulloch, 1993) also identified the presence of positive Eu
summers and silica precipitation during winters), led subsequent anomalies in many BIFs, which are important characteristics of
researchers to suggest that seasonal upwellings of Fe2+-rich water submarine hydrothermal fluids (Kato et al., this volume). These
from deep oceans were important in BIF formation. However, findings have prompted the idea that most Fe2+ in the Archean
their recognition of close temporal associations between vol- oceans was supplied by submarine hydrothermal fluids that
canic activity and BIFs, and their suggestion of a genetic link acquired Fe2+ from basalts on mid-ocean ridges (MORs), rather
between the source of Fe in BIFs and local volcanism, has not than from the weathering of continental rocks (e.g., Jacobsen
been taken seriously by supporters of Cloud’s (1968) model. This and Pimentel-Klose, 1988; Holland and Petersen, 1995; Klein
was because Trendall and Blockley (1970) inferred that the Fe2+ and Beukes, 1992). However, these researchers have retained
was derived by magmatic fluids, but the total mass of Fe in the other important elements of Cloud’s (1968) model, especially
Hamersley Basin BIF is much greater than that expected from that the global oceans were anoxic and Fe2+-rich (~10 ppm),
reasonable-sized magmas (e.g., Holland, 1984). except possibly in the photic zone, and that periodic upwell-
ing of Fe2+-rich ocean water to nearshore areas produced BIFs
Drever-Holland Model (ca. 1974) under an anoxic atmosphere (Fig. 2A).
Mass balance calculations of Fe supply in the Hamersley Over the years, the proposed date at which the O2 con-
Basin led Holland (1973, 1984) and Drever (1974) to reiterate tent of the atmosphere began to rise has moved from ca. 1.8
Cloud’s (1968) suggestion that the Fe in BIFs came from upwell- Ga (Cloud, 1968) to ca. 2.0 Ga (Cloud, 1973) to ca. 2.35 Ga
A
Atmosphere
pO2 = <0.01 PAL
pCO2 = ~10 PAL
Oxygenated Surface Water

+ Mixing zone Fe2+ ⇒ Fe3+
+ +
+
+ + Lake Superior
upwelling upwelling
AlgomaType BIF Type BIF
Anoxic Deep Ocean
Continental
Island Arc Fe2+ = 3 ppm
Shelf
Crust
Oceanic Crust +
+ +
+
+ +
Figure 2. (A) Jacobson-Holland-Klein-

> 1000s km Beukes model for the formation of BIFs
(Holland and Petersen, 1995). (B) Ohmoto’s
(1997) model for the formation of BIFs.
B
Atmosphere
pO2 = ~1 PAL
pCO2 = >100 PAL
Brine Pool-type BIF
Local Anoxic Basin

Continental
Crust Smoker-type BIF Smoker-type BIF
. .. . ..Plume Oxygenated Deep Ocean > 1000 m
Seawater . . ... .. . ...... (Fe poor) Plume
. .
........... ..
..
.
..... ..
...... .........
....
..... ...
+ + + + Hydrothermal BIF .
.......
.......
.......
++ Circulation
v
Continental Rift v
v v
v v
Crust ++ v v v v
++
++ Oceanic Crust ++
+ ++
Mantle +
Mantle
(e.g., Kasting, 2001; Holland, 2002; Kasting and Seifert, 2002; of iron was oxidized and precipitated close to the vents, as they
Bekker et al., 2004). This has presented a serious contradic- are today.” In contrast, Canfield and his associates (Canfield
tion to the Cloud-Walker-Holland-Kasting’s atmospheric et al., 2000; Habicht et al., 2002) have argued that the rise of
evolution model, which has linked BIFs to an anoxic atmo- atmospheric pO2 around 2.2 Ga changed the ocean from sul-
sphere. Because many BIFs, including those in the Lake Supe- fate-poor (SO42– content <1/100 of the present level) to slightly
rior region, clearly formed at least 500 m.y. after the 2.35 Ga sulfate-rich (SO42– content ~1/10 of the present level); they also
date for the rise of O2, BIFs no longer provide evidence for suggested that the ocean became rich in biogenic H2S, which
an anoxic atmosphere. Realizing this contradiction, Holland scavenged Fe2+ in the oceans as pyrite to cease the formation
(1999, p. 20) stated, “The formation of BIFs tells us more about of BIFs ca. 1.8 Ga.
the oxidation of the deeper parts of the oceans than about the
atmosphere. The cessation of BIF deposition ca. 1.8 Ga may be Dimroth’s Model (1974)
a signal that the deep ocean basins became oxygenated at that Although most geologists have linked BIFs to an anoxic
time, and that during the following 1 Ga, the hydrothermal flux atmosphere, some geologists (e.g., Dimroth and Kimberley,
1976; Clemmey and Badham, 1982; Ohmoto, 1996; Peter, Temperature of Hydrothermal Fluids versus Metal Types
2001; Phillips et al., 2001) have argued that BIFs formed of Ore Deposits
throughout geologic history (including the Phanerozoic) under
an oxic atmosphere, because the mineralogy and geochemistry The temperature of fluid-rock interaction is an important
of Archean and Proterozoic BIFs are essentially the same as parameter that determines the metal, H2S, and silica contents of
those for Phanerozoic BIFs. fluids (Ohmoto, 1996; Ohmoto and Goldhaber, 1997) (Fig. 3).
In general, fluids with salinity levels like normal seawater may
Ohmoto’s Model (1997) become rich (>1 mM) in H2S, as well as in metals (Zn, Pb, and
Ohmoto (1993) has recognized strong similarities in the Fe), at T >~250 °C; temperatures >~350 °C are required to gener-
geologic settings, mineralogy, and geochemistry of BIFs in ate Cu-rich fluids. Fluids rich in both H2S and metals may pre-
many districts with chert-hematite beds that are associated cipitate sulfides (ZnS, PbS, FeS2, and CuFeS2) by cooling, which
with VMSDs. Subsequently, Ohmoto (1997, 2004) suggested results when seawater mixes at discharge sites. If hydrothermal
that all BIFs formed by locally discharged submarine hydro- fluids are heated only to <~200 °C, they typically remain poor
thermal fluids under a fully oxygenated atmosphere and oceans (<1 mM) in H2S, Zn, Pb, and Cu, but they can be rich (>1 mM)
(except in local basins) since ca. 3.8 Ga (Fig. 2B). Most of this in ∑Fe2+ ( = FeCl20 + FeCl- + Fe2+). (These temperature values
paper presents supportive evidence for this BIF model. become lower with increasing salinity of fluid). Therefore, the
temperature of VMSD-forming hydrothermal fluids is generally
GENERAL CHARACTERISTICS OF SUBMARINE higher than ~250 °C, whereas BIF-forming fluids are generally
HYDROTHERMAL SYSTEMS lower than ~200 °C.
An important constraint on the formation of Cu-rich VMSDs
Origins of Submarine Hydrothermal Fluids is seawater depth. If the depth is <~2.5 km (p <~250 atm), hydro-
thermal fluids of T >~350 °C will separate into liquid and vapor
Many geologists have termed BIFs and VMSDs associ- phases (i.e., boiling) before they reach the seafloor, which will
ated with submarine volcanic rocks “exhalites” (e.g., Goodwin, cause sulfide minerals to precipitate and form vein- and dissemi-
1973; Spry et al., 2000), implying that locally discharged sub- nated-type mineralization, but not massive sulfide ores (Ohmoto,
marine hydrothermal fluids formed them. Recent geochemi-
cal investigations (especially H, O, S, and Sr isotopes and rare
earth elements [REEs]) of modern submarine hydrothermal
systems (such as black smokers and chimneys on MORs [Bar- 0
rett and Jambor, 1988] and metalliferous sediments in the Red
Sea [Shanks and Bischoff, 1980; Pottorf and Barnes, 1983]) and
VMSDs (such as the Japanese Kuroko deposits [e.g., Ohmoto -1
and Skinner, 1983]) have revealed many important character-
istics of submarine hydrothermal processes (see reviews by
Ohmoto, 1996; Peter and Goodfellow, 1996; Scott, 1997; Spry -2
Zn+Fe Cu+Zn+Fe
et al., 2000).
The most important characteristic is that submarine
log mi
hydrothermal fluids are essentially the product of seawater-

py
-3
o+
rock interactions in regions of submarine igneous activity (i.e., Fe

+p
greenstone belts) in a range of tectonic settings (Fig. 2B). Except

for some gaseous components (e.g., CO2, rare gases, and N2), -4
mt
magmatic contributions to submarine hydrothermal fluids are

generally negligible. Seawater that percolated through frac- H 2S
-5
tures in the underlying rocks is heated by a local intrusive rock
and/or magma and leaches metals (e.g., Fe, Cu, Zn, Pb, Ba),
silica, sulfide-S, and sulfate-S from the country rocks, whereas
-6
some (or all) of the seawater SO42– and country-rock SO42– are 0 100 200 300 400
converted to H2S. The heated submarine hydrothermal fluids T (°C)
become buoyant and discharge onto the seafloor. Whether the
discharged hydrothermal fluids immediately mix with cool Figure 3. Typical concentration ranges of metals and H2S in submarine
local seawater to precipitate minerals, or form brine pools and hydrothermal fluids based on thermodynamic calculations (Ohmoto et
al., 1983; Ohmoto and Goldhaber, 1997) and analyses of submarine
slowly mix with the overlying seawater to nucleate minerals, hydrothermal fluids (Scott, 1997). The lower H2S line represents the
depends on the temperature-composition-density relationships stability condition for the assemblage of pyrite+pyrrhotite+magnetite,
between the hydrothermal fluids and local seawater. a common assemblage in basalts.
1996). Therefore, when BIFs are closely associated with Cu-rich latitude, distance from the continent) (Ohmoto, 1996). Three rep-
massive sulfide deposits, we may conclude that both VMSDs and resentative cases for the hydrology of submarine hydrothermal
BIFs formed in a deep (>2.5 km) water body. systems are described below (Fig. 4A–C).
The first scenario (Fig. 4A) applies to the hydrothermal sys-
Hydrological Characteristics of Submarine Hydrothermal tem of an average VMSD. It is generated by a relatively small
Systems (~10 km3) intrusion (stock, dike) at a shallow depth (~2 km from
the seafloor) and has a fluid-circulation of ~5 km in radius (hori-
Circulating seawater acquires heat from rocks. The normal zontal and vertical dimensions). The fluid temperature at a dis-
geothermal gradient and heat from an igneous source control charge site above the intrusion may attain a maximum of ~400
the temperature of rocks in the plumbing system; the circulat- °C after ~4000 yr following the intrusion, and gradually decline
ing seawater gradually cools the intrusive source. Therefore, at to ~100 °C after ~10,000 yr (Cathles, 1983).
a discharge site, the temperature of the discharging fluids typi- The second scenario (Fig. 4B) applies to a hydrothermal
cally increases with time from near ambient to a maximum tem- system associated with a large (>100 km3), deep-seated (>3
perature, and then decreases back to ambient seawater. Although km from the seafloor) basaltic sill (and/or magma), which is
most sulfide mineralization occurs during the waxing stage, BIF periodically replenished by new magmas under an extensional
mineralization may occur during both the waxing and waning tectonic regime (e.g., MORs, continental rifts, and backarc
stages (Cathles, 1983; Ohmoto, 1996). The maximum tempera- rifts). Many hydrothermal circulation cells, whose geometries
ture of the discharging fluids generally decreases with increasing are largely controlled by the major fracture patterns, may be
distance from the heat source. established above the sill. Because new reactive rock surfaces
The thermal history (i.e., temperature versus time) and are created by episodic (rather than continuous) tectonic activ-
geometry (i.e., sizes and shapes of the total and subsidiary cir- ity and magma injections, the fluid discharge onto the seafloor
culation systems, including spacing between subsidiary systems) also becomes episodic; but the overall duration of hydrothermal
of a hydrothermal system are largely controlled by the following activity in such an environment may become several orders of
parameters: (1) the characteristics of the heat source (e.g., depth, magnitude longer (i.e., ~1–100 m.y.) than that associated with a
size, geometry, and temperature); (2) the geometry of fracture solitary small intrusion.
systems; and (3) submarine topography. These three param- The third scenario (Fig. 4C) is similar to the second one
eters are controlled by (a) tectonic setting (e.g., extension ver- in tectonic and igneous environments, but here the submarine
sus compression stress regime; ocean ridges versus subduction hydrothermal fluid flow is largely dictated by the large-scale sub-
zone versus backarc rifts versus mantle plumes; and thickness marine topography and lithology of major aquifers: fluid flow is
of continental crust under BIF-hosting basins), (b) magma type largely confined to permeable beds (e.g., sandstones, conglomer-
(e.g., basalt versus granite), and (c) paleogeography (e.g., paleo- ates, and tuffs) that gently dip toward a large heat source. Such
A B C
sea sea sea
Figure 4. Schematic illustrations of rep-

resentative sub-seafloor flow patterns of
hydrothermal fluids. (A) Around a small,
shallow pluton. (B) Over a large, deep sill
in a spreading center. (C) Topography-
+ +
+ ++ + + + + + + +
and lithology-controlled.
+ ++ + + + + + + + + + + + + + + + + ++ +
+ +
+ ++ + + +
+ +
older igneous older igneous
rocks rocks
Fluid circulation around a shallow new magma new magma
felsic pluton
(1 - 10 km) Fluid circulation over a deep, large, Large-scale topography driven fluid
continuously supplied basaltic magma circulation over a deep, large,
(10 - 100 km) continuously supplied basaltic magma
(> 100 km)
hydrothermal systems may develop in continental rift systems “white smokers”; these terms are mineralogy dependent) in a
(e.g., the Red Sea), rather than in MOR systems. hydrothermal plume until the plume is thermally and chemically
homogenized by the cold seawater; the mineral particles continue
Geometry of Submarine Hydrothermal Deposits to settle around the discharge sites as the plume rises though the
seawater.
VMSDs are frequently overlain by shales, whose aerial From observations of modern submarine hydrothermal
extents are limited essentially to the VMSDs (Ohmoto, 1996). plumes, Peter and Goodfellow (1996) estimated that hydrother-
Such a spatial association indicates that submarine depressions mal minerals may be carried as high as ~500 m above the sea-
(e.g., calderas, troughs, valleys, and basins) are favorable dis- floor and may settle as far as 5–10 km from the fluid discharge
charging sites for hydrothermal fluids; compared to the zone site. Therefore, a BIF that formed from such a discharge point
underneath submarine hills, the head pressure is less (due to a might extend up to ~10 km away, but more typically between
shorter distance to the seafloor) and the permeability is higher 1 and 5 km (Fig. 5A). If fluids were discharged from multiple
(due to a higher fracture density), thus focusing hydrothermal sites along a lineament (e.g., trough), the deposits could be much
fluid flow to depressions (Ohmoto et al., 1983). Because sub- larger (Fig. 5B). Here we term VMSDs and BIFs that formed
marine depressions also provide physical protection (i.e., traps) by mixing low-density, high-temperature fluid with cold sea-
for the deposited minerals from being dispersed by the bottom water near the vents as “smoker-type” deposits; we suggest that
currents, they become favorable sites for hydrothermal mineral most small and linear BIFs (<~100 km in longitudinal direction)
accumulation. Thus, the shape and size of a hydrothermal deposit belong in this category.
may be constrained by the depression in which they form. When a hydrothermal fluid acquires high salinity, either by a
More importantly, the geometry of a submarine hydrother- high degree of evaporation before it circulates through hot rocks
mal deposit (VMSD or BIF), whether the precipitated minerals or by passage through an evaporite bed, the fluid density may
are concentrated around discharging sites as mounds or spread become >1.0 g/cm3 even at elevated temperatures. For example, if
over a large area as a continuous bed, largely depends on how a fluid contains 25 wt% salts, its density at P = 200 bars becomes
hydrothermal fluids mix with the overlying local water, which 1.15 g/cm3 at 100 °C, 1.08 g/cm3 at 200 °C, and 0.99 g/cm3 at
is ultimately determined by the density contrast between these 300 °C (Potter and Brown, 1977). Such a fluid will not readily
fluids. When normal seawater (∑salts = 3.5 wt%) is heated dur- rise through the overlying seawater column and cool quickly, but
ing water-rock interaction and the amounts of dissolved salts may instead create a high-temperature brine pool in a depression,
remain basically the same as in normal seawater, the density of such as the hot brine pools in the Red Sea (Pottorf and Barnes,
the hydrothermal fluids decreases with increasing temperature: 1983). The nucleation of minerals primarily occurs gently and
e.g., from 1.01 at 25 °C to 0.98 at 100 °C, and 0.75 g/cm3 at 300 slowly at and/or near the interface of the brine pool and transi-
°C at p = 200 bars (Potter and Brown, 1977). When such a hydro- tion zone, which exhibits thermal and chemical characteristics
thermal fluid is discharged onto the seafloor, it almost instanta- that are intermediate to the brine and surface water (Fig. 5C).
neously mixes with local cold seawater at the vents to cause the The submarine hydrothermal deposits formed by such processes,
nucleation of fine-grained minerals (i.e., amorphous and crystal- termed here “brine pool-type” deposits, may be continuous over
line). These minerals are carried upward as “black smokers” (or a very large area (>100 km × 100 km).
atmosphere
ocean surface zone
transition zone
mineralization
mineralization Figure 5. Schematic illustrations of

mineralization high temp.
(1 - 5km) smoker- and brine pool-type subma-
(10 - 100 km) brine pool
rine mineralization.
(100 -1000 km)
metalliferous
sediment
Smoker type Brine pool type

