See

discussions, stats, and author profiles for this publication at:

https://www.researchgate.net/publication/279543802

Precipitation of precious and base metal-

bearing scales at the Berlín geothermal
field, El Salvador

Article in Transactions - Geothermal Resources Council · January 2003

CITATION READS

1 52

3 authors:

Jasmin Raymond Anthony E. Williams-Jones

Institut National de la Recherche Scien… McGill University
41 PUBLICATIONS 347 CITATIONS 290 PUBLICATIONS 6,411 CITATIONS

SEE PROFILE SEE PROFILE

James R Clark
McGill University
13 PUBLICATIONS 125 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Multi-scale analysis of multiparameter geophysical & geochemical data from active

volcanic system View project

All content following this page was uploaded by Anthony E. Williams-Jones on 03 December 2015.

The user has requested enhancement of the downloaded file.

 Journal of Volcanology and Geothermal Research 145 (2005) 81 – 96
www.elsevier.com/locate/jvolgeores

Mineralization associated with scale and altered rock and pipe

fragments from the Berlı́n geothermal field, El Salvador;
implications for metal transport in natural systems
Jasmin Raymond1, Anthony E. Williams-JonesT, James R. Clark
Dept. Earth and Planetary Sciences, McGill University, 3450 University St., Montréal, Québec, Canada H3A 2A7
Received 8 October 2003; accepted 17 January 2005

Abstract

Composite fragments sampled at solid collectors and drains of two-phase, re-injection, and vapour pipelines of the Berlı́n
geothermal field, El Salvador, consist mainly of sulphide- and electrum-bearing aluminium-rich amorphous silica scale,
sulphide- and electrum-bearing saponitic/vermiculitic clay from the reservoir, and altered metallic pipe linings containing As–S-
bearing iron oxide–oxyhydroxide grains. Siliceous and clay-rich precipitates contain concentrations of gold and silver in excess
of 180 and 8000 ppm, respectively, and appreciable concentrations of copper, lead, zinc, and antimony. Altered iron fragments
contain substantial arsenic.
Copper, lead, and zinc occur mainly as chalcopyrite, galena, and sphalerite, respectively, in amorphous silica and clay; near
the surface, chalcopyrite transported from depth alters to bornite. Gold and silver occur mainly as electrum, which deposited
with base metal sulphides in the clay precipitates, and amorphous silica at higher levels in the well. Electrum precipitates in the
wells due to the rapid drop in temperature and loss of H2S associated with boiling. The concentration of gold in vapour is ~4
times greater than that in water from associated wellheads. This suggests that gold can be transported efficiently by vapour, and
implies that such transport may be important in the formation of some hydrothermal ore deposits.
D 2005 Elsevier B.V. All rights reserved.

Keywords: geothermal scale; gold; amorphous silica; clay; vapour; boiling

1. Introduction

Geothermal systems were first recognized as

analogues of ore-forming hydrothermal systems by
Skinner et al. (1967), who reported the occurrence of
T Corresponding author. Tel.: +1 514 398 1676.
E-mail address: [email protected] (A.E. Williams-Jones). copper- and silver-bearing minerals in opaline silica
1
Presently at Département de Géologie et de Génie Géologique, scales from surface installations of the Salton Sea field,
Université Laval. California, and by Weissberg (1969), who discovered
0377-0273/$ - see front matter D 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jvolgeores.2005.01.003
82 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96

significant Au–Ag concentrations in well cuttings oxyhydroxide, quartz, amorphous silica, native sul-
from the Broadlands field, New Zealand. Subsequent phur, pyrite, pyrrhotite, halite, and sylvite in similar
studies by Brown (1986) and Clark and Williams- fragments (GESAL, 2002), but except for our
Jones (1990) established that the processes which preliminary work (Raymond et al., 2003), did not
control precipitation of gold and silver in geothermal describe textural relationships among these phases or
wells can be used to model directly the formation of interpret scale formation. The objectives of the current
epithermal ore deposits. Geothermal scales and fluids study are thus to fully characterize the mineralogy and
have been particularly valuable sources of information geochemistry of sampled fragments, to evaluate their
for understanding metal behaviour in hydrothermal origin, and to relate these findings to natural ore-
systems (Henley, 1984; Krupp and Seaward, 1987; forming systems.
Gallup, 1998a), and have demonstrated the importance
of phase separation (boiling) as a depositional mech-
anism. However, the possibility that vapour may be an 2. Geological setting
agent of base and precious metal transport was not
considered in these studies and, in view of recent The Berlı́n geothermal field (Fig. 1), El Salvador,
experiments demonstrating that these metals are is located 100 km east of San Salvador, on the flanks
relatively soluble in aqueous vapour (cf. Williams- of the Berlı́n–Tecapa Quaternary basaltic–andesite
Jones et al., 2002), it is something that merits stratovolcano, along the southern fault system of the
investigation in geothermal systems. Central American graben (Wiesemann, 1975). The
In this study, we report results of analysis and Berlı́n–Tecapa volcanic complex is part of the San
modelling of scale, rock, and corroded pipe fragments Salvador Formation, and is one of nine volcanoes
sampled from liquid and vapour pipelines of the occurring along a northwesterly trend parallel to the
Berlı́n geothermal field, El Salvador. Previous scale trench axis of the Cocos plate, which is subducting
studies at Berlı́n noted the presence of iron oxide and under the Caribbean plate (Carr and Stoiber, 1990).

Fig. 1. Simplified geological map of the Berlı́n–Tecapa volcanic complex (modified from GESAL, 2001), showing the location of the Berlı́n
geothermal field and the distribution of studied wells.
 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96 83

The Berlı́n–Tecapa volcanic complex (Fig. 1) is collectors and drains (Fig. 2) attached to liquid-vapour
characterized by basaltic to andesitic lava flows and and re-injection lines of wells TR-2/TR-9 (WS-1 to 8),
scoria, and andesitic to dacitic ignimbrites, which the TR-4B/TR-4C water tank (WS-9), and vapour
were produced during two major eruptions that lines (WS-10 to 14), during a maintenance operation
accompanied the development of the Berlı́n and in February 2000. The system had been cleaned
Blanca Rosa calderas (GESAL, 2001). The Berlı́n previously in August 1999, and the wells remained
caldera forms a keyhole-shaped crater that opens open throughout the period between the two cleaning
toward the north–northwest, and encloses the geo- operations (L. Barrios, pers. comm., 2001). The
thermal field. Three distinct aquifers have been fragments constitute mixtures of material that origi-
identified in the field: (1) a shallow, low salinity nated in surface pipelines, at depth in the wells, and/or
aquifer (1600 ppm NaCl equiv.; 200 to 300 m a.s.l.); in the reservoir. The fragments were detached and
(2) a moderately saline intermediate aquifer (6600 transported by geothermal fluids to the solid collectors
ppm NaCl equiv.; near sea level) with temperatures of and drains, where they were cemented by amorphous
150 to 200 8C, and a thickness of 500 to 700 m; and silica. Samples were analysed by X-ray fluorescence
(3) a deep (800 to 1200 m b.s.l.), saline aquifer that is for major elements, by INAA for Au, As, and Sb, and
exploited for geothermal energy (Monterrosa, 1993; by ICP-OES (combined hydrochloric–nitric–perchlo-
Santos, 1995) and is responsible for scaling in wells ric–hydrofluoric acid digestion) for Ag, Cu, Pb, and
and pipelines. Fluids in the deep aquifer contain 8000 Zn. Selected samples, especially those with high
to 12 000 ppm NaCl equivalent (Monterrosa, 1993), concentrations of precious and base metals, were
and have a pH of ~5 and a temperature of ~290 8C evaluated in detail by optical microscopy (transmitted
(D’Amore and Mejia, 1999). The reservoir is hosted and reflected light), scanning electron microscopy
by basaltic–andesite and andesite lavas interbedded (SEM), energy dispersive spectrometry (EDS), and
with minor tuff layers; these rocks have undergone electron microprobe (EMP) analyses. In addition, one
propylitic alteration, characterized by abundant quartz sample was analysed by X-ray diffraction to assess
and epidote, moderate chlorite and sericite, and minor clay mineralogy. Four water samples from the geo-
calcite and wairakite (CEL, 1999; GESAL, 2001). thermal reservoir, collected at wellheads TR-2, TR-9,
TR-4B, and TR-4C, one water sample from the re-
injection line at wellhead TR-8, and a sample of
3. Sample material and analytical methodology condensed vapour from the separator connecting wells
TR-4C and TR-4B to the vapour line, were analysed
Fourteen composite samples, consisting of mixed for cations by ICP-MS. These data were supple-
scale, rock, and corroded pipe fragments, and ranging mented with Cl
, SO4=, HCO3
, and pH values for the
in mass from ~2 to 22 g, were obtained from solid same samples obtained by La Geo S.A. de C.V.

