Ore Geology Reviews 39 (2011) 63–74

Contents lists available at ScienceDirect

Ore Geology Reviews

j ourna l homepage: www.e lsev ie r.com/ locate /oregeorev

Magmatic-vapor expansion and the formation of high-sulfidation gold deposits:Chemical controls on alteration and mineralization

Richard W. Henley a, Byron R. Berger b,⁎a Research School of Earth Sciences, Australian National University, Canberra, Australiab U.S. Geological Survey, Federal Center MS 964, Denver, CO 80225-0046, United States

⁎ Corresponding author.E-mail address: bberger@usgs.gov (Byron Berger).

0169-1368/$ – see front matter. Published by Elsevierdoi:10.1016/j.oregeorev.2010.11.003

a b s t r a c t

a r t i c l e i n f o

Article history:Received 14 July 2010Received in revised form 8 November 2010Accepted 8 November 2010Available online 23 November 2010

Keywords:High-sulfidationGoldMagmatic vaporEnargiteSilica-aluniteSulfosalt meltTennantiteSolfatara

Large bulk-tonnage high-sulfidation gold deposits, such as Yanacocha, Peru, are the surface expression ofstructurally-controlled lode gold deposits, such as El Indio, Chile. Both formed in active andesite–dacitevolcanic terranes. Fluid inclusion, stable isotope and geologic data show that lode deposits formed within1500 m of the paleo-surface as a consequence of the expansion of low-salinity, low-density magmatic vaporwith very limited, if any, groundwater mixing. They are characterized by an initial ‘Sulfate’ Stage of advancedargillic wallrock alteration±alunite commonly with intense silicification followed by a ‘Sulfide’ Stage — asuccession of discrete sulfide–sulfosalt veins that may be ore grade in gold and silver. Fluid inclusions inquartz formed during wallrock alteration have homogenization temperatures between 100 and over 500 °Cand preserve a record of a vapor-rich environment.Recent data for El Indio and similar deposits show that at the commencement of the Sulfide Stage,‘condensation’ of Cu–As–S sulfosalt melts with trace concentrations of Sb, Te, Bi, Ag and Au occurred atN600 °C following pyrite deposition. Euhedral quartz crystals were simultaneously deposited from the vaporphase during crystallization of the vapor-saturated melt occurs to Fe-tennantite with progressivenon-equilibrium fractionation of heavy metals between melt-vapor and solid. Vugs containing a range ofsulfides, sulfosalts and gold record the changing composition of the vapor.Published fluid inclusion and mineralogical data are reviewed in the context of geological relationships toestablish boundary conditions through which to trace the expansion of magmatic vapor from source tosurface and consequent alteration andmineralization. Initially heat loss from the vapor is high resulting in theformation of acid condensate permeating through the wallrock. This Sulfate Stage alteration effectivelyisolates the expansion of magmatic vapor in subsurface fracture arrays from any external contemporaryhydrothermal activity. Subsequent fracturing is localized by the embrittled wallrock to provide high-permeability fracture arrays that constrain vapor expansion with minimization of heat loss. The Sulfide Stagevein sequence is then a consequence of destabilization of metal-vapor species in response to depressurizationand decrease in vapor density.The geology, mineralogy, fluid inclusion and stable isotope data and geothermometry for high-sulfidation,bulk-tonnage and lode deposits are quite different from those for epithermal gold–silver deposits such asMcLaughlin, California that formed near-surface in groundwater-dominated hydrothermal systems wheremagmatic fluid has been diluted to less than about 30%. High sulfidation gold deposits are better termed‘Solfataric Gold Deposits’ to emphasize this distinction. The magmatic-vapor expansion hypothesis alsoapplies to the phenomenology of acidic geothermal systems in active volcanic systems and equivalentmagmatic-vapor discharges on the flanks of submarine volcanoes.

B.V.

Published by Elsevier B.V.

1. Introduction

High-sulfidation gold deposits range in scale from small, high-gradelodedeposits tovery large, low-gradedeposits. Theyhave in commonanenvelope of pre-ore advanced argillic alteration assemblages – Sulfate

Stage – that are commonly, but not exclusively, characterized byoxidized sulfur as sulfate in alunite (KAl3(SO4)2(OH)6), and a subsequentcontrasting sequence of ore-stage sulfide assemblages – Sulfide Stage –

that typically include pyrite, a range of sulfosalt minerals (includingenargite and tennantite), gold, and silver (Arribas, 1995; Berger andHenley, 2011).

Since the late 1990s, a number of studies of high-sulfidation gold–silver deposits have consistently shown that the temperatures recordedby fluid inclusions in quartz in the Sulfate Stage alteration assemblages

64 R.W. Henley, B.R. Berger / Ore Geology Reviews 39 (2011) 63–74

range from about 100 to over 500 °C and that vapor-rich inclusions areabundant (see Section 2.1.1). In turn recent FESEM data for sulfosaltassemblages from El Indio, Summitville and Lepanto (Mavrogenes et al.,2010) strongly suggest near-magmatic temperatures for the initialcopper-rich ore stage of high-sulfidation deposits (see Section 2.1.2).These data show that, following initial pyrite deposition, arsenic-rich

Fig. 1. Euhedral quartz and Fe-tennantite (Cu10.6, Fe1.4As3.9Sb0.1S13; Tn 0–5%, Fe 0.8 to 2.Fe-tennantite (Cu10.6Fe1.1As4S13) infill a void in concentrically banded and fractured pyriteFe-tennantite around quartz from Tn 4.0 to 77% (bright zones), d) curved interfaces between lthe vug-rich rims of the low Tn phase, the distribution of vugs (black) and euhedral quartFe-tennantite (Cu10FeAs4S13) with transient development of a metal-deficient chalcopyritegrowing in an isolated vug within enargite. Abbreviations: Qz, quartz, Fe–Tn, Fe-tennantite (location data: El Indio; EI1300, Enargite vein, DDH1300, 490.5 m, EI4216F; DDH4216, 27.7sample, Longfellow Mine, Red Mountain, Silverton, Colorado). Images: a), b) and e) are reflecimage.

sulfosaltmeltswith trace contents of other heavymetals (tin, silver, gold,bismuth, selenium, and tellurium)were formed from a vapor phase. Thehigh-temperature melts partially crystallized first to enargite andFe-tennantite and distinctive euhedral quartz crystals (Fig. 1a–c)whose oxygen isotope composition (Mavrogenes and Henley, unpub-lished data) is consistentwith that demonstrated formagmaticfluid at El

2 at.%) infill a void and associated fractures in earlier pyrite, b) euhedral quartz and, c) euhedral quartz fills in a void in earlier pyrite and with fractional crystallization ofow-antimony Fe-tennantite (Tn 3.4%) and fractionated Fe-tennantite (Tn 4.0–77%). Notez, and the minor Sb-enrichment adjacent to one vug, e) dissolution of early pyrite bycorona and isolated bornite and with quartz infilling void space, f) emplectite (CuBiS2)numbers are atomic percent Sb/(As+Sb)), ccp, chalcopyrite, Py, pyrite, v, vug. (Samplem; Lepanto; L4=3.29.3 of Mancano (1994, Fig. 8); Red Mountain, CO6, Undergroundted light images, c) and d) are ASB FESEM images and f) is a BSE in-lens detector FESEM

65R.W. Henley, B.R. Berger / Ore Geology Reviews 39 (2011) 63–74

Indio (Jannas, 1995). Fractionationofheavymetals (Sb, Bi, Ag, Te, andAu)occurred as the residual melt crystallized in the presence of vapor (Fig. 1c–d) anddissolved earlier pyrite (Fig. 1e) and enargite.Metal transport inthe vapor phase is recorded through the preservation of micron-scalevugs containing awide range of discrete sulfides and sulfosalts includingtennantite and tetrahedrite, chalcocite, bismuthinite and emplectite(CuBiS2) (Fig. 1f) as well as quartz, fluorapatite and gold (Mavrogenes etal., 2010).