A submarine hydrothermal
mineralization
B submarine hydrothermal C
mineralization
GEOLOGIC ENVIRONMENTS FOR BIFS history. The tables also reveal an interesting correlation between
footwall rock type and metal type of associated sulfide ore bod-
Classification of Oxide/Carbonate BIFs Based on Footwall ies: sulfide ore bodies associated with BIFs in volcanic sequence
Rock Types are characteristically Cu-rich, whereas those associated with BIFs
in sedimentary sequences are Zn- and Pb-rich. This relationship
The lithology, thickness, and aerial extent of BIF footwall coincides with the comparison between the metal characteristics
and hanging wall rocks provide the most useful information on of VMSDs versus shale/carbonate-hosted massive sulfide depos-
the geologic, tectonic, and chemical environments of BIF depo- its (Ohmoto et al., 1990), suggesting that BIFs in volcanic asso-
sition. Here we define footwalls as rocks that deposited shortly ciation were generally formed by higher temperature fluids than
before the BIF, excluding much older rocks that are bounded by those in sedimentary sequences.
a fault or unconformity. The thickness of a footwall rock may
range from ~1 m to >1000 m. On an outcrop scale (~1 km), the VMSD-Footwall Type Oxide/Carbonate BIFs
dominant footwall rock type for oxide- and siderite-type BIFs is Although VMSD-footwall type oxide BIFs are generally
either (1) VMSD, (2) submarine volcanic rock (tuffs and lavas), not familiar to researchers of sediment-hosted BIFs, they are
(3) shale (often tuffaceous), (4) dolomite and/or limestone, or (5) well known among VMSD-exploration geologists. Important
conglomerate and/or arkosic sandstone (Fig. 6). Each of the five examples are (1) the oldest BIFs (ca. 3.8 Ga) in Isua, Green-
footwall rocks indicates a specific geologic environment: a deep land, (2) the 2.7 Ga Geco deposits in the Manitowage district,
(>2.5 km) sea for Cu-rich VMSDs, a shallow evaporating water Abitibi, Canada, (3) the 1.8 Ga Jerome deposits, Arizona, (4)
body for dolomite/limestones, and nearshore and shallow water the Ordovician Bathurst deposits, New Brunswick, Canada,
for sandstone/conglomerates. Volcanic- and shale-footwall type (5) the Miocene Kuroko deposits, Japan, and (6) the youngest
BIFs, however, have formed over a wide range of water depths BIFs (25,000 yr old) in the Red Sea. In some cases, such as
(from the photic zone to >2 km basins). the Lyon Lake District of the Abitibi district, oxide BIFs occur
The type of BIF footwall rock may change over a regional beneath massive sulfide deposits (Koopman, 1993). The ca. 2.7
scale. For example, in the Algoma district, Ontario, Canada Ga oxide BIF in the Hemlo gold deposit in Ontario, Canada,
(Fig. 7), footwall rocks change from felsic volcanics to shales to is not associated with massive sulfide ores, but a thick bar-
sandstone-conglomerates over a distance of ~70 km (Goodwin, ite (BaSO4) bed occurs in a lower stratigraphic horizon (Lin,
1973). The ca. 1.9 Ga BIFs in the Lake Superior region are long 2001; Davis and Lin, 2003), suggesting the 2.7 Ga oceans were
thought to have formed on a tectonically stable, igneous-free, not poor in sulfate as previously suggested (e.g., Canfield et
continental slope. Although the dominant footwall rock types al., 2000).
in the Gunflint-Mesabi districts of the Lake Superior region are Most VMSDs, such as the ca. 2.7 Ga Kidd Creek deposit
quartz-arenites, basalts with minor felsic volcanics (bimodal in the Abitibi district, Canada (Ohmoto, 1996), are overlain by
volcanics) dominate the Menominee-Iron River-Crystal Fall dis- pyrite- and organic C-rich black shales (e.g., the “sulfide BIFs”
tricts (Morey and Southwick, 1995). For these reasons, we avoid of James’s [1954] classification) rather than by oxide BIFs. This
the terms Algoma- and Lake Superior-type BIFs and instead use is possibly because oxide BIF accumulation was interrupted by
terms such as “VMSD-footwall type BIFs” and “shale-footwall tuff accumulation and/or seawater conditions in the local sub-
type BIFs.” marine depressions became favorable for SRB after the cessa-
Examples of each of the BIF footwall types are presented tion of hydrothermal activity. VMSD-footwall type oxide BIFs
in Tables 1 and 2; contrary to the popular view that all BIFs are undoubtedly formed during the waning stage of local submarine
older than ca. 1.8 Ga, every BIF type formed throughout geologic hydrothermal activity because they are associated with well-
deep sea
A B C D E
sandstone/ shale shale shale volcanics volcanics
conglomerate carbonate shale shale
BIF BIF BIF BIF BIF
VMS
sandstone/ carbonate tuffaceous shale
conglomerate volcanics volcanics
examples
Gunflint IF Kuruman IF Brockman IF Abitibi district, Ontario
Figure 6. A proposed classification scheme of oxide/carbonate BIFs based on representative footwall rock types.
James Bay
: Oxide BIF
: Carbonate BIF
: Sulfide BIF (Pyritic shale)
: VMS
Chibougamau
Matagami
Manitouwadge
Superior
Timmins Grenville
Kidd Creek
Noranda
Wawa Karkland Lake 48°
Wawa Val d'Or
Michipicoten greenstonebelt
Gogama
Abitibi
Lake Superior greenstonebelt
Temagami
Goulais
Sudbury
84° 80° 100 km 76°
Figure 7. Occurrences of BIFs and major VMSDs in the 2.7 Ga Abitibi and Wawa Greenstone Belts of Ontario, Canada. Compiled from Goodwin
(1973) and Fyon et al. (1992).
developed footwall alteration zones (see references in Table 1). 1). Felsic volcanic rocks may be minor components of bimodal
Because the associated VMSDs are mostly Cu-rich, these BIFs volcanism in some areas (e.g., the ca. 2.3 Ga Menominee, Iron
must have formed in deep (>2.5 km) oceans. River–Crystal Falls, and Cuyuna North Range of the Lake Supe-
rior region) or major volcanic types in other areas. This close
Volcanic-Footwall Type Oxide/Carbonate BIFs association between volcanic-footwall type BIFs and felsic vol-
These BIFs are by far the most abundant throughout geo- canism is the same as that between VMSDs and felsic volcanism
logic history (Table 1); they are typically 30–300 m in thickness throughout the world (Ohmoto, 1996). Such associations suggest
and extend 2–100 km in longitudinal direction (Goodwin, 1973; that BIF(±VMSD)-forming hydrothermal systems are more fre-
Gross, 1990). VMSD-footwall type oxide BIFs are typically ~1 quently developed in continental-rift zones and subduction zones
m to ~100 m in thickness and 1–10 km2 in aerial extent; they are than in MOR regions. From size, shape, and distribution, we may
in fact a subsidiary type of volcanic-footwall type BIFs, because also conclude that most volcanic-footwall type oxide/siderite
VMSDs are hosted in submarine volcanic rocks, and VMSD- BIFs belong to the smoker-type mineralization group, but some
footwall type BIFs generally become volcanic-footwall type in (including Red Sea BIFs) are clearly brine pool-type.
areas outside VMS ore bodies. For this reason, BIFs have been a
useful exploration guide for VMSDs (Spry et al., 2000). The best Shale-Footwall Type Oxide/Carbonate BIFs
examples of this type of BIF are found in the ca. 2.7 Ga Abitibi The best-known examples of this BIF type are the 2.6–2.4
Greenstone Belt, Ontario, Canada (Fig. 7). Ga BIFs in the Hamersley Basin, including the ca. 2.6 Ga Mara
A previously unnoted important characteristic of volcanic- Mamba, ca. 2.5 Ga Brockman, and 2.45 Ga Weeli Wolli IFs (Table
footwall type BIFs is that they are more commonly associated 2). They range from ~100 to ~500 m in thickness and occur con-
with felsic volcanic rocks than with mafic volcanic rocks (Table tinuously over a > 60,000 km2 area. These footwall black shales
contain major components of tuffs and basalt sills, and younger (Fig. 7). Such a spatial relationship suggests that these BIFs
BIFs in the Hamersley Basin (i.e., the ca. 2.35 Ga Boolgeeda IF) were formed by low-temperature (T < 200 °C) submarine
are volcanic-footwall type BIFs. hydrothermal fluids that discharged at peripheral sites of hydro-
thermal systems, whereas the VMSDs were formed by higher
Dolomite/Limestone-Footwall Type Oxide/Carbonate BIFs temperature fluids located near the center of heat sources. The
The best example of this BIF type is the ca. 2.5 Ga Kuruman close spatial association of BIFs with Cu-rich VMSDs, and the
IF in the Transvaal district, South Africa (Beukes, 1983; Klein fact that the BIFs are overlain by thick (>2 km) successions of
and Beukes, 1989). These BIFs (200–400 m thick) overlie the submarine volcanic rocks, without interbedded units of con-
Campbellrand Dolomite (500–2000 m thick) and crop out conglomerates and sandstones, suggests that most (if not all) volca-
tinuously over a distance of >450 km. They may represent sedi- nic-footwall type BIFs in the Abitibi Greenstone Belts formed
ments that deposited in shallower and more marginal parts of the under >2.5 km of seawater.
Hamersley Basin (see a later section). BIFs in the Michipicoten-Wawa areas of Ontario (Fig. 7)
An oxide-type BIF at Tynagh, Ireland (Lower Carbonifer- belong to a different greenstone terrane (the 2.75–2.7 Ga Wawa
ous), is another example of a dolomite/limestone-hosted BIF Greenstone Belt) that was autochthonously or allochthonously
(Table 2). It is ~30 m thick and ~2 km in extension, and is closely assembled with the Abitibi belt ca. 2.7 Ga (Thurston, 1991,
associated with massive Pb-Zn sulfide deposits (Irish-type mas- 2002). These BIFs were probably deposited at shallow (<500 m)
sive sulfide) that formed on the seafloor, probably by hydro- water depths, because of the occurrence of stromatolites in some
thermal brines that generated from a deep sedimentary basin siderite BIFs (Hofmann et al., 1991) and the change in footwall
(Ohmoto et al., 1990). lithology from volcanic rocks to sandstone/conglomerates within
a distance of ~70 km (Goodwin, 1973). Although no VMSD has
Sandstone/Conglomerate-Footwall Type BIFs been found within a ~100 km radius of these BIFs (Fig. 7), a dis-
Examples of this BIF type include the 1.85 Ga Sokoman tinct footwall alteration zone developed underneath the BIFs in
IF at the Howell River area (Klein and Fink, 1976) and the these areas (e.g., Morton and Nebel, 1984; Gross, 1991), suggest-
200,000-yr-old Fe-Mn oxide deposits in the Afar Rift, Ethiopia ing that they also formed from locally discharged hydrothermal
(Table 2). But the best-known examples are the 1.87 Ga BIFs fluids. The absence of VMSDs in the Michipicoten-Wawa areas
in the Gunflint and Mesabi ranges of the Lake Superior region; is probably the consequence of a shallow-sea environment.
they were most likely a continuous formation that extended
over a distance of >400 km, but because of the intrusion of BIFs in the Hamersley Basin, Western Australia
the Duluth Gabbro they appear as two separate BIFs. These
two BIFs are 100–300 m thick and are underlain by a ~10- The prevailing geologic model for the 2.7–2.3 Ga BIFs in the
m-thick unit of quartz arenites, conglomerates, and carbon- Hamersley Basin, which was developed before the theory of plate
ates with an algal structure that deposited unconformably on tectonics, suggests that they formed in a basin that developed on
Archean basement rocks (e.g., Bayley and James, 1973; Morey a stable continental shelf where the basin water was in free com-
and Southwick, 1995). Abundant geological, petrological, and munication with the global oceans (e.g., Cloud, 1968; Holland,
paleontological evidence (e.g., stromatolites) exists to sug- 1984). However, during the past 20 years or so, the geological
gest that these BIFs deposited in shallow waters; granules and and geochemical data accumulated on BIFs and associated rocks
öölites are typical features in these BIFs. suggest that the tectonic, geologic, geochemical, and hydrologi-
cal processes that took place in the Hamersley Basin, and in other
BIFs in the Abitibi-Wawa Greenstone Belts, Ontario, brine pool-type BIFs, were similar to those occurring in the Red
Canada and Black Seas during the past ~50 m.y. period (Ohmoto, 1993,
Ohmoto et al., 2004). Thus, we name our theory for brine pool-
The Abitibi Greenstone Belt, which is ~500 × 1000 km in type BIFs as the “Red and Black Seas Hybrid Model.” First we
size and hosts ~150 BIFs and 30 VMSDs (Fig. 7), comprises will introduce a review of the geologic and hydrologic processes
several autochthonous submarine volcanic terranes that are in the Red and Black Seas, and then present geochemical data on
2.75–2.67 Ga in age (e.g., Jackson and Fyon, 1991; Thurston, the Hamersley BIFs to justify our model.
2002; Ayer et al., 2002). The submarine volcanic rocks consist
of bimodal compositions (mostly basalts and rhyolites, and much Mineralization in the Red and Black Seas
less, andesites). Recent studies (e.g., Hollings and Kerrich, 2000; Whereas the Red Sea rift system is the site of active tectonic,
Polat and Kerrich, 2001) suggest that these volcanisms were igneous, and hydrothermal activities, the Black Sea represents a
related to both arc basalts (from slab dehydration wedge melting) terminal (or resting) stage of a failed rift system with no current
and dacites/adakites (from later slab melting). igneous or hydrothermal activity, and is filled with a ~15-km-
The Abitibi BIFs typically occur in clusters that are spaced thick sequence of sedimentary and igneous rocks (Fig. 8A). Both
at ~20–100 km intervals and the volcanic-footwall type oxide/ the Red and Black Seas host stratified water bodies composed
siderite BIFs typically occur 3–30 km from known VMSDs of denser (saline) anoxic bottom water that is overlain by lighter
Deep, is only 5 × 14 km in size, and hosts a ~150-m-deep anoxic,

hot brine pool (T = ~60 °C, salinity = ~23%, and ∑Fe2+ = ~80
ppm at depths of ~2200 m), which is overlain by a ~50-m-thick
transitional zone and a ~2000 m-deep oxic, evaporated seawater
(T = ~25 °C, salinity = ~5%) (Pottorf and Barnes, 1983). The
submarine hydrothermal fluids feeding the brine pool originated
from evaporated seawater and became more saline, hotter, and
metal-rich while circulating through Tertiary evaporite beds
and underlying basalts before discharging into depressions on
the seafloor. A ~4-m-thick bed of metalliferous sediments with
average compositions of ~20% Fe, 2.1% Zn, and 0.45% Cu has
accumulated in the Atlantis II Deep (Shanks and Bischoff, 1980;
Pottorf and Barnes, 1983; El Shazly, 1990).
The temperature of discharging fluids in the Atlantis II Deep
brine pool has fluctuated between ~100 and >300 °C during the
past ~25,000 yr; iron oxide-rich sediments (i.e., oxide BIFs)
developed at T ≈100 °C, whereas alternating beds of Fe-Cu-Zn
sulfide-rich sediments (i.e., VMSD) precipitated at T > 300 °C
(Pottorf and Barnes, 1983; Ohmoto, 1996). High-temperature
(T >~250 °C) fluids can be rich in H2S, as well as Fe, Cu, Zn,
and other metals (Fig. 3), thus precipitating metalliferous sulfide-
rich sediments when the hot fluids mix with cooler basin bottom
water. Lower-temperature hydrothermal fluids (T = ~100 to ~200
°C) are typically Fe2+-rich, but too poor in H2S to precipitate all
the metals as sulfide minerals, thus creating a hot brine pool rich
in Fe2+- and silica (Fig. 3). As this hot brine and cooler overlying
O2-rich water mixed, ferric hydroxide and amorphous silica crys-
tals nucleated and settled on the seafloor to form goethite-silica-
rich layers (i.e., equivalent to oxide BIFs).
The Black Sea is the world’s largest anoxic basin (423,000
km2; ~500 × 1000 km); approximately half of the area is deeper
than 2000 m (Fig. 1B). It is land-locked and only connects to
the Mediterranean Sea by a narrow waterway (Bosporus Strait).
The Black Sea is also stratified; the surface zone (0 to ~50 m in
depth) is diluted by river water to ~1.8% salinity, and the bottom
water (~100 to ~2000 m in depth) is supplied mostly by oxic
Mediterranean seawater (~3.5% salinity) that is diluted by fresh
water to ~2.2% salinity. The bottom water is anoxic and rich in
H2S and CH4 that was produced by sulfate reducers and metha-
nogens, respectively. The transition zone between the oxic and
anoxic waters is ~50 m thick (Degens and Ross, 1974; Eremeev,
1992; Murray, 1991).
The nucleation of iron sulfide minerals occurs mostly in the
transition zone by utilization of the biogenic H2S from the bottom
Figure 8. (A) Simplified schematic of the crustal structure beneath
the Black Sea (Degens and Ross, 1974). (B) Comparison of the areal water and the particles of Fe3+-bearing minerals (e.g., goethite)
dimensions of the Hamersley Basin and the Black Sea. (C) The Red that were transported by river water and wind (aerosols) (Mura-
and Black Seas hybrid model for the Hamersley BIFs, highlighting moto et al., 1991; Wilkin and Arthur, 2001; Jørgensen et al.,
the involvements of seawater and river water in the formation of Su- 2004). The deposition of fine-grained iron sulfide particles was
perior-type BIFs. periodically interrupted by the deposition of shales via turbidites
to form finely laminated sulfide-rich sediments (e.g., Degens and
(less saline), cooler, oxic surface water. However, significant dif- Ross, 1974). During the active rifting period of the Black Sea (ca.
ferences exist in the origins of these stratified water bodies. 15 Ma), Mediterranean seawater could have circulated through
In the Red Sea, many isolated brine pools are created at the hot rocks underneath the Black Sea and discharged onto the
depths > 2000 m. The most famous brine pool, the Atlantis II seafloor to create a metalliferous brine pool. Such a scenario
is the essence of our Red and Black Seas hybrid model for the
Hamersley BIFs.
The Red and Black Seas Hybrid Model for the Hamersley
BIFs
The Hamersley Basin, which hosts the largest known BIFs
in Earth’s history, is only ~60,000 km2 in size today (~100,000
km2 originally; Trendall and Blockley, 1970), which corresponds
to less than one-fourth the size of the Black Sea (Fig. 8B). Our
model for the Hamersley BIFs incorporates the following sugges-
tions:(1) The Hamersley Basin was an intracontinental (failed)
rift basin, created by mantle upwelling, that experienced large-
scale, but intermittent, igneous and hydrothermal activities from
ca. 2.8 to ca. 2.4 Ga (Fig. 8C). (2) During BIF formation, the
Hamersley Basin was a land-locked sea, connected to the open
ocean only by (a) narrow inlet(s), and density stratified like the
Black Sea. (3) Large-scale, intermittent submarine hydrothermal
systems discharged Fe2+ through major fracture systems in (not
outside) the Hamersley Basin and created a basin-scale, hot (T =
~50 to ~100 °C) metalliferous brine pool. (4) Ferric (hydr)oxides
and silica nucleated primarily in the transition zone between the
surface oxic water and underlying hot brine pool, and then settled
onto the basin floor. (5) The separation of silica- and iron-rich
microbands occurred mostly during the early diagenesis of Fe-
and Si-rich sediments. Continuous reactions between the ferric
(hydr)oxides and overlying Fe2+-rich brines caused the ferric
(hydr)oxides to transform to magnetite. (6) The nucleation of sil-
ica also occurred by surface water evaporation, especially during
Figure 9. A simplified stratigraphic column of the Mount Bruce
periods of no hydrothermal activity. And (7) when the brine pool Supergroup of the Hamersley Basin, Australia. Compiled from data
temperatures decreased below ~50 °C during the waning peri- in Trendall and Blockley (1970), Krapez et al. (2003), and Pickard
ods of hydrothermal activity, SRB became active and produced (2003).
pyrite- and organic-rich shales (i.e., sulfide BIFs).
Geologic Evidence for the Red and Black Seas Hybrid Model bimodal volcanism) and the deposition of BIFs (e.g., the Weeli
Recent geological and petrological investigations of the Wolli and Boolgeeda IFs, and the Woongarra Volcanics) from
Hamersley Basin (e.g., Barley and Groves, 1992; Barley, 1993; 2460 to 2410 Ma. Note that bimodal volcanism, including rhyo-
Barley et al., 1997; Blake, 2001; Eriksson et al., 2002) suggest lite that generated by partial melting of lower crustal rocks by
it was a major igneous province, possibly related to mantle mantle-derived basalt magmas, is a characteristic of intraconti-
plume(s), during the period between ca. 2.8 and ca. 2.4 Ga. Its nental rift-related magmatism.
stratigraphic record (Fig. 9) suggests the following history: (1) Thick accumulations of shallow-water-facies carbonate
the initial rifting of a thick continental crust (composed mostly beds of the Tumbiana and Paraburdoo Members of the Wittenoon
of granite batholiths and greenstones older than 2.9 Ga) accom- Dolomite show that the Hamersley Basin was at times highly
panied by eruptions of large volumes of mostly subaerial flood evaporitic and (semi-)closed into one or more basins. The abun-
basalts (e.g., the Mount Roe, Kylena, and Maddina Basalts) dance of Na-rich silicates (e.g., riebeckite, crocidolite), which
and depositions of mostly fluvio-lacustrine sediments (e.g., the probably formed by reactions between tuff-rich shale units and
Hardey and Tumbiana Formations) from 2775–2680 Ma; no alkaline-rich hydrothermal solutions during regional hydrother-
record was preserved from 2680 to 2630 Ma; (2) the deposition mal events ca. 2.3–2.2 Ga (Hagemann et al., 1999; Brown et al.,
of alternating sequences of mostly black shales (e.g., the Jeeri- 2004), also suggests that rocks enriched in NaCl and KCl (i.e.,
nah, Mount Sylvia, and Mount McRae Formations), carbon- evaporite beds) existed in the pre-Brockman sequence. Although
ates (e.g., the Paraburdoo Member of the Wittenoon Dolomite), such evaporite beds are not found in the Hamersley district, prob-
cherts (e.g., the Bee Gorge Member of the Wittenoon Dolomite), ably because they were dissolved away during a long history of
and BIFs (e.g., the Mara Mamba and Brockman IFs), with minor fluid-rock interactions, they could have played an important role
interbedded tuffs, basalt, and rhyolite, from 2630 to 2460 Ma; in the development of hydrothermal brine pools, much like those
and (3) eruptions of large volumes of basalts and rhyolites (i.e., in the Red Sea.
The sedimentary units between the 2630 Ma Jeerinah Forma- Considering that the average thickness of these shales is ~100
tion and the 2420 Ma Boolgeeda IF total ~2.5 km in thickness m each, and that they were deposited throughout the Hamersley
and do not exhibit any evidence of subaerial erosion during their Basin, we suggest that these two black shales represent the larg-
deposition. On the basis of sedimentological analyses, Simonson est accumulations of organic matter in geologic history. By com-
et al. (1993) and Krapez et al. (2003) suggested that carbonates, parison, the average organic C content of Phanerozoic marine
as well as shales, were transported as turbidites from the northeast shales is only ~1 wt% (Holland, 1984).
margins (shelf) of the Hamersley Basin and deposited at depths The enormous accumulations of organic matter in the
> ~400 m (i.e., the depth below the storm wave base). Krapez et McRae and Jeerinah shales indicate that the primary productivity
al. (2003) also suggested that the precursor sediments to BIFs in the surface zone was extremely high and the bottom water of
(hydrothermal muds) were deposited on the flanks of submarine the Hamersley Basin was anoxic to facilitate the preservation of
volcanoes, which were located in the deeper, southwestern parts organic matter. The great primary productivity must have been
of the Hamersley Basin, and resedimented by density currents. a consequence of very efficient nutrient recycling (nitrate and
phosphates) to the surface zone, which would have been possible
Geochemical Evidence for the Red and Black Seas Hybrid if the basin was a nearly closed system where river waters sup-
Model plied nutrients to the basin. Therefore, a land-locked sea like the
Discoveries of molecular biomarkers of cyanobacteria (2- Black Sea, rather than a wide-open basin, is a better explanation
methyl hopanes) and eukaryotes (stranes) from shales of the for the organic-rich shales of the Hamersley Basin.
Mount McRae and Jeerinah Formations (Brocks et al., 1999, Extensive evidence indicates that large-scale hydrothermal
2003; Eigenbrode et al., 2004), and the experimentally deter- activity most likely occurred intermittently before, during, and
mined minimum pO2 requirements for the key enzymes in after the deposition of BIFs in the Hamersley Basin. The Mount
eukaryotes to produce sterols (Jahnke and Klein, 1983), suggest McRae Shale, which forms the footwall of the Brockman IF,
that the surface water layer of the Hamersley Basin contained exhibits several features typical of feeder zones, especially in
free O2 molecules equivalent to pO2 > 0.01 atm (i.e., >5% of areas of large iron deposits. For example, shale samples from
the present atmospheric pO2 level [PAL]) during the 2.7–2.5 Ga the Mount Whaleback mine (DDH186 and RD1) (Fig. 8B) are
period. This pO2 value corresponds to 18 μM as the minimum enriched in Zn (up to ~2000 ppm) and Fe (~20 wt%), whereas
dissolved oxygen (DO) content of the surface zone of the Ham- some shale samples show depletions of Fe (Fig. 11A). The K2O
ersley Basin. contents of the McRae shales decreased from an original value
The redox stratification of the Hamersley Basin is suggested of ~4 wt% to <1 wt% (Fig. 11B). High Zn contents (~2000 ppm)
by the common occurrence of kerogen with δ13C values <<−30‰ are also reported in samples from the Roy Hill Member of the
(Hayes et al., 1983; Brocks et al., 2003). Organic matter with Jeerinah Formation (Davy, 1985), which is the footwall of the
such low δ13C values is produced in environments where the Mara Mamba IF (2590 Ma).
remnants of primary producers (cyanobacteria and eukaryotes The apparent Rb-Sr age of the Mount McRae shales at
[aerobes]) with δ13C values ~−30‰ were partially converted the Mount Tom Price Mine decreased from the depositional
to CH4 (δ13C values <<−60‰) by methanogens (anaerobes) in age of ca. 2500 Ma to 1950 ± 290 Ma by hydrothermal altera-
an anoxic water body, which was subsequently oxidized to CO2 tion, whereas that at the Mount Whaleback Mine (2460 ± 130
(δ13C values <<−60‰) by methanotrophs (aerobes) that lived at Ma) is essentially the same as the depositional age (Fig. 12A).
the interface between the anoxic and oxic waters (Hayes, 1994). The apparent Nd-Sm ages of the Mount McRae Shale are 2100
One of the characteristic features of BIF-associated black ± 600 Ma for samples from the Mount Tom Price and Mount
shales is their high content of organic carbon (kerogen). For Whaleback Mines (Fig. 12B). Similarly, Alibert and McCulloch
example, the organic C contents of the 2.5 Ga Mount McRae (1993) reported that the Nd-Sm isotope systematics of many BIF
and 2.7 Ga Jeerinah shales typically range from 1 to 15 wt%, samples from the Brockman IF were disturbed from the original
with an average of ~5 wt% (Fig. 10A). The S-C contents of these ages of ca. 2500 Ma to ca. 2100 Ma by hydrothermal processes.
shales are similar to those of euxinic sediments in the Black Sea All these data suggest that hydrothermal activity in the Hamers-
(Fig. 10C). The H/C atomic ratio of kerogen in the Hamersley ley Basin persisted, probably intermittently, during a ~600 m.y.
shales are typically reduced to <0.1, compared to a ratio of ~1.5 period from ca. 2.6 to ca. 2.0 Ga.
for fresh organic matter (CH2O) that deposited onto the seafloor. The Mount McRae Shale contains abundant pyrite nodules
This reduction in the H/C ratio occurred because of the thermal that are locally surrounded by a quartz overgrowth layer that is
maturation of organic matter (i.e., losses of H, O, and C atoms as up to ~5 mm in thickness (Haruna et al., 2003). Textural rela-
petroleum and natural gases) during a long history of hydrother- tionships of the nodules with their host shales suggest that these
mal, diagenesis, and metamorphic processes. Our calculations, pyrite nodules and the overgrowth quartz crystals grew during
using an equation that relates the H/C ratio of kerogen to the the early diagenetic stage (i.e., before solidification) of the Mount
losses of H, O, and C compounds (Watanabe et al., 1997), suggest McRae Shale, which would correspond to the depositional period
that the original contents of organic matter in the Mount McRae of the overlying Brockman IF (Kakegawa et al., 1998; Haruna et
and Jeerinah shales was as high as 50–70 vol% of the sediments. al., 2003).
A B
20 20
2.7 Ga Jeerinah Formation 1.8 Ga Rove Formation
y = 0.787x + 1.166 r = 0.788 y = 0.615x + 1.055 r = 0.248
15 15
S (wt%)
S (wt%)
10 10
5 5
0 0
0 5 10 15 0 5 10 15
org. C (wt%) org. C (wt%)
C 10
Black Sea sediments : <200 m depth zone oxic bottom water
x : 300 - 1000 m depth zone
8 : >1000 m depth zone
] anoxic bottom water
6
S (wt%)
4
x
xx xx
xx x
2 x x x x x
xx x x
x
x x
0
0 2 4 6 8 10 12 14 16 18 20
org. C (wt%)
Figure 10. Sulfide-S and organic-C contents of black shales. (A) The 2.7 Ga Roy Hill Member of the Jeerinah Formation, Hamersley Basin,
Australia (Davy, 1985). (B) The 1.8 Ga Rove Formation of Ontario, Canada (Poulton et al., 2004). (C) The Black Sea (Ohmoto et al., 1990).
Fluid inclusions are abundant in the quartz overgrowth of The samples we investigated for alteration and fluid inclu-
pyrite nodules from the Mount Tom Price Mine. Two popula- sions came from iron ore mines (e.g., the Mount Tom Price and
tions of fluid inclusions are recognized: a high-salinity type (T Mount Whaleback Mines). We have not systematically inves-
= 130–220 °C; salinity = 5–12 wt%), and a low-salinity type (T tigated the differences in the footwall alteration in other areas.
= 150–200 °C; salinity = 0.5–3 wt%) (Fig. 13). These data sug- However, field observations suggest that large pyrite nodules are
gest that the Mount Tom Price Mine was one of the discharging more abundant in the mine areas than in other areas, which in
sites for hydrothermal fluids, and that both evaporated seawater turn suggests that the present mine sites were also the main fluid
(salinity > 3.5%) and meteoric water (salinity < 3.5 wt%) were discharge locations for BIFs. In many ore districts, major frac-
involved in hydrothermal mineralization. This recognition of two ture systems in the basement rocks become the main conduits for
water types in the Hamersley Basin was an important reason for fluid discharge during and after ore formation (e.g., Ohmoto and
our suggestion that the basin was stratified with a metalliferous Skinner, 1983).
hydrothermal brine pool (salinity = 5–12%) and an overlying Using an in situ laser ablation method, we have analyzed
diluted (salinity < 3.5%) water body (Fig. 8C). millimeter-scale variation in the δ18O values of quartz in the Fe-
A
10000
A
: DDH 289 4.0
: DDH RD-1 Mt. Whaleback Mine 2460 ± 126 Ma
1000 enriched
3.0
87Sr/86Sr
Zn (ppm)
2.0
100 Mt. Tom Price Mine 1952 ± 289 Ma
1.0
original
depleted shales
10
0.0
0.1 1 10 100 0 20 40 60 80 100
total Fe2O3 (wt%) 87Rb/86Sr
0.4 B
y = -1.268x + 0.507 r 2 = 0.791 B 0.512
0.5114
K2O/Al2O3 (wt%/wt%)
0.3
0.5112
0.511
143Nd/144Nd
0.2
0.5110
0.110 0.120 0.130
0.510
0.1 Mt. Whaleback and Mt. Tom Price Mines
: DDH 493 2010 ± 620 Ma
Al2O3 >10wt%
0.0 0.509
0.10 0.20 0.30 0.40 0.00 0.05 0.10 0.15
147Sm/144Nd
MgO/Al2O3 (wt%/wt%)
Figure 12. Rb-Sr and Nd-Sm ages of the Mount McRae Shale (Uyeda,
Figure 11. Geochemical data of the 2.55 Ga Mount McRae Shale 1994).
(Hamersley, Australia), which shows the characteristics of footwall-
rock hydrothermal alteration underneath VMSDs (cf. Ohmoto et al.,
1983). (A) Some samples are enriched in both Zn and Fe, whereas 15
some are depleted in both Zn and Fe. (B) Depletion of K was accom-
Salinity (wt%, NaCl equivalent)
panied by an increase in Mg (this study).
10
High salinity type
poor (silica-rich) and Fe-rich layers of a ~30 cm section of the 5