Fig. 2. Schematic sections showing solid collectors and drains from which composite samples were taken, and typical flow paths for Berlı́n
fluids: (A) location of samples related to wells TR-2 and TR-9, and (B) location of samples related to wells TR-4B and TR-4C. Composite
samples WS-1 to 7 and WS-9 to 14 are located by numbers 1 to 7 and 9 to 14, respectively. Geothermal infrastructure is labelled by letters: A,
back-pressure plate; B, solid collector; C, drain; D, separator; E, water tank; F, silencer; G, weirbox; and H, turbine.
84 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96

Fig. 2 shows typical flow paths for Berlı́n fluids. At Table 1

approximately 2000 m depth, the fluids boil at a Mineralogy of scale, rock and corroded pipe fragments in composite
samples W-1 to 14
temperature of ~290 8C in a radius of several metres
Minerals Siliceous Type 1 clay Type 2 Vapour line
around the wells (L. Barrios, pers. comm. 2002).
scale clay pipe frag.
Fluids enter the screened well, and continue boiling as
Early AS-1 VC
they rise rapidly to the surface under high pressure,
Later AS-2 VC VC VC
exiting the well at a temperature of ~145 8C. At the SVC VC VC
back pressure plate, pressure drops, and fluids flow at Qza C C C
lower pressures to the steam separator. Steam flows Amp R
through the vapour line to the turbine, and waste water Rt R
Ksp R
is eventually re-injected, either directly into the deep
Sl R
aquifer or the intermediate reservoir. Cp C C C
Sp C C C
Gl C C C
4. Mineralogy of scale, rock, and corroded pipe Pg R R
Bn C C R
fragments Tt R R
Py C C
Two dominant types of fragment, siliceous and Po C
clay-rich, both locally rimmed by amorphous silica, Act R R R
occur in composite samples WS-1 to 10 (Table 1). Gf R R
El R R R
Siliceous and clay fragments occur with clasts of
IOH C C
altered porphyritic lava and silicified tuff, monomi- AsS C
neralic plagioclase, quartz and epidote fragments VC=very common, C=common, and R=rare.
(from the reservoir), and corroded pipe fragments. Mineral abbreviations: AS=aluminium-rich amorphous silica,
Scale, rock, monomineralic, and corroded pipe frag- SVC=saponitic/vermiculitic clay, Qz=quartz, Amp=amphibole,
ments are locally cemented together by late-stage Rt=rutile, Ksp=K-feldspar, Sl=sylvite, Cp=chalcopyrite, Sp=spha-
aluminium-rich amorphous silica (or by quartz in lerite, Gl=galena, Pg=phosgenite(?), Bn=bornite, Tt=tetrahedrite,
Py=pyrite, Po=pyrrhotite, Act=acanthite, Gf=geffroyite, El=elec-
sample WS-10). Scale, rock, and corroded pipe trum, IOH=iron oxide/oxyhydroxide, AsS=As- and S-bearing iron.
fragments from vapour line samples WS-11 to 14 a
Only occurs in sample WS-10.
have been altered by acidic vapour, and are described
separately below.
generation of aluminium-rich amorphous silica (Fig.
4.1. Siliceous scales 3B) that has a similar composition but appears
fresher (not cracked and stained) than the primary
Approximately 94 vol.% of composite samples, silica. Siliceous scales in sample WS-10 are rimmed
except those from the vapour line, consist of by quartz. Based on semi-quantitative EDS analyses
siliceous scale fragments, comprising aluminium-rich (n=15), the average composition (anhydrous) of
amorphous silica that has two distinct modes of amorphous silica is 83 wt.% SiO2, 9 wt.% Al2O3,
occurrence: early cracked and stained, and later 5 wt.% K2O, and 3 wt.% CaO, and the silica/
bfreshQ. Siliceous scale fragments in samples WS-1, aluminium molar ratio is ~5.4; the early and later
WS-2, WS-5, and WS-10 consist of translucent, varieties of amorphous silica are essentially identical
aluminium-rich amorphous silica that is stained by in composition.
iron oxide/oxyhydroxide and exhibits colloform Early aluminium-rich amorphous silica scales host
(Fig. 3A) and cracked (Fig. 3B) textures. The dendritic and disseminated sulphides and electrum,
siliceous scales commonly have a base of altered variably constituting 1 to 5 vol.% of the siliceous
iron (pipe fragments) that is replaced locally by fragments. In general, the most abundant sulphide
pyrite. Siliceous scale and pipe fragments in samples minerals are chalcopyrite and sphalerite (~1 vol.%
WS-1, WS-2, and WS-5 are rimmed by a second each), followed by galena, bornite, Se-acanthite,
 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96 85

Fig. 3. Photomicrographs and backscattered electron images of scale and corroded pipe fragments. (A) Photomicrograph in plane-polarized
transmitted light of chalcopyrite dendrites (Cp; black) hosted by early colloform aluminium-rich amorphous silica (AS-1; translucent) on an
altered iron pipe fragment (IOH; black), and rimmed by the second generation of aluminium-rich amorphous silica (AS-2; translucent). (B)
Photomicrograph in plane-polarized transmitted light of early cracked and stained aluminium-rich amorphous silica (AS-1; translucent), on an
altered iron pipe fragment (IOH; black). The entire scale fragment, including the base, is rimmed by a second generation of bfreshQ aluminium-
rich amorphous silica (AS-2; translucent). (C) Backscattered electron image of Se-acanthite (Act; white) in a chalcopyrite dendrite (Cp; medium
gray), hosted by early aluminium-rich amorphous silica (black). (D) Backscattered electron image of galena (Gl; white), with sphalerite (Sp;
medium gray), chalcopyrite (Cp; dark gray), and Se-acanthite (Act; light gray), hosted by early aluminium-rich amorphous silica (black). (E)
Backscattered electron image of spherulitic type 1 saponitic/vermiculitic clay (SVC; medium gray) rimmed by late-stage amorphous silica (AS-
2; medium gray). This fragment contains small sylvite grains (Sl; white), and is adjacent to a large grain of probable phosgenite (Pg; white). (F)
Backscattered electron image of electrum and galena (El and Gl; both white) in sphalerite (Sp; medium gray), locally rimmed by chalcopyrite
(Cp; dark gray), hosted by type 1 saponitic/vermiculitic clay (black). (G) Backscattered electron image of chalcopyrite (Cp; dark gray) replaced
by bornite (Bn; medium gray), both with inclusions of galena and later probable phosgenite (Gl and Pg; white), hosted by type 1 saponitic/
vermiculitic clay (black). (H) Backscattered electron image of bornite (Bn; dark gray) with galena inclusions (Gl; white), replaced by
tetrahedrite (Tt; light gray) at grain margins; hosted by type 1 saponitic/vermiculitic clay (black).

