©2020 Society of Economic Geologists, Inc.

Economic Geology, v. 115, no. 3, pp. 489–503

Mapping Magmatic and Hydrothermal Processes from Routine

Exploration Geochemical Analyses
Scott Halley1,2,†
1Mineral Mapping Pty Ltd., 109 Joyce Street, Hawley Beach, Tasmania 7307, Australia
2Centre for Ore Deposit and Earth Sciences (CODES), University of Tasmania, Hobart, Tasmania 7001, Australia

Abstract
Analytical methods used by commercial assay laboratories have improved enormously in recent years. Induc-
tively coupled plasma-atomic emission spectroscopy and inductively coupled plasma-mass spectrometry meth-
ods now report analyses for half of the periodic table with exceptional detection limits and precision. It is
becoming commonplace for mining companies to use such methods routinely for the analysis of drill samples
throughout mineral deposits. Improvements in software and computing power now allow rapid interrogation
of upward of 100,000 assay samples. Geochemical analyses are quantitative, are independent of observer bias,
and can form the basis for robust geologic and mineralogical models of mineral deposits, as well as shed light
on scientific questions. In particular, consistently collected, high-quality geochemical analyses can significantly
improve and systematize logging of lithological and hydrothermal alteration mineralogic changes within drill
core. In addition, abundant, high-quality geochemical data provide insights into magmatic and hydrothermal
processes that were previously difficult to recognize and that have obvious applications to mineral exploration
and improved genetic models of ore deposits. This paper describes a workflow that mining industry geolo-
gists can apply to their multielement analysis data to extract more information about magma compositions and
gangue mineralogy.

Introduction sets. It is only just in the last five years that the software and
Over the last 40 years, commercial analytical methods for desktop computing power has been sufficient to deal with da-
multielement geochemistry have evolved from X-ray fluores- tabases of 50,000 analyses with more than 40 elements in the
cence (XRF) and atomic absorption spectroscopy (AAS), to assay suite. Such large volumes of data allow recognition of
instrumental neutron activation analysis (INAA), to inductive- patterns and processes in mineral deposits that could never
ly coupled plasma-atomic emission spectroscopy (ICP-AES) have been visualized in the past. Even though vast amounts of
and inductively coupled plasma-mass spectroscopy (ICP-MS). data are being generated, many geologists only plot commod-
Adjusted for inflation, commercial analytical laboratories now ity elements (e.g., Cu, Au) and a few pathfinder elements rath-
deliver analyses for half of the periodic table for the same er than taking advantage of the full element suite that is now
price as a limited suite of base metals by AAS 40 years ago. available. This paper presents a set of geochemistry plots that
The analytical data that mining and exploration companies mining and exploration geologists can apply to their ICP analy-
can routinely obtain from commercial laboratories have de- ses to help discriminate rock compositional groups (lithology)
tection limits and precision that are superior to those typically and hydrothermal alteration mineralogy and processes.
obtained in research laboratories 20 years ago. This analytical Commercial assay laboratories offer a choice of digestion
evolution has enabled the routine collection of large volumes and assay methods for multielement geochemistry. Aqua regia
of high-quality compositional data sets by mining companies, (mixture of nitric acid and hydrochloric acid in a molar ratio of
triggering a development in the methods that are commonly 1:3) dissolves sulfide minerals, carbonates, most oxides, and a
few clay minerals, but it does not dissolve framework silicates or
employed to interrogate compositional data sets with explora-
resistate accessory minerals. Most laboratories offer a four-acid
tion and mining objectives.
digest, sometimes called a multiacid digest. This is a combina-
Analytical methods: Strengths and weaknesses tion of nitric, perchloric, hydrofluoric, and hydrochloric acid.
When it is done well, a four-acid digest achieves complete di-
An analytical method that uses a four-acid digest with a com-
gestion of nearly all minerals. An alternative to the four-acid
bination of ICP-AES and ICP-MS is gradually becoming an
digest is a lithium borate fusion prior to dissolution in acid. The
industry standard. When every drill hole sample in a deposit
lithium borate fusion ensures a more complete digestion. How-
is analyzed with this method, the resulting database is likely
ever, there is a trade-off. The more complete digestion adds
to have greater than 50,000 geochemical analyses. As a com-
significant sample matrix, increasing the interference effects
parison, in the past, Ph.D. studies of mineral deposits might
during the analysis, so it increases the lower detection limit and
have generated a few hundred analyses at most. Numerous
reduces the precision. A four-acid digest will resolve patterns in
improved freeware and commercial products, as well as ma-
immobile trace elements present near detection limits better
chine learning approaches, can now take advantage of these
than a lithium borate fusion, even if the digestion is not total,
very large multidimensional data sets. The examples presented
but the fusion method will give more accurate analyses of major
here were plotted in ioGASTM, which is now an industry stan-
elements. Mining companies will generally prefer the four-acid
dard for interrogating large multielement geochemistry data
digest ICP-AES/-MS package because of the relative cost and
the quality of the ore metal analyses, but this method can also
†E-mail, [email protected] provide excellent major and trace element data.

ISSN 0361-0128; doi:10.5382/econgeo.4722; 15 p. 489 Submitted: August 14, 2019 / Accepted: January 20, 2020