Brockman IF. Results (Fig. 14A) show a very large variation, Low salinity type
from +11.2‰ to +23.0‰, with a general tendency for higher δ18O
values of quartz in the Fe-rich microbands, compared to those in ?
the silica-rich microbands. The δ18O values of quartz crystals in 0

the silica-overgrowth zone of pyrite nodules in the Mount McRae 100 150 200 250 300 350
Shale (~+20‰; Haruna et al., 2003) are similar to those of silica Homogenization temperature (°C)
in Fe-rich bands of BIFs.
The δ18O value of quartz depends on the temperature and Figure 13. Homogenization and freezing temperatures of fluid inclu-
sions in the quartz overgrowths of pyrite nodules from the Mount
δ O value of the water involved in quartz formation, whether the
18
McRae Shale. An additional ~30 fluid inclusions were investigated for
quartz and water attained isotopic equilibrium during formation, homogenization temperatures only, most of which were between 120
and whether the δ18O value of quartz was not modified during and 220 °C. Data from Haruna et al. (2003).
A B 250
210
200
δ18Oqtz = +11 ‰
208 Fluids for
150
T (°C)
Fe-poor bands
206 100 Fluids for

Fe-rich bands
50 δ18Oqtz = +23 ‰
204
Height from the Base of BIF4 (cm)
0
202 -30 -20 -10 0 10 20
δ18Ow (‰)
200
198
190
188
Figure 14. (A) δ18O profile of a ~30 cm section of a micro-banded BIF sample of the Brock-
man IF, as determined by an in situ laser ablation method. Gray and white bands represent
186 Fe-rich and Fe-poor cherts, respectively. Note the δ18O values are generally higher in Fe-rich
bands than in Fe-poor bands. (B) T - δ18OH2O relationships for δ18Oquartz ranging from +11‰ to
+23‰. Note that the δ18OH2O values of ~0‰ (i.e., seawater) and ~–10‰ (i.e., meteoric water)
184
at T ~100 °C can best explain the δ18Oquartz values of Fe-rich versus Fe-poor cherts, respectively
(this study).
182
180
10.0 12.5 15.0 17.5 20.0 22.5
δ18Oqtz (‰)
diagenesis and regional metamorphism. The δ18OH2O – T relation- quartz range of +11‰ to +23‰. In this case, the data can be
ships are constructed for quartz with a δ18O range of +11‰ and interpreted as follows: the precipitation of quartz with lower
+23‰ (Fig. 14B) using the quartz-water fractionation factor by δ18Oqtz values (e.g., +11‰) occurred from meteoric water with
Matsuhisa et al. (1978). This figure shows that many combina- an δ18OH2O value ~−10‰, whereas the precipitation of Fe-oxides
tions of δ18OH2O and T values can explain the observed δ18O range and quartz with δ18Oqtz values ~+23‰ occurred from hydrother-
for quartz. For example, if the δ18OH2O value was constant at 0‰ mal fluids with an δ18OH2O value ~0 to +8‰; the positive δ18OH2O
(i.e., today’s seawater value), a temperature variation from ~90 values suggest the hydrothermal fluids were seawater modified
to ~200 °C can explain the observed δ18O variation from +11‰ through evaporation and/or water-rock interaction. Therefore,
(T = ~200 °C) for quartz in a silica-rich mesoband to +23‰ (T variations in the mixing ratio of brine-pool water and the fresh-
= ~90 °C) for quartz in a Fe-rich band; however, these tempera- water-dominated upper zone (Fig. 8C) best explain our δ18O data.
ture relationships between Fe-rich or -poor bands are opposite to Note that our oxygen isotope data are not compatible with the
what we would expect from a hydrothermal model. Furthermore, conventional BIF model, which postulates an essentially constant
brine pool temperatures and quartz nucleation were probably δ18OH2O value of ~0‰ and T <50 °C for BIF mineralization.
lower than those of the discharging hydrothermal fluids (i.e., T Hamade et al. (2003) have found that the Ge/Si ratios of Fe-
< 200 °C). poor silica bands in the Brockman BIFs are essentially identical
If we assume that the probable temperature range of quartz to present-day seawater (~0.8 × 10−6), suggesting that the silica in
precipitation for these samples was between 50 and 150 °C, a Fe-poor cherts formed from silica derived from the weathering of
δ18OH2O range of ~–10‰ to +8‰ can explain the observed δ18O continental crust (i.e., riverine input), rather than from hydrother-
mal fluids. The Ge/Si ratios in the Fe-rich bands are as high as 20 (positive Eu anomalies) was overlain by diluted seawater (no Eu
× 10−6, which are similar to those in MOR hydrothermal fluids. anomalies), and periodic mixing of the two waters resulted in
Such data suggest that the Si, Ge, and Fe in Fe-rich bands were BIF formation.
derived from hydrothermal fluids, although Hamade et al. (2003) The absence, or only minor presence, of Ce anomalies in the
have concluded that the Ge and Fe were of hydrothermal origin, Brockman IF (Fig. 15C; also see Kato et al., this volume) is also
whereas the Si was from normal seawater. consistent with the suggestion that the BIFs accumulated in an
The neodymium isotopic compositions (εNd values) of the anoxic basin. The Fe-oxides probably had negative Ce anomalies
Brockman IF are slightly negative (~−1.0), which suggests sig- when they formed in the transition zone, but were subsequently
nificant REE contributions (about 1/3 ) in the Hamersley Basin modified during and after settling on the seafloor by reactions
came from the weathering of old continental crust (i.e., riverine with anoxic basin water and/or brine-pool water, which were
inputs); the other ~2/3 of the REEs were input from submarine both characterized by the absence of Ce anomalies (Fig. 15A).
hydrothermal fluids (Alibert and McCulloch, 1993). Although
many smoker-type BIFs show distinctly positive Eu anomalies The Kuruman IF, South Africa
(Fig. 15B; also see Kato et al., this volume), the Hamersley BIFs
show only weakly positive Eu anomalies (Fig. 15C). This is also The Kuruman IF in the Transvaal district of South Africa
consistent with our model that hydrothermal brine-pool water is another large BIF that has many similarities to the Brockman
IF in the Hamersley Basin, Australia, including deposition age
(Eriksson et al., 2002; Pickard, 2003). In the Northern Cape
Province area of the Transvaal district, BIFs (200–400 m thick)
overlie the Campbellrand Dolomite (500–2000 m thick) and crop
out continuously over a strike (~N-S) distance >450 km; in the
adjacent Transvaal area, the BIFs outcrop discontinuously (due
to erosion) over another ~400 km distance (E-W). The exposed
parts of these BIFs most likely deposited in shallow water along
the basin margins, as suggested from various sedimentological
features in the underlying dolomite (e.g., stromatolites) and in
the overlying units (e.g., öölites) (e.g., Beukes, 1973, 1983; Klein
and Beukes, 1989). However, in the unexposed parts of the sedi-
mentary basin, a thick (~100 m) black shale unit has been found
between the BIFs and the Campbellrand Dolomite, indicating that
the deposition of oxide BIFs continued in a deep water body.
Recent zircon SHRIMP (sensitive high-resolution ion
microprobe) dating of the Kuruman and Brockman IFs shows
excellent correlations between several key tuff beds in these two
districts (Pickard, 2003). These data suggest that the two BIFs,
which are currently separated by the Indian Ocean, formed in the
same sedimentary basin ca. 2.5 Ga; the Hamersley district repre-
sents deeper parts and the Transvaal district represents the basin
margins. Therefore, we classify the Kuruman IF, as well as the
Brockman IF, as examples of brine pool-type mineralization.
The isotopic compositions of Fe2+ in MOR hydrothermal
fluids are very uniform, with δ56Fe values in the range of –0.5‰
± 0.2‰, and the isotopic fractionation factors between siderite
and Fe2+ are ~0‰ (Yamaguchi et al., 2005). Therefore, we would
expect the δ56Fe values of siderite in BIFs to be quite uniform
(~−0.5‰), if the Fe in BIFs was supplied from a large Fe2+ res-
ervoir, such as Jacobson-Holland-Klein-Beukes’s model oceans.
However, the δ56Fe values of siderite in the Kuruman IF, and also
those in the 2.60 Ga Carawine Dolomite in the Hamersley Basin
(stratigraphically equivalent to the Mara Mamba IF in Fig. 9),
range from ~–2‰ to ~+1‰ (Yamaguchi et al., 2005). Such large
Figure 15. Comparisons of the REE chemistry of (A) modern seawa-
ters; (B) 2.7 Ga Algoma-type BIFs from Temagami, Ontario, Canada δ56Fe variations are, in fact, strong evidence for the brine-pool
(M. Bau, 2003, personal commun.); and (C) 2.5 Ga Brockman IF from model, in which the total amount of Fe2+ in a brine pool increased
the Hamersley Basin, western Australia (Kato et al., this volume). by new fluxes of hydrothermal fluids or decreased with time by
the precipitation of minerals. For example, the δ56Fe value of Fe2+ Ga); (2) formation of small BIFs in the intracontinental basin (ca.
in a brine pool may become less than −0.5‰ by the continuous 2.2–2.0 Ga); (3) continuous rifting, mafic volcanism, and forma-
precipitation of ferric hydroxides or greater than –0.5‰ by the tion of the Menominee, Iron River-Crystal Falls, Marquette, and
continuous precipitation of pyrite and/or siderite, because the iso- Cuyuna North Range BIFs (ca. 2.0–1.9 Ga); (4) change from
topic fractionation factors between ferric (hydr)oxide and Fe+ are extensional to compressional tectonics, development of an island
positive, whereas those between pyrite (and siderite) and Fe2+ are arc and forearc basin (Animikie Basin), and formation of the
negative (Yamaguchi et al., 2005). Gunflint, Mesabi, Emily, Gogebic, and Baraga BIFs (ca. 1.9–1.85
Ga); (5) continent-arc collision, deepening of the forearc basin,
BIFs in the Lake Superior Region, United States-Canada and deposition of the Rove-Virginia black shales (ca. 1.85 Ga);
and (6) folding, thrusting, and metamorphism (ca. 1.85 Ga).
The Lake Superior region, the type locality of the Lake Supe- However, a recent geochronological and geochemical study
rior-type BIFs in Gross’s (1965) classification, hosts many major of volcanic rocks and iron formations by Schneider et al. (2002)
BIFs (Fig. 16). Stratigraphic correlations of BIFs in this region indicates that all iron formations in the Lake Superior district
are difficult, because they are highly deformed and metamor- formed at 1874 ± 9 Ma, coeval with arc-related volcanism, in one
phosed. Morey and Southwick (1995) suggested the existence of or more forearc basins during arc accretion from the south.
two distinct BIF groups that formed during different times and The lithostratigraphy of the ca. 1.9–1.85 Ga Animikie Group
under different geologic environments. The first group (e.g., the in the Lake Superior region, which includes basal quartzite, BIFs,
Menominee, Iron River-Crystal Falls, Marquette, and Cuyuna and an overlying shale formation, suggests that whereas the
North Range BIFs), belonging to the volcanic-footwall type of exposed sections of the Mesabi-Gunflint BIFs most likely depos-
our classification, was suggested to have formed ca. 2.0–1.9 Ga ited in shallow water, possibly the photic zone, the Marquette
in association with bimodal volcanism. The second group (e.g., region was probably under a deep brine pool. Although Morey
the Gunflint, Mesabi, Emily, Gogebic, and Baraga BIFs), belong- and Southwick (1995) have not speculated on this, their tectonic
ing to our sandstone/conglomerate-footwall type, were suggested model suggests that the Animikie Basin (ca. 1.9–1.85 Ga) was
to have formed between ca. 1.9 and 1.85 Ga. not completely open to the oceans, a setting that would have
According to the Morey-Southwick (1995) reconstruction allowed submarine hydrothermal fluids that discharged from the
of the tectonic and ore-forming histories of the Lake Superior basin floor to create large brine pool-type BIFs (e.g., the Gunflint
region, BIFs in this region were deposited throughout the pro- and Mesabi BIFs).
gressive growth and ultimate destruction of a rifted continental The geological and mineralogical characteristics of the ca.
margin. The suggested sequence of events is as follows: (1) ini- 1.9–1.85 Ga BIFs in the Lake Superior region (e.g., the Gun-
tiation of continental rifting and mafic volcanism (ca. 2.3–2.2 flint and Mesabi BIFs) are more similar to those of the Kuru-
Figure 16. BIFs in the Lake Superior