geffroyite(?), tetrahedrite, and electrum (b1 vol.% Sulphide dendrites contain individual grains (Fig.
each). Energy dispersive spectrometric analyses (n=4 3A,C) that are V5 Am in diameter, and consist
acanthite; n=3 geffroyite) indicate that the stoichiom- dominantly of chalcopyrite, with lesser fine-grained
etry of acanthite is Ag1.8[S0.7Se0.3], and that for sphalerite, galena, acanthite (Fig. 3C), and geffroy-
geffroyite(?) is (Ag2.3Cu4.0Fe3.7)(S7.4Se0.6); the iden- ite(?). Sphalerite and galena are commonly replaced
tification of geffroyite is tentative, as the quality of the by chalcopyrite, and acanthite and geffroyite(?) occur
analyses was limited by the extremely fine grain-size at chalcopyrite grain margins. Late-stage bornite and/
of this mineral (V3 Am). or tetrahedrite locally replace the sulphide dendrites.
86 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96

Sphalerite (1 to 20 Am), chalcopyrite (1 to 30 Am), Table 2

galena (1 to 10 Am), and pyrite (10 to 35 Am) also Average composition (wt.%) and stoichiometry of saponitic/
vermiculitic clay
occur as disseminated grains in the early aluminium-
wt.% Type 1 (n=24) Type 2 (n=6) Type 3a (n=7)
rich amorphous silica; pyrite grains are typically
isolated and are generally not intergrown with other SiO2 39.37 38.25 37.52
TiO2 0.01 0.01 0.01
sulphide minerals. Chalcopyrite locally rims galena,
Al2O3 5.91 6.98 9.84
sphalerite, and electrum grains, and is locally replaced Fe2O3 11.78 13.61 4.40
by bornite and tetrahedrite. The main gold- and silver- MnO 0.37 0.60 0.14
bearing minerals in early amorphous silica are MgO 16.53 15.13 26.17
electrum, acanthite, and geffroyite(?). Acanthite CaO 1.25 1.28 0.13
Na2O 0.20 0.14 0.04
(Fig. 3D) and geffroyite(?) (both 1 to 3 Am) occur at
K2 O 0.19 0.31 0.02
the margins of chalcopyrite and sphalerite grains. Total 75.61 76.31 78.27
Electrum occurs as 1 to 3 Am diameter inclusions in
chalcopyrite, bornite, and sphalerite grains. Stoichiometry
Late-stage aluminium-rich amorphous silica hosts Si 1.31 1.27 1.25
Al 0.17 0.21 0.29
minor disseminated chalcopyrite, sphalerite, galena Ti 0.00 0.00 0.00
and pyrite, and traces of electrum. Chalcopyrite grains Fe3+ 0.22 0.26 0.08
range from 30 to 100 Am in diameter, enclose galena Mn 0.01 0.01 0.00
and electrum, and are locally replaced by bornite, Mg 0.41 0.38 0.65
acanthite, geffroyite(?), and tetrahedrite. Sphalerite Ca 0.02 0.02 0.00
Na 0.00 0.00 0.00
grains (5 Am diameter) enclose galena and electrum,
K 0.00 0.00 0.00
and are locally replaced by acanthite and geffroyite(?). Sum 10.24 10.25 9.67
Pyrite grains are locally intergrown with chalcopyrite Si/Al 7.53 6.20 4.31
and galena, and are typically 10 to 40 Am in diameter. Mg/Fe 1.85 1.47 7.84
Electrum, acanthite, and geffroyite(?) are the main Jeol JXA-8900L electron microprobe WDS analyses; ZAF correc-
gold and- silver-bearing minerals in the later amor- tion; 15 kV, 20 nA; 20 s peak counts; 5 Am beam; total Fe as Fe3+;
phous silica, but are less abundant than in the early stoichiometry based on 22 oxygen equivalents; analytical totals are
less than 100 wt.% due to the likely presence of water, which cannot
amorphous silica. be analysed by this method.
a
From vapour line (sample WS-13).
4.2. Clay-rich precipitates
indicate that the clay does not contain Na (thus
Clay-rich fragments comprise approximately 5% of eliminating loughlinite), and that compositions are
the composite samples. Most of these fragments similar to saponite or vermiculite. Thus, the clay has a
appear to have formed in the geothermal reservoir as structure that is similar to sepiolite or montmorillonite,
they are not attached to altered pipe substrate. Three but a stoichiometry that more closely resembles
types of clay have been distinguished based on mode saponite or vermiculite. Similar clays have been
of occurrence, composition, and associated minerals; reported to occur in wells and pipes at other geo-
all types are characterized by high concentrations of thermal fields (cf. Reyes and Cardile, 1989).
magnesium and iron and high silicon/aluminium
ratios, and stoichiometries that resemble saponite or 4.2.1. Massive to well-crystallized clay-rich fragments
vermiculite (Table 2). Separates containing type 1 clay (type 1)
were hand-picked from sample WS-2 under a Most (~95%) clay-rich fragments are brown to
binocular microscope, and were analysed qualitatively yellow–brown coloured and massive to well-crystal-
by powder diffraction using a Rigaku D/MAX-2400V lized with conspicuous to distinctive spherulites (Fig.
X-ray diffractometer. The resulting spectrum is con- 3E), and are locally rimmed by late-stage aluminium-
sistent with a clay mineral having a structure similar rich amorphous silica (samples WS-5 and WS-7) or
to sepiolite, loughlinite, or montmorillonite, with a quartz (WS-10). Type 1 clay fragments contain traces
major reflection at ~7 2. However, EMP analyses of sylvite and up to 45 vol.% of sulphide minerals.
 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96
Table 3
Compositions of composite samples (scale, corroded pipe, and rock fragments) from solid collectors and drains at Berlı́n
Sample Location Major elements (wt.%) Trace metals (ppm)
SiO2 TiO2 Al2O3 Fe2O3 MnO MgO CaO Na2O K2O LOI Au* Ag As* Cu Pb Sb* Zn Total
WS-1 TR-2/9 S. collector #1 L+V line 65.99 0.02 9.25 6.40 0.07 0.71 1.22 1.73 3.23 9.99 31.100 1840 158 3133 3474 40.5 2920 99.80
WS-2 TR-2/9 S. collector #2 L+V line 62.08 0.02 8.63 8.30 0.12 2.29 1.10 1.64 2.85 10.48 187.000 1140 150 5707 7278 49.7 6100 99.59
WS-3 TR-2/9 Drain #2 L+V line 59.57 0.04 8.57 17.18 0.14 0.19 1.43 1.66 3.00 8.34 2.530 358 162 451 347 25.8 364 100.33
WS-4 TR-2/9 Drain #3 L+V line 70.84 0.02 8.57 5.48 0.06 0.10 1.08 1.82 2.62 9.27 9.320 580 94 680 379 31.8 332 100.09
WS-5 TR-8 Drain #2 re-injection line 62.54 0.10 8.98 7.31 0.16 2.55 1.78 1.75 2.70 9.99 149.000 1210 345 6203 4261 28.5 4260 99.54
WS-6 TR-8 Drain #1 re-injection line 72.78 0.02 8.35 2.03 0.07 1.34 1.16 1.74 2.92 9.15 18.800 206 69 895 1041 66.4 829 99.90
WS-7 TR-8 Drain #2 re-injection line 74.87 0.02 7.72 1.41 0.06 1.24 1.03 1.79 2.70 8.75 17.800 79 49 738 843 52.9 705 99.86
WS-9 TR-4B/4C Water tank – – – – – – – – – – 7.920 61 50 290 119 6.6 188 –
WS-10 TR-4B/4C Drain #1 V line 67.56 0.02 8.19 2.85 0.05 1.61 1.21 1.47 2.82 9.06 148.000 8100 393 9772 10490 240.0 14500 99.23
WS-11 TR-4B/4C Drain #3 V line – – – – – – – – – – 1.100 22 1040 1878 27 17.2 3330 –
WS-12 TR-4B/4C Drain #7 V line 1.01 0.02 0.60 96.28 0.31 0.04 0.08 0.15 0.02 3.00 0.915 39 740 544 25 18.8 212 101.72
WS-13 TR-4B/4C Drain #9 V line 1.38 0.09 0.51 94.36 0.32 0.04 0.09 0.14 0.02 4.59 0.439 bd/l 1860 779 bd/l 36.0 387 101.91
WS-14 TR-4B/4C Drain #14 V line 1.41 0.04 0.64 80.36 0.24 0.12 0.06 0.21 0.02 18.22 4.500 20 11300 1587 92 67.9 175 102.70
D. L. (ppm): 60 35 120 30 30 95 15 75 25 100 0.002 5 0.5 10 3 0.2 30
–: insufficient sample material for analysis.
Major elements by XRF, trace metals by ICP-OES except where noted by * INAA. Note that totals may be high due to reporting of total Fe as Fe2O3, and that the sums include data
for additional analysed elements that are not included in the table.
Abbreviations: S=solid, L=liquid, V=vapour, D.L.=detection limit, and bd/l=less than detection limit.