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio
 490 SCOTT HALLEY

A typical protocol for a four-acid digestion is to add ni- Many mineral deposits are genetically related to fractionat-
tric acid and perchloric acid to precisely weighed portions ing magmas (Guilbert and Park, 2007). The mineral phases that
of pulverized samples in Teflon test tubes. The samples are crystallize early and fractionate are determined by the source
then heated for a period of time to promote reaction. Next, region composition, temperature, pressure, oxidation state,
a precisely measured volume of HF is added. At this stage water content, and sulfur content of the melt. All these fac-
the Teflon test tubes are lowered into a heating block, where tors contribute to the potential of a given magma to generate a
the temperature and reaction time is precisely controlled. The metal-bearing hydrothermal fluid, and those factors determine
aim is to heat the samples until they reach a state of incipi- which metals are most enriched in that fluid. Loucks (2014)
ent dryness. At this stage the samples are removed from the described a set of geochemical proxies for magmatic processes
heat and HCl is added. The four-acid digestion procedure is that indicate prospectivity for porphyry Cu magmas. By un-
a compromise between minimizing volatilization of As, Sb, derstanding the fractionation signatures in analytical data, pro-
and Tl in particular by not overheating the samples but at the spectivity indicators for all types of magmatic and hydrothermal
same time minimizing retention of Al and Ti in particular. The systems, not just porphyry Cu systems, can be derived.
most common source of error from a four-acid digestion is The same basic workflow can also be applied to sedimenta-
from samples that have not been taken to incipient dryness. ry rocks, although it is easier to interpret the chemistry of co-
Under these conditions, Al, Ti, and other trace elements tend herent facies igneous rocks than volcaniclastic or sedimentary
to be retained in the test tube as insoluble fluoride complexes. rocks. Coherent igneous rocks are usually quite composition-
There is significant variation between commercial laborato- ally homogeneous, whereas volcaniclastic and sedimentary
ries in the way they control the volume of acids added, control rocks generally have a mixed provenance and do not plot as
the temperature and heating time, and monitor the extent of distinct clusters on immobile trace element plots. Sorting of
evaporation. Some laboratories are conservative in minimiz- minerals and grain size fractions during transport and deposi-
ing the loss by volatilization, but in so doing, they compromise tion of sediments leads to large variability in trace element
the completeness of dissolution of other elements. For mining contents of sedimentary rocks across small spatial scales.
companies routinely acquiring analytical data via this method, Geologists typically think about rock compositions initially
it is advisable to have a small percentage of samples reana- in terms of SiO2 content—basalt, andesite, rhyolite, and so
lyzed using a lithium borate digestion so that the effectiveness on. Silicon dioxide is not reported from a four-acid digest ICP
of the four-acid digestion can be assessed. analysis, because it forms a volatile SiF4 complex during the
Zircon is the most difficult mineral to dissolve (Magaldi et reaction with hydrofluoric acid in the digestion process. Other
al., 2019). When there is radiation damage in the zircon crys- major elements like K and Na, traditionally used in petrology
tals—for example, in old rocks >1 Ga in age—the zircons to characterize rocks, are very mobile during hydrothermal
will dissolve. However, zircons in young rocks or with low U alteration and weathering processes and are therefore com-
content will not readily dissolve, and the zircon-contained monly not useful to characterize the rock composition. Rather
elements (Zr, Hf, Y, and heavy rare earth elements [REEs]) than plotting elements against Si to make a first-pass subdi-
will be significantly underreported. Cr by four-acid diges- vision of compositions, it is recommended that scandium is
tion ICP-MS is commonly 50 to 75% of Cr reported by XRF used instead. Sc is highly correlated with the silicate-hosted
(Yamasaki et al., 2016). The extent of Cr underreporting is component of Fe and is hosted in Fe-Mg minerals such as
dependent on the mineralogy of the samples, with chromite pyroxenes and amphibole. It always has a valency of 3+ and
being the least soluble. The degree to which many elements is immobile in most weathering and hydrothermal alteration
are incompletely dissolved is commonly related to alteration environments (Williams-Jones and Vasyukova, 2018).
types. Refractory minerals, such as titanite and rutile, are The elements that are least mobile in hydrothermal envi-
more difficult to dissolve in rocks from acidic alteration fa- ronments are high field strength elements (HFSEs) that have
cies compared to fresh rock. a valency of 3+ or 4+. Of the elements that are routinely ana-
lyzed with a four-acid digestion ICP-MS/-AES package, the 3+
Realizing the value of modern geochemical data and 4+ suite includes Sc, Al, P, Ti, V, Cr, Y, Zr, Nb, La, and Th
The interpretation of geochemical data always has a degree (Railsback, 2003).
of subjectivity. Interpreting analytical data should always in- Most of the first-row transition metals, including Sc, V, Mn,
volve reexamining the drill core megascopically and by hand Fe, Co, Ni, Cu, and Zn, have quite similar chemical proper-
lens to test the ambiguities and to confirm the boundaries and ties. In unaltered rocks the first transition-row metals usually
discontinuities in the data. The data presented in this paper plot with a linear relationship against Sc, which decreases with
are from various real-world case studies. Permission has been magmatic differentiation. Sc is usually the least mobile. A de-
granted to present the data but not to display it in geospatial parture from the linear trend against Sc is usually an indication
or geologic position. of fractionation of opaque oxide minerals, sulfide saturation in
melts, or leaching of metals during hydrothermal alteration.
Mapping Rock Compositions Some of the metals, most notably V, have multiple valencies
The following workflow is designed to allow a geologist to and are redox sensitive. While scandium is typically hosted
track magmatic evolution processes through the assessment only in silicate minerals, some of the metals substitute into spi-
of four-acid ICP-AES-/MS data and to allow the recognition nels and primary oxides as well as silicates. Mn and Zn are the
of things such as fractional crystallization of particular mineral most soluble of the transition metals. Plotting these against Sc
phases, magma mixing, and other processes that might be im- highlights areas of metal leaching during hydrothermal altera-
portant contributing factors to ore genesis. tion or metal enrichment, for example in chlorite (Wilkinson et

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio
 GEOCHEMICAL PLOTS FOR MAPPING MAGMATIC AND HYDROTHERMAL PROCESSES 491

al., 2015). Some of these metals are sequestered by immiscible ings that the points are assigned to should have a spatial clus-
magmatic sulfides, particularly Ni and Cu. tering that makes sense in terms of the geologic maps or cross
Geochemistry identifies compositions of rocks but does not sections; if they don’t match the underlying geology, then the
distinguish rock types. For example, a single magmatic event data has been grouped in the wrong way. The advantage of
could produce intrusive rocks, lava, and volcaniclastic and epi- this approach is that the user has a more explicit view of which
clastic rocks, all with the same composition but obviously in elements define the groups and which petrologic processes
different emplacement environments. Logging information may be needed to produce the compositional diversity. Sta-
should be added to assay tables so that geologic information tistical methods of geochemical data analysis such as princi-
can be integrated with the geochemical interpretation. pal component analysis or k-means clustering may simplify
A recommended starting point is to plot Sc versus Nb, P, the workflow and speed up data assessment. However, these
Th, Ti, V, and Zr (Fig. 1). The spread from high Sc (corre- methods that focus on a numerical analysis of the data have
sponding to high-Fe rocks) to low Sc (corresponding to low- a disadvantage of disconnecting the geologic and mineralogic
Fe rocks) indicates compositional ranges from mafic to felsic. thinking from the data interrogation process.
A point density contour overlay will highlight the composi- Another recommendation is to plot Cr against Sc or Ti. In
tional groups in the data. Each cluster in the data can be as- magmatic hydrothermal systems, it is quite common to see
signed a color, shape, or size attribute and displayed in 2-D evidence of magma mixing. The mixing processes commonly
maps or 3-D models. The point density contours usually show involve an injection of mafic magma derived from the mantle
quite distinct groupings for most of the data, but there is some at depth. Dacitic melts typically have 10 to 30 ppm Cr (Win-
ambiguity where the clusters overlap. ter, 2010). Chromium levels much above this usually point to
The second recommended step is to plot the same list of Cr pyroxene as a phenocryst phase, which must have come
elements but use Ti as the y-axis versus Sc, Nb, P, Th, V, and from a mafic composition melt (Mair et al., 2011).
Zr. Titanium is a compatible element, and Ti minerals such as Most magmatic rocks contain somewhere between 6 to 9%
Fe-Ti oxides tend to form early in a crystallization sequence, Al (Winter, 2010). Low Al values are usually a result of dilution
whereas Zr, Nb, and Th are incompatible and stay in the melt as a result of hydrothermal processes. In rocks that are strongly
until late in the cooling sequence (Pearce and Norry, 1979). veined or cemented breccias, the infilling mineral dilutes the
Compositional clusters that are defined within one x-y bivari- immobile trace element component of the rock (due to volume
ate plot should remain as distinct clusters in every other plot. change of the system), and all the immobile element scatter-
Making a series of plots like this allows the data to be viewed plots will have trends toward the origin (Barrett and MacLean,
from many different perspectives. The compositional group- 1999). This will be evident on a plot of Al against Ti or Sc.
Sc (ppm)

Sc (ppm)

Th (ppm) Ti (wt. %)
Sc (ppm)

Sc (ppm)

V (ppm) Nb (ppm)
Fig. 1. Sc versus Th, Ti, V, and Nb. This data set contains 30,000 assay points. The point density contour overlay highlights
compositional clusters in the data.