district, United States-Canada (Morey
and Southwick, 1995).
man IF (shallow water), than the Brockman IF (deep water). The Several recent investigators (e.g., Widdel et al., 1993; Kon-
presence of an anoxic Fe2+-rich, H2S-poor brine at very shallow hauser et al., 2002; Brown, this volume) have suggested that
depths (<10 m) in the Hamersley and Animikie Basins under iron-oxidizing bacteria were responsible for the precipitation of
an oxygenated atmosphere was not unusual, because the redox ferric hydroxides in BIFs. Two groups of iron-oxidizing bacteria
boundary in many saline lakes (e.g., the Great Salt Lake) occurs are known: (1) aerobic chemolithoautotrophs (e.g., Gallionella
at depths <5 m. ferruginea), and (2) anaerobic anoxygenic photoautotrophs (e.g.,
Chromatium sp.). Aerobic iron-oxidizers basically promote reac-
CHEMICAL PROCESSES FOR THE FORMATION OF tion (2), which is the oxidation of Fe2+ by dissolved O2 mole-
BIFS AND SECONDARY IRON ORES cules, whereas anaerobic photoautotrophs promote reaction (5)
by utilizing the H2 generated by reaction (5) in non-oxygenic
Precipitation Mechanisms for Ferric (Hydr)oxides in BIFs photosynthesis (such as reaction 3). Therefore, Konhauser et al.
(2002) expressed the overall reactions involving iron oxidizers
General as follows:
Prior to the 1980s, the only known mechanism for the pre-
cipitation of ferric (hydr)oxides from aqueous Fe2+ was by reac- 6Fe2+ + 1/2O2 + CO2 + 16H2O ⇒ CH2O
tion with molecular O2, which was generated by the photodis- + 6Fe(OH)3 + 12H+ (6)
sociation of H2O either with or without biological factors (e.g.,
cyanobacteria), to facilitate the following reactions: 4Fe2+ + CO2 + 11H2O ⇒ CH2O + 4Fe(OH)3 + 8H+ (7)
2H2O + hν ⇒ 2H2 + O2 (1) Any or all of the above three mechanisms (reactions 5–7)
could have caused the nucleation of ferric (hydr)oxides in BIFs,
and if nucleation was confined to the photic zone and under an anoxic
atmosphere. However, if nucleation occurred below the photic
Fe2+ + 1/4O2 + 5/2H2O ⇒ Fe(OH)3 + 2H+ (2) zone, the only plausible oxidation mechanism would have been
the reaction of hydrothermal Fe2+ with molecular O2 in deep
or (Fe2+ + 1/4O2 + H2O ⇒ 1/2Fe2O3 + 2H+). (2′) ocean water. The presence of free O2 molecules below the photic
zone can be linked to an oxygenated atmosphere (see below).
If the atmosphere was rich in H2 from volcanic gas and oxy-
genic photosynthetic organisms (e.g., cyanobacteria) were absent, Ferric (Hydr)oxides in Smoker-type BIFs
O2 production by reaction (1) would become insignificant because If pre–1.8 Ga oceans were Fe2+-rich, as suggested by many
of the backward reaction. However, if H2 is removed through dif- previous investigators (e.g., Cloud, 1968; Drever, 1974; Holland
fusion into space or biosynthesis (i.e., CO2 reduction), and Petersen, 1995), we would predict pre–1.8 Ga volcanic-asso-
ciated BIFs (1) to occur as large continuous beds, rather than as
CO2 + 2H2 ⇒ CH2O(organic matter) + H2O, (3) small individual bodies and clusters of variable sizes and shapes;
(2) to occur in close association with basalts, rather than with
O2 production from reaction (1) could build an oxygen-rich envi- less abundant rhyolites; (3) not to occur in close association with
ronment. The combination of reactions (1) and (3), which occurs hydrothermally altered footwall rocks; (iv) to have uniform min-
in cells of oxygenic photoautotrophs (e.g., cyanobacteria), is a eralogy (e.g., siderite/hematite ratio) and uniform compositions
well-known overall reaction for oxygenic photosynthesis: (e.g., Eu and Ce anomalies) among different microbands in a
single hand specimen and among different BIF bodies; and (v) to
CO2 + H2O + hν ⇒ CH2O(organic matter) + O2. (4) have different mineralogy, textures, and geochemistry than post–
1.8 Ga BIFs. However, none of these predictions were found to
Therefore, early BIF models (e.g., Cloud, 1968) related the be true.
formation of oxide-BIFs to the evolution of cyanobacteria, the In a previous section, we have suggested that many (if not
size of oxide-BIFs to the amount of O2 generation, and the pre- most) VMSD- and volcanic-associated BIFs are the products of
cipitation of iron-rich microbands to summer seasons (e.g., Tren- smoker-type mineralization, in which the nucleation of minerals
dall and Blockley, 1970). occurred by the rapid mixing of hydrothermal fluids with local
Supported by laboratory experiments, Braterman and Cairns- cold seawater in deep (>2.5 km) oceans, although some volca-
Smith (1987) and Anbar and Holland (1992) proposed that the nic-associated BIFs (e.g., the 2.7 Ga Michipicoten BIFs in the
photolysis of Fe2+ by UV lights in the absence of molecular O2 Wawa Greenstone Belt) clearly formed at shallow depths. One
might have been the cause of ferric hydroxide precipitation in may argue that the Fe2+ that discharged at a depth >2.5 km did
BIFs before the emergence of cyanobacteria: not immediately form ferric oxides, but instead was carried by a
hydrothermal plume to the photic zone, where it was converted
Fe2+ + 3H2O + hν ⇒ Fe(OH)3 + 2H+ + 1/2H2. (5) to ferric hydroxide minerals by photochemical and/or microbial
reactions (see reactions 1–7); these minerals then sank through the open oceans. Such a fluid circulation scenario, however, would
>2.5-km-deep water column and accumulated on top of a VMS have produced Fe-rich microbands that were very uneven in
body (i.e., the discharging spot). However, such a scenario is thickness and distribution but had a general trend of thickening
unlikely, because the ferric hydroxide particles would have been toward shallower parts of the basin. Such large-scale circulation
widely dispersed by both surface and deep currents and therefore also would have disturbed the sediments, especially in shallow
would not have accumulated near fluid discharge sites (i.e., areas basins. However, two of the important characteristics of BIFs
of footwall alteration) to form a well-defined BIF body. in the Hamersley Basin are that (1) the thicknesses of a Fe-rich
Individual grains of iron oxides and carbonates in BIFs show microband and mesoband are essentially uniform (averaging
large variations in size and shape (Trendall and Blockley, 1970; ~100 μm and ~1 cm, respectively) throughout the basin, and (2)
Ahn and Buseck, 1990) because of the coagulation of amorphous the sediments were not disturbed by bottom currents (Trendall
minerals during settlement through a water column and recrys- and Blockley, 1970). These BIF textures suggest that Fe2+-rich
tallization during diagenesis and metamorphism. However, the water mixed with O2-rich water gently, uniformly, and simultane-
abundance of very fine grained (<1 μm in size) crystals of hema- ously throughout the Hamersley Basin as suggested in our Red
tite and siderite in many BIFs and the confined geometry of both and Black Seas hybrid model (Fig. 8C).
VMSD- and volcanic-footwall type BIFs suggest that the nucle- As indicated by the molecular fossils of cyanobacteria and
ation of iron-bearing minerals and amorphous silica occurred eukaryotes, the surface water of the Hamersley Basin contained
very rapidly at (or near) the discharging sites, much like the black at least 18 μM of DO. Reaction (2) and (2)’ indicate that 1 mol
smokers on MOR hydrothermal systems. DO can precipitate 4 mol Fe2+ as ferric (hydr)oxide. Therefore,
The simple cooling of low-temperature (<~250 °C) subma- a simple mass balance calculation indicates that mixing a ~2-
rine hydrothermal fluids does not form sulfide-rich ores, because m-thick layer of this water and a 1.3-cm-thick layer of Fe2+-rich
they are poor in H2S (Fig. 3), or iron oxides, because most hydro- brine (∑Fe2+ = 1 mM) will form a microband that is 100 μm
thermal fluids do not contain any dissolved O2 molecules. For thick and 30% Fe2O3. For comparison, the thickness of a Fe-rich
example, a typical submarine hydrothermal fluid at ~250 °C may microband in the Brockman IF—and also in most other BIFs—
have a fO2 value of ~10−40 atm and contain ~10−3 moles/kg H2O of falls in a range from ~20 μm to ~5 mm with an average of ~100
∑Fe2+ (e.g., Ohmoto et al., 1983a). According to thermodynamic μm (Trendall and Blockley, 1970) (Fig. 17A). If the DO content
data, simple cooling may cause this fluid to become supersatu- of the surface water was 180 μM (i.e., ~50% that of the present
rated with respect to magnetite and/or hematite. However, no ocean surface water), the required thickness of O2-rich water to
oxide will form unless mechanisms exist at the depositional site precipitate the same amount of iron oxides is reduced to only
to both continuously dissociate H2O and remove (oxidize) the ~20 cm. Clearly, the thickness of a Fe-rich microband varies with
dissociated H2 molecules so that the generated O2 molecules can the DO concentration of the surface water, ∑Fe2+ concentration
react with the Fe2+ to form iron oxides. of the brine, and the thicknesses of the two mixed water bodies.
Anaerobic microbes are one possible mechanism to remove However, our calculations suggest that the microbands in brine-
H2 in submarine environments (see reaction 7). Although such a pool type BIFs were produced by the basinwide, gentle distur-
mechanism may have been important in the formation of some bance (mixing) of thin layers (centimeters to meters) of brine and
BIFs that formed in the photic zone, it was not significant for surface water.
most volcanic-associated BIFs that formed in deep oceans. Fur- Such mixing of the bottom layer of the surface water (mostly
thermore, kerogen (a remnant of organisms) is essentially absent river water) and the upper layer of the underlying brine pool is
in oxide/carbonate BIFs and footwall rocks of most volcanic- expected to occur when the surface water flows from a land-
associated BIFs, although it is abundant in the overlying sedi- locked basin toward an open ocean (Fig. 8C). The flux of river
ments (black shales). This suggests that areas of fluid discharge water to the surface zone of the basin varies seasonally with the
were too hostile for microbes to participate in BIF mineraliza- climate and geography of the fluvial systems. The depth and
tion, but microbes flourished in the local submarine depressions thickness of the transition (mixing) zone also depend on the vol-
after the cessation of hydrothermal events. ume of the brine pool, which is influenced by the intensity of sub-
Another possible mechanism to remove H2 from the hydro- marine hydrothermal activity, sea level, and basin topography. If
thermal fluid is to react it with the O2 in local seawater, but this the basin is completely isolated from an open ocean because of
is basically the same as precipitating Fe2+ by reactions with O2 a sea level drop and/or a significant decrease in riverine flux, the
molecules in local seawater. thickness of the oxygenated surface water zone may be signifi-
cantly reduced and the transition zone may become anoxic, ceas-
Ferric (Hydr)oxides in Brine Pool-type BIFs ing the formation of Fe oxides. If the sea level rises and the basin
The current popular model for BIF formation (e.g., Holland, becomes more open to an ocean, normal seawater may invade
1984; Klein and Beukes, 1992; Morris, 1993) proposes that a the basin and change the characteristics of the O2-rich surface
large mass of Fe2+-rich deep ocean water periodically invaded zone. For all these reasons, we suggest that the mixing of Fe2+-
a basin, overturned the entire basin water, mixed with O2-bear- rich brine and O2-rich surface water occurred periodically, rather
ing surface water to nucleate ferric oxides, and returned to the than continuously, over time to generate Fe-rich microbands.
A Whether iron oxide, siderite, or iron sulfide precipitates dur-

ing the mixing of hydrothermal fluids and local seawater depends
on which of the following reactions dominates (Fig. 18):
Fe2+(hyd) + 1/4O2(s.w.) + 5/2H2O ⇒ Fe(OH)3 + 2H+ (2) at T <~80 °C

5 mm
2Fe2+(hyd) + 1/2O2(s.w.)+ 2H2O ⇒ Fe2O3 + 4H+ (2’) at T > ~80 °C
Fe2+(hyd) + HCO3–(s.w.) ⇒ FeCO3 + H+ (8)
Fe2+(hyd) + 2H2S(s.w.) ⇒ FeS2 + 2H+ + H2 (9)
B The relative importance of the above reactions is determined

1 cm by the following three parameters: (1) the chemistry of the hydro-
thermal fluids, especially the concentrations of ∑Fe2+ (= Fe2+ +
FeCl+ + FeCl2 + FeHCO3–); (2) the chemistry of the local seawa-
ter, especially the concentrations of DO, ∑CO32–, and ∑S2–; and
(3) the kinetics of the above reactions. For example, the O2 mol-
ecules in seawater inside a local depression (basin) may be totally
consumed by reactions with hydrothermal ∑Fe2+ to form iron
oxides, whereas the remaining ∑Fe2+ is consumed by ∑CO32– to
form siderite. Thus, the abundance ratio of iron oxide to sider-
ite in the precipitates depends on the concentrations of DO and
∑CO32– in the seawater, the ∑Fe2+ concentration of hydrothermal
fluids, and the mixing ratio of hydrothermal fluid to seawater to
consume all the ∑Fe2+ in the discharged fluid (Fig. 18).
siderite + quartz If the local water becomes anoxic and the temperature falls
below ~50 °C, SRB may create an H2S-rich local environment
and mixing of such local seawater with Fe2+-bearing hydrother-
mal solutions will precipitate iron sulfides. The relationships
magnetite + quartz between the mineralogy of dominant Fe-bearing minerals and
the compositions of hydrothermal fluid and local seawater are
illustrated in Figures 18 and 19.
Figure 17. (A) A thin section photo of an oxide BIF from the 2.7 Ga
Temagami BIF, Ontario, Canada. Red bands are silica-rich, ∑Fe-poor, Atmosphere CO2 (350 ppm) O2 (0.2 atm)
and high in hematite/magnetite ratio; black bands are silica-poor, Oxygenated Ocean
∑Fe-rich, and high in magnetite/hematite ratio. (B) A hand specimen Euxinic Sea
showing alternating mesobands of oxide BIFs (magnetite+quartz) CO2 (0.01 mM)
SO42- (28 mM) O2 (0.3 mM)
and siderite BIFs (siderite+quartz) from the Helen Mine, Wawa, On- (SRB) HCO3- (2 mM)
tario, Canada.
H2S (0.3 mM)
FeS2 FeCO3 Fe(OH)3
Seafloor
Formation Mechanisms of Siderite- and Pyrite-Rich BIFs FeS2

Fe2+ (2 mM) (cooling)
In every BIF we investigated, the changes in BIF types from H2S (0.5 mM)
oxide-, to carbonate-, to sulfide-type occurred with time (i.e., Hydrothermal Fl.
stratigraphy) during the history of a sedimentary basin, rather
than laterally as suggested by James (1954). Alternation of BIF Figure 18. Schematic illustrations of the relationships between local
types over a centimeter-vertical scale is very common (Fig. 17B). seawater chemistry and BIF Fe-mineralogy: the relative abundance
of O2, ΣCO32–, and H2S in local seawater determines whether iron
These features suggest that changes in BIF type resulted from (hydr)oxide, siderite, or pyrite precipitates when the seawater mixes
changes in the concentrations of DO, HCO3–, and H2S in local with Fe2+-rich hydrothermal solutions. The concentration values are of
seawater over time. the modern atmosphere and oceans.
A 10 B
Present Pre-2.2-Gyr 10
atmosphere atmosphere
0
Rye & Holland -40 0
(1998) Ohmoto (1999)
pO2 = 1 atm
-10 Pavlov et al. oxic
-10 seawater
(2001) Kasting (1987) Fe3+
-20
log pO2 (atm)
-30 -20 10-6 m
log pH2 (atm)

no free O2
log aO2 (aq)

-30
-30 (a)
10-4 m Hematite
-40 no free H2
-20 -40
Goethite
-50 Fe2+
-50 BIF Magnetite
subsurface anoxic water
-60
-10 -60 hydrothermal anoxic
fluids
(b)
seawater Siderite
pCH4 = 100 ppm
-70
pCH4 = 10000 ppm -70
Minnesotaite
-80 (greenalite) Siderite T = 100ûC
0 -80
lower limit of water 0 2 4 6 8 10 12 14
-90
-4 -3 -2 -1 0 pH
log pCO2 (atm) at pCO2 = 10-0.5, SiO(aq) = 10 ppm
Figure 19. The stability relationships of Fe-bearing minerals. (A) logaO2 - logaCO2 diagram at SiO2(aq) = 10–3.4 M (24 ppm) at 25 °C. (B) logaO2
– pH diagram showing the stability fields of Fe-bearing minerals at pCO2 = 10–0.5 atm, SiO2(aq) = 10 ppm, and T = 100 °C. Also shown are the
changes in fluid chemistry and formational minerals when hydrothermal fluid is mixed with O2- or HCO3–-rich local seawater.
The concentration of ∑Fe2+ in submarine hydrothermal fluids seawater was free of O2 and H2S, as suggested by many inves-
at ~100 to ~200 °C typically ranges from 0.1 mM to 5 mmol per tigators (e.g., Walker and Brimblecombe, 1985; Canfield and
kilogram of H2O (Fig. 3). The maximum concentrations of DO, Raiswell, 1999), siderite would have been the only Fe-bearing
∑CO32–, and ∑S2– in modern oceans also fall within this range. mineral in volcanic-associated BIFs. However, the abundance of
For example, the average DO concentration in high-latitude ocean hematite and magnetite indicates that the bottom seawaters near
surface water is ~0.35 mM. However, the DO content becomes these BIFs contained free O2 molecules.
zero in anoxic waters. The concentration of ∑CO32– (≈HCO3–) in
today’s oceans averages 2.3 mM but is higher in anoxic basins, Precipitation Mechanisms of Silica in BIFs and Silica
such as ~4 mM in the Black Sea (Fry et al., 1991), due to the Chemistry of the Archean Oceans
addition of CO2 (+HCO3–) from the oxidation of organic mat-
ter in the water column. The ∑S2– content of normal seawater is Essentially all researchers agree that the cherts in BIFs pre-
<<0.01 mM but is much higher in anoxic waters due to bacterial cipitated as amorphous silica from the aqueous silica in seawater
sulfate reduction, such as ~0.3 mM in the anoxic water body of (regardless of whether it was supplied mostly by rivers or by sub-
the Black Sea (Ohmoto and Goldhaber, 1997). marine hydrothermal fluids) and were subsequently transformed
The similarities in the concentration of ∑Fe2+ in hydrother- to opal, cristobalite, and quartz during diagenesis and metamor-
mal fluids and the maximum concentrations of DO, ∑CO32–, and phism:
∑S2– in seawater are important because the ∑Fe2+ in hydrothermal
fluids may be quantitatively precipitated as either oxide, carbon- H4SiO4(aq) ⇒ Si(OH)4(amorphous silica) (10)
ate, or sulfide during mixing with a variety of seawaters at ratios
ranging from ~1/100 to 100. This is probably the main reason the Si(OH)4(amorphous silica) ⇒ SiO2(quartz) + 2H2O (11)
abundance of Fe-bearing minerals in a BIF is typically unimodal,
i.e., oxide-, carbonate-, or sulfide-rich. The H4SiO4(aq) in today’s oceans is supplied almost entirely
The above discussions illustrate that the relative concentra- by rivers. The silica concentration of normal seawater is ~5 ppm
tions of DO, ∑CO32–, and ∑S2– in local seawater are essential to as SiO2 (Drever, 1982), which is about the saturation value for
control the dominant mineralogy (oxide, carbonate, or sulfide) of quartz at ~4 °C and approximately an order of magnitude less
submarine hydrothermal deposits. A corollary is that if Archean than the saturation value for amorphous silica at ~25 °C (i.e.,
~100 ppm) (Fig. 20). The formation of chert occurs in today’s mal fluid with normal seawater may cause the precipitation of
oceans because silica-accreting organisms (e.g., radiolarians and quartz, but not amorphous silica, if the precipitation of silica-rich
siliceous sponges) extract silica from seawater to form amor- minerals is controlled solely by the solubility of quartz and/or
phous silica shells; the dead organisms accumulate on the sea- amorphous silica (Fig. 20). However, laboratory experiments
floor as amorphous silica-rich sediments, which are subsequently (e.g., Bazilevskaya, 2004) have demonstrated that the precipita-
transformed to less soluble minerals (e.g., cristobalite, tridymite, tion of amorphous silica occurs from silica-undersaturated solu-
and quartz) during diagenesis. tions when ferric hydroxides are formed, due to the adsorption
Silica-accreting organisms probably first appeared ca. 750 of amorphous silica on ferric (hydr)oxide particles. Thus, co-
Ma (Bengtson, 1994). Knauth and Lowe (2003) have suggested precipitation of amorphous silica and ferric (hydr)oxide can be
that, in the absence of silica-accreting organisms, the silica con- explained. Similarly, mixing brine-pool water that is undersatu-
centrations of Precambrian oceans must have been much higher rated with amorphous silica, but supersaturated with quartz (e.g.,
(e.g., ~300 ppm) than today and were controlled by the solubility SiO2 = 200 ppm at 100 °C), with cooler, silica-poor surface water
of amorphous silica, rather than quartz. The precipitation rate of (SiO2 ≈50 ppm) would have caused the precipitation of amor-
silica is greatly increased when the availability of solid and/or phous silica, if the nucleation of ferric (hydr)oxides occurred.
colloid surfaces (e.g., clays, glass, quartz, silicate minerals, and The separation of amorphous silica from ferric (hydr)oxide
colloids in sediments) for silica nucleation increases (Rimstidt, particles, and migration and reprecipitation of amorphous silica
1997). Therefore, it is probable that the silica concentrations of onto the seafloor, would have occurred during the early diagen-
the Archean oceans were maintained at levels about the satura- esis of Fe-Si-rich sediments, creating Fe-rich (lower) and silica-
tion values of quartz, much like modern oceans, although the rich (upper) microbands.
actual SiO2(aq) values could have been ~20 ppm, instead of ~5 The nucleation of amorphous silica may also occur in the
ppm (today’s oceans), if the Archean oceans were ~50 °C. upper zone of a hot brine pool through simple conductive cooling
If the Archean oceans were saturated or supersaturated with by the overlying cooler water body, because the solubilities of
respect to amorphous silica, silica would have precipitated from both quartz and amorphous silica decrease with decreasing tem-
high-latitude ocean water during winters and created essentially perature (Fig. 20). For example, fluid saturated with quartz at 200
uniform layers of cherts over a large (>>1000 × 1000 km) area. °C (SiO2 = ~200 ppm) may precipitate amorphous silica at ~70
However, the lateral extents of cherts in greenstone belts are typi- °C during conductive cooling. Such a precipitation mechanism
cally ~1 to ~100 km. may explain the frequent occurrence of thick (>1 mm), Fe-poor
Furthermore, the lateral extents of Fe-poor, silica-rich micro- silica bands in the Hamersley BIFs, as well as others. Evapora-
and mesobands are essentially the same as those of Fe-rich, sil- tion of the surface water zone, especially during periods of no
ica-poor bands in most BIFs. These features suggest that most of brine pool development, would have been another mechanism to
the SiO2 in BIFs, as well as Fe, was supplied by local submarine cause the precipitation of silica.
hydrothermal fluids, rather than from normal ocean waters.
The silica concentrations of hydrothermal solutions are typi- Transformation from Hematite-Rich to Magnetite-Rich BIFs
cally controlled by the solubility of quartz (Rimstidt, 1997), such
as ~250 ppm at ~250 °C (Fig. 20). Simply mixing hydrother- A VMS ore body typically alters its mineralogical and geo-
chemical characteristics as it grows in size because the earlier
minerals are continually transformed by reactions with later
10 600 hydrothermal fluids (Fig. 21A). Similarly, during the growth of
a volcanic-associated BIF body, goethite and/or hematite crys-
tals could be subjected to later Fe2+-rich hydrothermal fluids (Fig.
SiO2 (milli moles/kg)
Amorphous Silica
Saturation 21B). As the crystals react with these “new” hydrothermal fluids,
hematite (or goethite) is converted to magnetite through the fol-
lowing reaction (Ohmoto, 2003):
5 conductive cooling 300
conductive cooling
Fe2O3(hm) + Fe2+ + H2O ⇒ Fe3O4(mt) + 2H+ (12)
& mixing
Quartz Saturation
mixing only Reaction (12) is not a redox reaction because there is no
change in the valence of any Fe or other atoms; it is simply an
0 0 acid-base reaction. This proposed mechanism differs from the
0 100 200 300 conventional model for hematite-magnetite transformation (e.g.,
T (°C) Perry et al., 1973), which suggests the following redox reaction
Figure 20. Solubilities of quartz and amorphous silica, and mecha- between the primary ferric (hydr)oxide minerals and organic
nisms of silica precipitation from submarine hydrothermal solutions carbon in BIF sediments to form magnetite during the metamor-
(modified after Spry et al., 2000). phic stage:
Figure 21. Growth model for submarine hydrothermal

deposits. (A) VMSD transformation. (B) BIF
transformation. Hm—hematite-chert ore, Mt—
magnetite-chert ore, BO—black ore, PO—massive
pyrite ore, YO—yellow ore (Ohmoto, 2003). (C)
Transformation of hematite to magnetite in brine
pool–type BIFs. (D) Transformation of magnetite-
rich BIFs to hematite-rich iron ores by subsurface
non-redox reactions.
6Fe2O3(hm) + CH2O(org) ⇒ 4Fe3O4(mt) + CO2 + H2O (13)