By contrast the homogenization data for fluid inclusions inenargite at Lepanto, Philippines, suggested amuch lower temperaturerange (less than 310 °C) for the ore stage (Mancano, 1994; Mancanoand Campbell, 1995). Although, as determined by Hedenquist et al.(1998, p. 386), no primary fluid inclusions were found in the Lepantoenargite or quartz samples, the acceptance of these fluid inclusionhomogenization data (b310 °C) as recording the temperature ofmineralization forced the conclusion that substantial system-scalecooling had occurred between the alteration and ore stages(Hedenquist et al. (1998, Fig. 19). In the absence of independenttemperature estimates for the Sulfide Stage, this hypothesis wassubsequently adsorbed into a number of generic models for theformation of high-sulfidation gold deposits (e.g. Heinrich, 2005;Muntean and Einaudi, 2001). However a number of fundamentaldifficulties with the geology, physics and chemistry attend thesemodels, not least of which is accounting for how the fluid in such largescale hydrothermal systems lost some 75% of its heat between itsevolution at magmatic temperatures (NN750 °C) and the assumedtemperature of mineralization (b310 °C). Each of these modelsimposes critical assumptions about pressure as well as temperaturein the depositional regime of high sulfidationmineral systems that arenot sustained by critical review of available geochemical and,crucially, geologic data. Neither can these models accommodate thenew high temperature data for the Sulfide Stage so that a freshapproach is necessary.

In this paper, in order to set boundary conditions, we first review theinterpretation of available fluid inclusion data in the context of thegeology and geochemistry of high-sulfidationdeposits.We thenprovidea new hypothesis for the origin of these important gold deposits basedon the expansion of a magmatic-vapor phase under low-pressureconditions (see Section 3) that is fully consistent with the geologicalrelationships (Berger and Henley, 2011) and stable isotope data andwith the active volcanic context in which these deposits formed.

2. The environment of formation of high-sulfidation lode-golddeposits

The advanced argillic alteration and ore stages in high-sulfidationlode deposits track active faults and branch fractures throughintrinsically low-permeability host rocks such as marine sedimentsequences or ash-flow tuffs. These fracture arrays crosscut host rocksthat commonly contain evidence of previous or concurrent weak,near-neutral pH groundwater-related alteration indicating that largerscale groundwater-dominated geothermal systems may have operatedcontemporaneously with the formation of the high-sulfidationmineralization. Based on field data, Berger and Henley (in press)suggested that the aggressive silica–alunite1 alteration localized bysyn-hydrothermal fault arrays effectively isolatedmagmatic fluid flow infractures from any external hydrothermal system—a discretization that

1 Silica–alunite. We are concerned in this paper with the reactions between acidiccondensate and rock during alteration and subsequent cooling in the Sulfate Stage. Asa convenience, we shall use the generic term “silica–alunite” to refer to the range ofadvanced argillic alteration assemblages (Meyer and Hemley, 1967) that areencountered in the high-sulfidation mineral districts whether or not alunite is aproduct.

characterizes acidic geothermal systems in active volcanic settings today(e.g., Ruaya et al., 1992) (see Section 5.2).

Sulfur isotope data consistently emphasize the predominance ofmagma-sourced fluid in both the alteration and mineralization stages(e.g., Rye, 1993, 2005) of these deposits. Other stable isotope and fluidinclusion data are also consistentwith a high-level volcanic environmentand the predominance of a low-salinity (b5 wt.% NaCleq) magmatic fluid(Bethke et al., 2005; Jannas et al., 1999; Rye et al., 1992). The involvementof groundwater has been investigated in a number of 18O–D studies, butits proportion in the ore setting is generally much less than 25% and, insome cases, is close to zero (Bethke et al., 2005; Deyell et al., 2004, 2005;Hedenquist et al., 1994, 1998; Jannas et al., 1999; Rye, 1993; Vennemannet al., 1993). This is the reverse of the relative proportions of magmaticfluid and groundwater in high-level adularia–sericite (so-called ‘lowsulfidation’ gold–silver) and analogous submarine massive sulfidesystems. In high-sulfidation deposits, evidence for the involvement ofgroundwater (and seawater in coastal settings) is strongest fordistinctive alunite veins and barite that post-date the sulfosalt-goldstage in many deposits and record waning thermal activity.

2.1. Constraints on temperature

In order to set boundary constraints on, for example, fluidtemperature and pressure, for themodeling of the physics and chemistryof the common and characteristic silica–alunite advanced argillic andsulfide stages of high-sulfidation gold deposits, it is first necessary tocarefully review the interpretation of fluid inclusion and other datawithin their geologic context. The difficulties in interpreting fluidinclusion data for deposits formed at shallow paleo-depths (seeSection 2.2) are well known and primarily relate to uncertainties withrespect to the timing of trapping relative to primary mineral deposition,the extent of cooling and involvement of groundwater (e.g., Foley et al.,1989) and as discussed below, phase changes, in the hostmineral, duringthe cooling stages of alteration and mineralization.

2.1.1. Silica–alunite stage temperaturesThere is general agreement that the silica–alunite stage of intense

leaching and alteration results from the in situ formation of acidiccondensate derived from magmatic vapor (Arribas, 1995; Chouinardet al., 2005; Hedenquist et al., 1994, 1998; Sillitoe and Hedenquist,2003; Stoffregen, 1987; Vennemann et al., 1993). At Chinkuashih,Wang et al. (1999) noted an abundance of vapor-rich inclusions inSulfate Stage quartz samples. At Furtei, Sardinia, Ruggieri et al. (1997)also comment on the abundance of vapor-rich inclusions, and report apopulation of distinctive high-salinity, high-temperature inclusions(390°–500 °C) aswell as lower temperature populations. At Summitville,Bethke et al. (2005) similarly report an abundance of vapor-richinclusions with homogenization temperatures greater than 500 °C.Hedenquist et al. (1994) also reported an abundance of vapor-richinclusions in alteration assemblages in theNansatsu district, Japan, as didJannas et al. (1999) in alunite in the Campaña fracture-vein system at ElIndio. Nevertheless, homogenization temperature data for vapor-richinclusions2 have been largely ignored in the interpretation of thesedeposits, even though their abundance confirms a vapor-richenvironment and a wide range of temperatures and temperaturegradients within the alteration zones during their development.

The mineralogy of the Sulfate Stage has been extensivelydocumented (e.g., Heald et al., 1987; Hedenquist et al., 1994, 1998;Jannas, 1995; Stoffregen, 1987). As well as alunite and a range of silicapolymorphs including quartz and chalcedony, it comprises theminerals that are typical of advanced argillic alteration; namely,

2 Homogenization temperature data for vapor-rich inclusions provide minimumtrapping temperatures for the inclusions.

66 R.W. Henley, B.R. Berger / Ore Geology Reviews 39 (2011) 63–74

kaolinite, pyrophyllite, diaspore, andalusite and disseminated pyriteand in rare instances, corundum (e.g., Famatina, Argentina, Losada-Calderon and Bloom, 1990). Alunite itself is stable up to at least 450 °C(Stoffregen et al., 1994) in accord with temperatures as high as 500 °Cdetermined through fluid inclusion measurements in silica–alunitealteration zones in many deposits despite the frequent assumptionthat the upper stability limit of alunite is 300 °C (e.g., Muntean andEinaudi, 2001). The alteration assemblage ormineralogical proportionscan change with depth. At El Indio, for example, pyrophyllitepredominates over alunite below 4000 m asl (Jannas et al., 1999) andat Chinkuashih, alunite simply disappears at deep levels (Sakazaki et al.,1964).