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio
 492 SCOTT HALLEY

As well as the general immobile element scatterplots listed the fertile porphyry Cu trend. The advantage of plotting
above, there are a series of specific x-y plots that should be V/Sc against Sc rather than SiO2 is that it will be less affected
examined. Pairs of trace elements that partition into a specific by alteration and vein density. The data shown in Figure 2 are
mineral phase but with slightly different partition coefficients derived from samples collected from a drill hole through the
are very sensitive indictors of fractionation of that mineral. Al- center of a porphyry Cu deposit in South America.
though this was recognized by igneous petrologists long ago, If fractionation of hornblende or pyroxene in anomalously
the analytical methods to measure this with any great preci- oxidized magmas leads to melts with an unusually high V/Sc
sion were not available at that time. Now that much lower ratio, then it follows that there ought to be preserved horn-
detection limits and higher precision can be obtained from blende and/or pyroxene cumulates with very low V/Sc ratios.
routine ICP-MS analyses, it is possible to characterize these There is a very different type of mineral system where such
fractionation processes in thousands of analyses through en- cumulates are observed. In the Archean greenstone belts of
tire mineral deposits. the Yilgarn craton in Western Australia, differentiated gabbro
Some of the plots from ICP-MS analyses that can be used sills with basal pyroxene cumulates are common. The lower
to map specific processes are listed in Table 1. portions of these sills are ilmenite bearing, and the pyroxenites
always have a V/Sc ratio around 7 (O’Connor-Parsons and
V versus Sc; amphibole versus magnetite fractionation Stanley, 2007). However, there are rare pyroxene cumulates
In reduced magmas, V3+ has behavior very similar to that of in gabbroic intrusions where the V/Sc ratios are consistently
Sc3+. Both metals substitute for Fe in amphibole, pyroxene, around 3. These intrusions host orthomagmatic Ni sulfides.
and biotite. In reduced magmas, the V/Sc ratio remains con- This suggests there is a link between higher magmatic sulfur
stant at around 7:1 from mafic through to felsic compositions contents and anomalous oxidation states in these Ni deposits.
(Li and Lee, 2004). In more oxidized magmas, the ion V4+
has a very strong affinity for magnetite (Shervais, 1982; Iveson Sr versus Y; high-pressure hydrous melting of plagioclase
et al., 2018). Fractional crystallization of magnetite causes a High Sr/Y and low total Y is a signature that is observed in por-
depletion of V relative to Sc. Accordingly, declining V/Sc with phyry Cu magmas (Fig. 3; Richards and Kerrich, 2007). De-
increasing SiO2 (or with decreasing Sc) is interpreted to re- pleted Y and middle to heavy REEs is interpreted as a result of
flect fractional crystallization of magnetite (Fig. 2). This is a garnet and hornblende fractionation. High Sr in intermediate
ubiquitous signature in magnetite-series melts. composition magmas is typical of hydrous melts where crystal-
Fertile porphyry Cu magmas show a trend of increasing lization of sodic plagioclase is suppressed until after the appear-
V/Sc with increasing SiO2 (Loucks, 2014). Loucks argues ance of hornblende and clinopyroxene. Because of its stability
that the effect of increasing water content in a melt is to at high pressure, fractionation of garnet is an indicator of thick
promote the crystallization of hornblende well ahead of crust (>30 km; Chiaradia et al., 2012). Most simply, this plot
magnetite. V acts incompatibly in strongly oxidized magmas, distinguishes magmas with a high water content, generated at
whereas scandium remains compatible in clinopyroxene and high pressure (Richards, 2011; Richards et al., 2012).
hornblende (Toplis and Corgne, 2002). Hence, in oxidized The data in Figure 3 include 30,000 geochemical analyses
and hydrous melts, crystallization of hornblende and clino- from an entire resource drill-out on a porphyry Cu deposit.
pyroxene lowers Sc in the melt, but V/Sc progressively in- It is inadvisable to use this plot on analyses from altered
creases. Figure 2 differentiates those melts where magnetite rocks, because Sr is quite mobile during hydrothermal al-
fractionates before Fe silicates from those where silicates teration. However, in this case the high Sr/Y rocks still plot
fractionate before magnetite. The more oxidized the mag- mostly as a tight cluster, indicating that in most of these
ma, the more pronounced the effect. rocks, there has been little mobility of Sr. Within the min-
Although hornblende fractionation is commonly implicated eralized zone, primary plagioclase was altered to albite plus
for the V/Sc trends (Toplis and Corgne, 2002; Loucks, 2014), anhydrite, and the Sr along with Ca was retained in anhy-
clinopyroxene has the highest partition coefficient for Sc of all drite. Significant depletion in Sr only occurs within domains
the likely host minerals (Williams-Jones and Vasyukova, 2018). of acid alteration.
Porphyry Cu magmas typically have 3 to 8 ppm Sc and
V/Sc ratios of 10 to 15. A four-acid digest ICP-MS/-AES does Sc versus Ni, Cu; sulfide saturation
not report Si; thus, some key discrimination diagrams from Figures 4, 5, 6, and 7 were created from the same data set.
Loucks (2014) cannot be produced from this assay method. These are analyses of outcrop samples from the Mount Read
As an alternative, plotting V/Sc versus Sc effectively maps Volcanics in western Tasmania, and these rocks host sever-

Table 1. Recommended Geochemical Plots for Identifying Magmatic Processes

Discriminators Process Reference

V/Sc vs. Sc To identify signatures of high-pressure melting of an amphibole-bearing This study

source or fractional crystallization of magnetite
Sr/Y vs. Y To identify high-pressure melting of plagioclase in hydrous environment Richards and Kerrich (2007)
Ni vs. Sc To identify sulfide-saturated magmas This study
Hf vs. Zr To identify fractional crystallization of zircon Claiborne et al. (2006)
Ti vs. Nb To characterize opaque oxide mineralogy This study
Ta/Nb vs. Ti To identify fractional crystallization of biotite Stepanov and Hermann (2013)

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio
 GEOCHEMICAL PLOTS FOR MAPPING MAGMATIC AND HYDROTHERMAL PROCESSES 493