cally exhibit black color due to high magnetite/hematite ratios
(Fig. 17A).
Reaction (13) may explain some (but not all) of the mag- The goethite/hematite crystals in brine pool-type BIFs,
netite crystals in BIFs that were subjected to high-grade meta- which nucleated at the top of brine pools and settled onto the
morphism in the Lake Superior region and other regions (French, floor of brine pools, would have continuously reacted with the
1973; Perry et al., 1973). However, reaction (12) best explains overlying Fe2+-rich brine and partially transformed to magnetite
the magnetite crystals in most volcanic-associated oxide BIFs, through reaction (12). This explains the high magnetite/hematite
because their metamorphic grades are not high enough to have ratios (>1) in brine pool-type BIFs.
caused reaction (13) and/or because the kerogen content in BIFs
and footwall rocks is too low to have accounted for the high Formation from Magnetite-Rich BIFs to Hematite-Rich
magnetite/hematite ratios (>1) that they contain. Reaction (12) Secondary Iron Ores
also better explains the general trend that larger BIFs have higher
magnetite/hematite ratios. It also explains the well-known fea- The major iron ores in the world are those that developed
ture that low-Fe, silica-rich bands typically exhibit red color due from BIFs by the loss of silica and transformation of magnetite
to high hematite/magnetite ratios, whereas Fe-rich bands typi- to hematite; the Fe contents typically increased from ~20 wt% in
precursor BIFs to ~50 wt%. A popular model for the formation oceans were already oxidized to produce negative Ce anomalies.
of these iron ores combines supergene enrichment and metamor- The DO content of the bottom water in oxide-BIF environments
phic processes (e.g., Morris, 1985). It postulates that magnetite- can be estimated from simple mass balance calculations based on
rich BIFs were uplifted by a tectonic process and subjected to reaction (2) or (2′).
weathering under an oxygenated atmosphere to form goethite- According to our studies of several volcanic-associated
rich ores: BIFs, a typical oxide-rich microband is ~1 mm thick and distrib-
uted over an area ~5 km in radius with an average composition
2Fe3O4(mt) + 1/2O2 + 3H2O ⇒ 6FeO(OH)(gt) (14), of ~30% Fe2O3 and 70% SiO2. Thus, the total amount of Fe in an
average microband can be calculated as:
and then subsequently buried to a great depth (>5 km) where they
metamorphosed to hematite-rich ores: Fe (mol) = (10−1cm) • π • (5 × 105cm)2 •
(3.4 gm/cm3) • 0.30 • (1/159.7 g/mol) • 2 = 1.0 • 109 moles
6FeO(OH)(gt) ⇒ 3Fe2O3(hm) + 3H2O (15)
where the density of the iron-rich microband was assumed to be
This paleoweathering-metamorphic model was developed 3.4 g/cm3 and the molecular weight of Fe2O3 is 159.7 g. The total
because the principal mechanism for the transformation of mag- amount of hydrothermal fluid required to provide 1.0 × 109 mole
netite to ferric (hydr)oxides was thought to be the oxidation of of Fe is 1.0 × 1012 kg ( = 1.0 • 1015 cm3 = 1.0 km3), if the concen-
Fe2+ atoms in magnetite to Fe3+ atoms (i.e., reaction 14). Most tration of ∑Fe2+ in the hydrothermal fluid is 1 mM (56 ppm) and
geologists have also accepted the theory for the rise of atmo- its density is 1.0 g/cm3.
spheric O2 ca. 2.2 Ga (e.g., Holland, 1994). For these two rea- Because 1 mole O2 can precipitate 4 moles Fe2+ as Fe2O3
sons, the exploration for iron ores has primarily focused on major (or goethite), 2.5 × 108 moles O2 are required to precipitate 1.0
unconformities younger than 2.2 Ga. Ohmoto (2003), however, × 109 moles Fe2+. If the DO content of local seawater was as low
suggested that the more common mechanism for the conversion as 10−15 moles/kg H2O, the total volume of seawater required to
of magnetite to hematite in most geologic settings is the reverse provide this amount of O2 was 2.5 × 1023 kg ( = 2.5 × 1011 km3).
of reaction (12), i.e., the leaching of Fe2+ from magnetite by Fe2+- Obviously, this is an unacceptable amount because it far exceeds
poor hydrothermal fluids (such as those developed in carbonate- the total volume of every ocean combined (1.35 × 109 km3). If
and/or chert-dominated terranes): the DO content of local seawater was 340 μM (the average O2
content of today’s high-latitude surface water), only 7.1 × 1011 kg
Fe3O4(mt) + 2H+ ⇒ Fe2O3(hm) + Fe2+ + H2O (16) (= 0.71 km3) of seawater is required to form one microband. This
corresponds to a 1.4 mixing ratio (Rmix) of hydrothermal fluid to
Reaction (16) is not a redox reaction, either. According to local seawater, where Rmix is defined as:
this model, the transformation of magnetite-rich BIFs to hematite-
rich ores in large ore bodies mostly occurred in subsurface condi- Rmix = (mass of hydrothermal fluid)/
tions without an oxidant (Fig. 21D). Therefore, the formation of (mass of local seawater) (17)
hematite-rich secondary ores cannot be linked to atmospheric O2
content. Another important implication of this model is that iron The typical Rmix values for the formation of volcanic-hosted
ore exploration targets are not confined to paleosurfaces younger BIFs are estimated to be between 0.01 and 100, because at Rmix
than ca. 2.2 Ga; the formation of secondary ores from magnetite- <~0.01, positive Eu anomalies would not appear in BIFs, and
rich BIFs probably occurred throughout geologic time. at Rmix > ~100, negative Ce anomalies would not appear. At
Otake et al. (2005) have experimentally confirmed that the the minimum Rmix value of 0.01, the volume of local seawater
non-redox transformations of hematite to magnetite (reaction 12) required to provide the necessary amount of O2 (2.5 × 108 moles
and magnetite to hematite (reaction 16) indeed occur rapidly (<1 O2) to precipitate an average iron-rich microband is 100 km3 and
day at 150 °C) in a pH range of ~4 to ~6, even under high pH2 the DO concentration is 2.5 μM (= 2.5 × 10−6 moles/kg H2O).
conditions (up to 50 bars). This DO value is, therefore, the minimum value for deep (>2.5
km) ocean water involved in the formation of volcanic-associ-
CONSTRAINING THE EVOLUTIONARY HISTORIES ated BIFs. It compares well with the minimum DO value of 18
OF THE ATMOSPHERE, OCEANS, AND BIOSPHERE μM for the photic zone of the ca. 2.7 Ga Hamersley Basin, which
FROM BIFS was estimated from molecular fossils of eukaryotes.
Dissolved O2 Contents of Deep Ocean Water Atmospheric pO2
REE data of smoker-type BIFs (Fig. 15B; also see Kato et al., The DO content of today’s oceans continuously decreases
this volume) clearly indicate that BIFs are products of the mix- from ~240 μM (i.e., the amount in equilibrium with an atmo-
ing of hydrothermal fluids and seawater, and that Archean global sphere of pO2 = 0.21 atm at 25 °C) to ~50 μM at depths of ~500
m (i.e., the oxygen minimum depth) and then increases to ~170 Using the equation by Sarmiento (1992) that relates the
μM at depths of ~4 km (Fig. 22). The DO decrease in the upper DO and PO43– contents of the bottom ocean water to the DO
part of the oceans is due to DO consumption during the oxidation and PO43– contents of higher latitude water and the solubility
of dead organic matter as it sinks through the water column. The of O2, Lasaga and Ohmoto (2002) derived the following simple
increase in DO toward the bottom of the oceans is due to the influx equation that relates atmospheric pO2 to the DO value of deep
of a cold O2-rich surface water mass from higher latitudes. water (DOd):
DOd (μM) = 340 • (pO2/ pO20) –172 (18)
Atmospheric pO2 (% PAL) where the pO2/ pO20 value is the atmospheric pO2 value relative
40 50 70 100 to the present atmospheric pO2 level (PAL). Thus, if the atmo-
0
spheric pO2 falls below ~0.50 PAL (i.e., 50% of today’s level),
500 the DOd value becomes zero (i.e., anoxic).
If the atmospheric pO2 was as low as ~0.40 PAL, only the
1000
Ocean Depth (m)
photic zone (areas <~100 m deep) could have been oxygenated

(Fig. 22). If Earth’s climate in the Archean was much warmer
than today (e.g., Karhu and Epstein, 1986; Ohmoto and Felder,
2000
1987), deep ocean circulation may have been the reverse of
today’s circulation; evaporated seawater may have sunk in the
equatorial region and circulated toward higher-latitude areas.
3000
If this was the case, an atmospheric pO2 level much higher
than 0.50 PAL was required to maintain an oxygenated bottom
water, because the solubility of O2 in warmer and more saline
0.1 0.2 0.3 water is less than that in cold water.
Dissolved O2 (mmole/kg H2O) Local “oxygen oases” (e.g., Kasting, 2001) may have been
possible in the photic zone, but not in deep oceans because
Figure 22. Relationships between the oceanic DO contents and atmo-
there is no mechanism to generate O2 below the photic zone,
spheric pO2 levels, estimated from equations by Lasaga and Ohmoto
(2002). or to transport a small (<100-m-thick) O2-rich water mass from
the photic zone to the bottom waters through a deep anoxic
1000 %
Ohmoto
(1 atm) 100 %
10 %
1%
1000 ppm
100 ppm
10 ppm
1 ppm
Holland
Figure 23. Proposed models for the evolution of atmospheric
pO2
0.1 ppm
oxygen by Kasting (Kasting, 2001), Holland and his
10-2 ppm associates (Rye and Holland, 1998; Holland, 2002; Bekker
10-3 ppm Kasting et al., 2004), and Ohmoto and his associates (Ohmoto, 1997;
Lasaga and Ohmoto, 2002).
10-4 ppm
10-5 ppm
10-6 ppm
10-7 ppm
10-8 ppm
10-9 ppm
4.5 4 3.5 3 2.5 2 1.5 1 0.5 0
Age (Ga)
water column. Therefore, in order to have >2.5 μM of DO siderably less than today: ~70% of today’s value at 4.0 Ga, 85%
below the photic zone to form an average Fe-oxide layer in a at 2.5 Ga, and 95% ca. 1 Ga (Kasting, 1987). To compensate for
volcanic-associated BIF, the atmospheric pO2 level must have this lower solar luminosity and maintain the liquid oceans, the
been greater than ~0.50 PAL since ca. 3.8 Ga (i.e., the time concentrations of greenhouse gases must have been much higher
of Isua BIF formation) (Fig. 23). Note that the minimum pO2 than today. Assuming CO2 was the only greenhouse gas besides
requirement becomes higher if the ocean temperatures were water vapor, Owen et al. (1979) and Kasting (1987) calculated
higher than today, because the solubility of O2 decreases with that the atmospheric pCO2 was as high as ~1 atm ( = 3000 PAL,
increasing temperature. Therefore, the Archean oceans must where 1 PAL = 350 ppm) at 4.5 Ga, then gradually decreased to
have been fully oxygenated, except in local (or regional) anoxic 0.1 atm (~300 PAL) at 2.5 Ga, and to 0.01 atm (~30 PAL) at 1.0
basins, much like today. Ga (curve A, Fig. 24).
An important question that follows our model of a nearly Rye et al. (1995) proposed that siderite was absent in pre–2.2
constant atmospheric pO2 level through geologic time is the Ga paleosols because the atmospheric pCO2 was less than the
O2-controlling mechanism. According to Holland (1984), the equilibrium pCO2 value for the greenalite + siderite assemblage.
atmospheric pO2 level has been controlled by the relative fluxes From thermodynamic calculations, they concluded that the atmo-
of (1) O2 production by the burial of organic matter that was spheric pCO2 value must have been lower than that calculated
produced primarily by oxygenic photoautotrophs, (2) O2 con- from previous climatic models, and suggested that CH4, as well
sumption by reduced volcanic gases (e.g., H2, CH4), and (3) O2 as CO2, was a major greenhouse gas prior to ca. 2.2 Ga. Subse-
consumption by kerogen, sulfides and Fe2+-bearing minerals quently, a revised climatic model (Pavlov et al., 2001a, 2001b)
during soil formation. Lasaga and Ohmoto (2002) have applied suggested values of ~1000 ppm CH4 and ~2500 ppm CO2 for the
this principle to their numerical simulations of the geochemi- atmosphere at 2.8 Ga (curve B, Fig. 24). By comparison, today’s
cal cycles of carbon-hydrogen-oxygen-phosphate through atmosphere contains only ~1 ppm CH4 and ~350 ppm CO2.
the atmosphere-hydrosphere-biosphere-crust-mantle system. The atmosphere cannot contain both high O2 and CH4 con-
Their results suggest that the atmospheric pO2 rose essentially centrations, because together they react to form CO2 and H2O by
to the present atmospheric level (1 PAL) within ~10 m.y. since the following photochemical reaction:
the emergence of oxygenic photoautotrophs (even if the vol-
canic flux of reduced gas was as much as five times greater 2O2 + CH4 ⇒ CO2 + 2H2O (19)
than today) and that the atmospheric pO2 level has remained
within ± 50% of the PAL throughout geologic history (Fig. 23). Therefore, a methane-rich atmosphere is compatible with an
According to Lasaga and Ohmoto (2002), the main reasons that anoxic Archean atmosphere. However, serious flaws exist in Rye
atmospheric pO2 has maintained this level are two strong nega- et al.’s (1995) assumption that siderite did not form in Archean
tive feedback systems: one between the burial flux of organic surface environments, because it is an abundant mineral in BIFs.
matter and atmospheric pO2 value, and another between the For the formation of siderite, all of the following conditions must
soil O2 consumption flux and atmospheric pO2 value. be satisfied: (1) a low pO2 (<10−60 atm); (2) a very high pCO2
(>10–2.5 atm); (3) a slightly acidic to near neutral pH (~5.5 to
Atmospheric pCO2 and pCH4 ~7.5); and (4) a high concentration of ∑Fe2+ in water (~0.1 to
~100 ppm) (Figs. 19A, 19B). Such a low pO2 value only occurs
Because nuclear fusion in the Sun has been increasing since when anaerobic organisms produce a substantial amount of H2 in
its formation, the solar flux on early Earth must have been con- organic C-rich environments (e.g., subsurface anoxic water) by:
1 1000 %
0 100 %
CO2 (A)
-1 10 %
CO2 (C)
Concentration
log pi (atm)
-2 1%
Figure 24. Proposed models for the evo-
-3 CO2 (B) 1000 ppm
lution of atmospheric pCO2 and pCH4.
CO2 100 ppm Curves (A) are from Kasting (1987), (B)
-4
CH 4 (A) CH 4 (B) from Pavlov et al. (2001a, 2001b), and (C)
-5 10 ppm from Ohmoto et al. (2004).
CH 4 (C)
-6 1 ppm
CH4
-7 0.1 ppm
4.5 4 3.5 3 2.5 2 1.5 1 0.5 0
Age (Ga)

40
CH2O + H2O ⇒ CO2 + 2H2 (20) Modern continental siderites 15 n = 177
30 mean = 6.0‰
N 20 σ = 5.7‰
In fact, siderite appears to have formed only in organic-rich
environments throughout geologic history (Ohmoto et al., 2004).
10
0
A
40
But there are distinct differences in the occurrence and carbon Modern marine siderites15 n = 188
30 mean = -4.7‰
isotopic compositions between siderite in Phanerozoic sediments N 20 σ = 2.2‰
and in pre–1.8 Ga sedimentary sequences. The former typically
occurs as minor, disseminated grains with variable δ13C values
10
0
B
(−25‰ to +15‰) (Fig. 25). In contrast, siderite in pre–1.8 Ga -25 -20 -15 -10 -5 0 5 10 15 20 25
sedimentary rocks is very abundant, occurring as an important 150
Pre-1.8 Gyr low Fe limestones n = 341
constituent of BIFs and as Fe-rich massive carbonate beds (e.g., 100 and dolomites 25 mean = 0.1‰
N σ = 2.2‰
the 2.6 Ga Wittenoon Dolomite in the Hamersley Basin). δ13C 50
values of these rocks range from +3‰ to –15‰, but are more
0
C
commonly between –5‰ and 0‰ (Fig. 25), suggesting that -25 -20 -15 -10 -5 0 5 10 15 20 25
10%–20% of the ∑CO32– used to form the siderites came from 40
1.9 Gyr Gunflint I. F. n = 122
the decomposition of organic matter by SRB, and 80%–90% of 30
(siderite & ankerite) 22,23 mean = -3.3‰
the ∑CO32– was from normal seawater that was in equilibrium N 20 σ = 1.8‰
with the atmosphere. Using these values, together with the ther- D
10
modynamic evaluation of the conditions for siderite formation, 0

40
1.9 Gyr Gunflint I. F. n = 60
Ohmoto et al. (2004) constrained the pre–1.8 Ga atmospheric 30
(siderite); this study mean = -2.8‰
pCO2 level to >10–1.4 ± 0.2 atm (>100 PAL) if T = 25–50 °C and N 20 σ = 0.6‰
SiO2 (aq) < 100 ppm (curve C, Fig. 24). With such a high pCO2 ,
the presence of >10 ppm CH4 would have made the surface tem-
10
0
E
peratures much warmer than 50 °C. Therefore, the pre–1.8 Ga 40
-25 -20 -15 -10 -5 0 5 10 15 20 25
atmosphere was most likely CO2-rich and CH4-poor; a low pCH4 30 2.3 Gyr Kuruman I. F. n = 35
(siderite & ankerite) 18 mean = -3.6‰
(<10 ppm) is compatible with a high pO2 atmosphere. N 20 σ = 2.3‰
10
Oceanic Sulfur Chemistry 0 F
-25 -20 -15 -10 -5 0 5 10 15 20 25
40
The sulfur chemistry of the ocean is closely linked to the 2.5 Gyr Brockman I. F. n = 50
30 mean = -9.6‰
evolution of SRB and the atmospheric pO2 level. Modern oceans (siderite)19,20,25
σ = 1.7‰
N 20
are sulfate-rich, with an average concentration of 28 mM SO42–
(900 ppm S), because 30%–50% of the total SO42– flux to the
10
G
oceans comes from the oxidative weathering of pyrite; the sul- -25 -20 -15 -10 -5 0 5 10 15 20 25
fate is in turn removed from the oceans as biogenic sulfides and 40
2.6 Gyr Wittenoon F. n = 49
gypsum. Under an anoxic atmosphere, the SO42– content of the 30
(ankerite & dolomite)19,20,25 mean = -2.4‰
N 20 σ = 3.2‰
ocean water is expected to be much lower than the present value,
although there is no consensus as to how low Archean seawater 10
H
0
sulfate contents could have been. For example, Walker and Brim- 50
blecombe (1985) have estimated that the SO42– in Archean oceans 40 2.6 Gyr Wittenoon F. n = 235
(siderite, ankerite, & mean = -4.9‰
was generated only by photochemical reactions of volcanic SO2 N 30 dolomite); σ = 2.7‰
I
20 this study
and its concentration was less than ~1/30 of the present value 10
(i.e., <~1 mM); but they also suggested that the SO42– content 0
could have been as high as 1/3 of the present value. 40
-25 -20 -15 -10 -5 0 5 10 15 20 25
A recent proposal by Canfield and associates (Canfield and 2.75 Gyr Helen I. F. n = 32
30 (siderite)21 mean = -1.6‰
Teske, 1996; Canfield and Raiswell, 1999; Canfield et al., 2000; N 20 σ = 2.7‰
Bjerrum and Canfield, 2002; Habicht et al., 2002) suggests that 10
J
the oceanic SO42– content remained ~200 μM (i.e., <1/100 of 0
40
the present level), except in local evaporitic basins, until ca. 2.2 2.75 Gyr Helen I. F. n = 86
30
Ga, then gradually increased to ~10 mM ca. 800 Ma (Fig. 26). (siderite); this study mean = -1.5‰
N 20 σ = 1.0‰
This proposal is based on an interpretation that the small differ-
ence in δ34S values between sulfates and biogenic sulfides (Δ34S
10
0
K
= δ34Ssulfate – δ34Spyrite = <10‰), which characterizes Archean -25 -20 -15 -10 -5 0 5 10 15 20 25
sedimentary rocks, only occurs when the sulfate concentration δ13C (‰: PDB)
100 mM
Ohmoto
Seawater Sulfide and Sulfate Concentrations
28 mM
SO42-
10 mM
SO42-
1 mM
Figure 26. Evolution curves for the sulfate,
H2S sulfide, and Fe2+ contents of the oceans, as
suggested by Canfield and his associates (Can-
field and Raiswell, 1999; Bjerrum and Can-
0.1 mM field, 2002; Habicht et al., 2002) and Ohmoto
Canfield (2004).
0.01 mM
Fe
Ohmoto
Fe and H2S
0.001 mM
4.5 4 3.5 3 2.5 2 1.5 1 0.5 0
Age (Ga)
is <200 μM; Δ34S values >25‰ (typical values in Phanerozoic of dominant organic matter (e.g., remnants of cyanobacteria,
sedimentary rocks) occur when the sulfate concentration is > eukaryotes), the ratio of sulfide S to organic C may vary signifi-
200 μM. Ohmoto (1992), however, has suggested that the small cantly among sediments (Ohmoto et al., 1990). For example, the
Δ34S values are not necessarily an indicator of low-sulfate oceans sulfide S content and S/C ratio of average normal marine sedi-
because Δ34S values are also influenced by temperature and the ments are ~0.2 wt% and 0.4, respectively, but those of freshwater
relative rates of sulfate supply versus sulfate reduction in a water lake sediments are ~0.05 wt% and 0.03, respectively. The S con-
column and underlying sediments. tents and S/C ratios of marine sediments in euxinic basins can be
Compared to sulfur isotope values, a more reliable indicator as high as ~10 wt% and ~1, respectively (Ohmoto et al., 1990).
of seawater sulfate concentration is the relationship between the Therefore, if the sulfate concentration in the Archean oceans was
sulfide and organic C contents of sediments since SRB produce <200 μM, the sulfide contents in Archean shales are expected to
pyrite through the following reaction: be uniformly <0.1 wt%, and the S/C ratios <<0.1, provided the
sulfides were all biogenic.
8SO42– + 16H+ + 15CNR-CH2O + 4Fe(OH)3 As mentioned earlier, pyrite-rich black shales (i.e., sulfide
⇒ 15CNR + 4FeS2 + 15CO2 + 29H2O (21) BIFs) are the most common hanging wall (i.e., younger) rock
type of all oxide/carbonate BIFs. This common temporal rela-
where CH2O refers to the organic molecules metabolizable by tionship between oxide/carbonate BIFs and black shales gener-
SRB and CNR to non-reactive organic carbon molecules (i.e., ates the following questions: (1) Was the deepening of BIF-form-
residual organic C in sediments). Reaction (21) indicates that the ing basins to accumulate black shales caused by tectonic subsid-
pyrite content of a sediment increases with an increasing sulfate ence of the region or by a global rise in sea level? (2) Was the
content of water and that a positive correlation may exist between development of an euxinic water body the cause or consequence
the contents of pyrite and organic C in sediments if all the pyrite of the cessation of oxide BIF-forming processes? (3) Does the
was formed by bacterial sulfate reduction. However, because chemistry of BIF-associated black shales reflect global or local
the ratio of CNR to CH2O depends on the environment and type seawater chemistry?
For example, Poulton et al. (2004) have concluded that the
pyritic black shales (e.g., the Rove Formation) overlying the ca.
1.85 Ga Gunflint IF (Fig. 10B) are excellent evidence that the
global oceans became SO42–- and H2S-rich for the first time ca.
1.8 Ga, and that the creation of such oceans was the cause of
Figure 25. Comparison of the carbon isotopic compositions of carbon-
ate in various rocks. (A, B) Disseminated siderite in modern sediments. BIF cessation worldwide. However, if the Rove Formation is evi-
(C) Pre–1.8 Ga Fe-poor limestones and dolomites. (D–K) Pre–1.8 Ga dence for global SO42–- and H2S-rich oceans, then other BIF-asso-
Fe-rich carbonates (Ohmoto et al., 2004). ciated pyritic black shales, such as the 2.7 Ga Jeerinah Formation
(Fig. 10A) and 2.5 Ga Mount McRae Shale of the Hamersley ses) from the 2.76 Ga Hardey Formation (lacustrine) and the 3.0
Basin, must also be evidence for globally SO42–- and H2S-rich Ga Mosquito Creek Formation (marine) in the Pilbara district,
oceans. But such an argument would create a serious problem for Australia. (4) Tachibana et al. (2004) have also found no evi-
the Cloud-Holland-Canfield Archean ocean model because the dence of MIF-S in the 2.45 Ga Livingstone Creek Formation in
oceans could not become rich in both Fe2+ and H2S at the same Elliot Lake, Ontario, Canada. (Note that the Livingstone Creek
time, owing to iron sulfide precipitation. Formation underlies uraniferous quartz-conglomerate beds that
We interpret the S-C relationships of the BIF-associated were previously used as evidence for an anoxic atmosphere [e.g.,
Archean black shales (Fig. 10A) as evidence for sulfate-rich Holland, 1994]).
oceans, and the occurrences of black shales to the periodic devel- These recent findings suggest the following possible inter-
opment of local (regional) euxinic basins, rather than globally pretations: (1) the Archean atmosphere fluctuated between oxic
anoxic oceans. The alternations of oxide- (and/or carbonate)-type and anoxic; (2) the atmosphere has been oxic since ca. 3.8 Ga,
BIFs and pyritic black shales in the Hamersley Basin (and other and the MIF-S record indicates periods of violent volcanic erup-
BIF-forming basins) are interpreted as the periodic development tions, which ejected large amounts of SO2 above the stratosphere;
of hot brine pools, which created unsuitable environments for or (3) large MIF-S values in some geologic samples were caused
SRB and other microbes, much like today’s hot (T > 60 °C) brine by mechanisms other than atmospheric photochemical reactions.
pool in the Atlantis II Deep, Red Sea (e.g., Kaplan et al., 1969). Possibilities (1) and (2) should be considered, if atmospheric pho-
tochemical reactions are the only cause of MIF-S. Possibility (3)
Atmospheric Sulfur Cycle must be considered because of (a) the recognition of strong cor-
relations between the MIF-S values and the degree of hydrother-
The presence of the mass independent fractionation of sulfur mal alteration in samples from the 2.5 Ga Mount McRae Shale
isotopes (MIF-S) in some pre–2.1 Ga sedimentary rocks and the and the 2.7 Ga Jeerinah Formation that are highly organic carbon
absence of MIF-S in younger rocks, which were first discovered rich (Fig. 27); and (b) the creation of significant MIF-S (∆33S
by Farquhar et al. (2000), have been regarded by many recent = ~0.5‰) during thermochemical reduction of sulfate by amino
investigators (Kasting and Siefert 2002; Bekker et al., 2004) as acids (Watanabe et al., 2006). Clearly, more investigations are
the “smoking gun” for the dramatic change from an anoxic to an needed to understand the causes of MIF-S in geologic samples
oxic atmosphere ca. 2.2 Ga. This is because the MIF-S has been before linking MIF-S to atmospheric chemistry.
linked to photochemical reactions of volcanic SO2 in the absence
of an ozone shield to produce S0 with large positive Δ33S values
(>50‰) and SO42– with negative Δ33S values (<−10‰) (where
Δ33S = δ33S – 0.515 δ34S); these Δ33S values are characteristically 8
y = -0.055x + 22.824 r 2 = 0.754
within the range of 0‰ ± 0.3‰ in Phanerozoic sediments.
Ono et al. (2003) recognized the largest variations in Δ33S val-
6
ues (–2‰ to +8‰) in pyrite crystals from pyrite-rich black shales
(sulfide BIFs) in the Hamersley Basin (e.g., the 2.5 Ga Mount
McRae Shale and 2.7 Ga Jeerinah Formation; Fig. 9). However, 4
Bekker et al. (2004) found no evidence of MIF-S (i.e., Δ33S = 0‰
Δ33S (‰)
± 0.3‰) in the 2.35 Ga Timeball Hill Formation, which hosts