Hemley et al. (1980) experimentally determined the phaserelationships of kaolinite, pyrophyllite, diaspore, andalusite, andcorundum as a function of temperature, pressure and the activity ofsilica in solution (Fig. 2). During advanced argillic alteration, thesephases may be substituted by alunite (Hemley et al., 1969) in thepresence of alkalis derived from volcanic glass or feldspar in the hostrock, and high sulfate and hydrogen ion concentrations derived fromefficient partitioning of highly soluble SO2 and HCl from the vapor intothe condensate fraction. When a strongly acidic condensate reactswith volcaniclastic or sedimentary rocks, it results in high silicaactivity in the liquid phase (Africano et al., 2002; Fournier, 1985) suchthat metastable polymorphs of silica precipitate with alunite orpyrophyllite, or with kaolinite (Fig. 2). In addition silica is added to thewallrock during alteration since the solubility of quartz in the vaporphase in themain fracture is, at 500 bars and 600 °C, about 1000 mg/kg(Morey and Hesselgesser, 1951; Fournier and Potter, 1982). When asmall condensate fraction is formed in response toheat loss, partitioningof silica to the liquid condensate leads immediately to very high silicaconcentrations that drive very rapid silica deposition (Fig. 2). Duringcooling of the alteration assemblage in the presence of condensatetrapped in pore space, as well as later with the infiltration ofgroundwater, progressive recrystallization of metastable silica occursthrough a sequence of polymorphs and ultimately to quartz (Africanoand Bernard, 2000; Okamoto et al., 2010). This sequence of processes

Fig. 2. Stability relations in the system Al2O3–SiO2–H2O at PH2O=250 bars with theaddition of silica activity-temperature relations for the common polymorphs of quartzto temperatures well above 300 °C (Okamoto et al., 2010). Deposition of silica occursinitially as metastable silica polymorphs that subsequently recrystallize to quartzduring cooling of the alteration assemblage with development of pseudo-secondaryfluid inclusions with a wide range of homogenization temperatures from ~100 toN500 °C. Point MH shows, for illustration, the concentration of silica when a liquidcondensate fraction of 10% separates from the vapor at 550 °C based on the vapor phasesolubility data of Morey and Hesselgesser (1951). Constructed from GeochemistsWorkbench data kindly provided by James Cleverley.

results in massive, highly silicified rocks, including “vuggy silica”textured rocks of the type described by Stoffregen (1987) at Summit-ville, Jannas et al. (1999) at El Indio and Wang (2010) at Chinkuashih.Consequently, the trapping of inclusion populations takes place over anextended period so that awide range of temperatures are recorded (e.g.,Bethke et al., 2005, p. 290). Furthermore, since alteration progressivelyreduces permeability, the altered rock becomes both a barrier to theingress of groundwater and, at the same time, an effective insulatorwithrespect to the dispersion of heat away from the fluid phase that ischanneled through the principal controlling fractures (Berger andHenley, 2011). The temperature gradients within such insulatingwallrock environments are consequently very large, an attribute that isreflected in the wide range of fluid inclusion homogenization data(N500 to 100 °C), from alteration assemblages in which silicarecrystallized.

During cooling, redistribution and recrystallization of silica in thealtered wallrock occur where available water penetrates or wets thesurface of the primary and successive silica precipitates.Fluid inclusions in quartz in this evolving assemblage do not thereforenecessarily record the presence of a continuum of liquid water or aboiling-point depth relationship through which to converthomogenization and freezing data to an estimate of depth. Ratherthe pore pressure during silica recrystallization is more reliablyrelated to lithostatic pressure (see Section 2.2). Importantly, sincemost of the homogenization temperature data record trapping duringcooling fom N500 °C the primary deposition temperature of the hostassemblage was necessarily hotter.

2.1.2. Sulfide Stage temperaturesIt is demonstrable that it is Sulfate Stage embrittlement of the

wallrock that localizes subsequent fracturing and consequent SulfideStage lode mineralization (Berger and Henley, 2011). The latter isassociated with little recognizable wallrock alteration other than, forexample, isolated replacement of early alunite by illite–sericite at ElIndio (Jannas, 1995). In most deposits, pyrite (±solid solution goldand silver; e.g., Deditius et al., 2009) is the first sulfide to precipitate inlodes and is followed by enargite and Fe-tennantite (±minorchalcopyrite and base-metal sulfides) assemblages that also containdistinctive euhedral quartz crystals (Fig. 1) (Mavrogenes et al., 2010;Mavrogenes and Henley, in preparation). Later stages of sulfide–sulfosalt deposition are characterized by more complex textures andmineral suites that include some tellurides, selenides and bismuthminerals as well as ore-grade silver and gold (e.g., Claveria, 2001;Jannas, 1995). The textural and compositional data (Fig. 1) also showthat high-sulfidation sulfide–sulfosalt assemblages do not showequilibrium relationships nor are they analogous to the sulfideassemblages encountered in epithermal adularia–sericite typedeposits.

Sulfide stage mineral assemblages are not amenable to standardmethods of transmitted light fluid inclusion analysis so that infraredtechniques have been progressively developed for the measurementof homogenization temperatures and salinities in enargite andsometimes pyrite. Moritz (2006) has cautioned that many infrareddeterminations of inclusion fluid salinity and homogenizationtemperatures may be in error and has since published new data forenargite sampled from a number of deposits in Europe and Peru(Moritz and Benkhelfa, 2009). These liquid-phase inclusion datadefine a range of salinities from very low to less than 5 wt.% NaCleqand have a range of homogenization temperatures from about150–300 °C. Kouzmanov et al. (2010) have reported a similar rangeof fluid inclusion homogenization temperatures for well-formedenargite crystals in drusy late-stage cavities in minor post-porphyryveins from the Rosia Poieni porphyry copper deposit, Romania.

The recrystallization of sulfosalts to complex sulfosalt assemblagestogether with a wide variety of tellurides (e.g., goldfieldite) and nativegold, during cooling to below about 300 °C, has been previously

Fig. 3. Phase relations in the system Cu–As–Sb–S redrawn from Posfai and Buseck(1998). The range of compositions of coexisting phases in the enargite–famatinitesystem at Chinkuashih, Goldfield and El Indio is indicated (Huang and Chou, 1975;Huang and Tsai, 1984; Mavrogenes et al., 2010; Sung, 2005). An estimate of the rangefor Lepanto ‘enargite–luzonite’ is provided based on a smelter concentrate assay(Filippou et al., 2007) and our unpublished FESEM data in order to compare with themaximum and minimum ranges of fluid inclusion homogenization data from enargite(Mancano, 1994). For comparison the compositional range of high temperaturevolcanic gas condensates from fumaroles with discharge temperatures N600 °C isshown based on the data compiled by Williams-Jones et al. (2002). The average crustalantimony and arsenic concentrations of granite and basalt are shown for reference.Abbreviations; En, enargite, luz, luzonite, Fam, Famatinite, hcp, hexagonal close packed,ccp, cubic close packed, int, interlayered polytype.

4 Fluid-phase terminology. The term “hydrothermal” has a commonly assumed connota-tion of referring to a liquid phase. Here we simply use it in reference to the context of thephenomenology of H2O-dominated fluid phases within an epizonal or sub-volcanic context.The term “magmatic vapor” has been widely used to refer to the low-density fluid phasereleased from a melt so that the term “vapor” has a connotation with respect to both phasestate and provenance. Strictly the term “vapor” refers to a gas phase in equilibrium with acondensed phase – solid or liquid – as for example in the term “vapor pressure”. Thus a“magmatic vapor” once released from direct interaction with a melt should be termed a“magmatic gas” or simply a “gas” or indeed more precisely a “gas mixture” because of thewide range of components that it contains. Since this gasmixture is dominated bywater it iscondensible and thereforeonexpansionmay intersect the two-phaseboundaryof the systemresulting in phase separation to fractions of liquid “condensate” and equilibrium “vapor” (agas). Whilst the distinctions may appear semantic, it is important to be aware of the preciseimplications of these terms, particularly with respect to the physical properties and behaviorof “fluid” phases with respect to fluid properties like density and viscosity in geological

67R.W. Henley, B.R. Berger / Ore Geology Reviews 39 (2011) 63–74

stressed by Sack and Goodell (2002) for the Julcani and other deposits.Fig. 3 shows phase relationships in the Cu–Sb–As–S system; enargite(orthorhombic, wurtzite-type, hexagonal close-packed structure) andluzonite (tetragonal, sphalerite-type cubic close-packed structure)are, sensu stricto, polytypes of Cu3AsS4 (Posfai and Sundberg, 1998).Enargite is the high-temperature form of luzonite above 280–300 °Cand an immiscibility gap is defined between enargite and the luzonite–famatinite solid solution series (Skinner, 1960) where these mineralsdisplay a wide range of ordered to disordered structures andintergrowths at the atomic scale. Posfai and Buseck (1998) noted thatthe occurrence of enargite is an indicator of its formationwell above thetransitionbetween luzonite and enargite (280 °C) and that luzonitemayform through solid-state recrystallization of enargite below thetransition temperature. In describing the enargite inclusion samplesfrom Lepanto, Mancano (1994) specifically noted heterogeneities andinternal structural defects,3 as well as the difficulties caused by varyingdegrees of IR transparency in establishing the primary or secondarynature of inclusions that enabled only 29% of the 97 enargite inclusionsin 17 samples to be interpreted as potentially primary (and two assecondaries) while all quartz inclusions were classified as secondaries.We suggest that these data record post-primary phase changes andfracturing during the cooling of the enargite assemblages in thepresence of groundwater.