87
88 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96

Chalcopyrite, sphalerite, and galena occur as dissemi- injection line samples, and therefore are probably of
nated grains (up to 60 Am diameter), and sphalerite local origin. Some fragments may also have been
also occurs as acicular grains filling cracks in clay detached and transported to the vapour line (i.e.,
fragments. Chalcopyrite commonly rims or is inter- originated from well casings/linings or two-phase pipe
grown with sphalerite (Fig. 3F), and fine-grained (V3 linings), but these are indistinguishable from in situ
Am) acanthite and geffroyite(?) occur near the margins fragments. The pipe fragments consist of iron oxide/
of chalcopyrite and sphalerite grains. Bornite locally oxyhydroxide and locally contain small grains (5 to
replaces chalcopyrite (Fig. 3G) and sphalerite, and 10 Am diameter) of an As–S-bearing iron phase.
tetrahedrite (Fig. 3H) locally replaces chalcopyrite, These grains have an average composition of 17 wt.%
sphalerite, and bornite. Chalcopyrite, bornite, and As and 10 wt.% S, but elemental proportions vary
sphalerite contain inclusions of galena and electrum widely; in some cases, the composition approximates
(average 2 to 5 Am in diameter; Fig. 3F). Galena is that of arsenopyrite. The altered iron fragments are
locally altered to a Pb- and Cl-bearing phase (EMP- locally replaced by intergrowths of pyrite and
WDS scans also suggest elevated peak counts for C pyrrhotite.
relative to adjacent grains), tentatively identified as Monomineralic chalcopyrite and covellite frag-
phosgenite (Pb2(CO3)Cl2). ments also occur in vapour line samples, and have
embayed rims. Covellite was observed only in sample
4.2.2. Rock fragments pervasively altered to clay WS-14, and is rimmed by pyrite. Gold- and/or silver-
(type 2) bearing minerals were not observed.
Massive saponitic/vermiculitic clay similar to type Quartz occurs as distinct monomineralic fragments
1 clay occurs in altered reservoir rock fragments, and as grains accreted to altered pipe fragments in
which contain remnants of K-feldspar, rutile, and sample WS-13; it also rims and cements composite
amphibole. This type of clay-bearing fragment is rare, fragments in sample WS-10. In sample WS-13, quartz
and occurs only in sample WS-5 (re-injection line). fragments are corroded and embayed, and thus appear
Chalcopyrite, sphalerite, and galena (up to 25 vol.%) to have been locally out of equilibrium with the
occur as disseminated grains (up to 30 Am in vapour.
diameter) in type 2 fragments. Chalcopyrite rims
sphalerite and galena. Fine-grained (1 to 3 Am)
acanthite occurs at the edges of chalcopyrite and 5. Paragenesis
sphalerite grains, and is locally intergrown with
bornite. Chalcopyrite is locally replaced by bornite, The mineral paragenesis for an idealized well/
and galena by phosgenite(?). Chalcopyrite, bornite, surface pipeline sequence is summarized in Fig. 4.
and sphalerite host inclusions of galena and fine- The first mineral to precipitate in response to geo-
grained electrum (2–3 Am in diameter). thermal production appears to have been saponitic/
vermiculitic clay, which filled open space and
4.2.3. Altered clay fragments from the vapour line pervasively altered reservoir rocks in the production
(type 3) horizon adjacent to the well. Traces of sylvite in type
Clay fragments from the vapour line sample are 1 clay minerals were likely formed by evaporation of
opaque and appear unstable. They contain more brine trapped in pores during sample collection.
magnesium and aluminium, but less iron and alkalis Amorphous silica and saponitic/vermiculitic clay did
(Table 3) than those from two-phase/re-injection lines not precipitate together, as they do not occur in the
samples. Type 3 clay fragments contain mineral same scale fragments (except where the latter forms
assemblages similar to those of type 1 and 2. composite clasts cemented by late-stage amorphous
silica or quartz in surface pipes). The precipitation of
4.3. Vapour line fragments late-stage amorphous silica likely occurred after fluids
entered surface pipelines.
Altered iron fragments are more abundant in Millimetre-size fragments originating in the reser-
vapour line samples than in two-phase and re- voir and the well column were transported by
 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96 89