pisolitic IFs, in South Africa. These data led Bekker et al. (2004) 2
to conclude that the “Great Oxidation Event” (the change from
an anoxic to oxic atmosphere) occurred ca. 2.35 Ga.
0
However, there are several very serious problems in link-
ing the presence or absence of MIF-S in geologic samples to the
atmospheric oxygen level: (1) The S0 produced by the UV radia- -2
tion of SO2 in laboratory experiments (Farquhar et al., 2001) has
very large positive and variable Δ33S values, but also very nega-
tive and variable δ34S values relative to SO2. In contrast, geo- -4
logic samples with positive Δ33S values show only positive δ34S 250 300 350 400 450 500 550
values. (2) Savarino et al. (2003) reported the discovery of the K/Rb (atomic)
presence of MIF-S (Δ33S = +0.67‰ and −0.50‰) in volcanic
ashes associated with the violent eruptions of Mount Pinatubo Figure 27. The relationships between the magnitude of MIF-S isotopes
and an unknown volcano, but also the absence of MIF-S in ashes (Δ33S) and K/Rb ratios of a 22 m core section of the 2.55 Ga Mount
McRae Shale, measured by Ono et al. (2003) and this study, respectively.
associated with minor eruptions. (3) Watanabe et al. (2005) found Other geochemical data (e.g., Zn contents) and mineralogy of the same
that the MIF-S is absent (i.e., Δ33S = 0‰ ± 0.3‰) in syngenetic core sample indicate that trends of increasing Δ33S value and decreasing
and diagenetic pyrite crystals in all shale samples (>40 analy- K/Rb ratio accompany increasing degrees of hydrothermal alteration.
The Biosphere A
BIFs
Our suggestion of the development of an oxygenated
atmosphere and oceans by ca. 3.8 Ga implies that oxygenic
photoautotrophs (i.e., cyanobacteria or their ancestors) had B VMS
already evolved. This suggestion is consistent with others by
Schidlowski and Aharon (1992) and Rosing (1999) based on
C No summary for
Mantle Plumes younger plumes
the carbon isotope data of Isua sedimentary rocks. Our sug-
gestion is also consistent with the discovery of Ce anomalies D Super continent
in Isua BIFs (Kato et al., this volume). Greenstones
The biomarkers, carbon isotopic compositions of kerogen, (Submarine
and S/C relationships of pyritic black shales (sulfide BIFs) in volcanism)
the Hamersley Basin suggest that methanogens and methano-
trophs, as well as SRB, cyanobacteria, and eukaryotes, have 4.0 3.0 2.0 1.0 0
been active since at least ca. 2.7 Ga. The recent discovery by Time (Ga)
Buick (2005) of biomarkers of cyanobacteria, methanotrophs,
Figure 28. Schematic illustrations of the temporal distributions of (A)
and eukaryotes in 3.2 Ga sedimentary rocks from Pilbara, BIFs, (B) VMSDs, (C) global plumes, and (D) greenstones. (A) and (B)
Australia, indicates the very early divergence of major organ- are modified after Barley and Groves (1992) and Ohmoto (1996) with
isms, and the importance of biological feedbacks in regulating additional data from literature. (C) is from Isley (1995) and Condie
oceanic and atmospheric chemistry. (2001). (D) is from Green (1992) and Condie (2002).
BIF Distribution through Geologic Time
Because BIFs are found in rocks of all ages (Tables 1, 2), essentially the same; (2) the major microbial constituents in the
we can unequivocally conclude that BIFs are not indicators of an oceans (e.g., cyanobacteria, sulfate reducers, methanogens) have
anoxic atmosphere or anoxic oceans, as many previous research- been essentially the same; (3) the atmospheric pCO2 and oce-
ers suggested. Instead, the formation of oxide- and sulfide-geo- anic concentration of ∑CO32– have continuously decreased; (4)
logic history is an excellent indicator that the atmospheric O2 the pHs of rain- and ocean waters have continuously increased;
and oceanic SO42– levels have been essentially constant since ca. and (5) the global heat flux from Earth’s interiors and submarine
3.8 Ga. The higher abundance of siderite-rich BIFs in pre–1.8 Ga volcanic activity has continuously decreased.
BIFs, however, was probably the result of a higher pCO2 (>100
PAL) atmosphere. ACKNOWLEDGMENTS
Isley (1995) and Isley and Abbott (1999) have recognized
that the formation of BIFs in geologic history was episodic and This paper represents our work over the past 15 years or so.
coincided with major mantle plume events. The close tempo- We express our sincere appreciation to the following investigators
ral associations of BIFs with VMSDs, greenstone belts, oceanic who have helped us in the field and laboratories, and in discus-
rift systems, and mantle plumes (Figs. 28A–D) further suggest sions: Munetomo and Yoko Nedachi of Kagoshima University;
that BIFs were related to regional-scale submarine hydrothermal Mark Barley, Brian Krapez, and April Pickard of the University
processes. The scarcity of BIFs, VMSDs, mantle plumes, and of Western Australia; Arthur Hickman, Al Trendall, and Rich-
greenstones during the ca. 1.8–1.0 Ga period was not because ard Morris of the Geological Survey of Western Australia; Nic
the oceans and/or atmosphere became oxygenated, as postu- Beukes and Jens Gutzmer of Rand Afrikaans University; Kazu-
lated by the popular model, but because of the change in mantle masa Kumazawa, Ritsuko Murata, Takahiro Hanamuro, Kaoru
dynamics. During this period, a supercontinent existed, and Uyeda, Hiroaki Fujimaki, Toshio Nagase, and Shuji Kojima of
large marine rift systems, which were necessary for the forma- Tohoku University; Takushi Yokoyama of Kyushu University;
tion of VMSDs and BIFs, were scarce. and Michael Bau, Denny Walizer, Katya Bazilevskaya, Kate
BIFs became generally smaller and less abundant after Spangler, Tsubasa Otake, and David Bevacqua of the Pennsyl-
ca. 1.8 Ga. These changes probably reflect the fact that mantle vania State University. This manuscript greatly benefited from
plumes became smaller and less frequent, and the tempera- critical reviews of an earlier manuscript by Jim Franklin, Bob
tures of the mantle and core decreased; such changes would Rye, Cornelius Klein, Rob Kerrich, Art Rose, and especially
have resulted in smaller and short-lived submarine hydrother- Steve Kesler (a perfectionist). This work was supported by the
mal systems. Japanese Ministry of Education (Grant 03102002), the National
In summary, BIFs are excellent indicators that, since ca. Science Foundation (EAR9706279, EAR-0229556), the NASA
3.8 Ga, (1) the redox structures of the atmosphere (pO2, pCH4) Astrobiology Institute (NCC2-1057; CA#NNA04CC06A), and
and oceans (concentrations of Fe2+, H2S, and SO42–) have been the NASA Exobiology Program (CA#NNG04GK00G).
Blake, T.S., 2001, Cyclic continental mafic tuff and flood basalt volcanism
REFERENCES CITED in the Late Archaean Nullagine and Mount Jope supersequences in
the eastern Pilbara, Western Australia: Precambrian Research, v. 107,
Abrams, C.E., and McConnell, K.I., 1984, Geologic setting of volcanogenic p. 139–177, doi: 10.1016/S0301-9268(00)00135-2.
base and precious metal deposits of the West Georgia Piedmont: A mul- Bonatti, E., Fisher, D.E., Joensuu, O., Rydell, H.S., and Beyth, M., 1972,
tiply deformed metavolcanic terrain: Economic Geology and the Bul- Iron-manganese-barium deposit from the Northern Afar Rift (Ethio-
letin of the Society of Economic Geologists, v. 79, p. 1521–1539. pia): Economic Geology and the Bulletin of the Society of Economic
Adamson, R.G., and Teichmann, R.F.H., 1986, The Matchless cupreous pyrite Geologists, v. 67, p. 717–730.
deposit, South West Africa/Namibia, in Anhaeusser, C.R., and Maske, Bowing, R.J., 1989, Rare earth and other geochemical studies of Archean
S., eds., Mineral deposits of South Africa: Johannesburg, Geological banded iron formation; Sherman and Adams mines, Ontario [Ph.D. the-
Society of South Africa, v. 2, p. 1755–1760. sis]: Hamilton, Ontario, Canada, McMaster University, 399 p.
Ahn, J.H., and Buseck, P.R., 1990, Hematite nanospheres of possible colloi- Braterman, P.S., and Cairns-Smith, A.G., 1987, Iron photoprecipitation
dal origin from a Precambrian banded iron formation: Science, v. 250, and the genesis of the banded iron-formations, in Appel, P.W.U., and
p. 111–113. LaBerge, G.L., eds., Precambrian iron-formations: Athens, Theophras-
Alibert, C., and McCulloch, M.T., 1993, Rare earth element and neodymium tus Publications, p. 215–245.
isotopic compositions of the banded iron-formations and associated Brocks, J.J., Logan, G.A., Buick, R., and Summons, R.E., 1999, Archean
shales from Hamersley, western Australia: Geochimica et Cosmochi- molecular fossils and the early rise of eukaryotes: Science, v. 285,
mica Acta, v. 57, p. 187–204, doi: 10.1016/0016-7037(93)90478-F. p. 1033–1036, doi: 10.1126/science.285.5430.1033.
Allen, R.L., Lundström, I., Ripa, M., and Christofferson, H., 1996, Facies Brocks, J.J., Buick, R., Summons, R.E., and Logan, G.A., 2003, A recon-
analysis of a 1.9 Ga, continental margin, back-arc, felsic caldera prov- struction of Archean biological diversity based on molecular fossils
ince with diverse Zn-Pb-Ag-(Cu-Au) sulfide and Fe oxide deposits, from the 2.78 to 2.45 billion-year-old Mount Bruce Supergroup, Ham-
Bergslagen region, Sweden: Economic Geology and the Bulletin of the ersley Basin, Western Australia: Geochimica et Cosmochimica Acta,
Society of Economic Geologists, v. 91, p. 979–1008. v. 67, p. 4321–4335, doi: 10.1016/S0016-7037(03)00209-6.
Anbar, A.D., and Holland, H.D., 1992, The photochemistry of manganese Brown, M.C., Oliver, N.H.S., and Dickens, G.R., 2004, Veins and hydrother-
and the origin of banded iron formations: Geochimica et Cosmochimica mal fluid flow in the Mt. Whaleback iron ore district, eastern Hamersley
Acta, v. 56, p. 2595–2603, doi: 10.1016/0016-7037(92)90346-K. Province, Western Australia: Precambrian Research, v. 128, p. 441–474,
Appel, P.W.U., 1987, Geochemistry of the early Archaean Isua iron-forma- doi: 10.1016/j.precamres.2003.09.013.
tions, West Greenland, in Appel, P.W.U., and LaBerge, G.L., eds., Buick, R., 2005, Paleontological and geochemical records of early life in
Precambrian iron-formations: Athens, Theophrastus Publications, p. Archean rocks from the Pilbara Craton, Australia: 2005 Japan Earth
31–67. and Planetary Science Joint Meeting, Session P074, abstracts 036.
Appel, P.W.U., and LaBerge, G.L., eds., 1987, Precambrian iron-formations: Canfield, D.E., and Teske, A., 1996, Late Proterozoic rise in atmospheric oxy-
Athens, Theophrastus Publications, 674 p. gen concentration inferred from phylogenetic and sulphur-isotope stud-
Ayer, J., Amelin, Y., Corfu, F., Kamo, S., Ketchum, J., Kwok, K., and Trowell, ies: Nature, v. 382, p. 127–132, doi: 10.1038/382127a0.
N., 2002, Evolution of the southern Abitibi greenstone belt based on Canfield, D.E., and Raiswell, R., 1999, The evolution of the sulfur cycle:
U-Pb geochronology: Autochthonous volcanic construction followed American Journal of Science, v. 299, p. 697–723.
by plutonism, regional deformation and sedimentation: Precambrian Canfield, D.E., Habicht, K.S., and Thamdrup, B., 2000, The Archean sul-
Research, v. 115, p. 63–95, doi: 10.1016/S0301-9268(02)00006-2. fur cycle and the early history of atmospheric oxygen: Science, v. 288,
Bayley, R.W., and James, H.L., 1973, Precambrian iron-formations of the p. 658–661, doi: 10.1126/science.288.5466.658.
United States: Economic Geology and the Bulletin of the Society of Cathles, L.M., 1983, An analysis of the hydrothermal system responsible for
Economic Geologists, v. 68, p. 934–959. massive sulfide deposition in the Hokuroku Basin of Japan, in Ohmoto,
Barley, M.E., 1993, Volcanic, sedimentary and tectonostratigraphic environ- H., and Skinner, B.J., eds., The Kuroko and related volcanic massive
ments of the ~3.46 Ga Warrawoona Megasequence: A review: Precam- sulfide deposits: Economic Geology Monograph 5, p. 439–487.
brian Research, v. 60, p. 47–67, doi: 10.1016/0301-9268(93)90044-3. Clemmey, H., and Badham, N., 1982, Oxygen in the Precambrian atmosphere:
Barley, M.E., and Groves, D.I., 1992, Supercontinent cycles and the distri- An evaluation of the geological evidence: Geology, v. 10, p. 141–146,
bution of metal deposits through time: Geology, v. 20, p. 291–294, doi: doi: 10.1130/0091-7613(1982)10<141:OITPAA>2.0.CO;2.
10.1130/0091-7613(1992)020<0291:SCATDO>2.3.CO;2. Clifford, J.A., Ryan, P., and Kucha, H., 1986, A review of the geological
Barley, M.E., Pickard, A.L., and Sylvester, P.J., 1997, Emplacement of a large setting of the Tynagh Orebody, Co. Galway, in Andrew, C.J., Crowe,
igneous province as a possible cause of banded iron formation 2.45 bil- R.W.A., Finlay, S., Pennell, W.M., and Pyne, J.F., eds., Geology and
lion years ago: Nature, v. 385, p. 55–58, doi: 10.1038/385055a0. genesis of mineral deposits in Ireland: Dublin, Irish Association for
Barrett, T.J., and Jambor, J.L., eds., 1988, Seafloor hydrothermal mineraliza- Economic Geology, p. 419–439.
tion: Canadian Mineralogist, v. 26, p. 429–888. Cloud, P.E., 1968, Atmospheric and hydrospheric evolution on the primitive
Bazilevskaya, E., 2004, Ferric hydroxide formation in silica-rich solutions earth: Science, v. 160, p. 729–736.
[M.S. thesis]: University Park, Pennsylvania State University, 105 p. Cloud, P.E., 1973, Paleoecological significance of the banded iron-formation:
Beeson, R., 1990, Broken Hill-type lead-zinc deposits—An overview of their Economic Geology and the Bulletin of the Society of Economic Geolo-
occurrence and geological setting: Transactions of the Canadian Insti- gists, v. 68, p. 1135–1143.
tute of Mining and Metallurgy, section B, v. 99, p. 163–175. Cloud, P.E., 1978, Cosmos, Earth, and man: A short history of the universe:
Bekker, A., Holland, H.D., Wang, P.-L., Rumble, D., Stein, H.J., Hannah, J.L., New Haven, Connecticut, Yale University Press, 372 p.
Coetzee, L.L., and Beukes, N.J., 2004, Dating the rise of atmospheric Cole, M.J., and Klein, C., 1981, Banded iron-formations through much of Pre-
oxygen: Nature, v. 427, p. 117–120, doi: 10.1038/nature02260. cambrian time: Journal of Geology, v. 89, p. 169–183.
Bengtson, S., 1994, The advent of animal skeleton, in Bengtson, S., ed., Early Collao, S., Alfaro, G., and Hayashi, K., 1990, Banded iron formation and
life on Earth: Nobel Symposium 84: New York, Columbia University massive sulfide orebodies, south-central Chile: Geologic and isotopic
Press, p. 412–425. aspects, in Fontboté, L., Amstuz, G.C., Cardozo, M., Cedillo, E., and
Beukes, N.J., 1973, Precambrian iron-formations of South Africa: Economic Frutos, J., eds., Stratabound ore deposits in the Andes: Society for Geol-
Geology and the Bulletin of the Society of Economic Geologists, v. 68, ogy Applied to Mineral Deposits Special Publication 8, p. 209–219.
p. 960–1004. Condie, K.C., 2001, Mantle plumes and their record in earth history: Cam-
Beukes, N.J., 1983, Paleoenvironmental setting of iron-formations in the dep- bridge, UK, Cambridge University Press, 306 p.
ositional basin of the Transvaal Supergroup, South Africa, in Trendall, Condie, K.C., 2002, Continental growth during a 1.9-Ga superplume event:
A.F., and Morris, R.C., Iron-formation: Facts and problems: Amster- Journal of Geodynamics, v. 34, p. 249–264, doi: 10.1016/S0264-
dam, Elsevier, p. 131–209. 3707(02)00023-6.
Bjerrum, C.J., and Canfield, D.E., 2002, Ocean productivity before about 1.9 Costa, U.R., Barnett, R.L., and Kerrich, R., 1983, The Mattagami Lake Mine
Gyr ago limited by phosphorus adsorption onto iron oxides: Nature, Archean Zn-Cu sulfide deposit, Quebec: Hydrothermal coprecipitation
v. 417, p. 159–162, doi: 10.1038/417159a. of talc and sulfides in a sea-floor brine pool—Evidence from geochem-
istry, 18O/16O, and mineral chemistry: Economic Geology and the Bul- Farquhar, J., Bao, H., and Thiemens, M., 2000, Atmospheric influence of
letin of the Society of Economic Geologists, v. 78, p. 1144–1203. Earth’s earliest sulfur cycle: Science, v. 289, p. 756–758, doi: 10.1126/
Davidson, G.J., 1992, Hydrothermal geochemistry and ore genesis of sea- science.289.5480.756.
floor volcanogenic copper-bearing oxide ores: Economic Geology and Farquhar, J., Savarino, J., Airieau, S., and Thiemens, M.H., 2001, Observa-
the Bulletin of the Society of Economic Geologists, v. 87, p. 889–912. tion of wavelength-sensitive mass-independent sulfur isotope effects
Davidson, G.J., Stolz, A.J., and Eggins, S.M., 2001, Geochemical anatomy during SO2 photolysis; implication for the early atmosphere: Journal
of silica iron exhalites: Evidence for hydrothermal oxyanion cycling in of Geophysical Research, ser. E, Planets, v. 106, p. 32829–32839, doi:
response to vent fluid redox and thermal evolution (Mt. Windsor Sub- 10.1029/2000JE001437.
province, Australia): Economic Geology and the Bulletin of the Society Farrell, L., 1990, Ancient banded iron formations: Athens, Theophrastus
of Economic Geologists, v. 96, p. 1201–1226. Publications, 462 p.
Davis, D.W., and Lin, S., 2003, Unraveling the geologic history of the Hemlo Fernández, A., and Moro, M.C., 1998, Origin and depositional environment
Archean gold deposits, Superior province, Canada: A U-Pb geochro- of Ordovician stratiform iron mineralization from Zamora (NW Ibe-
nological study: Economic Geology and the Bulletin of the Society of rian Peninsula): Mineralium Deposita, v. 33, p. 606–619, doi: 10.1007/
Economic Geologists, v. 98, p. 51–67. s001260050176.
Davy, R., 1985, The mineralogy and composition of a core which intersects French, B.M., 1973, Mineral assemblages in diagenetic and low-grade met-
the Marra Mamba iron formation and the Roy Hill Shale Member: Geo- amorphic iron-formation: Economic Geology and the Bulletin of the
logical Survey of Western Australia Record, v. 6, 60 p. Society of Economic Geologists, v. 68, p. 1063–1074.
Degens, E.T., and Ross, D.A., eds., 1974, The Black Sea—Geology, chemis- Fry, B., Jannasch, H.W., Molyneaux, S.J., Wirsen, C.O., Muramoto, J.A., and
try, and biology: Tulsa, Oklahoma, American Association of Petroleum King, S., 1991, Stable isotope studies of the carbon, nitrogen and sulfur
Geologists, Memoir 20, 633 p. cycles in the Black Sea and the Cariaco Trench, in Murray, J.W., ed.,
Derry, L.A., and Jacobsen, S.B., 1990, The chemical evolution of Precam- Black Sea oceanography: Results from the 1988 Black Sea expedition:
brian seawater: Evidence from REEs in banded iron formations: Geo- Deep-Sea Research, part A, Oceanographic Research Papers, v. 38,
chimica et Cosmochimica Acta, v. 54, p. 2965–2977, doi: 10.1016/0016- Suppl. 2A, p. S1003–S1019.
7037(90)90114-Z. Fyon, J.A., Breaks, F.W., Heather, K.B., Jackson, S.L., Muir, T.L., Stott, G.M.,
DeWitt, E., 1979, New data concerning Proterozoic volcanic stratigraphy and Thurston, P.C., 1992, Metallogeny of metallic mineral deposits in
and structure in central Arizona and its importance in massive sulfide the Superior Province of Ontario, in Thurston, P.C., Williams, H.R.,
exploration: Economic Geology and the Bulletin of the Society of Eco- Sutcliffe, R.H., and Stott, G.M., eds., Geology of Ontario: Ontario Geo-
nomic Geologists, v. 74, p. 1371–1382. logical Survey Special Volume 4, part 2, p. 1091–1174.
Dimroth, E., 1976, Aspects of the sedimentary petrology of cherty iron- Gauthier, M., Brown, A.C., and Morin, G., 1987, Small iron-formations as a
formation, in Wolf, K.H. ed., Au, U, Fe, Mn, Hg, Sb, W and P deposits: guide to base- and precious-metal deposits in the Grenville Province of
Amsterdam, Elsevier, Handbook of Strata-bound and Stratiform Ore southern Quebec, in Appel, P.W.U., and LaBerge, G.L., eds., Precam-
Deposits, v. 7, p. 203–254. brian iron-formations: Athens, Theophrastus Publications, p. 297–327.
Dimroth, E., and Kimberley, M.M., 1976, Precambrian atmospheric oxygen: Gnaneshwar Rao, T., and Naqvi, S.M., 1995, Geochemistry, depositional
Evidence in the sedimentary distributions of carbon, sulfur, uranium, environment and tectonic setting of the BIF’s of the Late Archaean Chi-
and iron: Canadian Journal of Earth Sciences, v. 13, p. 1161–1185. tradurga Schist Belt, India: Chemical Geology, v. 121, p. 217–243, doi:
Dorr, J., II, and Van, N., 1973, Iron-formation in South Africa: Economic 10.1016/0009-2541(94)00116-P.
Geology and the Bulletin of the Society of Economic Geologists, v. 68, Goodwin, A.M., 1973, Archean iron-formations and tectonic basins of the
p. 1005–1022. Canadian shield: Economic Geology and the Bulletin of the Society of
Drever, J.I., 1974, Geochemical model for the origin of Precambrian Banded Economic Geologists, v. 68, p. 915–933.
Iron formations: Geological Society of America Bulletin, v. 85, p. 1099– Green, J.C., 1992, Proterozoic rifts, in Condie, K.C., ed., Proterozoic crustal
1106, doi: 10.1130/0016-7606(1974)85<1099:GMFTOO>2.0.CO;2. evolution: Amsterdam, Elsevier, p. 97–149.
Drever, J.I., 1982, The geochemistry of natural waters: Englewood Cliffs, Grenne, T., and Vokes, F.M., 1990, Sea-floor sulfides at the Høydal volcano-
New Jersey, Prentice Hall, 388 p. genic deposit, central Norwegian Caledonides: Economic Geology and
Duhig, N.C., Stolz, J., Davidson, G.J., and Large, R.R., 1992, Cambrian the Bulletin of the Society of Economic Geologists, v. 85, p. 344–359.
microbial and silica gel textures in silica iron exhalites from the mount Gross, G.A., 1965, Geology of iron deposits in Canada: General geology and
Windsor Volcanic Belt, Australia: Their petrography, chemistry, and evaluation of iron deposits: Geological Survey of Canada Economic
origin: Economic Geology and the Bulletin of the Society of Economic Geology Report 22, v. 1, 181 p.
Geologists, v. 87, p. 764–784. Gross, G.A., 1990, Geochemistry of iron-formation in Canada, in Chauvel,
Eigenbrode, J.L., Freeman, K.H., Love, G.D., Snape, C.E., and Summons, J.J., Cheng Yuqi, El Shazly, E.M., Gross, G.A., Laajoki, K., Markov,
R.E., 2004, Hydropyrolytic release of hopane and sterane biomarkers M.S., Rai, K.L., Stulchikov, V.A., and Augustithis, S.S., eds., Ancient
from 2.7 Ga kerogens: Program and Abstracts, 13th Annual V.M. Gold- banded iron formations (regional presentations): Athens, Theophrastus
schmidt Conference, Hot Springs, Virginia, p. A240. Publications, p. 3–26.
El Shazly, E.M., 1990, Red Sea deposits, in Chauvel, J.J., Cheng Yuqi, El Gross, G.A., 1991, Genetic concepts for iron-formation and associated metal-
Shazly, E.M., Gross, G.A., Laajoki, K., Markov, M.S., Rai, K.L., Stul- liferous sediments, in Hutchinson, R.W., and Grauch, R.I., eds., His-
chikov, V.A., and Augustithis, S.S., eds., Ancient banded iron forma- torical perspectives of genetic concepts and case histories of famous
tions (regional presentations): Athens, Theophrastus Publications, p. discoveries: Economic Geology Monograph 8, p. 51–81.
157–222. Habicht, K.S., Gade, M., Thamdrup, B., Berg, P., and Canfield, D.E., 2002,
Eremeev, V.N., ed., 1992, Complex oceanographic research of the Black Sea: Calibration of sulfate levels in the Archean ocean: Science, v. 298,
Utrecht, VSP, 130 p. p. 2372–2374, doi: 10.1126/science.1078265.
Eriksson, P.G., Condie, K.C., van der Westhuizen, van der Merwe, R., de Hagemann, S.G., Barley, M.E., Folkert, S.L., Yardley, W.D., and Banks, D.A.,
Bruiyn, H., Nelson, D.R., Altermann, W., Catuneanu, O., Bumby, A.J., 1999, A hydrothermal origin for the giant BIF-hosted Tom Price iron
Lindsay, J., and Cunningham, M.J., 2002, Late Archaean superplume ore deposit, in Stanley, C.J., et al., Mineral deposits: Processes to pro-
events: A Kaapvaal-Pilbara perspective: Journal of Geodynamics, v. 34, cessing: Proceedings of the fifth biennial SGA meeting and the tenth
p. 207–247. quadrennial IAGOD symposium, London: Rotterdam, A.A. Balkema,
Eriksson, P.G., Altermann, W., Nelson, D.R., Mueller, W.U., and Catuneanu, p. 41–44.
O., eds., 2004, The Precambrian Earth: Tempos and events: Amster- Hamade, T., Konhauser, K.O., Raiswell, R., Goldsmith, S., and Morris,
dam, Elsevier, Developments in Precambrian Geology, v. 12, 941 p. R.C., 2003, Using Ge/Si ratios to decouple iron and silica fluxes in
Eugster, H.P., and I-Ming Chou, 1973, The depositional environments of Precambrian banded iron formations: Geology, v. 31, p. 35–38, doi:
Precambrian banded iron-formations: Economic Geology, v. 68, p. 10.1130/0091-7613(2003)031<0035:UGSRTD>2.0.CO;2.
1144–1168. Hartlaub, R.P., Heaman, L.M., Ashton, K.E., and Chako, T., 2004, The
Ewers, W.E., and Morris, R.C., 1981, Studies of the Dales Gorge Member of the Archean Murmac Bay Group: Evidence for a giant Archean rift in the
Brockman Iron Formation, Western Australia: Economic Geology and the Rae Province Canada: Precambrian Research, v. 131, p. 345–372, doi:
Bulletin of the Society of Economic Geologists, v. 76, p. 1929–1953. 10.1016/j.precamres.2004.01.001.
Haruna, M., Hanamuro, T., Uyeda, K., Fujimaki, H., and Ohmoto, H., 2003, Kakegawa, T., Kawai, H., and Ohmoto, H., 1998, Origins of pyrites in the 2.5
Chemical, isotopic, and fluid inclusion evidence for the hydrothermal alter- Ga Mt. McRae shale, the Hamersley district, western Australia: Geochi-
ation of the footwall rocks of the BIF-hosted iron ore deposits in the Hamer- mica et Cosmochimica Acta, v. 62, p. 3205–3220, doi: 10.1016/S0016-
sley district, western Australia: Resource Geology, v. 53, p. 75–88. 7037(98)00229-4.
Haydon, R.C., and McConachy, G.W., 1987, The stratigraphic setting of Pb-Zn- Kalogeropoulos, S.I., and Scott, S.D., 1983, Mineralogy and geochemistry
Ag mineralization at Broken Hill: Economic Geology and the Bulletin of of tuffaceous exhalites (tetsusekiei) of the Fukazawa Mine, Hokuroku
the Society of Economic Geologists, v. 82, p. 826–856. district, Japan, in Ohmoto, H., and Skinner, B.J., eds., The Kuroko and
Hayes, J.M., 1994, Global methanotrophy at the Archean-Proterozoic transi- related volcanic massive sulfide deposits: Economic Geology Mono-
tion, in Bengtson, S., ed., Early life on earth: Nobel symposium no. 84: graph, v. 5, p. 412–432.
New York, Columbia University Press, p. 220–236. Kalugin, A.S., 1973, Geology and genesis of the Devonian banded iron-forma-
Hayes, J.M., Kaplan, I.R., and Wedeking, K.W., 1983, Precambrian organic tion in Altai, western Siberia and eastern Kazakhstan, in Genesis of
geochemistry: Preservation of the record, in Schopf, J.W., ed., Earth’s ear- Precambrian iron and manganese deposits, Proceedings, Kiev Sympo-
liest biosphere: Its origin and evolution: Princeton, New Jersey, Princeton sium, 1970: Paris, UNESCO, Earth Science (Sciences de la Terre), v. 9,
University Press, p. 93–134. p. 159–165.
Helvaci, C., 1984, Apatite-rich iron deposits of the Avnik (Bingöl) region, Kanehira, K., and Tatsumi, T., 1970, Bedded cupriferous iron sulphide deposits
southeastern Turkey: Economic Geology and the Bulletin of the Society in Japan, a review, in Tatsumi, T., ed., Volcanism and ore genesis: Tokyo,