Prospective porphyry
copper deposit field
V/Sc

Fractional crystallization
of magnetite

Sc (ppm)
Fig. 2. V/Sc versus Sc. Decreasing V/Sc with decreasing Sc and increasing SiO2 is a signature of fractional crystallization of
magnetite. The classification of these points was derived from a combination of visual logging and scatterplots of Sc, Th, Ti,
V, Nb, Cr, and P.

al significant volcanogenic massive sulfide (VMS) deposits. the most evolved rocks have additionally fractionated biotite
These data were selected as a type example of fractionation in and zircon (Figs. 5–7).
moderately oxidized midcrustal magmas. The dacitic to rhyo- In unaltered rocks Cu is highly correlated with Sc with a ratio
litic rocks are predominantly volcaniclastic. One of the VMS of about 2.5:1, and in igneous rocks that do not contain olivine
horizons in the Mount Read Volcanics is located at the top of or clinopyroxene, Ni is highly correlated with Sc with a ratio of
the most evolved unit. This suite exhibits a very strong trend about 1.5:1. In melts that have become sulfide saturated, mono-
of fractional crystallization of magnetite and plagioclase, and sulfide solid solution crystallizes as a solid FeS containing signif-

Ore-related rocks
Sr/Y

Normal arc rocks

Y (ppm)
Fig. 3. Sr/Y versus Y showing the typical ranges for fertile porphyry Cu magmas.

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio
 494 SCOTT HALLEY
Ni (ppm)

Cu (ppm)
Sc (ppm) Sc (ppm)

Fig. 4. Sc versus Ni and Sc versus Cu. The classification of rock compositions was derived from visual logging and scatterplots
of Sc, Th, Ti, V, Zr, and Nb. The arrows show the expected Ni and Cu contents at a given Sc level. Sulfide saturation in the
melt results in stripping of Ni and Cu from the silicate melt.

icant amounts of Ni and Co, leaving the silicate melt depleted midcrustal magma chambers and fractionated magnetite prior
in Ni and Co. With further sulfide saturation, an immiscible to being emplaced in the upper crust show V/Sc ratios lower
Fe-Cu-Au-S (intermediate solid solution) phase is formed. It than 7 and Cu and Ni contents that are depleted relative to Sc.
is very common for melts that have undergone fractional crys- Once magmas have formed immiscible sulfides, Cu and Au are
tallization of magnetite to show depletion of Ni and Cu in the stripped from the melt, and the ore-forming potential of those
remaining silicate melt as a result of sulfide saturation (Fig. 4). magmas is reduced (Richards, 2015). This can be recognized
Jenner et al. (2010) referred to this process as “the magnetite as depletion due to sulfide saturation in the melt (rather than
crisis.” This phenomenon is observed in a diverse range of mag- leaching of the metals by hydrothermal fluids) because the Cu
ma compositions in many different geologic environments. For and Ni are depleted, but Zn and Mn are not.
example, the coupling of magnetite fractionation and sulfide
saturation is easily mapped during in situ fractionation in lay- Hf versus Zr; fractional crystallization of zircon
ered gabbro sills (O’Connor-Parsons and Stanley, 2007). More Other trace elements are proxies for fractionation of specific
commonly, batches of magma that have had residence time in minerals. The primary mineral where Hf resides is zircon

Hf/Zr = 36:1

Hf/Zr = 20:1 Hf/Zr = 25:1

Hf (ppm)

Basalt
Andesite
Dacite 1
Dacite 2
Weakly fractionated dacite
Weakly fractionated rhyolite
Moderately fractionated rhyolite
Strongly fractionated rhyolite

Zr (ppm)
Fig. 5. Hf versus Zr, showing decreasing total Zr and increasing Hf/Zr as a result of fractional crystallization of zircons.

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio
 GEOCHEMICAL PLOTS FOR MAPPING MAGMATIC AND HYDROTHERMAL PROCESSES 495

Trend for midcrustal

Trend for midcrustal
melts fractionating
melts fractionating
ilmenite-biotite
magnetite
Ti (ppm)

Basalt
Andesite
Dacite 1
Dacite 2
Weakly fractionated dacite
Typical range
for porphyry Weakly fractionated rhyolite
copper deposit Moderately fractionated rhyolite
magmas
Strongly fractionated rhyolite

Nb (ppm)
Fig. 6. Ti versus Nb.

(Claiborne et al., 2006), where it substitutes for Zr. In the Fractional crystallization of zircon is strongly enhanced in
vast majority of data sets, Hf versus Zr plots with a remark- water and fluorine-rich magmas. In very hydrous melts, and
ably linear correlation, with a ratio of 1:36. Hf is less compat- especially in fluorine-rich melts, silicate chains are depoly-
ible and less abundant than Zr, so as zircons crystallize, the merized, because the hydroxylation breaks the polymerizing
Hf/Zr ratio in the melt increases. Zircon crystals are typi- Si-O-Al bonds (Burnham, 1979; Stolper, 1982). This results
cally zoned, with higher Hf/Zr in the rims (Claiborne et al., in a decrease of viscosity and lowering of the solidus. In these
2010). Depending on the composition of the melt, zircon circumstances, fractional crystallization of zircon is more pro-
saturation temperatures are generally around 750° to 800°C nounced. Felsic magmas that are related to hydrothermal
(Dilles et al., 2015). Crystal settling of zircons in a magma mineral deposits, for example tin granites, have low Zr and
chamber results in lower total Zr in the remaining melt, but much higher Hf/Zr.
fractionation leads to an increased Hf/Zr ratio. Reduced to- The Hf versus Zr plot should be used with caution with
tal Zr content but high Hf/Zr ratio is a clear indicator of frac- four-acid digestion analyses due to the difficulty of dissolving
tional crystallization of zircons in silicic hydrous melts (Fig. zircons. As a guide for considering the effectiveness of the
5). This is a ubiquitous signature in very high silica (>74% digestion, basalts typically contain 50 to 100 ppm Zr and gran-
SiO2) magmas. ites around 200 ppm (Berkmann, 2011).

Fractional
crystallization of biotite
Ta/Nb

Basalt
Andesite
Dacite 1
Dacite 2
Weakly fractionated dacite
Weakly fractionated rhyolite
Moderately fractionated rhyolite
Strongly fractionated rhyolite

Ti (ppm)

Fig. 7. Ta/Nb versus Ti. An increase in Ta/Nb with decreasing Ti is an indication of fractional crystallization of biotite.