3 An IR image of enargite in Mancano (1994, Fig. 3a), for example, clearly shows arelation between an inclusion and a subsolidus structural, not compositional, attributeof the crystal.

2.1.3. Summary of temperatures for alteration and mineralizationThe Sulfate Stage fluid inclusion and stable isotope data reliably

identify a) awide range of temperatures (N500 to 100 °C) consistentwithhigh thermal gradients and the recrystallization of metastable silicapolymorphs to quartz during cooling, and b) derivation from partialcondensation of a magmatic vapor with very limited involvement ofgroundwater. Since condensation of vapor occurs through heat loss, it isimplicit that the primary fluid4 had a higher enthalpy and temperatureandwas therefore, necessarily, also a vapor. Temperature estimates for theSulfide Stage cannot be reliably basedonfluid inclusiondata in enargite orother sulfides. Micro-analytical data (see Section 1) from El Indio,Summitville and Lepanto provide independent evidence that sulfide–sulfosalt deposition along with quartz in the earliest part of the SulfideStage were at least 550 °C, consistent with the high temperaturesindicated for thevapor condensate responsible for Sulfate Stage alteration.

2.2. Constraints on fluid pressure

Since many of the well-known high-sulfidation deposits areyoung, pressure estimates for alteration and mineralization can bemade from field data. For example at Chinkuashih, Taiwan, the ruggedtopography and geological setting of the ≈1 Ma Chinkuashih depositsuggest that, at most, only one or two hundred meters of rock havebeen removed by erosion. Mining extended to about 850 m below thepresent surface with decreasing enargite and increasing Cu/Au ratio.As discussed above (see Section 2.1.1), fluid inclusion formation islikely to have occurred during recrystallization of silica in thealteration zones so that interpreted trapping pressures record porefluid pressures consistent with the reconstructed paleo-depth atwhich samples were obtained. Jannas et al. (1999) reconstructed theoriginal topography at El Indio to suggest a paleo-depth range formineralization of −500 to −1100 m (or less if post-mineral faultmovement occurred). Hedenquist et al. (1998) suggested shallowdepth for the Lepanto mineralization and at Goldfield, Nevada, debrisflow sediments (Berger, 2007) containing alunite pebbles occur at thepaleosurface in the eastern part of the district indicating thatmineralization and alteration developed at ≈200 m depth.

2.2.1. Summary of pressure constraints on alteration and mineralizationSince alteration and mineralization at both Chinkuashih and El

Indio were formed from a vapor phase (see Section 2.1) to depthsbelow the paleosurface of about 1100 m and stable isotope dataindicate limited, if any, incursion of groundwater, the fluid pressure inthe developing vein system was evidently greater than that due to anexternal groundwater system (≤~110 bars at this depth). At thisdepth, the lithostatic pressure would have been about 290 bars andsimilar pressure estimates apply for other high-sulfidation veindeposits. The pressure constraints provided by geologic and fluid

environments. Here, the term “vapor” is used generically where the argument has notreached a point of defining physical properties and then ceases to have value in much thesame way as Verhoogen (1949) advocated the avoidance of the non-specific term“supercritical”. We have used the term fluid as a non-specific term prior to defining itsphase state in line with the recommendations of Bodnar (pers.comm., 2010).

Fig. 4. The relationship between pressure and depth in a vapor expanding to the surfacethrough a fracture array. MV represents the flow through low permeability fracturedrock from a high pressure magma source, M, to a low pressure sink, V, provided by thehighly transmissive fracture array. VF is a simplified pressure-depth profile where thevapor pressure gradient is greater than hydrostatic (see Section 2.2.1). Hydrothermalbrecciation occurs above point S as shown. At surface rapid open-system cooling occursdue to atmospheric interactions and mixing with shallow groundwater or perchedlayers of condensate providing local regions with hydrostatic gradients that may bepenetrated by fumarolic discharges. For further discussion see text and Berger andHenley (in press). A useful modern example of the contrast of vapor and liquid pressureprofiles is detailed by Moore et al. (2008) for the Karaha-Telaga Bodas volcanicgeothermal system, Indonesia.

68 R.W. Henley, B.R. Berger / Ore Geology Reviews 39 (2011) 63–74

inclusion data do not therefore sustain the high pressure conditionsrequired by Heinrich et al. (2004) to maintain a liquid phase conditionfor high-sulfidation mineralization.

As discussed later (see Section 5.2), where limited erosion hasoccurred, high-sulfidation lode mineralization is expressed as low gradebulk-tonnage deposits associated with extensive advanced argillicalteration, indicating that vapor flow occurred freely to and throughthe paleo-surface. Ignoring small changes in vapor density due tofractional heat loss, the pressure-depth curve for a vapor ismuch steeperthan that of liquid water (the boiling point-depth curve or coldgroundwater hydrostatic gradient) or of rock (lithostatic gradient)(Fig. 4). The steep vapor gradient intersects the lithostatic gradient at ashallow level so that pressure release after choking bymineral depositioncould produce shallow breccias bodies as is observed in high leveldeposits such as Yanacocha, Peru.5 Below this crossover, lithostaticpressure greatly exceeds the vapor pressure in anopen fracture. This is anunstable situation with respect to the relative fluid pressure within theopen fracture system which must therefore progressively close bywallrock fragmentation due to rock stress as well as mineral deposition.Maintenance of fracture permeability is thus dependent on renewedfracturing due to the active stressfield and the condition (Cox, 2010) thatPfluidNProck+T, where T is the tensile strength of the rock (about 10 barsmaximum).

5 Many of the so-called hydrothermal breccias in shallow deposits can however beshown to have formed during the Sulfate Stage, for example at Tambo, Chile (Jannas etal., 1999), in the presence of aggressively reactive acid condensate indicatingdissolution and steam explosion as their formation process.

A review of data for alteration and ore mineralization tempera-tures (see Section 2.1) indicates that the Sulfate Stage fluid transitingto the surface through faults and associated fractures was acondensible vapor (i.e., condensate liquid formed as fluid transitedpressure and temperature gradients in the wallrock) at a temperatureof N550 °C with a density of about 0.03 g/cm3.

3. Magmatic-vapor expansion

The recognition, through stable isotope data, that the fluid phaseresponsible for both alteration and oremineralization in high-sulfidationgold deposits was predominantly of magmatic origin forcesconsideration of the systematics of magmatic-fluid expansion from itssource through a fracture array in the upper few kilometers (b1.5 to2 km) of the Earth's crust towhere it discharges at the surface. Pressure isthe critical control on fluid flow and the phase state of the fluid, the lattercrucially controlling chemical processes such as mineral deposition.Critical examination of available fluid inclusion and other data (seeSection 2) confirms that the fluid responsible for alteration andore mineralization was a low-salinity, high-temperature (N550 °C),low-pressure (bb500 bars) vapor phase released from a magmaticreservoir comprising multiple sub-volcanic intrusions with alterationand ore mineralization occurring within 1500 m of the surface atpressures below about 110 bars.

Sodium chloride (b5 wt.% NaCleq) dominates other non-volatilecomponents in both volcanic gases andfluid inclusions inhigh-sulfidationdeposits. By analogy with volcanic gases, CO2 and other insoluble gasesoccur at less than 5 mol% such that their contribution on an ideal gas basisto the total pressure of the vapor phase is less than 5%. The phase behaviorof the magmatic fluid can appropriately be approximated with respect topressure, enthalpy (heat content), temperature and solute content on thebasis of data for the simple NaCl–H2O system (Driesner, 2007). Fig. 5ashows the extent of the immiscibility gap between liquid and vapor up to1000 °C in this system and Fig. 5b is a projection of the two-phase surfaceand isopleths of NaCl concentration for a selection of illustrativetemperatures. In the high-enthalpy, low-pressure, single-phase regionand below 5 wt.%. NaCleq, the properties of the vapor phase can beapproximated using values for pure water under the same conditions inorder to provide quantification of, for example, the density of the vaporphase.