of Economic Geologists, v. 79, p. 354–371. University of Tokyo Press, p. 51–76.
Hoffmann, D., 1994, Geochemistry and genesis of manganiferous silicate-rich Kaplan, I.R., Sweeney, R.E., and Nissenbaum, A., 1969, Sulfur isotope studies
iron formation bands in the Broken Hill deposit, Aggeneys, South Africa: on Red Sea geothermal brines and sediments, in Degens, E.T., and Ross,
Exploration and Mining Geology, v. 3, p. 407–417. D.A., eds., Hot brines and recent heavy metal deposits in the Red Sea:
Hofmann, H.J., Sage, R.P., and Berdusco, E.N., 1991, Archean stromatolites in A geochemical and geophysical account: New York, Springer-Verlag,
Michipicoten Group siderite ore at Wawa, Ontario: Economic Geology p. 474–498.
and the Bulletin of the Society of Economic Geologists, v. 86, p. 1023– Karhu, J., and Epstein, S., 1986, The implication of the oxygen isotope records
1030. in coexisting cherts and phosphates: Geochimica et Cosmochimica Acta,
Holland, H.D., 1973, The oceans: A possible source of iron in iron-formations: v. 50, p. 1745–1756, doi: 10.1016/0016-7037(86)90136-5.
Economic Geology and the Bulletin of the Society of Economic Geolo- Kasting, J.F., 1987, Theoretical constraints on oxygen and carbon dioxide con-
gists, v. 68, p. 1169–1172. centrations in the Precambrian atmosphere: Precambrian Research, v. 34,
Holland, H.D., 1984, The chemical evolution of the atmosphere and oceans: p. 205–229, doi: 10.1016/0301-9268(87)90001-5.
Princeton, New Jersey, Princeton University Press, 582 p. Kasting, J.F., 2001, Earth history: The rise of atmospheric oxygen: Science,
Holland, H.D., 1994, Early Proterozoic atmospheric change, in Bengtson, S., v. 293, p. 819–820, doi: 10.1126/science.1063811.
ed., Early life on Earth: Nobel Symposium 84: New York, Columbia Uni- Kasting, J.F., and Siefert, J.L., 2002, Life and the evolution of Earth’s atmo-
versity Press, p. 237–244. sphere: Science, v. 296, p. 1066–1068, doi: 10.1126/science.1071184.
Holland, H.D., 1999, When did the Earth’s atmosphere become oxic?: Reply: Kimberley, M.M., 1989a, Nomenclature for iron formations: Ore Geology
The Geochemical News, v. 100, p. 20–23. Reviews, v. 5, p. 1–12, doi: 10.1016/0169-1368(89)90002-4.
Holland, H.D., 2002, Volcanic gases, black smokers, and the great oxidation Kimberley, M.M., 1989b, Exhalative origins of iron formations: Ore Geology
event: Geochimica et Cosmochimica Acta, v. 66, p. 3811–3826, doi: Reviews, v. 5, p. 13–145, doi: 10.1016/0169-1368(89)90003-6.
10.1016/S0016-7037(02)00950-X. Klein, C., 1983, Diagenesis and metamorphism of Precambrian banded iron-
Holland, H.D., and Petersen, U., 1995, Living dangerously: Princeton, New formations, in Trendall, A.F., and Morris, R.C., Iron-formation: Facts and
Jersey, Princeton University Press, 490 p. problems: Amsterdam, Elsevier, p. 417–460.
Hollings, P., and Kerrich, R., 2000, An Archean arc basalt-Nb-enriched basalt- Klein, C., and Beukes, N.J., 1989, Geochemistry and sedimentology of a
adakite association: The 2.7 Ga Confederation assemblage of the Birch- facies transition from limestone to iron-formation deposition in the
Uchi greenstone belt, Superior Province: Contributions to Mineralogy early Proterozoic Transvaal Supergroup, South Africa: Economic
and Petrology, v. 139, p. 208–226. Geology and the Bulletin of the Society of Economic Geologists, v. 84,
Imai, H., 1978, Geology and genesis of the Okuki mine, Ehime Prefecture, and p. 1733–1774.
other related cupriferous pyrite deposits in southwest Japan, in Imai, H. Klein, C., and Beukes, N.J., 1992, Proterozoic iron formations, in Condie, K.C.,
ed., Geological studies of the mineral deposits in Japan and East Asia: ed., Proterozoic crustal evolution: Amsterdam, Elsevier, p. 383–418.
Tokyo, University of Tokyo Press, p. 233–256. Klein, C., and Beukes, N.J., 1993, Sedimentology and geochemistry of the gla-
Isley, A.E., 1995, Hydrothermal plumes and the delivery of iron to banded iron ciogenic Late Proterozoic Rapitan Iron-Formation in Canada: Economic
formation: Journal of Geology, v. 103, p. 169–185. Geology and the Bulletin of the Society of Economic Geologists, v. 88,
Isley, A.E., and Abbott, D.H., 1999, Plume-related mafic volcanism and p. 542–565.
the deposition of banded iron formation: Journal of Geophysical Klein, C., and Fink, R.P., 1976, Petrology of the Sokoman Iron Formation in
Research, ser. B, Solid Earth and Planets, v. 104, p. 15,461–15,477, doi: the Howells River area, at the western edge of the Labrador Trough: Eco-
10.1029/1999JB900066. nomic Geology and the Bulletin of the Society of Economic Geologists,
Jackson, S.I., and Fyon, J.A., 1991, The western Abitibi subprovince in Ontario, v. 71, p. 453–487.
in Thurston, P.C., Williams, H.R., Sutcliffe, R.H., and Stott, G.M., eds., Knauth, L.P., and Lowe, D.R., 2003, High Archean climatic temperature
Geology of Ontario: Toronto, Ontario Geological Survey Special Volume inferred from oxygen isotope geochemistry of cherts in the 3.5 Ga
4, p. 405–482. Swaziland Supergroup, South Africa: Geological Society of America
Jacobsen, S.B., and Pimentel-Klose, M.R., 1988, A Nd isotopic study of the Bulletin, v. 115, p. 566–580, doi: 10.1130/0016-7606(2003)115<0566:
Hamersley and Michipicoten banded iron formations: The source of REE HACTIF>2.0.CO;2.
and Fe in Archean oceans: Earth and Planetary Science Letters, v. 87, Konhauser, K.O., Hamade, T., Raiswell, R., Morris, R.C., Ferris, F.G.,
p. 29–44, doi: 10.1016/0012-821X(88)90062-3. Southam, G., and Canfield, D.E., 2002, Could bacteria have formed the
Jahnke, L., and Klein, H.P., 1983, Oxygen requirements for formation and Precambrian banded iron formations?: Geology, v. 30, p. 1079–1082, doi:
activity of the squalene epoxidase in Saccharomyces cerevisiae: Journal 10.1130/0091-7613(2002)030<1079:CBHFTP>2.0.CO;2.
of Bacteriology, v. 155, p. 488–492. Koopman, E.R., 1993, Stratigraphy, structural geology, and structural controls
James, H.L., 1954, Sedimentary facies of iron-formation: Economic Geology and of ore distribution of the Lyon Lake massive sulphide deposit, Sturgeon
the Bulletin of the Society of Economic Geologists, v. 49, p. 235–293. Lake, Ontario [M.Sc. thesis]: Ottawa, Carleton University, 170 p.
James, H.L., 1983, Distribution of banded iron-formation in space and time, Krapez, B., Barley, M.E., and Pickard, A.L., 2003, Hydrothermal and resedi-
in Trendall, A.F., and Morris, R.C., Iron-formation: Facts and problems: mented origins of the precursor sediments to banded iron formation:
Amsterdam, Elsevier, p. 471–490. Sedimentological evidence from the early Palaeoproterozoic Brockman
Jørgensen, B.B., Böttcher, M.E., Lüschen, H., Neretin, L.N., and Volkov, I.I., Supersequence of western Australia: Sedimentology, v. 50, p. 979–1011,
2004, Anaerobic methane oxidation and a deep H2S sink generate isotopi- doi: 10.1046/j.1365-3091.2003.00594.x.
cally heavy sulfides in Black Sea sediments: Geochimica et Cosmochi- LaBerge, G.L., Robbins, E.I., and Tsu-Ming Han, 1987, A model for the bio-
mica Acta, v. 68, p. 2095–2118, doi: 10.1016/j.gca.2003.07.017. logical precipitation of Precambrian iron-formations – A: Geological
evidence, in Appel, P.W.U., and LaBerge, G.L., eds., Precambrian iron- Ohmoto, H., and Goldhaber, M.B., 1997, Sulfur and carbon isotopes, in Barnes,
formations: Athens, Theophrastus Publications, p. 69–96. H.D., ed., Geochemistry of hydrothermal ore deposits (3rd edition): New
Langmuir, D., 1997, Aqueous environmental geochemistry: Upper Saddle York, John Wiley and Sons, p. 517–612.
River, New Jersey, Prentice Hall, 600 p. Ohmoto, H., Kajiwara, Y., and Date, J., 1970, The Kuroko ores in Japan: Prod-
Lasaga, A.C., and Ohmoto, H., 2002, The oxygen geochemical cycle: Dynam- ucts of seawater? Geological Society of America Abstracts with Pro-
ics and stability: Geochimica et Cosmochimica Acta, v. 66, p. 361–381, grams, v. 1, p. 640.
doi: 10.1016/S0016-7037(01)00685-8. Ohmoto, H., Mizukami, M., Drummond, S.E., Eldridge, C.S., Pisutha-Arnold,
Liaghat, S., and MacLean, W.H., 1992, The Key tuffite, Matagami mining dis- V., and Lenagh, T.C., 1983, Chemical processes of Kuroko formation,
trict: Origin of the tuff components and mass changes: Exploration and in Ohmoto, H., and Skinner, B.J., eds., The Kuroko and related vol-
Mining Geology, v. 1, p. 197–207. canogenic massive sulfide deposits: Economic Geology Monograph 5,
Lin, S., 2001, Stratigraphic and structural setting of the Hemlo gold deposits, p. 570–604.
Ontario, Canada: Economic Geology and the Bulletin of the Society of Ohmoto, H., Kaiser, C.J., and Geer, K.A., 1990, Systematics of sulphur iso-
Economic Geologists, v. 96, p. 477–507. topes in recent marine sediments and ancient sediment-hosted base metal
Lottermoser, B.G., and Ashley, P.M., 1996, Geochemistry and exploration deposits, in Herbert, H.K., and Ho, S.E., eds., Proceedings of the confer-
significance of ironstones and barite-rich rocks in the Proterozoic Wil- ence on stable isotopes and fluid processes in mineralization: Perth, Uni-
lyama Supergroup, Olary Block, South Australia: Journal of Geochemical versity of Western Australia Special Publication 23, p. 70–120.
Exploration, v. 57, p. 57–73, doi: 10.1016/S0375-6742(96)00016-7. Ohmoto, H., Watanabe, Y., and Kumazawa, K., 2004, Evidence from massive
Matsuhisa, Y., Goldsmith, J.R., and Clayton, R.N., 1978, Mechanisms of siderite beds for a CO2-rich atmosphere before approximately 1.8 billion
hydrothermal crystallization of quartz at 250 °C and 15 kbar: Geochi- years ago: Nature, v. 429, p. 395–399, doi: 10.1038/nature02573.
mica et Cosmochimica Acta, v. 42, p. 173–182, doi: 10.1016/0016- Ono, S., Eigenbrode, J.L., Pavlov, A.A., Kharecha, P., Rumble, D., Kasting,
7037(78)90130-8. J.F., and Freeman, K.H., 2003, New insights into Archean sulfur cycle
Mel’nik, Y.P., 1982, Precambrian banded iron-formations: Physicochemical from mass-independent sulfur isotope records from the Hamersley Basin,
conditions of formation: Amsterdam, Elsevier, Developments in Precam- Australia: Earth and Planetary Science Letters, v. 213, p. 15–30, doi:
brian Geology, v. 5, 310 p. 10.1016/S0012-821X(03)00295-4.
Morey, G.B., and Southwick, D.L., 1995, Allostratigraphic relationships of Ootes, L., and Lentz, D.R., 2002, Occurrence of bleached mafic flows and their
early Proterozoic iron-formations in the Lake Superior region: Economic association with stockwork sulphides and banded iron-formation in the
Geology and the Bulletin of the Society of Economic Geologists, v. 90, Crestaurum Formation of the later Archean Yellowknife greenstone belt,
p. 1983–1993. Northwest Territories: Ottawa, Geological Survey of Canada, Current
Morris, R.C., 1985, Genesis of iron ore in banded iron-formation by super- Research Report 2002-E5, p. 1–12.
gene and supergene-metamorphic processes: A conceptual model, in O’Rourke, J.E., 1962, The stratigraphy of Himalayan iron ores: American Jour-
Wolf, K.H., ed., Regional studies and specific deposits: Amsterdam, nal of Science, v. 260, p. 294–302.
Elsevier, Handbook of strata-bound and stratiform ore deposits, v. 13, Otake, T., Wesolowski, D.J., Anovitz, L.M., Hayashi, K., and Ohmoto, H.,
p. 73–235. 2005, Magnetite transformation to hematite under high H2 pressure at 150
Morris, R.C., 1993, Genetic modeling for banded iron-formation of the Ham- °C: Geochimica et Cosmochimica Acta, v. 69, S10, A424.
ersley Group, Pilbara Craton, Western Australia: Precambrian Research, Owen, T., Cess, R.D., and Ramanathan, V., 1979, Enhanced CO2 greenhouse to
v. 60, p. 243–286, doi: 10.1016/0301-9268(93)90051-3. compensate for reduced solar luminosity on early Earth: Nature, v. 277,
Morton, R.L., and Nebel, M.L., 1984, Hydrothermal alteration of felsic vol- p. 640–642, doi: 10.1038/277640a0.
canic rocks at the Helen siderite deposit, Wawa, Ontario: Economic Page, R.W., and Laing, W.P., 1992, Felsic metavolcanic rocks related to the
Geology and the Bulletin of the Society of Economic Geologists, v. 79, Broken Hill Pb-Zn-Ag orebody, Australia: Geology, depositional age, and
p. 1319–1333. timing of high-grade metamorphism: Economic Geology and the Bulletin
Muramoto, J.A., Honjo, S., Fry, B., Hay, B.J., Howarth, R.W., and Cisne, J.L., of the Society of Economic Geologists, v. 87, p. 2138–2168.
1991, Sulfur, iron and organic carbon fluxes in the Black Sea: Sulfur Parák, T., 1991, Volcanic sedimentary rock-related metallogenesis in the Kiruna-
isotopic evidence for origin of sulfur fluxes, in Murray, J.W., ed., Black Skellefte belt of northern Sweden, in Hutchinson, R.W., and Grauch, R.I.,
Sea oceanography: Results from the 1988 Black Sea expedition: Deep- eds., Historical perspectives of genetic concepts and case histories of
Sea Research, part A, Oceanographic Research Papers, v. 38, Suppl. 2A, famous discoveries: Economic Geology Monograph 8, p. 20–50.
p. S1151–S1187. Pavlov, A.A., Kasting, J.F., Eigenbrode, J.L., and Freeman, K.H., 2001a,
Murray, J.W., ed., 1991, Black Sea oceanography, Results from the 1988 Black Organic haze in Earth’s early atmosphere: Source of low-13C Late
Sea expedition: Deep-Sea Research, part A, Oceanographic Research Archean kerogens?: Geology, v. 29, p. 1003–1006, doi: 10.1130/0091-
Papers, v. 38, Suppl. 2A, p. S655–S1266. 7613(2001)029<1003:OHIESE>2.0.CO;2.
Ohmoto, H., 1992, Biogeochemistry of sulfur and the mechanisms of sulfide- Pavlov, A.A., Brown, L.L., and Kasting, J.F., 2001b, UV shielding of
sulfate mineralization in Archean oceans, in Schidlowski, M., Golubic, NH3 and O2 by organic hazes in the Archean atmosphere: Journal
S., Kimberley, M.M., McKirdy, D.M., and Trudinger, P.A., eds., Early of Geophysical Research, E, Planets, v. 106, p. 23267–23287, doi:
organic evolution: Implications for mineral and energy resources: Berlin, 10.1029/2000JE001448.
Springer-Verlag, p. 378–397. Perry, E.C., Jr., Tan, F.C., and Morey, G.B., 1973, Geology and stable isotope
Ohmoto, H., 1993, The banded iron formations: Products of the oxygen-rich geochemistry of the Biwabik Iron Formation, Northern Minnesota: Eco-
Archean atmosphere?: Geological Society of America Abstracts with Pro- nomic Geology and the Bulletin of the Society of Economic Geologists,
grams, v. 25, no. 6, p. A-89. v. 68, p. 1110–1125.
Ohmoto, H., 1996, Formation of volcanogenic massive sulfide deposits: The Peter, J.M., 2001, Ancient iron-rich metalliferous sediments (iron forma-
Kuroko perspective: Ore Geology Reviews, v. 10, p. 135–177, doi: tions): Their genesis and use in the exploration for stratiform base metal
10.1016/0169-1368(95)00021-6. sulphide deposits, with examples from the Bathurst Mining Camp,
Ohmoto, H., 1997, When did the Earth’s atmosphere become oxic?: Geochemi- in Lentz, D. R., ed., Geochemistry of sediments and sedimentary
cal News, v. 93, p. 12–13 and p. 26–28. rocks: Secular evolutionary considerations to mineral deposit-forming
Ohmoto, H., 2003, Non-redox transformation of magnetite-hematite in hydro- environments: St. John’s, Newfoundland, Geological Association of
thermal systems: Economic Geology and the Bulletin of the Society of Canada, p. 1–38.
Economic Geologists, v. 98, p. 157–161. Peter, J.M., and Goodfellow, W.D., 1996, Mineralogy, bulk and rare earth ele-
Ohmoto, H., 2004, Archean atmosphere, hydrosphere, and biosphere, in P. Erick- ment geochemistry of massive sulphide-associated hydrothermal sedi-
son et al., eds. The Precambrian Earth: Tempos and events: Amsterdam, ments of the Brunswick Horizon, Bathurst Mining Camp, New Bruns-
Elsevier, Developments in Precambrian Geology, v. 12, p. 361–368. wick: Canadian Journal of Earth Sciences, v. 33, p. 252–283.
Ohmoto, H., and Skinner, B.J., eds., 1983, The Kuroko and related volcano- Peter, J.M., and Scott, S.D., 1998, Windy Craggy, northwestern British Colum-
genic massive sulfide deposits: Economic Geology Monograph 5, 604p. bia: The world’s largest Besshi-type deposits, in Barrie, C.T., and Han-
Ohmoto, H., and Felder, R.P., 1987, Bacterial activity in the warmer, sul- nington, M.D., eds., Volcanic-associated massive sulfide deposits:
phate-bearing, Archaean oceans: Nature, v. 328, p. 244–246, doi: Processes and examples in modern and ancient settings: Reviews in
10.1038/328244a0. Economic Geology, v. 8, p. 261–295.
Phillips, G. N., Law, J. D. M., and Myers, R. E., 2001, Is the redox state of Scott, S.D., 1997, Submarine hydrothermal systems and deposits, in Barnes,
the Archean atmosphere constrained?: Society of Economic Geologists H.L., ed., Geochemistry of hydrothermal ore deposits (3rd edition): New
Newsletter, v. 47, 1, 9–18. York, John Wiley and Sons, p. 797–875.
Pickard, A.L., 2003, SHRIMP U-Pb zircon ages for the Palaeoproterozoic Shanks, W.C., III, and Bischoff, J.L., 1980, Geochemistry, sulfur isotope com-
Kuruman iron formation, northern Cape Province, South Africa: Evi- position, and accumulation rates of Red Sea geothermal deposits: Eco-
dence for simultaneous BIF deposition on Kaapvaal and Pilbara cra- nomic Geology and the Bulletin of the Society of Economic Geologists,
tons: Precambrian Research, v. 125, p. 275–315, doi: 10.1016/S0301- v. 75, p. 445–459.
9268(03)00113-X. Shegelski, R.J., 1987, The depositional environment of Archean iron forma-
Piercey, S.J., Paradis, S., Murphy, D.C., and Mortensen, J.K., 2001, Geochem- tions, Sturgeon-Savant greenstone belt, Ontario, Canada, in Appel,
istry and paleotectonic setting of felsic volcanic rocks in the Finlayson P.W.U., and LaBerge, G.L., eds., Precambrian iron-formations: Athens,
Lake volcanic-hosted massive sulfide district, Yukon, Canada: Economic Theophrastus Publications, p. 329–344.
Geology, v. 96, p. 1877–1905. Simonson, B.M., Schubel, K.A., and Hassler, S.W., 1993, Carbonate sedi-
Polat, A., and Kerrich, R., 2001, Geodynamic processes, continental growth, mentology of the early Precambrian Hamersley Group of western Aus-
and mantle evolution in late Archean greenstone belts of the southern tralia: Precambrian Research, v. 60, p. 287–335, doi: 10.1016/0301-
Superior Province, Canada: Precambrian Research, v. 112, p. 5–25, doi: 9268(93)90052-4.
10.1016/S0301-9268(01)00168-1. Sleep, N.H., 1978, Thermal structure and kinematics of mid-ocean ridge axis,
Potter, R.W., and Brown, D.L., 1977, The volumetric properties of aqueous some implications to basaltic volcanism: Geophysical Research Letters,
sodium chloride solutions from 0° to 500°C at pressures up to 2000 bars v. 5, p. 426–428.
based on a regression of available data in the literature: U.S. Geological Soriano, C., and Marti, J., 1999, Facies analysis of volcano-sedimentary suc-
Survey Bulletin, Report B 1421-C, 36 p. cessions hosting massive sulfide deposits in the Iberian pyrite belt, Spain:
Pottorf, R.J., and Barnes, H.L., 1983, Mineralogy, geochemistry, and ore gen- Economic Geology and the Bulletin of the Society of Economic Geolo-
esis of hydrothermal sediments from the Atlantis II Deep, Red Sea, in gists, v. 94, p. 867–882.
Ohmoto, H., and Skinner, B.J., eds., The Kuroko and related volcanogenic Spry, P.G., Peter, J.M., and Slack, J.F., 2000, Meta-exhalites as exploration
massive sulfide deposits: Economic Geology Monograph 5, p. 198–223. guides to ore, in Spry, P.G., Marshall, B., and Vokes, E.M., eds., Meta-
Poulton, S.W., Fralick, P.W., and Canfield, D.E., 2004, The transition to a morphosed and metamorphogenic ore deposits: Reviews in Economic
sulphidic ocean approximately 1.84 billion years ago: Nature, v. 431, Geology, v. 11, p. 163–201.
p. 173–177, doi: 10.1038/nature02912. Tachibana, S., Hirai, T., Goto, K., Yamamoto, S., Isozoki, Y., Tada, R., Tajika,
Quade, H., 1976, Genetic problems and environmental features of volcano- E., Shimoda, G., Marishita, Y., and Kita, N.T., 2004, Sulfur isotopic
sedimentary iron-ore deposits of the Lahn-Dill type, in Wolf, K.H., ed., compositions of sulfides from the lower Huronian Supergroup, Ontario,
Geochemical studies: Amsterdam Elsevier, Handbook of strata-bound Canada [abs.]: Eos (Transactions, American Geophysical Union), v. 85,
and stratiform ore deposits, v. 2, p. 255–294. p. F1897.
Rimstidt, J.D., 1997, Gangue mineral transport and deposition, in Barnes, H.L., Thurston, P.C., 1991, Archean geology of Ontario: Introduction, in Thurston,
ed., Geochemistry of hydrothermal ore deposits: New York, John Wiley P.C., Williams, H.R., Sutcliffe, R.H., and Stott, G.M., eds., Geology of
and Sons, p. 487–516. Ontario: Ontario Geological Survey Special Volume 4, p. 73–78.
Rosing, M.T., 1999, 13C-depleted carbon microparticles in >3700-Ma seafloor Thurston, P.C., 2002, Autochthonous development of Superior Province
sedimentary rocks from West Greenland: Science, v. 283, p. 674–676, greenstone belts?: Precambrian Research, v. 115, p. 11–36, doi: 10.1016/
doi: 10.1126/science.283.5402.674. S0301-9268(02)00004-9.
Rozendaal, A., 1986, The Gamsberg zinc deposit, Namaqualand district, in Torres-Ruiz, J., 1983, Genesis and evolution of the Marquesado and adjacent
Anhaeusser, C.R., and Maske, S., eds., Mineral deposits of South Africa: iron ore deposits, Granada, Spain: Economic Geology and the Bulletin of
Johannesburg, Geological Society of South Africa, v. 2, p. 1477–1488. the Society of Economic Geologists, v. 78, p. 1657–1673.
Rucker, E.A., Dudley, M.A., and Nold, J.L., 2003, Ore textural studies and Trendall, A.F., and Blockley, J.G., 1970, The iron formations of the Precam-
origin of the Pilot Knob hematite deposit, Iron County, MO: Geological brian Hamersley group Western Australia: Geological Survey of Western
Society of America, North-Central Section, 37th Annual Meeting, War- Australia Bulletin, v. 119, 366 p.
rensburg, Montana, Abstracts with Programs, v. 35, p. 20. Trendall, A.F., and Morris, R.C., eds., 1983, Iron-formation: Facts and prob-
Russell, M.J., 1975, Lithogeochemical environment of the Tynagh base-metal lems: Amsterdam, Elsevier, 558 p.
deposit, Ireland, and its bearing on ore deposition: Transactions of the Tsikos, H., Beukes, N.J., Moore, J.M., and Harris, C., 2003, Deposition, dia-
Canadian Institute of Mining and Metallurgy: Section B: Applied Earth genesis, and secondary enrichment of metals in the Paleoproterozoic
Science, v. 84, p. B128–B133. Hotazel Iron Formation, Kalahari manganese Field, South Africa: Eco-
Rye, R., and Holland, H.D., 1998, Paleosols and the evolution of atmospheric nomic Geology and the Bulletin of the Society of Economic Geologists,
oxygen: a critical review: American Journal of Science, v. 298, p. 621– v. 98, p. 1449–1462.
672. Uyeda, K., 1994, Sr and Nd geochemistry of Mt. McRae Shale in the Ham-
Rye, R., Kuo, P., and Holland, H.D., 1995, Atmospheric carbon dioxide con- ersley District, Australia [M.S. thesis]: Sendai, Japan, Tohoku Univer-
centrations before 2.2 billion years ago: Nature, v. 378, p. 603–605, doi: sity, 64 p.
10.1038/378603a0. van Staal, C.R., and Williams, P.F., 1984, Structure, origin, and concentration of
Sarmiento, J., 1992, Biogeochemical Ocean Models, in Trenberth, K.E., ed., the Brunswick 12 and 6 ore bodies: Economic Geology and the Bulletin
Climate system modeling: Cambridge, UK, Cambridge University Press, of the Society of Economic Geologists, v. 79, p. 1669–1692.
p. 519–564. Walker, J.C.G., and Brimblecombe, P., 1985, Iron and sulfur in the pre-biologic
Savarino, J., Romero, A., Cole-Dai, J., Bekki, S., and Thiemens, M.H., 2003, ocean: Precambrian Research, v. 28, p. 205–222, doi: 10.1016/0301-
UV induced mass- independent sulfur isotope fractionation in strato- 9268(85)90031-2.
spheric volcanic sulfate: Geophysical Research Letters, v. 30, p. 2131– Watanabe, Y., Naraoka, H., Wronkiewicz, D.J., Condie, K.C., and Ohmoto,
2134, doi: 10.1029/2003GL018134. H., 1997, Carbon, nitrogen, and sulfur geochemistry of Archean and
Schidlowski, M., and Aharon, P., 1992, Caron cycle and carbon isotope record: Proterozoic shales from the Kaapvaal Craton, South Africa: Geochi-
Geochemical impact of life over 3.8 Ga of earth history, in Schidlowski, mica et Cosmochimica Acta, v. 61, p. 3441–3459, doi: 10.1016/S0016-
M., Golubic, S., Kimberley, M.M., McKirdy, D.M., and Trudinger, 7037(97)00164-6.
P.A., eds., Early organic evolution: Implications for mineral and energy Watanabe, Y., Ikemi, H., and Ohmoto, H., 2005, Absence of mass independent
resources: Berlin, Springer-Verlag, p. 378–397. sulfur isotope fractionation in 2.76 Ga freshwater and 3.0 Ga marine sedi-
Schneider, D.A., Bickford, M.E., Cannon, W.F., Schulz, K.J., and Hamilton, ments: Astrobiology, v. 5, p. 272.
M.A., 2002, Age of volcanic rocks and syndepositional iron formations, Watanabe, Y., Naraoka, H., and Ohmoto, H., 2006, Mass independent fraction-
Marquette Range Supergroup: Implications for the tectonic setting of ation of sulfur isotopes during thermochemical reduction of native sulfur,
Paleoproterozoic iron formation of the Lake Superior region: Canadian sulfite and sulfate by amino acids: American Geophysical Union Spring
Journal of Earth Sciences, v. 39, p. 999–1012, doi: 10.1139/e02-016. Meeting, 23–26 May, Baltimore, Maryland (in press).
Widdel, F., Schnell, S., Heising, S., Ehrenreich, A., Assmus, B., and Schink, Young, T.P., and Taylor, W.E.G., eds., 1989, Phanerozoic ironstones: London,
B., 1993, Ferrous iron oxidation by anoxygenic phototrophic bacteria: The Geological Society, 251 p.
Nature, v. 362, p. 834–836, doi: 10.1038/362834a0. Zalenski, E., and Peterson, V., 1995, Depositional setting and deformation of
Wilkin, R.T., and Arthur, M.A., 2001, Variations in pyrite texture, sulfur iso- massive sulfide deposits, iron-formation, and associated alteration in the
tope composition, and iron systematics in the Black Sea: Evidence for Manitouwadge greenstone belt, Superior Province, Ontario: Economic
late Pleistocene to Holocene excursions of the O2-H2S redox transition: Geology and the Bulletin of the Society of Economic Geologists, v. 90,
Geochimica et Cosmochimica Acta, v. 65, p. 1399–1416, doi: 10.1016/ p. 2244–2261.
S0016-7037(01)00552-X. Zantop, H., 1981, Trace elements in volcanogenic manganese oxides and iron
Yamaguchi, K.E., Johnson, C.M., Beard, B.L., and Ohmoto, H., 2005, Bio- oxides: The San Francisco manganese deposit, Jalisco, Mexico: Eco-
geochemical cycling of iron in the Archean-Paleoproterozoic Earth: nomic Geology and the Bulletin of the Society of Economic Geologists,
Constraints from iron isotope variations in sedimentary rocks from the v. 76, p. 545–555.
Kaapvaal and Pilbara Cratons: Chemical Geology, v. 218, p. 135–169,
doi: 10.1016/j.chemgeo.2005.01.020.
Young, T.P., 1989, Phanerozoic ironstones: An introduction and review, in
Young, T.P., and Taylor, W.E.G., eds., Phanerozoic ironstones: London,
Geological Society Special Publication 46, p. ix–xxv. MANUSCRIPT ACCEPTED BY THE SOCIETY 29 OCTOBER 2005
Printed in the USA