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio
 496 SCOTT HALLEY

Ti versus Nb; opaque mineral phases Alteration reactions result in an existing mineral phase be-
ing converted to a different phase, usually with the addition
Minerals that incorporate Ti, such as Ti magnetite, ilmenite,
or removal of one or more components. As a reaction runs
biotite, and hornblende, commonly crystallize early in hy-
toward completion, the molar proportions of components
drous intermediate magmas. Nb has chemistry that is very
trend toward the stoichiometry of the mineral phases that are
much like Ti, but it has much lower compatibility. In any one
being formed in the reaction. Scatterplots of the major ele-
batch of magma, Ti and Nb are highly correlated, but the way
ments in these minerals typically show an alteration trajec-
in which the Ti/Nb ratio changes with fractionation depends
tory from the composition of the unaltered rock toward the
on which mineral phase is the main repository for Ti. The ex-
projected stoichiometry of the newly formed mineral. Some
ample shown in Figure 6 has strong fractional crystallization
of the scatterplots designed for tracking these reactions are
of magnetite and lesser fractionation of biotite. With frac-
shown in Table 2.
tional crystallization of magnetite, Ti content in the magma
may fall to around 1,000 ppm, but Nb contents remain about K/Al versus Na/Al (molar); sericitization
the same. Late in the evolution of very felsic magmas (as in
A useful way to characterize phyllic alteration in porphyry
Sn-related granites), biotite may become the dominant frac-
copper deposits is to plot K/Al versus Na/Al calculated on
tionating Ti host mineral, and the melts show a strong enrich-
a molar basis, which relates the bulk geochemical response
ment in Nb and Ta relative to Ti. In contrast, typical porphyry
to mineral stoichiometry (Stanley and Madeisky, 1996). In
copper deposit-related magmas always have around 2,000 to
this plot, most fresh magmatic rock compositions will plot
4,000 ppm Ti and 2 to 4 ppm Nb and show limited evolution
with fractionation. between 0.3 to 0.5 Na/Al and 0.1 to 0.4 K/Al, forming a trend
from mafic at low Na/Al and K/Al to felsic at higher Na/Al
Ta/Nb versus Ti; fractional crystallization of biotite and K/Al (Fig. 8). This is in response to the proportions of the
common Na-K-Al rock-forming minerals (K-feldspar, biotite,
Ti, Nb, and Ta have similar chemical properties. Ti is compat-
plagioclase, etc.). In fresh rock, progressively lower Na/Al
ible; Nb is less compatible; Ta is the least compatible. Nb and
ratios respond to the increasing anorthite component of the
Ta are least compatible in biotite; therefore, granites that have
plagioclase, as Na decreases as Ca increases. Different rocks
the greatest amount of biotite fractionation have the largest
have different starting compositions, which affects how the
amount of enrichment in Nb and Ta (Stepanov and Hermann,
altered composition is interpreted.
2013). Increasing Ta/Nb with decreasing Ti is an indicator for
Consider a rock that is totally sericitized. The mineralogy
fractional crystallization of biotite (Fig. 7).
of the rock might be muscovite-quartz-carbonate-pyrite. All
Mapping Alteration Mineralogy of the K and Al in that rock will be within sericite. Therefore,
the ratio of K/Al in the sericitized rock is 1:3. Similarly, K-
Consistently logging mineralogy and alteration intensity in silicate–altered rocks can be identified. A totally K-feldspar
drill core or drill chips is very subjective. Variation in skills and (KAlSi3O8)–altered rock will have a K/Al ratio of 1:1. The
opinions from one geologist to the next makes it very difficult amount of K that is added in a biotite-altered rock is limit-
to create reliable 3-D alteration models of orebodies based ed by the availability of Fe in that rock, but it will plot on a
on visual logging. Routine four-acid digest ICP analyses pro- trajectory toward K/Al = 1 and Na/Al = 0. In the same way,
vide a quantitative and consistent data set through an entire albitization (Na-Ca alteration) can also be tracked. Albite is
orebody. This provides the basis for improved 3-D resource NaAlSi3O8: Na/Al =1:1. Therefore, in a data set that includes
domain models. Where correlations are established between unaltered and altered rock compositions, samples will plot in
metallurgical test results and mineralogy estimated from the trends from the least altered composition in the center of the
chemistry, the analyses can be used as a proxy to populate an diagram toward different mineral composition on the edges of
entire block model with predicted rock properties. The eco- the diagram (e.g., Fig. 8).
nomic value of an ore block is determined not only by the In quartz-sericite-pyrite-albite-chlorite-carbonate altera-
amount of contained metal but also by the cost of mining, tion systems, it is quite common to see altered rocks plot-
milling and extracting the metal from the ore, and dealing ting on a tie line between muscovite and albite, as shown
with deleterious elements after mining. If mineralogical prox- in Figure 8. This example is from altered basalts in an Ar-
ies for cost components can be established, then the block chean orogenic gold system. These rocks have had potas-
model can be populated with a predicted value per block
rather than just a grade estimate per block. This will lead to
Table 2. Recommended Geochemical Plots for Identifying Hydrothermal
improved financial models and mine scheduling optimization. Alteration Mineralogy Trends
Consider the compositions of the common alteration
minerals: Discriminators Process
1. Kaolinite Al4(Si4O10)(OH)8;
K/Al vs. Na/Al (molar) To map sericite alteration
2. White mica KAl3Si3O10(OH)2; Ca-K-Na ternary To map feldspar compositions
3. Chlorite (Fe,Mg)5Al2Si3O10(OH)8; Al-K-Mg ternary To map advanced argillic alteration and define
4. Carbonate Ca(Fe,Mg)(CO3)2; relative proportions of sericite vs. chlorite
5. Albite NaAlSi3O8; Rb vs. K To map alunite
Ca vs. Mg To map calcite vs. dolomite
6. Potassium feldspar KAlSi3O8; Ca-Fe-Mg ternary To pick carbonate compositions
7. Pyrite FeS2; and Ca-Fe-S ternary To map anhydrite vs. pyrite
8. Anhydrite CaSO4.

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio
 GEOCHEMICAL PLOTS FOR MAPPING MAGMATIC AND HYDROTHERMAL PROCESSES 497

K-Feldspar
Kaolinite
Strong sericite
Moderate sericite
Sericite albite
Sericite chlorite
Albite
K/Al (molar)

Muscovite
Unaltered rhyolite

Unaltered dacite

Albite
Kaolinite Unaltered basalt

Na/Al (molar)
Fig. 8. K/Al versus Na/Al molar ratio plot. The mineralogical classification of these analyses was assigned using the K/Al versus
Na/Al plot, the Al-K-Mg ternary plot, and SWIR spectra.