elevations in the well than sphalerite and galena, as it

generally rims the latter minerals. Electrum co-
precipitated with these base metal sulphides. The
deposition of precious and base metals occurred
throughout the well and surface pipelines, but late-
stage amorphous silica hosts only minor abundances
of these minerals.
Late-stage replacement minerals occurring close to
the wellhead and/or the back-pressure plate, where
transported fragments experienced abrupt changes in
pressure and temperature, include acanthite, geffroy-
ite(?), bornite, and tetrahedrite after chalcopyrite and
sphalerite, tetrahedrite after bornite, and phosgenite(?)
after galena. Acanthite co-precipitated with bornite
and the latter with tetrahedrite. Tetrahedrite to bornite
volume ratios increase with distance from the well
along surface pipelines, indicating not only that
tetrahedrite formed later than other sulphide minerals,
but also that it formed mainly after the fragments had
been transported to the surface. The presence of
tetrahedrite in sample WS-2 indicates that some
tetrahedrite also formed in the two-phase line.
Fig. 4. Schematic mineral paragenesis of composite samples from Additional deposition of sulphides occurred when
surface pipelines. The paragenetic sequence is plotted versus fragments were transported into the vapour lines.
distance from well and surface structures, but is not drawn to scale.
Pyrite replaced iron oxide/oxyhydroxide, and co-
Mineral precipitates from two-phase and re-injection lines are
plotted on the right, and corrosion products and minerals forming in precipitated locally with pyrrhotite. Covellite, which
the vapour line on the left. Geothermal infrastructure is labelled with occurs only in vapour line samples, may have
letters: A, back-pressure plate; D, separator; and E, water tank. replaced transported chalcopyrite fragments; however,
Mineral abbreviations are the same as those used in Table 1. chalcopyrite and covellite were not observed together.
Sample WS-10 contains the greatest abundance of
geothermal fluids to the re-injection and vapour lines. tetrahedrite, suggesting that some tetrahedrite formed
Quartz probably formed in the vapour line because it in the vapour lines. Arsenic- and sulphur-bearing iron
only occurs in samples WS-10 and WS-13. Since minerals are also interpreted to have formed in the
quartz has the same mode of occurrence as late-stage vapour lines because they were only observed in
amorphous silica (i.e., rims and cements fragments), sample WS-14.
and the enthalpy of the steam is substantially higher
than that of the corresponding liquid, it is possible that
quartz formed by re-crystallization of amorphous 6. Bulk compositions of composite fragment
silica. samples
Chalcopyrite, galena, sphalerite, pyrite, and elec-
trum precipitated throughout the well column and Concentrations of SiO2, Al2O3 and Fe2O3 in
occur in both siliceous and clay-rich scales. These samples from drains and solid collectors of two-phase
minerals started to precipitate early, probably in the and re-injection lines (WS-1 to 7) range from 60 to 75
reservoir, and continued to do so in the well and wt.%, 8 to 9.5 wt.%, and 1.5 to 17 wt.%, respectively
surface pipelines. Pyrite also precipitated throughout (Table 3). The proportions of these oxides reflect the
the well column, as it replaces metallic pipe substrates dominance of amorphous silica in the samples, and to
and occurs as disseminated grains in amorphous a lesser extent, the presence of metallic pipe frag-
silica. Chalcopyrite precipitated later and/or at higher ments. Sample WS-10 is from a drain located at the
90 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96

entry of the vapour line connecting wells TR-4C and samples, but correlates strongly with Se (R 2=0.94)
TR-4B to the turbine, and has a similar composition to reflecting the control of Ag by Se-acanthite and
samples WS-1 to 7, suggesting that most fragments in geffroyite(?).
sample WS-10 were imported into the vapour line. Vapour line samples WS-11 to 14 consist mainly of
Samples WS-1 to 10 contain significant concentra- corroded pipe fragments (~97 vol.%) that appear to
tions of gold and silver; WS-2 contains the highest have formed predominantly in situ due to the highly
concentration of gold (187 ppm), and WS-10 the acidic environment, as well as minor scale and rock
highest concentration of silver (8100 ppm). Copper, fragments that were transported to the vapour line (~3
lead, and zinc concentrations are also elevated in these vol.%). The principal oxide is thus Fe2O3 (80 to 96
samples; up to ~1 wt.% Cu, ~1 wt.% Pb, and ~1.5 wt.%). The concentration of arsenic is significantly
wt.% Zn in WS-10. Gold correlates positively with higher in vapour line samples (up to 11 300 ppm) than
Cu, Pb, and Zn (R 2N0.86) in two-phase and re- in those from the two-phase/re-injection lines. Con-
injection line samples collected from wells TR-2/9 centrations of Sb are comparable in vapour and two-
(WS-1 to WS-7), which reflects the co-precipitation of phase/re-injection line fragments; the two highest
electrum with chalcopyrite, sphalerite, and galena concentrations, 240 ppm and 68 ppm, are both from
evident in textural relationships. Silver has only a samples collected from drains in vapour lines, WS-10
weak positive correlation with Au (R 2=0.23) in these and WS-14, respectively. Gold concentrations in

Table 4
Compositions of Berlı́n geothermal waters and vapour condensates (ppb) at 25 8C
Sample TR-9 TR-2 TR-4B TR-4C TR8A TR-4B/4C D.L.
Location P Wellhead P Wellhead P Wellhead P Wellhead RI Wellhead Separator
Na 3 440 000 4 230 950 2 380 000 3 070 000 4 470 000 1825 5
Mg 286 202 1230 155 118 220 1
Al 335 196 1935 89 105 9 2
Si 272 000 369 000 347 000 286 000 372 000 8160 50
K 702 000 923 000 543 000 652 000 942 000 666 10
Ca 110 000 94 200 29 600 45 900 117 000 bd/l 50
Ti 119.3 131.1 194.0 107.7 127.6 8.0 0.1
Mn 116.6 117.2 181.5 70.9 124.6 66.5 0.1
Fe 409 bd/l 8080 117 429 130 5
Ni 14.4 34.6 604.9 117.5 82.3 45.9 0.3
Cu 64.4 bd/l 119 48.3 13.2 3.6 0.2
Zn 23.1 21.1 27.4 145 86.0 17.6 0.5
As 9490 11 800 16 700 7750 12 800 10 1
Se bd/l bd/l bd/l bd/l bd/l 0.9 0.2
Mo 37.1 35.3 69.3 61.6 40.0 0.2 0.1
Ag 1.1 0.5 bd/l 0.6 0.3 bd/l 0.2
Cd 0.18 0.15 0.24 0.27 0.13 0.02 0.01
Sb 34.5 77.7 524.7 57.8 298.7 1.6 0.5
Au 0.013 0.005 0.002 0.003 0.004 0.012 0.002
Hg bd/l bd/l bd/l bd/l bd/l bd/l 0.2
Pb 10.2 15.1 57.2 21.4 85.5 0.1 0.1
*Cl
4 164 000 5 889 800 3 560 000 5 916 000 8 191 000 3040 NA
*SO4
8800 4420 2000 10 200 12 000 15 100 NA
*HCO
3 9030 5092 15 130 7320 21 000 NA NA
*pH 6.32 6.43 6.98 6.06 6.31 4.56
Abbreviations: P=production, RI=re-injection, NA=not available, D.L.=detection limit, bd/l=less than detection limit.
All data (except those noted *) by ICP-MS; * data from La Geo S.A. de C.V. Samples were acidified (pH b2) after pH was measured in the field.
Note that anion concentrations are insufficient to balance total cation charge in data columns 1 to 3, and may be under-reported. The fifth data
column represents a boiled water sample collected at re-injection well TR-8A, and the sixth data column is a vapour condensate from the TR-
4B/TR-4C separator.
 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96 91

vapour line samples WS-11 to WS-14 are appreciable 8. Discussion

(ranging from 0.4 to 4.5 ppm), although lower than in
two-phase/re-injection line fragments, and approach 8.1. Formation of aluminium-rich amorphous silica
those of some economic epithermal deposits. and clay-rich precipitates