The melting relations of water-saturated granite (Fig. 5b) withrespect to pressure and temperature define the enthalpy range of thelow-salinity fluid exsolved from a granitic magma. For example, amagmatic fluid exsolving from melt at ~850 °C in relatively shallowvolcanic settings has an enthalpy in excess of 4000 kJ/kg and a densityof about 0.2 g/cm3 or less. It follows that, for the shallow (b1.5 km)volcanic setting of enargite–gold deposits, the low-salinity fluididentified by fluid inclusion and stable isotope analyses is of magmaticorigin and has the properties of a gas; it expands to fill all availablevolume.

It is essential in considering thebehavior of hydrothermal systems totrack the fate of all the conservative components in the fluid phaseincluding salt content, stable isotopes and heat energy. The thermaldynamics of a number of common depressurization paths for magmaticvapor are shown in Fig. 5c in order to contrast that for high-sulfidationdeposits with others in more permeable environments wheregroundwater interactions dominate ore-forming processes. In settingssuch as geothermal systems (M–G) where high permeability ismaintained by extensional fracturing, mixing between coldgroundwater and expanding magmatic vapor occurs across the extentof the system and gives rise to the development of magmatic vaporplumes as described by Henley and McNabb (1978). In these settings,fluid enthalpy (and therefore temperature), isotope ratios and soluteconcentration are determined by the progressive change in the ratio ofmagmatic fluid to groundwater. Intersection of the two-phase regionmay occur on particular enthalpy–pressure mixing paths depending on

Fig. 5. (a) Three dimensional perspective of the 2-phase region in the systemNaCl–H2Obased on the data compilation of Driesner (2007, Supplementary Material). (b)Enthalpy–pressure–temperature relations for phases in the system NaCl–H2O calcu-lated using SOWAT code kindly provided by Thomas Driesner.). The figure comprises aprojection of a selection of isopleths and the two-phase boundary at 400°, 750° and1000 °C to illustrate the expansion of the two-phase field as temperature increases, theeffect of salt-content on the enthalpy of liquid (L) and gas phases (V) outside the two-phase field and the relative location of the halite and halite (H)+vapor (V) fields atthese temperatures. For discussion purposes, density isochors for pure water are alsoprovided. The water-saturated minimum and 2% water content melting curves forgranite are shown in order to identify the enthalpy–pressure field for magmatic vaporderived from epizonal intrusions. c) Illustrative paths for mixing and heat loss flow inhydrothermal systems; for detail see Section 3.

Fig. 6. Halite solubility in the halite-vapor field of the NaCl–H2O system, redrawn fromthe data of Sourirajan and Kennedy (1962). An illustrative quasi-isenthalpicdepressurization and expansion path is shown for a magmatic vapor emphasizing thedecrease in water vapor pressure and the rapid decrease in halite solubility in the haliteplus vapor field both of which lead to rapid sulfide and sulfosalt deposition as melt orsolid phase.

69R.W. Henley, B.R. Berger / Ore Geology Reviews 39 (2011) 63–74

structural-permeability and in these settings porphyry-stylemineralisation (ϕ) may eventuate (see Section 5.1). In the single-phaseregion near surface, phase separation occurs as the mixed fluidintersects the two-phase region and, in favorable structural settings,may generate adularia–sericite styles of epithermal precious and basemetalmineralization (Henley and Berger, 2000; Henley and Ellis, 1983).This mixing scenario differs from path M–FHE where the isenthalpicexpansion6 of the magmatic vapor is confined to discrete high-permeability fracture arrays through impermeable rocks and heat lossis limited by the conductivity of thewallrock (Toulmin and Clark, 1967).It is this setting that corresponds to the formation of high-sulfidationgold deposits and high-enthalpy fumaroles (FHE) in actively degassingvolcanoes.Where expanding vapor loses a significant proportionof heatby conduction (M–V), a small fraction of vapor condenses as pressuredrops below M–V–C (Fig. 5c). Through partitioning of acid gases,primarily SO2, to the condensate in wallrock fractures, aggressivealteration ensues and gives rise eventually to the characteristic silica–alunite alteration of high-sulfidation deposits (see Section 2.1.1). PathS–S′ is illustrative of conductive cooling where a limited flow of vaporloses all of its heat to wallrock; such total condensation retains thecomposition of the primary vapor phase and the liquid condensate maydeposit sulfide minerals in isolated cavity fillings.7 A high ΔH/ΔPtrajectory also occurs where vapor condenses directly into a perchedgroundwater or condensate liquid aquifer very close to the surface and

6 Heat-balance terminology. We here use the term “isenthalpic” to specifyexpansion processes where there is no exchange of heat energy with the surroundings.The alternative term “irreversible adiabatic expansion” distinguishes the non-equilibrium aspect of such processes (for example, rapid expansion through a seriesof throttles or dilations in a fracture array) and their rate relative to “reversibleadiabatic expansion” processes. The latter refer to the slow constrained expansion of afluid where there is an inter-conversion of internal energy and work energy. In such aconstrained isentropic process, a far more substantial temperature decrease occursrelative to that due to irreversible isentropic expansion over the same pressure drop(Anderson and Crerar, 1993). Both these processes may control the range of fumaroletemperatures within the La Fossa volcano on Vulcano, Italy (Nuccio et al., 1999).

7 An example (see Section 2.1.2) is the enargite and pyrite (TH ~317 °C) within latedrusy cavities in isolated veins cutting the Rosia Poieni porphyry copper deposit(Kouzmanov et al., 2010).

70 R.W. Henley, B.R. Berger / Ore Geology Reviews 39 (2011) 63–74

results in the well-known solfataric environment of active volcanoesassociated with low-enthalpy fumaroles (FLE).

Fig. 6 provides the detail of the low-pressure environment wherequasi-isenthalpic expansion of magmatic vapor enters the halite plusvapor field in which the vapor-phase solubility of halite decreasesrapidly with pressure and temperature (or more precisely enthalpy).As will be discussed later, it is this rapid decrease in concentration ofcomponents such as chloride that, in association with the decrease inwater pressure during expansion, is likely to be the principal controlon the deposition of metal and metal sulfides and sulfosalts as well asgold (Mavrogenes et al., 2010).

4. Chemical processes during alteration and mineralization

In the magmatic-vapor expansion hypothesis proposed in thispaper, it is assumed, in commonwithmodels proposed by Hedenquistet al. (1998), Henley and McNabb (1978), Williams-Jones andHeinrich (2005) and earlier by Krauskopf (1964), that metals aretransferred from a magma source as vapor-phase species. The abilityof a vapor phase to transport significant quantities of metals does,however, remain one of the most controversial areas in the discussionof the formation of magmatic-hydrothermal ore deposits despiteprima facie evidence such as the transport and deposition of a widerange of metals including gold, iron, and arsenic in high-temperaturefumarole gases (e.g., Africano et al., 2002; Churakov et al., 2000;Fulignati and Sbrana, 1998; Larocque et al., 2008; Taran et al., 2000,2001; Williams-Jones et al., 2002; Yudovskaya et al., 2006). Furthersupport is provided by fluid inclusions in porphyry copper–goldsystems (e.g., Ullrich et al., 1999) and heavy metal sulfosalts andsulfides in vugs in high-sulfidation deposits (Mavrogenes et al., 2010).Chemical-vapor transport and deposition at low pressures are alsovery well established in the semiconductor and solar technologyindustries for a wide spectrum of metals (cf. Hastie, 1975; Kodas andHampden-Smith, 1994). However, as stressed by Pokrovski et al. (2008),there is a frustrating data vacuum for the high-temperature speciation ofmetals in a high-temperature gas phase under geological conditions, andthis reflects howdifficult such experimental determinations are given theproblems of containment, sampling and analysis of highly corrosive,low-density fluids.