sium added until compositions reach the muscovite-albite The example in Figure 10 shows a trend of points between
tie line. In the acid part of the system, compositions follow the projected compositions of muscovite and chlorite, a
the tie line along to the muscovite node. In neutral-pH parts trend of points from muscovite to the Al apex (kaolinite and
of the system, compositions trend toward the albite node. other clays), and a trend of points from K-feldspar to chlo-
The relative proportions of sericite to albite are easily read rite. This example is from an unconformity-related uranium
from this diagram. deposit. This plot is also very useful in porphyry copper de-
posits, because advanced argillic alteration assemblages are
Ca-K-Na ternary; hydrothermal feldspar always depleted in Mg as well as Na and, therefore, plot
The data in Figure 9 comes from a collection of surface rock along the Al-K join.
chip samples from the Yerington porphyry Cu district (Halley It is impossible to differentiate between different types
et al., 2015). The ellipse in the center shows where the least of advanced argillic alteration using compositional data, as
altered rocks plot. There are a range of hydrothermal feld- the key advanced argillic minerals formed at different tem-
spar compositions from oligoclase to albite to alkali feldspar peratures are only composed of Al and Si as major elements
to orthoclase. These map out very distinct zones around the (e.g., pyrophyllite, dickite, kaolinite). Likewise, it is difficult
porphyry Cu environment. The red points in Figure 9 include to differentiate weathering that typically has kaolinite from
samples with both orthoclase and muscovite. A plot that in- advanced argillic alteration that can have a higher-tempera-
cludes K/Al ratios is required to differentiate orthoclase-rich ture mineralogy. Short-wave infrared (SWIR) mineralogy is
from muscovite-rich samples within this group. High-salinity particularly useful for this purpose, because it is an indepen-
fluids favor the formation of feldspars rather than phyllosili- dent data set compared to the compositional data set and
cates. Visually identifying the various feldspar species can be relies on spectral absorptions to identify minerals. If spectral
difficult (e.g., Escolme et al., 2019). data are available, they should be merged with the compo-
sitional data at this stage, to cross-check the mineralogy of
Al-K-Mg ternary; sericite-chlorite advanced argillic clays the inferred alteration types and help discern the advanced
Although the K/Al versus Na/Al molar ratio plot is useful, it argillic mineralogy.
can be ambiguous, especially since it says nothing about the
proportion of chlorite or other Fe-Mg–bearing minerals in a Rb versus K; alunite
sample. Furthermore, mixtures of feldspar and chlorite can Rubidium readily substitutes for potassium in most silicate
plot with the same K/Al ratio as sericite, and, therefore, the minerals, but it does not substitute into sulfates. This proper-
K/Al versus Na/Al plots (Fig. 8) should always be used in con- ty is very useful for distinguishing alunite-bearing samples in
junction with an Al-K-Mg ternary plot (Fig. 10). a table of geochemical analyses. Figure 11 shows two differ-

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio
 498 SCOTT HALLEY

Ca (wt. %)

Albite
Albite-Kspar-sericite
Background
Sodic-calcic
Pyroph/alun/topaz
Sericite
Sericite-albite
Ore zones
Oligoclase Kspar-albite
(sodic-calcic)

Least altered
rocks

K-feldspar Albite
Alkali feldspar

K (wt. %) Na (wt. %)
Fig. 9. Ca-K-Na ternary plot. The mineralogical classification of these analyses was assigned using a K/Al versus Na/Al plot,
the Ca-K-Na ternary plot, Cu analyses, and SWIR spectra. Alun = alunite, Kspar = K-feldspar, Pyroph = pyrophyllite.

Al (wt. %)
Kaolinite

Muscovite
Chlorite

K-feldspar

K (wt. %) Mg (wt. %)
Fig. 10. Al-K-Mg ternary plot showing physical mixtures of kaolinite, muscovite, chlorite, and potassium feldspar.

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio
 GEOCHEMICAL PLOTS FOR MAPPING MAGMATIC AND HYDROTHERMAL PROCESSES 499

ent suites of rocks with different linear K versus Rb trends. chlorite, present in the rock. Thus, having geologic informa-
The data points that have K with no Rb are samples where tion from logging is critical to making the most robust geo-
alunite was identified from SWIR spectra collected with a chemical interpretations and testing ideas developed from
TerraSpec instrument. evaluation of geochemical data.
Ca versus Mg; hydrothermal carbonate Ca-K-Na ternary; sericite-carbonate
Figure 12 shows a plot of Ca versus Mg from a sediment- Figure 14 shows a data set of about 5,000 analyses from the
hosted base metal system. The dolomite line shows a Ca to stringer zone in the Hellyer VMS deposit (Gemmell and Large,
Mg molar ratio of 1:1. Samples with pure dolomite (± quartz- 1992). The least altered samples plot approximately near the
illite-pyrite) would plot on this line. The samples in blue show projected compositions of a standard dacite or andesite. A
a strongly linear trend, suggesting a carbonate control, but characteristic of alteration in VMS systems is depletion of Na.
they plot above the dolomite line, because the dolomite is Fe This plot shows that in many samples, Na is depleted, but Ca
bearing (ferroan). If ferroan dolomite is present, then a plot of is retained (as carbonate). Depletion of Na but retention of Ca
Ca (molar) versus Mg + Fe (molar) will show a linear trend at causes the data points to move directly away from the Na apex
1:1, which forms a bounding line to the data points. This is a on to the Ca-K edge of the ternary plot. Samples with more
common pattern in carbonate-bearing systems. In sediment- intense alteration plot as points that move away from the Ca
hosted systems, it is also common to see data points trending apex until they reach the K apex. Spatially, this compositional
away from the projected compositions of calcite or dolomite, trend maps a zoning pattern from relict plagioclase to sericite-
as the primary carbonate minerals are replaced. carbonate to (quartz)-sericite-pyrite. The carbonate halo is not
easy to quantify from visual logging. It is difficult to achieve
Ca-Fe-Mg ternary; hydrothermal carbonate consistency in visual logging of core like this, but it is easily
Figure 13 is based on the same data as Figure 8. The miner- mapped from the chemical composition. In these rocks, the
alogy in these rocks is quartz-sericite-albite-carbonate-pyrite. carbonate content becomes most obvious after the drill core
All the calcium, magnesium, and iron in these rocks is in car- has been exposed to the atmosphere for a period of time and
bonate and pyrite. These data plot on a tight linear trend that the surface of the ferroan carbonate becomes stained by orange
runs from a point on the ankerite-dolomite tie line to the Fe iron oxides.
apex. This linear trend is controlled by mixture of magnesian
ankerite and pyrite. From this, the solid solution composi- Fe versus S; sulfidation
tion of the carbonates can be approximately determined. This Although Fe is quite mobile, in most hydrothermal sys-
would not work if there was another Fe-Mg mineral, such as tems involving low- to moderate-salinity fluids, a plot of Fe
Rb (ppm)

K (wt. %)
Fig. 11. Rb versus K, with data points colored by the dominant mineral measured from a SWIR spectrum.

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio
 500 SCOTT HALLEY

Ca (wt. %)

ine
mite l
Dolo

Mg (wt. %)

Fig. 12. Ca versus Mg. Kspar=K-feldspar.

Ca (wt. %)
Kaolinite
Strong sericite
Moderate sericite
Sericite albite
Sericite chlorite
Albite

Dolomite

Ankerite

Pyrite

Fe (wt. %) Mg (wt. %)
Fig. 13. Ca-Fe-Mg ternary plot, showing a mixing trend between magnesian ankerite and pyrite.