Aluminium-rich amorphous silica scales are known

7. Fluid compositions to form in many geothermal fields around the world
from near-neutral pH, moderately saline brines (Gal-
Geothermal waters (Table 4) sampled at production lup, 1997). The average temperature at which pure
wellheads TR-9, TR-2, TR-4B, and TR-4C (at 25 8C) amorphous silica saturates during one-step adiabatic
have moderate salinity, with Cl
, SO4=, and HCO3
boiling of Berlı́n fluid is ~170 8C (Martı́nez, 1997).
concentrations averaging 4882 ppm, 6.4 ppm, and 9.1 However, aluminium-rich amorphous silica is less
ppm, respectively. The principal dissolved cations are soluble than its aluminium-free counterpart, and has
sodium (3280 ppm), potassium (705 ppm), silicon been observed to precipitate at temperatures ~25 8C
(3185 ppm), and calcium (70 ppm). All wellhead above the saturation point for the latter (Gallup,
liquids have a near-neutral pH at 25 8C, averaging 6.5. 1998b). By contrast, the average wellhead temper-
Concentrations of Au, Ag, Cu, Pb, Zn, As, and Sb ature is ~145 8C. We therefore conclude that early
average 0.006, 0.6, 58.0, 26.0, 54.2, 11435, and 173.7 aluminium-rich amorphous silica formed in the wells,
ppb, respectively. which is consistent with the mineral paragenesis.
Vapour condensates sampled at the separator The colloform texture of aluminium-rich amor-
connecting wells TR-4B and TR-4C contain much phous silica in Berlı́n wells suggests that amorphous
less Cl
(3 ppm), but considerably more SO4= (15 silica was supersaturated (cf. Rimstidt, 1997) and
ppm) than corresponding wellhead liquids. These precipitated as a gel. It is thus possible that aluminium
differences are interpreted to reflect preferential and alkalis were incorporated directly into the sili-
partitioning of sulphur species into the gas phase, ceous gel by adsorption; the gel then dehydrated after
and chlorine into the liquid phase during boiling (cf. the fragments were sampled, producing aluminium-
Drummond and Ohmoto, 1985). Concentrations of rich amorphous silica with desiccation cracks as
Na, K, Ca Si, Ag, Cu, Pb, Zn, Sb, and As are much described earlier. Alternatively, it is possible that
lower in the mechanically separated vapour than in aluminium reacts with dissolved silica and forms
corresponding wellhead liquids (TR-4B and TR4-C). precipitates through reactions (Gallup, 1997, 1998b),
The very low concentration of As is particularly such as:
surprising given the relative abundance of this
element in vapour line fragments, and may indicate 2AlðOHÞ83 þ 14SiðOHÞ4 YAl2 O3 d14SiO2 þ 31H2 O
loss of As during sampling. By contrast, the gold
concentration in vapour is considerably higher than 2AlðOHÞ

4 þ 14SiðOHÞ8YAl
4 2 O3 d14SiO2 þ 31H2 O
those in all wellhead liquids except for TR-9, and þ 2OH

about four times higher than in corresponding well-
head liquids (TR-4B and TR4-C). However, the low concentration of aluminium
The respective concentrations of Cl
, SO4=, HCO3
, (~100 to 2000 ppb; Table 4), the relatively high
Na, K, Ca, and Si in boiled water sampled at the head solubility of aluminium–silicate scale (500 ppm at 150
of re-injection well TR-8A are 3165, 5.4, 14, 635, 8C; Gallup, 1998b), and the comparable solubility of
130, 15, and 52 ppm higher than in corresponding amorphous silica scale make direct precipitation of
liquids sampled at production wellheads TR-2 and Al-bearing silica scale at Berlı́n unlikely. We therefore
TR-9. The behaviour of the trace metals is more conclude that incorporation of aluminum in amor-
irregular. Zinc, Pb, As, and Sb concentrations are 64, phous silica by adsorption is a more plausible
73, 2155, and 243 ppb higher, respectively, in the explanation for the formation of these scales. Late-
boiled water, but Cu, Ag, and Au are 51, 0.5, and stage aluminium-rich amorphous silica that rims
0.005 ppb lower than in wellhead liquids. fragments does not exhibit colloform textures, which
92 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96

may indicate lower degrees of supersaturation in the similar mineral assemblages, suggesting that such
two-phase line. fragments were transported and altered in situ by
The saponitic/vermiculitic clay is similar to that late-stage acidic liquid/vapour.
reported from the Philippines by Reyes and Cardile
(1989) and from Iceland by Kristmannsdottir (1983). 8.2. Precious and base metal deposition
The clays are massive to spherulitic, have the optical
and chemical characteristics of saponite/vermiculite Chalcopyrite, bornite, galena, sphalerite, and elec-
(relatively high Mg and low Al), and a structure similar trum have been reported from numerous geothermal
to sepiolite/montmorillonite. Kristmannsdottir (1983) fields, including Salton Sea, USA (Skinner et al.,
suggested that such clays precipitate as Fe- and Mg- 1967), Ohaaki and Rotokawa, New Zealand (Brown,
rich and Al-poor silicate gels that recrystallize to a 1986; Warren et al., 1998; Reyes et al., 2002), and
mineraloid. This hypothesis may not be applicable at Cerro Prieto, Mexico (Clark and Williams-Jones,
Berlı́n, as most clay appears to be well crystallized, 1990), where they are interpreted to have precipitated
suggesting formation near equilibrium conditions. due P to boiling-induced changes in temperature, pH
Reyes and Cardile (1989) suggested that this type of and [S]. Geochemical modelling shows that gold at
clay could form by the decomposition of bentonite Berlı́n also precipitated due to fluid boiling.
(Na-rich montmorillonite used in drilling mud), and its Log fO2-pH diagrams (Fig. 5) were constructed at
reconstitution into saponitic/vermiculitic clay. This reservoir (average 290 8C for wells TR-2 and TR-9;
hypothesis could apply at Berlı́n, as the clay has a Table 5) and wellhead temperatures (145 8C) to model
structure similar to montmorillonite. However, type 1 the deposition of gold at Berlı́n; calculations were
clay fragments exhibit a spherulitic texture, suggesting based on fluid and gas chemistry from well TR-2
that some clay precipitated in fracture and void space (Tables 4 and 6), which is the best candidate to model
in the reservoir near the well column. Furthermore, gold deposition since most siliceous scales were
type 2 clay-rich fragments clearly formed by alteration collectedPfrom well suite TR-2/TR-9. The diagrams
of the reservoir rock, as indicated by preserved textures assume [S]=0.003 and 0.0003 P m at 290 and 145 8C,
and relicts of K-feldspar, amphibole, and rutile. Clay- respectively. The values for [S] and pH (4.7) at 290
rich clasts from the vapour lines (type 3) have lower Si/ 8C (reservoir conditions) were calculated using gas
Al and higher Mg/Fe ratios than type 1 and 2 clays, but (Table 6) and liquid (Table 4) compositions for well

Fig. 5. Log fO2-pH diagrams for 290 and 145 8C at saturated vapour pressure, showing predominance fields for aqueous sulphur species (dashed
lines), stability fields for Fe–O–S minerals and bornite–chalcopyrite (solid grey lines), and solubility contours (solid black lines) for gold
(Au(HS)


2 ); the data for Au(HS)2 were taken from Benning and Seward (1996), calculated using the HCh software package (Shvarov Yu and
Bastrakov, 1999). The grey shaded bar for 290 8C (pH 4.7), and the grey shaded dot for 145 8C represent possible fO2-pH conditions for
mineralizing fluids at Berlı́n.
 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96 93