Pokrovski et al. (2008, p. 357) note that the formation of solvationshells around core metal species “is the essential factor controllingvapor-phase transport and vapor–liquid partitioning” in the vaporphase at high temperatures. Species such as Cu(HS)g, nH2O have highhydration numbers8 that respond rapidly to decreasing water pressureas shown by vapor–liquid partition coefficients: in a sulfur-free systemfor example, base-metal partition coefficients at 500 °C decrease by twoorders of magnitude between 580 and 480 bars, while partitioncoefficients for gold and copper decrease even more rapidly in systemscontaining reduced sulfur (Pokrovski et al., 2008). The solubility ofquartz as SiO2·2H2O is strongly pressure dependent (Fournier andPotter, 1982; Morey and Hesselgesser, 1951) as is that of halite asNaCl·2H2O (Hastie, 1975, p. 82).

Particularly pertinent for high-sulfidation systems, Pokrovskiet al. (2005) have shown that the dominant arsenic species in low-density magmatic vapor at 500 °C is As(OH)3 the concentration ofwhich is therefore strongly dependent on PH2O. For illustrative

8 ) n, the number of solvation sphere water molecules, is a statistical value extractedfrom experimental data for the effective number of water or other molecules surroundinga coremolecular species. The bindingof eachsolvationmolecule through, for example,Vander Waals bonding involves a solvation free energy that contributes to the overallsolubility equilibrium constant for the dissolution reaction in the vapor phase.

purposes the solubility of enargite, as a solid or melt, may then bewritten,

3CuðHSÞ0g þ AsðOHÞ03g þ H2Sg ¼ Cu3AsS4 þ 3H2Og þ H2g ð1aÞso that, with a=activity, the solubility equilibrium constant, Ks,g

is,

Ks;g ¼ ½a3H2Og • a H2;g�=½a3CuðHSÞog • a AsðOHÞo3g • a H2Sg

� ð1bÞ

The solubility of enargite in vapor therefore at least a cubicdependence on the activity of water or, to a reasonable approximation,the pressure of water in the expanding vapor. Its solubility in amagmatic vapor, initially at say 1000 bars consequently decreases by afactor of 1000 in expanding to 100 bars pressure simply due to thechanging pressure of thewater-dominated vapor phase. Similarly, thereis an overall solubility decrease of 106 in expanding to a surface pressureof 1 bar. Inclusion of solvation water molecules (Eq. (1c)) shows ahigher order dependence on the activity or pressure of water; here H2Odistinguishes the properties of solvation water from free water.

3CuðHSÞ0g ; nH2Og þ AsðOHÞ03;g þ H2Sg ¼ Cu3AsS4 þ 3H2Og þ H2;g þ nH2Og

ð1cÞThe release of water molecules from the solvation sphere due to

vapor expansion and decrease in the dielectric constant has a multipliereffect on solubility decrease. Copper, and othermetals in thevapor phasemay also be present as chloride species so thatwhen vapor expands intothe halite plus vapor field (Fig. 6) the removal of available chloride tohalite significantly reduces their solubility; this is potentially mostsignificant for a ‘hard’metal such as iron in controlling pyrite deposition.

The solubility of enargite in a vapor is also, through Eq. (1b),dependent upon the rate of change of aH2S/aH2

. While wallrockreactions are commonly assumed to control the redox state (aH2

) ofthe fluid phase in many ore-forming systems, the chemical processesin a vapor expanding within a fracture array through impermeablesilica–alunite altered wallrocks are dominated by homogenous gasreactions. The activity of hydrogen, aH2,g

, is controlled by a matrix ofreactions but primarily those involving sulfur (Giggenbach, 1987).

SO2ðgÞ þ 3H2ðgÞ ¼ H2SðgÞ þ 2H2OðgÞ ð2aÞ

K2a ¼ aH2Sd a2

H2O=aSO2d a

3H2

ð2bÞ

SO2(g) dominates the gaseous sulfur species at magmatic tempera-tures and, through thenegative volume change of Reaction 2a aswritten,is also favored by decreasing pressure. Equilibrium calculations showthat aH2S/aH2

does not vary by more than a factor of 10 over thetemperature range 350 to 900 °C or between 10 and 1000 bar pressuresat 600 °C and is accordingly far less significant than thevariation ofwaterpressure. Giggenbach (1987),Martin et al. (2009) andMontegrossi et al.(2008) have drawn attention to the non-equilibrium behavior of theH2S–SO2 reaction which in turn implies unfavorable kinetics as a redoxcontrol reaction.

The rapid change in water fugacity and density consequent onvapor expansion leads to major changes in the solubility of metals –

and silica – in the vapor phase. As a consequence, decompressionresults in an immediate very high supersaturation of metals that, inturn, triggers a cascade of far-from-equilibrium reactions (Henley andBerger, 2000). As shown in Fig. 1, at El Indio, following pyrite and localenargite deposition, Cu–Fe–As–S melt condensed from the vaporphasewith a range of trace heavymetals. As progressive crystallizationproceeded the sulfosalt melts dissolved earlier pyrite9 and enargite and

9 The sulfosalt melt+pyrite+vapor reaction occurs via chalcopyritesp (Fig. 1e) withultimate resorption of the chalcopyrite (Fig. 1e) but leaving islands of pyrite within theFe-tennantite to produce the disease texture described by Imai (1999) at Lepanto. Thecomplex redox reaction process is described elsewhere (Mavrogenes et al., in prep).

10 Crater lakes form through the complete condensation of magmatic vapor andadmixture of meteoric water. Their mineralogy and geochemistry are therefore acomposite of sulfate and sulfide geochemistry with clays, alunite, enargite, pyrite, etc.carried in suspension, ejecta and sediment (e.g., Kawah Ijen, Indonesia, Delmelle andBernard, 1994; Kawah Putih, Indonesia, Sriwana et al., 2000).

71R.W. Henley, B.R. Berger / Ore Geology Reviews 39 (2011) 63–74

fractionation of antimony, bismuth, tin, selenium, tellurium, silver (andgold) occurred between the melt and vapor phase to producechemically zoned Fe-tennantite and other sulfosalts, tellurides andbismuth minerals (Mavrogenes and Henley, in preparation) in ananalogousmanner to that described by Tomkins (2010) for sulfosalts inmagmatic sulfide ore deposits. The simultaneous growth of euhedralquartz during melt crystallization and fractionation (Fig. 1c and d) is aconsequence of high initial vapor phase silica solubility and the limitedsolubilityofwater in sulfide–andbyanalogy sulfosalt–melts (Wykes andMavrogenes, 2005). The end product of the sulfosalt–sulfide crystalliza-tion sequence is a vapor phasewith relatively high gold and silver contentwhich ultimately separates to form silica-rich high grade gold veins.Elsewhere vapor deposition may have occurred directly to solid statesulfosalt phases with partitioning of heavy metals between solid andvapor.

5. Discussion

It is well established that during the crystallization of melts, metalspartition into magmatic vapor (e.g., Candela and Holland, 1984;Simon et al., 2008). This process occurs early during crystallization todevelop an aqueous fluid phase within sub-volcanic intrusions atlithostatic pressure. At 4 km depth, for example, the maximumpressure is about 1000 bars. Where leakage occurs directly, or morelikely indirectly via a magmatic-vapor reservoir, into active fracturesystems open to the surface (1 to 110 bars, see Section 2.2) the drop influid pressure is more than 890 bars. This disparity in fluid pressuresdrives the expansion and discharge of magmatic vapor and theconsequent vapor-phase solubility disequilibria that result in thesequence of sulfosalt–sulfide stages typical of high-sulfidationdeposits. Since rapid vapor expansion constrained by active fracturearrays within 1500 m or so of the paleosurface involves only minorheat losses and temperature decrease the Sulfide Stage is initiated byhigh temperature sulfide–sulfosalt assemblages as described byMavrogenes et al. (2010). Crystalline enargite and pyrite may formdirectly from the vapor phase at temperatures above about 650 °C. AtEl Indio and elsewhere these phases are subsequently dissolved bya Cu–As–S melt that condensed from the expanding vapor phaseand which progressively crystallizes with fractionation of heavymetals between melt and vapor giving rise, through the possiblemixing of groundwater in open fracture arrays, to the later complextelluride–bismuth–tin–tungsten enriched vein assemblages with goldand silver. The recognition of the sulfosalt melt stage is stronglysupported by the combination of several lines of evidence of evidenceas is illustrated in Fig. 1. Key evidence includes the following: (a)euhedral quartz crystals arewholly enclosedwithin cavities andveinletswith the sulfosalt (Fig. 1a–c), (b) symmetrical zonation of arsenic,antimony and tellurium in Fe-tennantite around quartz crystals, (c)Fe-tennantite demonstrably dissolves early pyrite (Fig. 1e),(d) preservation of vugs and vug-filling sulfides and sulfosalts,(e) fluid inclusions in quartz in vugs in Fe-tennantite crystalshomogenize to a vapor at N450 °C indicating a minimum trappingtemperature (Mavrogenes et al., 2010).