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio
 GEOCHEMICAL PLOTS FOR MAPPING MAGMATIC AND HYDROTHERMAL PROCESSES 501

Ca (wt. %)

Se
r
Ca icite
rbo -
na
te

Least altered

e
rit
-P
y andesite
ite
ric
Se

K (wt. %) Na (wt. %)
Fig. 14. Ca-K-Na ternary plot.

against S shows that pyrite is formed by adding sulfur and rate representation of the intensity of pyritic alteration than
utilizing Fe that is already present in the rock (Fig. 15). just total S content.
When all of the available Fe in a rock has reacted to form
pyrite, samples plot on the pyrite line, with a molar ratio of Ca-Fe-S ternary; anhydrite-pyrite
Fe/S = 1:2. Where pyrite-bearing veins are present in the A Ca-Fe-S ternary plot is a useful way to look at the relative
rock samples, the data points plot at higher values on the importance of pyrite and anhydrite in porphyry Cu systems
pyrite line. An S/Fe ratio (sulfidation index) is a more accu- (Fig. 16). With the addition of sulfur to a rock, the data points

Pyrite veins
Fe (wt.%)

Disseminated pyrite line

ite
utilizing host-rock Fe Pyr
Background
Strong sericite
Sericite-chlorite
Moderate sericite
Albite
K-feldspar
Weak sericite

S (wt. %)
Fig. 15. Fe versus S. The mineralogical classification of these analyses was assigned using a K/Al versus Na/Al plot and
SWIR spectra.

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio
 502 SCOTT HALLEY

move from the least altered composition toward the S apex. V/Sc with increasing SiO2. Decreasing Ti/Nb ratio can be used
In some parts of the porphyry Cu environment, anhydrite is to track fractionation of Fe-Ti oxides. Potassic magmas that
formed without sulfides, but feldspars (Na-K) remain stable. In fractionate biotite will show a trend of increasing Ta/Nb. In-
that case, the amount of S that can be added is limited by the creasing Hf/Zr with lower total Zr is an indicator of fractional
amount of Ca in the rocks. These samples would plot on the tie crystallization of zircon. Fractional crystallization of magne-
line between anhydrite and the Fe apex (biotite, magnetite). tite leads to sulfur saturation in a melt, and this is evident from
In other parts of the porphyry Cu environment, pyrite is low Ni contents relative to Sc.
formed without anhydrite, and feldspars are destroyed. In Scatterplots of major elements can be designed to dem-
this case, the amount of S that can be added is limited by the onstrate wall-rock alteration reaction paths dictated by the
amount of Fe in the rock. These samples would plot along mineral compositions in the altered rocks and to quantify the
a trend from the projected position of the least altered rock intensity of alteration. Using the chemical composition as a
toward the pyrite node on the Fe-S join. supplement to visual logging overcomes much of the subjec-
In the example shown in this plot, the data points mostly tivity and observer bias that is inherent in logging. This pro-
plot along the join between anhydrite and pyrite. The amount vides the opportunity to build more robust models of mineral
of S that was added was limited by the amount of Ca + Fe in deposits and additionally improves the understanding of ore
the rock. Points that lie on the sulfur-rich side of the pyrite- deposit formation, ore type characterization, and metallurgi-
anhydrite tie line contain alunite—a mineral phase that con- cal domaining. The next logical development that should fol-
tains sulfur—but no Fe or Ca. low on from this should be quantitative estimations of miner-
alogy from assay data.
Conclusions
Many mining and exploration companies are now routinely Acknowledgments
acquiring four-acid digest ICP-AES/-MS geochemical analy- The author has greatly benefited from many years of men-
ses from drill samples throughout entire orebodies. Immobile toring and collaboration with Simon Gatehouse, Greg Hall,
trace elements can be used to identify compositional groups John Walshe, John Dilles, and Dick Tosdal. The manuscript
within the data. Pairs of trace elements that have a common was vastly improved by reviews from Kevin Byrne, Jeremy
host mineral provide sensitive indicators of magmatic pro- Vaughn, John Dilles, Shaun Barker, and Federico Cernuschi.
cesses such as fractional crystallization. V/Sc ratios are redox The ideas presented in this paper have evolved from working
sensitive. Reduced magmas retain relatively constant V/Sc on many thousands of data sets from innumerable mineral de-
with increasing SiO2. Moderately oxidized magmas show de- posits and exploration projects in a variety of commodities and
creasing V/Sc with SiO2 if they fractionate magnetite. Magmas geologic settings. All the companies involved are thanked for
that fractionate Fe silicates but not Fe oxides evolve to higher the opportunity to learn from their data and from their staff.

Ca (wt. %)
Alunite
Topaz
Pyrophyllite
Dickite
Kaolinite
Sericite
Chlorite sericite
Carbonate
Smectite
Least-altered Anhydrite
Albite chlorite sericite
rock

Add
ition
of s
ulfu
r

Pyrite
Fe (wt. %) S (wt. %)
Fig. 16. Ca-Fe-S ternary plot. The mineralogical classification of these analyses was assigned using a K/Al versus Na/Al plot,
Rb versus K plot, Al-K-Mg ternary plot, and SWIR spectra.

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio
 GEOCHEMICAL PLOTS FOR MAPPING MAGMATIC AND HYDROTHERMAL PROCESSES 503