Table 5 WS-1 to 7. These metals are assumed to be

Calculated reservoir temperatures (8C) for studied wells using transported predominantly as chloride complexes,
various cation geothermometers
and to have deposited in response to the increase in
Well (Na–K) (Na–K) (Na–K) (Na–K) (Na–K) T average
pH and decrease in temperature that accompanied
1 2 3 4 5
boiling.
TR-9 281 296 287 286 280 286
TR-2 291 303 294 293 291 294
TR-4B 298 308 300 298 298 300 8.3. Comparison of Berlı́n to epithermal precious
TR-4C 287 301 291 290 286 291 metal systems
1: T (8C)=933/[0.993+log (Na/K)]
273.15 (Arnórsson et al.,
1983). Base and precious metal concentrations in Berlı́n
2: T (8C)=1390/[1.75+log (Na/K)]
273.15 (Giggenbach, 1988). wellhead fluids are generally lower than those
3: T (8C)=1217/[1.483+log (Na/K)]
273.15 (Fournier, 1979). reported from many other geothermal fields (e.g.,
4: T (8C)=1289/[1.615+log (Na/K)]
273.15 (Verma and Santoyo,
Gallup, 1998a,b; Henley, 1984). Nevertheless,
1997).
5: T (8C)=856/[0.857+log (Na/K)]
273.15 (Truesdell, 1976). Berlı́n fluids are actively precipitating Au and Ag
in scales and reservoir rocks in concentrations
comparable to those of economic ore deposits,
and the siliceous scales forming in the wells are
TR-2, adjusted to represent equilibrium with quartz,
analogous to low-sulphidation epithermal veins.
sericite, epidote, and chlorite (which occur in the
Amorphous silica precipitates instead of quartz
reservoir; GESAL, 2001), using the programs SOL-
because the Berlı́n geothermal fluids are super-
VEQ
P and CHILLER (Reed, 1982). The values for saturated and deposition occurs rapidly. The
[S] and pH (5.7) at 145 8C (wellhead conditions)
absence of adularia and sericite in Berlı́n wells,
were assumed to decrease and increase, respectively,
which are common in low-sulphidation epithermal
from the reservoir values by 1 log unit due to boiling
systems (Cooke and Simmons, 2000), is likely due
of fluids. Gold was assumed to be transported as
to the incorporation of Al and K by amorphous
Au(HS)2
, the most stable complex for conditions in
silica. Although smectite is common in alteration
the well (pH ranges from ~4.7 to ~5.7 from reservoir
assemblages formed by weakly acidic fluids in
to wellhead, respectively), and to precipitate accord-
hydrothermal systems, it generally does not host
ing to the reaction:
much mineralization. However, significant Au and
AuðHSÞ
þ
2 þ H þ 1=2H2 OYAuðsÞ þ 2H2 SðaqÞ
Ag precipitated in the saponitic/vermiculitic clay
þ 1=4O2 from the reservoir at Berlı́n. Such saponitic/vermic-
ulitic clays appear to be unique to geothermal
In the reservoir, the upper and lower limits for environments, and may form in response to the
fO2 (between
30 and
35) are constrained by the extremely accelerated boiling process around the
pyrite–hematite and pyrite–pyrrhotite boundaries, for wells.
which calculated gold solubilities range from 1 to
10 ppb. At the wellhead, fO2 (near
48) is 8.4. Vapour transport of metals
constrained by the co-existence of bornite, pyrite,
and chalcopyrite, and the calculated gold solubility One of the most interesting features of the
is b1 ppb. The decrease P in temperature, increase of Berlı́n scales is the occurrence of appreciable Au,
pH, and decrease in [S] due to boiling reduced
gold solubility sufficiently to cause its precipitation
in well scales at depth. Other metals, such as Ag, Table 6
Cu, Pb, and Zn, are also interpreted to have Dry gas compositions for well TR-2 (mol%)
precipitated due to boiling, as electrum co-precipi- Ar H2 N2 CH4 H2S He CO2 NH3 Ys XH2O
tated with sphalerite, galena, and chalcopyrite, and 0.214 1.186 2.068 0.209 9.881 0.196 86.697 0.091 0.276 0.997
gold and silver concentrations correlate positively Data obtained from La Geo S.A. de C.V.
with those of Cu, Pb, and Zn in composite samples Abbreviation: Ys=steam fraction.
94 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96

As, and Sb in composite samples from the vapour and electrum-bearing aluminium-rich amorphous
lines. Silver, Cu, Pb, and Zn also occur in silica scale and saponitic/vermiculitic clay precip-
significant concentrations in vapour line samples, itates. Siliceous scales formed in the wells as a gel,
but appear to reflect deposition prior to transport of which incorporated aluminium and alkalis by adsorp-
the fragments into the vapour lines. Greater tion. Clay precipitated in the reservoir, filling void
concentrations of Au, As, and Sb in samples space, and altering reservoir rock. Scale, rock, and
further along the vapour line from wells TR-4B/ corrosion fragments were cemented by a second
4C (compare WS-14 with WS-11 to 13; Fig. 2B) generation of amorphous silica, after transport into
suggest that some Au, As, and Sb were precipitated two-phase and re-injection lines. Gold deposited as
directly from vapour. Arsenic in particular has been electrum in well scales due to decreases in temper-
previously reported to be precipitated from vapour ature, H2S depletion, and increases in pH that
in geothermal systems (e.g., Reyes et al., 2002). accompanied fluid boiling.
The fluid chemistry at Berlı́n supports the idea that Fragments transported into the vapour lines and in
significant gold is also transported in the vapour, as situ pipe linings were altered by steam acidity.
the condensate sample taken at the separator Antimony and arsenic were transported in the vapour
contains about four times as much Au as the line due to their relatively high volatilities, and
corresponding production wellhead liquids (TR-4B/ precipitated as tetrahedrite (after bornite) in trans-
4C). Arsenic and antimony precipitated as an As– ported scales, and as an As–S-bearing iron phase in
S-bearing iron phase (sample WS-14) and as altered pipe fragments. Gold was transported into the
tetrahedrite (sample WS-10), respectively; the form vapour lines as a hydrated gas species by the slightly
of precipitated gold in the vapour line samples is acidic chloride vapour, as indicated by higher
unknown. concentrations of Au in condensed steam relative to
Volatility data (cf. Williams-Jones et al., 2002) wellhead fluids.
indicate that As and Sb are the most volatile ore- This study of mineralized fragments from drains
related metals after Hg, and suggest that these and solid collectors of the Berlı́n field shows that
metals were transported by vapour. By contrast, precious metal-bearing siliceous well scales are
Au has very low volatility suggesting that it should analogous to epithermal veins and are deposited by
have negligible solubility in the vapour. However, boiling. The significant Au, As, and Sb concentrations
experiments by Archibald et al. (2001) indicate that in vapour line fragments suggest that geothermal
hydration greatly enhances Au solubilities in aque- vapour pipelines provide an active environment in
ous vapour, and that significant concentrations can which to study the role of vapour in metal transport.
be transported as hydrated chloride complexes. Future research in such systems may furnish addi-
Considering the importance of H2S in enhancing tional information with which to evaluate mineral
Au solubility in aqueous liquids, and the fact that deposit models.
H2S partitions strongly into vapour, it is also likely
that complexes involving H2S play an important role
in increasing Au solubility in aqueous vapour. Acknowledgements
Finally, we note that the evidence presented above
for the transport of gold in vapour is consistent with We wish to acknowledge support from Ing. Luz A
reports of high gold concentrations in vapour-rich Barrios, Elizabeth de Enriquez and La Geo S.A. de
fluid inclusions (Ulrich et al., 1999) in some ore C.V. during this study. Funding was provided by
deposits. grants to A.E. Williams-Jones from NSERC and
FQRNT, and to J. Raymond from GEOTOP, the
Society of Economic Geologists, and the Geological
9. Conclusions Society of America. We are also grateful for the
comments of reviewers D. Gallup and T. Ulrich
Mineralized fragments originating in the Berlı́n which significantly improved the manuscript. Special
geothermal field are composed mainly of sulphide- thanks go to S. Archibald who provided valuable
 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96 95