5.1. The volcanic context of high-sulfidation gold deposits

It is a matter of common observation that, between eruptive stages,magmas beneath composite volcanoes continue to releasemagmatic-gasmixtures that expand via fracture arrays to form high-temperaturefumaroles and extensive near-surface alteration. Field data for enargite–gold lode deposits define an analogous magmatic-vapor expansioncontext punctuated by eruption and dome extrusion. Following hisdetailed study of Goldfield, Nevada, Ransome (1909, 1910) suggestedthat the ores there may have been formed by the ascension of hot, acidicsolutions emanating from a deep magma reservoir. The role of volcanicgases in the origin of gold deposits was subsequently debated

intermittently until substantially brought into focus in a review ofvolcanic-gas chemistry by Hedenquist (1995). This review showed thatdespite thedifficulties in samplingvolcanic gasesduring eruptionor fromactive high-temperature fumaroles between eruptive stages, theobserved flux of metals could be argued as sufficient to account for oreformation within geologically reasonable time periods (104–105 years)depending of course on the efficiency of fluid focusing and metaldeposition in the superstructure of the volcanic system.

Whether or not there is a direct link between the formation ofenargite–gold lode deposits and porphyry copper–gold deposits isimmaterial to our hypothesis for the fracture-controlled expansion ofmagmatic vapor in the formation of high-sulfidation gold deposits. Theone is not necessarily tied to the other. At Lepanto, for example, theranges of the K–Ar data for porphyry and high-sulfidation mineraliza-tion (Arribas et al., 1995, GSA Data Repository Item 9517) do notdemonstrate contemporaneity of mineralization but simply allow thateach developed as discrete events of approximate duration 50,000–100,000 years during the history of the Mankayan volcanic complexover 700,000 years. The relative structural timing of mineralization andalteration events is discussed by Berger and Henley (in press). AtFamatina, Pudack et al. (2009) carefully detailed the development of amagmatic-vapor phase associatedwith earlier high-level intrusions thatare spatially related to weakly-mineralized but high-grade, enargite–famatinite–gold–silver veins associated with alunite. However, thechain of evidence between the degassing of the earlier epizonalintrusions and the development of the enargite–famatinite veins isbroken due to extensive erosion of at least 800 m of the host-rocksequence prior to the sulfosalt mineralization stage, and the rarity ofusable fluid inclusions in the sulfosalt stage samples.

Porphyry copper–gold deposits occur in large but still locallyconstrained areas of extensional fracturing in contrast to high-sulfidationdeposits that are restricted to much more isolated and irregular fracturearrays through impermeable rocks. For both, isotope data consistentlyidentify the role of magmatic vapor sensu latu, but it is more likely in acomposite volcanic system that they record a “magmatic-vaporreservoir” containing a mixture of magmatic gases derived from suitesof discrete melt reservoirs that may also include basalts (e.g., Robergeet al., 2009) rather than any single or specific reservoir. The dynamics ofthefluid phase in such a reservoir are determinedby crustal stresswhich,in conjunction with pore-fluid pressure, leads at some intermediatedepth to the accumulation ofmagmatic vapor fromavailable sources andcontrols higher-level fracture formation and reactivation (Cox et al.,2001) as suggestedby seismic data at Yellowstone (Husen et al., 2004). Inturn this suggests that high-sulfidation and porphyry Cu–Au depositsoccupy different structural regimes at different times as stress changesprogress during the history of a large volcanic complex, directlydetermining the available structures within which mixing betweenmagmatic vapor and groundwater can occur (Fig. 5c, MG); the ΔP/ΔHtrajectory determines whether the two-phase fluid regime is intersectedwith the potential for the formation of porphyry copper–gold stylemineralisation.

5.2. The near-surface discharge regime

At and very near the surface in “degassing” volcanic environments,condensation of high-enthalpy volcanic vapor occurs very efficientlythrough heat transfers due to interaction with the atmosphere, contactwith highly permeable, epiclastic sediments, or by condensation andmixing into groundwater regimes and crater lakes (Brantley et al.,1987).10 These processes inevitably lead to the formation of localized

Fig. 7. The relationship between pressure, temperature and depth for a magmatic vapor expanding to the surface through a fracture array, in relation to the scale and geologicelements of a typical high-sulfidation deposit (modified after Berger and Eimon, 1983). At the surface, rapid open-system cooling occurs due to atmospheric interactions andmixingwith shallow groundwater or perched layers of condensate to produce a solfataric environment (note that the Solfatara Zone is shown as extending above the zero surface referencein order to accommodate topographic variations on hydraulic head). For further discussion see text and Berger and Henley (in press).

72 R.W. Henley, B.R. Berger / Ore Geology Reviews 39 (2011) 63–74

bodies of advanced argillic and associated more extensive zones ofadvanced argillic and argillic alteration as well as perched layers ofacidic “steam-heated” water near the surface that mimic topographyand which are the basic elements of solfatara as observed in active ormore strictly quiescent volcanic environments today. These phenomenaconstitute a highly efficient trap for the remaining metals exitingthe array of ‘feeder’ fractures and result in the shallow formationof bulk-tonnage, but low-grade, gold deposits. The bulk-tonnagePascua–Lama deposit, for example, has a depth extent of about 500 mand is distinguished by the extraordinary dynamic complexity of therelationships between an active structural array and relationshipsbetween silica–alunite and pyrite–enargite deposition (Chouinard et al.,2005). The occurrence of both alunite and sulfosalts in some veins is aconsequence of rapidly changing pressures and temperatures in thenear surface part of the system due to complex interplay betweencondensate and magmatic vapor on multiple flow paths in theextensive, active fracture system (Berger and Henley, 2011). Inmore permeable systems, magmatic vapor expanding on isolatedfracture arrays may become trapped in groundwater flow systemswith consequent dilution and dispersion of metals, as is commonlyobserved for arsenic, for example, in groundwater around activevolcanoes.

6. Summary

High-sulfidation gold deposits occur in andesite–dacite volcaniccomplexes and relate closely to the discharge of magmatic-vapor gasas it expands at depths of only a few hundred meters beneathhigh-temperature fumaroles and solfatara. Their association withsurficial volcanic rocks however led Lindgren (1913) to classify thesedeposits as epithermal implying formation temperatures on theorder of180–300 °C. Recent stable isotope, fluid inclusion and micro-analyticaldata however have demonstrated a clearer relationship tomuch highertemperature fumarolic discharges as indeed was suggested by Cross(1891, 1896), and a magmatic fluid component of at least 75%mass.

These deposits are therefore more appropriately termed “Solfataricgold deposits” in order to distinguish them from epithermal (sensustricto) gold deposits (e.g., McLaughlin, California) whose geological

context and alteration and mineralization processes relate to theshallow ‘hot-springs’ discharge regime of paleo-geothermal systemswhere, due tomixing, the magmatic fluid component is less than about30%. The time-integrated elements of Solfataric gold deposits aresummarized in Fig. 7 in relation to the pressure and temperaturewithinanexpanding columnofmagmatic vapor, the spatial control imposedbyactive fracture arrays and the formation of solfataras with extensiveadvanced argillic alteration and low-grade mineralization solfatara atthe paleosurface. The narrowness of the isotherms localized by flowingfractures and the high thermal gradients are important to note, alongwith the expansion of the advanced argillic alteration due to heat lossand ‘ponding’ of mixed acid condensate–groundwater aquifers belowthe surface. Recognition of the paleo-environment and controls onvapor expansion is of course significant in exploration for thesedeposits.This synopsis also applies to solfataras inactivevolcanoesandassociatedacidic geothermal systems (Ruaya et al., 1992) and equivalent sea floorvolcanic gas discharges, such as at Brothers Volcano in the southernKermadec Arc, New Zealand (de Ronde et al., 2005).