Evolution of Early Earth's Atmosphere, Hydrosphere, and Biosphere, Constraints From Ore Deposits

Uploaded by

Copyright:

Available Formats

Evolution of Early Earth's Atmosphere, Hydrosphere, and Biosphere, Constraints From Ore Deposits

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Evolution of Early Earth's Atmosphere, Hydrosphere, and Biosphere, Constraints From Ore Deposits

Uploaded by

Copyright:

Available Formats

How did early life emerge on Earth according to Russell and Hall?

How did early life emerge on Earth according to Russell and Hall?

What does the study by Thiart and de Wit demonstrate about the development of early continents?

What does the study by Thiart and de Wit demonstrate about the development of early continents?

Downloaded from memoirs.gsapubs.

org on February 23, 2010

SUMMARY OF THE VOLUME

1. The onset and early evolution of life ...................................................................................................1

Evolution of the Early Continents

5. Secular variations of N-isotopes in terrestrial reservoirs and ore deposits .....................................81

Evolution of Uranium Deposits and Atmospheric Evolution

6. The sedimentary setting of Witwatersrand placer mineral deposits

7. Witwatersrand gold-pyrite-uraninite deposits do not support a reducing

Evolution of Seawater and Basinal Sulfur Geochemistry

10. Proterozoic sedimentary exhalative (SEDEX) deposits and links to evolving

11. Precambrian Mississippi Valley–type deposits: Relation to changes in composition

12. Superheavy S isotopes from glacier-associated sediments of the Neoproterozoic

16. Rare-earth elements in Precambrian banded iron formations: Secular changes

Geological Society of America

The onset and early evolution of life

The tension between CO2 dissolved at relatively high atmospheric pressure in

2 M.J. Russell and A.J. Hall

Embedded in fresh manganiferous exhalites, early photosynthetic bacteria could fur-

Keywords: CODH/ACS, greigite, origin of life, oxygenic photosynthesis, ranciéite.

—Ernst Haeckel 1892, p. 414

INTRODUCTION fides, mackinawite and greigite—sulfides which accommodate

The onset and early evolution of life 3

4 M.J. Russell and A.J. Hall

The onset and early evolution of life 5

3MgFeSiO3 + 7H2O → 3SiO2 + Fe3O4 +

Thus these moderate temperature springs are buffered to a pH

6 M.J. Russell and A.J. Hall

The onset and early evolution of life 7

Figure 5. Model environment for the

8 M.J. Russell and A.J. Hall

The onset and early evolution of life 9

Figure 7. Structure of mackinawite Fe1+xS.

Figure 8. Supposed emergence of chemi-

The onset and early evolution of life 11

Figure 11. The hydrothermal mound as an acetate generator. The de-

12 M.J. Russell and A.J. Hall

thioacetate (CH3COSCH3) in the presence of an FeS/NiS slurry

CO + 2CH3SH → CH3.COSCH3 + H2S (9)

Methane thiol has been synthesized from CO2 at 100 °C in the

CH3.COSCH3 + OH- → CH3.COO- + CH3SH (10)

The acetate reaction, augmented by the proton gradient

The onset and early evolution of life 13

14 M.J. Russell and A.J. Hall

(Russell et al., 2003). Other species such a Fe(OH)3 have lower

The onset and early evolution of life 15

16 M.J. Russell and A.J. Hall

THE ORGANIC TAKEOVER

Recalling our inorganic predilection, we use the way metals

The onset and early evolution of life 17

“peptide nests” the active centers were partially protected from

COENZYMES WITH ORGANIC RING COMPONENTS

One of the first requirements for organic molecules dur-

18 M.J. Russell and A.J. Hall

Figure 20. Table of the “universal” genetic