help and guidance, and to E. Labrecque who helped GESAL, 2002. Historia sobre anàlisis mineralógico en muestras
initiate the project in El Salvador. colectadas en el sistema de acarreo y generación, central
geotérmica de Berlı́n. Unpubl. report, La Geo S.A. de C.V.
10 pp.
Giggenbach, W.F., 1988. Geothermal solute equilibria. Derivation
References of Na–K–Mg–Ca geoindicators. Geochim. Cosmochim. Acta
52, 2749 – 2765.
Arnórsson, S., Gunnlaugsson, E., Svavarsson, H., 1983. The Henley, R.W., 1984. Metals in hydrothermal fluids. In: Robertson,
chemistry of geothermal waters in Iceland: III. Chemical J.M. (Ed.), Fluid–Mineral Equilibria in Hydrothermal Systems,
geothermometry in geothermal investigations. Geochim. Cos- Reviews in Economic Geology, vol. 1. Society of Economic
mochim. Acta 47, 567 – 577. Geology, Boulder, CO, pp. 115 – 129.
Archibald, S.M., Migdisov, A.A., Williams-Jones, A.E., 2001. The Kristmannsdottir, H., 1983. Chemical evidence from Icelandic
stability of Au-chloride complexes in water vapour at elevated geothermal systems as compared to submerged geothermal
temperatures and pressures. Geochim. Cosmochim. Acta 65, systems. In: Rona, P.A., Bostrom, K., Laubier, L., Smith, K.L.
4413 – 4423. (Eds.), Hydrothermal Processes at Seafloor Spreading Centers.
Benning, L.G., Seward, T.M., 1996. Hydrosulphide comple- Plenum Press, New York, pp. 291 – 320.
xing of Au(I) in hydrothermal solutions from 150–400 Krupp, R.E., Seaward, T.M., 1987. The Rotokawa geothermal
degrees C and 500–1500 bar. Geochim. Cosmochim. Acta system, New Zealand; an active epithermal gold-depositing
60, 1849 – 1871. environment. Econ. Geol. 82, 1109 – 1129.
Brown, K.L., 1986. Gold deposition from geothermal discharges in Martı́nez, F.M.A., 1997. Chemical Interpretation and Scaling
New Zealand. Econ. Geol. 81, 983 – 997. Potential of Fluid Discharge from Wells in the Berlı́n
Carr, M.J., Stoiber, R.E., 1990. Volcanism. In: Dengo, G., Case, Geothermal Field, El Salvador. Geothermal Training Program
J.E. (Eds.), The Geology of North America, The Caribbean Report, vol. 12. United Nations University, pp. 295 – 309.
Region, vol. H. Geological Society of America, Boulder, CO, Monterrosa, V.M.E., 1993. A 3-D Natural State Modeling and
pp. 375 – 391. Reservoir Assessment for the Berlı́n Geothermal Field in El
CEL, 1999. Informe final de caracterización microscopia y Salvador, CA. Geothermal Training Program Report, vol. 11.
correlaciones litoestratigraficas, pozos perforandos 1997–1999, United Nations University. 45 pp.
campo geotérmico de Berlı́n. Unpubl. report, La Geo S.A. de Raymond, J., Williams-Jones, A.E., Clark, J.R., 2003. Precipitation
C.V. 46 pp. of precious- and base metal-bearing scales at the Berlin
Clark, J.R., Williams-Jones, A.E., 1990. Analogues of epithermal geothermal field, El Salvador. Trans.-Geotherm. Resour. Counc.
gold–silver deposition in geothermal well scales. Nature 346, 27, 331 – 334.
644 – 645. Reed, M.H., 1982. Calculation of multicomponent chemical
Cooke, D.R, Simmons, S.F., 2000. Characteristics and genesis of equilibria and reaction processes in systems involving minerals,
epithermal gold deposits. In: Hagemann, S.G., Brown, P.E. gases, and an aqueous phase. Geochim. Cosmochim. Acta 46,
(Eds.), Gold in 2000, Reviews in Econ. Geol. vol. 13. Society of 513 – 528.
Economic Geology, Boulder, CO, pp. 221 – 244. Reyes, A.G., Cardile, C.M., 1989. Characterization of clay
d’Amore, F., Mejia, J.T., 1999. Chemical and physical scales forming in Philippine geothermal wells. Geothermics
reservoir parameters at initial conditions in Berlı́n geo- 18, 429 – 446.
thermal field, El Salvador: a first assessment. Geothermics Reyes, A.G., Trompetter, W.J., Britten, K., Searle, J., 2002. Mineral
28, 45 – 73. deposits n the Rotokawa geothermal pipelines, New Zealand. J.
Drummond, S.E., Ohmoto, H., 1985. Chemical evolution and Volcanol. Geotherm. Res. 119, 215 – 239.
mineral deposition in boiling hydrothermal systems. Econ. Geol. Rimstidt, J.D., 1997. Gangue mineral transport and deposition. In:
80, 126 – 147. Barnes, H.L. (Ed.), Geochemistry of Hydrothermal Ore Depos-
Fournier, R.O., 1979. A revised equation for the Na/K geo- its, 3rd edition. J. Wiley & Sons, New York, pp. 487 – 516.
thermometer. Trans.-Geotherm. Resour. Counc. 3, 221 – 224. Santos, L., 1995. One and Two Dimensional Interpretations of
Gallup, D.L., 1997. Aluminum silicate scale formation and DC-resistivity Data from the Berlı́n Geothermal Field, El
inhibition (1): scale characterization and laboratory experiments. Salvador. Geothermal Training Program Report, vol. 11. United
Geothermics 26, 483 – 499. Nations University, pp. 269 – 302.
Gallup, D.L., 1998a. Geochemistry of geothermal fluids and well Shvarov Yu, V., Bastrakov, E.N., 1999. HCh: Software Package for
scales, and potential for mineral recovery. Ore Geol. Rev. 12, Geochemical Equilibrium Modeling. Record 1999/25. Austral-
225 – 236. ian Geological Survey Organisation, pp. 1 – 61.
Gallup, D.L., 1998b. Aluminum silicate scale formation and Skinner, B.J., White, D.E., Rose, H.J., Mays, R.E., 1967. Sulphides
inhibition (2): scale solubilities and laboratory and field associated with the Salton Sea geothermal brine. Econ. Geol. 62,
inhibition tests. Geothermics 27, 485 – 501. 316 – 330.
GESAL, 2001. Sı́ntesis y modelo conceptual del campo Truesdell, A.H., 1976. Geochemical Techniques in Exploration.
geotérmico de Berlı́n. Unpubl. report, La Geo S.A. de Proceedings of the 2nd United Nations Symposium on the
C.V. 22 pp. Development and Use of Geothermal Resources, pp. 3 – 29.
96 J. Raymond et al. / Journal of Volcanology and Geothermal Research 145 (2005) 81–96

Ulrich, T., Gqnther, D., Heinrich, C.A., 1999. Gold concentrations Wiesemann, H.G., 1975. Remarks on the geologic structure of the
of magmatic brines and the metal budget of porphyry copper republic of El Salvador, Central America. Mitt. Geol.-Paleontol.
deposits. Nature 339, 676 – 679. Inst. Univ. Hamb. 44, 557 – 574.
Verma, S.P., Santoyo, E., 1997. New improved equations for Na/K, Williams-Jones, A.E., Migdisov, A.A., Archibald, S.M., Xiao,
Na/Li and SiO2 geothermometers by outlier detection and Z., 2002. Vapour-transport of ore metals. In: Hellmann, R.,
rejection. J. Volcanol. Geotherm. Res. 79, 9 – 23. Wood, S. (Eds.), Water–Rock Interactions, Ore Deposits and
Warren, N.A., Mauk, J.L., Brown, K.L., 1998. Mineralogy of Environmental Geochemistry: A Tribute to David. A. Crerar,
Sulfide Scale Deposited in Geothermal Pipelines, Ohaaki, New Special Publication, vol. 7. Geochemical Society, Houston, TX,
Zealand. AusIMM Annual Conference vol. 31, pp. 135 – 144. pp. 279 – 306.
Weissberg, B.G., 1969. Gold–silver ore-grade precipitates from
New Zealand thermal waters. Econ. Geol. 64, 95 – 108.

View publication stats