REFERENCES Richards, J.P., 2011, High Sr/Y arc magmas and porphyry Cu ± Mo ± Au
Barrett, T.J., and MacLean, W.H., 1999, Volcanic sequences, lithogeochem- deposits: Just add water: Economic Geology, v. 106, no. 7, p. 1075–1081.
istry, and hydrothermal alteration in some biomdal volcanic massive sulfide ——2015, The oxidation state, and sulfur and Cu contents of arc magmas:
systems: Reviews in Economic Geology, v. 8, p. 101–131. Implications for metallogeny: Lithos, v. 233, p. 27–45.
Berkmann, D.A., 2011, Field geologists’ manual, 5th edition: Australasian Richards, J.P., and Kerrich, R., 2007, Special paper: Adakite-like rocks: Their
Institute of Mining and Metallurgy, Monograph 9, 382 p. diverse origins and questionable role in metallogenesis: Economic Geology,
Burnham, C.W., 1979, The importance of volatile constituents, in Yoder, H.S. v. 102, no. 4, p. 537–576.
Jr., ed., The evolution of the igneous rocks: 50th anniversary perspectives: Richards, J.P., Spell, T., Rameh, E., Razique, A., and Fletcher, T., 2012, High
Princeton, New Jersey, Princeton University Press, p. 439–482. Sr/Y magmas reflect arc maturity, high magmatic water content, and por-
Chiaradia, M., Ulianov, A., Kouzmanov, K., and Beate, B., 2012, Why large phyry Cu ± Mo ± Au potential: Examples from the Tethyan arcs of central
porphyry Cu deposits like high Sr/Y magmas?: Scientific Reports, v. 2, arti- and eastern Iran and western Pakistan: Economic Geology, v. 107, no. 2, p.
cle 685. 295–332.
Claiborne, L.L., Miller, C.F., Walker, B.A., Wooden, J.L., Mazdab, F.K., and Shervais, J.W., 1982, Ti-V plots and the petrogenesis of modern and ophiolitic
Bea, F., 2006, Tracking magmatic processes through Zr/Hf ratios in rocks lavas: Earth and Planetary Science Letters, v. 59, no. 1, p. 101–118.
and Hf and Ti zoning in zircons: An example from the Spirit Mountain Stanley, C.R., and Madeisky, H.E., 1996, Lithogeochemical exploration for
batholith, Nevada: Mineralogical Magazine, v. 70, no. 5, p. 517–543. metasomatic zones associated with hydrothermal mineral deposits using
Claiborne, L.L., Miller, C.F. and Wooden, J.L., 2010, Trace element compo- Pearce element ratio analysis: British Columbia, Canada, Mineral Deposit
sition of igneous zircon: A thermal and compositional record of the accu- Research Unit, Short Course Notes on Pearce Element Ratio Analysis, 99
mulation and evolution of a large silicic batholith, Spirit Mountain, Nevada: p.
Contributions to Mineralogy and Petrology, v. 160, no. 4, p. 511–531. Stepanov, A.S., and Hermann, J., 2013, Fractionation of Nb and Ta by biotite
Dilles, J.H., Kent, A.J., Wooden, J.L., Tosdal, R.M., Koleszar, A., Lee, R.G., and phengite: Implications for the “missing Nb paradox”: Geology, v. 41,
and Farmer, L.P., 2015, Zircon compositional evidence for sulfur-degassing no. 3, p. 303–306.
from ore-forming arc magmas: Economic Geology, v. 110, no. 1, p. 241–251. Stolper, E., 1982, Water in silicate glasses: An infrared spectroscopic tech-
Escolme, A., Berry, R.F., Hunt, J., Halley, S., and Potma, W., 2019, Predictive nique: Contributions to Mineralogy and Petrology, v. 81, p. 1–17.
models of mineralogy from whole-rock assay data: Case study from Produc- Toplis, M.J., and Corgne, A., 2002, An experimental study of element parti-
tora Cu-Au-Mo deposit, Chile: Economic Geology, v. 114, p. 1513–1542. tioning between magnetite, clinopyroxene and iron-bearing silicate liquids
Gemmell, J.B., and Large, R.R., 1992, Stringer system and alteration zones with particular emphasis on vanadium: Contributions to Mineralogy and
underlying the Hellyer volcanogenic massive sulfide deposit, Tasmania, Petrology, v. 144, no. 1, p. 22–37.
Australia: Economic Geology, v. 87, no. 3, p. 620–649. Wilkinson, J.J., Chang, Z., Cooke, D.R., Baker, M.J., Wilkinson, C.C., Inglis,
Guilbert, J.M., and Park Jr., C.F., 2007, The geology of ore deposits: Illinois, S., Chen, H., and Gemmell, J.B., 2015, The chlorite proximitor: A new tool
Waveland Press, 985 p. for detecting porphyry ore deposits: Journal of Geochemical Exploration,
Halley, S.W., Dilles, J.H., and Tosdal, R.M., 2015, Footprints: Hydrothermal v. 152, p. 10–26.
alteration and geochemical dispersion around porphyry copper deposits: Williams-Jones, A.E., and Vasyukova, O.V., 2018, The economic geology of
SEG Newsletter, no. 100, p. 12–17. scandium, the runt of the rare earth element litter: Economic Geology, v.
Iveson, A.A., Rowe, M.C., Webster, J.D., and Neill, O.K., 2018, Amphibole-, 113, no. 4, p. 973–988.
clinopyroxene- and plagioclase-melt partitioning of trace and economic Winter, J.D., 2010, An introduction to igneous and metamorphic petrology:
metals in halogen-bearing rhyodacitic melts: Journal of Petrology, v. 59, no. Upper Saddle River, New Jersey, Prentice Hall, 697 p.
8, p. 1579–1604. Yamasaki, S.I., Takeda, A., Kimura, K., and Tsuchiya, N., 2016, Underesti-
Jenner, F.E., O’Neill, H.S.C., Arculus, R.J., and Mavrogenes, J.A., 2010, mation of chromium and zirconium in soils by hydrofluoric acid digestion
The magnetite crisis in the evolution of arc-related magmas and the ini- and inductively coupled plasma-mass spectrometry: Soil Science and Plant
tial concentration of Au, Ag and Cu: Journal of Petrology, v. 51, no. 12, p. Nutrition, v. 62, no. 2, p. 121–126.
2445–2464.
Li, Z.X.A., and Lee, C.T.A., 2004, The constancy of upper mantle fO2 through
time inferred from V/Sc ratios in basalts: Earth and Planetary Science Let-
ters, v. 228, no. 3–4, p. 483–493.
Loucks, R.R., 2014, Distinctive composition of copper-ore-forming arc mag-
mas: Australian Journal of Earth Sciences, v. 61, no. 1, p. 5–16.
Magaldi, T.T., Navarro, M.S., and Enzweiler, J., 2019, Assessment of disso-
lution of silicate rock reference materials with ammonium bifluoride and
nitric acid in a microwave oven: Geostandards and Geoanalytical Research,
v. 43, no. 1, p. 189–208.
Mair, J.L., Farmer, G.L., Groves, D.I., Hart, C.J., and Goldfarb, R.J., 2011,
Petrogenesis of postcollisional magmatism at Scheelite dome, Yukon,
Canada: Evidence for a lithospheric mantle source for magmas associated Scott Halley received his B.Sc. (Hons) degree
with intrusion-related gold systems: Economic Geology, v. 106, no. 3, p. from the University of Tasmania (1982) and his
451–480. Ph.D. degree from The Australian National Univer-
O’Connor-Parsons, T., and Stanley, C.R., 2007, Downhole lithogeochemical sity (1986). He worked as an exploration geologist
patterns relating to chemostratigraphy and igneous fractionation processes with Australian mining companies from 1986 until
in the Golden Mile dolerite, Western Australia: Geochemistry: Exploration, 2001, and then with Placer Dome until 2005. Since
Environment, Analysis, v. 7, no. 2, p. 109–127.
then, Scott has worked as a consulting geochem-
Pearce, J.A., and Norry, M.J., 1979, Petrogenetic implications of Ti, Zr, Y, and
Nb variation in volcanic rocks: Contributions to Mineralogy and Petrology, ist in mineral exploration and mining for around 150 different companies
v. 69, no. 1, p. 33–47. in 40 different countries. Since 2008, Scott has run an annual geochemistry
Railsback, L.B., 2003, An earth scientist’s periodic table of the elements and short course at CODES, University of Tasmania, and he is now a part-time
their ions: Geology, v. 31, p. 737–740. researcher at CODES.

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/115/3/489/5015550/4722_halley.pdf

by Ánghelo A. Baldoceda Dionisio