Acknowledgments

We thank Richard Arculus, Paul Barton, Phil Bethke, James Cleverley,Graham Hughes, Gary Landis, Sue Kieffer, Laurie Madden, JohnMavrogenes, Agnes Reyes, Cornel de Ronde, and Bob Rye for the helpfulcomments and discussion. Dana Bove, Jeff Hedenquist, Kalin Kouzmanovand Francois Robert kindly provided some of the samples illustrated inFig. 1. We thank Steve Kesler and Bob Bodnar for providing constructivereviews on the original manuscript and Thomas Driesner for providingaccess to the programs used in construction of the NaCl–H2O phasediagrams.

References

Africano, F., Bernard, A., 2000. Acid alteration in the fumarolic environment of Usuvolcano, Hokkaido, Japan. Journal of Volcanology and Geothermal Research 97,475–495.

Africano, F., van Rompaey, G., Bernard, A., Le Guern, F., 2002. Deposition of traceelements from high temperature gases at Satsuma-Iwojima volcano. Earth PlanetSpace 54, 275–286.

73R.W. Henley, B.R. Berger / Ore Geology Reviews 39 (2011) 63–74

Anderson, G.M., Crerar, D.A., 1993. Thermodynamics in Geochemistry. OxfordUniversity Press, Oxford. 588 pp.

Arribas Jr., A., 1995. Characteristics of high-sulfidation epithermal deposits, and theirrelation to magmatic fluid. Mineralogical Association of Canada Short Course 23,419–454.

Arribas Jr., A., Hedenquist, J.W., Itaya, T., Okada, T., Rogelio, A.C., Garcia Jr., J.S., 1995.Contemporaneous formation of adjacent porphyry and epithermal Cu–Au depositsover 300 ka in northern Luzon, Philippines. Geology 23, 337–340.

Berger, B.R., 2007. The 3D fault and vein architecture of strike–slip releasing andrestraining bends: evidence from volcanic-centre-related mineral deposits. In:Cunningham,W.D.,Mann, P. (Eds.), Tectonics of Strike–Slip Restraining and ReleasingBends: Geological Society of London, Special Publication, 290, pp. 447–471.

Berger, B.R., Eimon, P.I., 1983. Conceptual models of epithermal precious metaldeposits. In: Shanks, W.C. (Ed.), Cameron Volume on Unconventional MineralDeposits. Society of Mining Engineers of AIME, pp. 191–205.

Berger, B.R., Henley, R.W., 2011. Magmatic-vapor expansion and the formation of high-sulfidation gold deposits: Structural controls on hydrothermal alteration and oremineralization. Ore Geology Reviews 39, 75–90.

Bethke, P.M., Rye, R.O., Stoffregen, R.E., Vikre, P.G., 2005. Evolution of the magmatic-hydrothermal acid-sulfate system at Summitville, Colorado: integration ofgeological, stable-isotope, and fluid-inclusion evidence. Chemical Geology 215,281–315.

Brantley, S.L., Borgia, A., Rowe, G., Fernandez, J.F., Reynolds, J.R., 1987. Poas volcanocrater lake acts as a condenser for acid metal-rich brine. Nature 330, 470–472.

Candela, P.A., Holland, H.D., 1984. The partitioning of copper and molybdenumbetween silicate melts and aqueous fluids. Geochimica et Cosmochimica Acta 48,373–380.

Chouinard, A., Williams-Jones, A.E., Leonardson, R.W., Hodgson, C.J., Silva, P., Tellez,J.V., Rojas, F., 2005. Geology and genesis of the multistage high-sulfidationepithermal Pascua Au–Ag–Cu deposit, Chile and Argentina. Economic Geology100, 463–490.

Churakov, S.V., Tkachenko, S.I., Korzhinskii, M.A., Bocharnikov, R.E., Shmuloviich, K.I., 2000.Evolutionof compositionof high-temperature fumarolic gases fromKudryavyVolcano,Iterup, Kuril Islands: the thermodynamic modeling. Geochemistry International 38,436–451.

Claveria, R.J.R., 2001. Mineral paragenesis of the Lepanto copper and gold and theVictoria gold deposits, Mankayan district, Philippines. Resource Geology 51,97–106.

Cox, S.F., 2010. The application of failure mode diagrams for exploring the roles of fluidpressure and stress states in controlling styles of fracture-controlled permeabilityenhancement in faults and shear zones. Geofluids 10, 217–233.

Cox, S.F., Knackstedt, M.A., Braun, J., 2001. Principles of structural control onpermeability and fluid flow in hydrothermal systems. Reviews in EconomicGeology 14, 1–24.

Cross, W., 1891. On alunite and diaspora from the Rosita Hills, Colorado. AmericanJournal of Science 48, 460–475 3rd Series.

Cross, W., 1896. Geology of Silver Cliff and the Rosita Hills, Colorado. U.S. GeologicalSurvey Seventeenth Annual Report, 2, pp. 263–403.

De Ronde, C.E.J., Hannington, M.D., Stoffers, P., Wright, I.C., Ditchburn, R.G., Reyes, A.G.,Baker, E.T.,Massoth, G.J., Lupton, J.E.,Walker, S.L., Greene,R.R., Soong, C.W.R., Ishibashi,J., Lebon, G.T., 2005. Evolution of a submarine magmatic-hydrothermal system:Brothers Volcano, Southern Kermadec Arc, New Zealand. Economic Geology 100,1097–1133.

Deditius, A.P., Utsunomiya, S., Ewing, R.C., Chryssoulis, S.L., Venter, D., Kesler, S.E., 2009.Decoupled geochemical behavior of As and Cu in hydrothermal systems. Geology37, 707–710.

Delmelle, P., Bernard, A., 1994. Geochemistry, mineralogy, and chemical modeling ofthe acid crater lake of Kawah Ijen Volcano, Indonesia. Geochimica et CosmochimicaActa 58, 2445–2460.

Deyell, C.L., Bissig, T., Rye, R.O., 2004. Isotopic evidence for magmatic-dominatedepithermal processes in the El Indio-Pascua Au–Cu–Ag Belt and relationship togeomorphologic setting. Society of EconomicGeologists Special Publication11, 55–73.

Deyell, C.L., Leonardson, R., Rye, R.O., Thompson, J.F.H., Bissig, T., Cooke, D.R., 2005.Alunite in the Pascua-Lama high-sulfidation deposit: Constraints on alteration andore deposition using stable isotope geochemistry. Economic Geology 100, 131–148.

Driesner, T., 2007. The system H2O–NaCl: II. Correlations for molar volume, enthalpy,and isobaric heat capacity from 0 to 1000 °C, 1–5000 bar and 0–1 XNaCl. Geochimicaet Cosmochimica Acta 71, 4902–4919.

Filippou, D., St-Germain, P., Grammatikopoulos, T., 2007. Recovery of metal values fromcopper—arsenic minerals and other related resources. Mineral Processing andExtractive Metallurgy Review 28, 247–298.

Foley, N.K., Bethke, P.M., Rye, R.O., 1989. A reinterpretation of the δDH2O of inclusionfluids in contemporaneous quartz and sphalerite, Creede Mining District, Colorado:a generic problem for shallow orebodies? Economic Geology 84, 1966–1977.

Fournier, R.O., 1985. The behavior of silica in hydrothermal solutions. In: Berger, B.R.,Bethke, P.M. (Eds.), Geology and Geochemistry of Epithermal Systems. : Reviews inEconomic Geology, 2. Society of Economic Geologists, pp. 45–61.

Fournier, R.O., Potter II, R.W., 1982. An equation correlating the solubility of quartz inwater from 25° to 900°C at pressures up to 10,000 bars. Geochimica etCosmochimica Acta 46, 1969–1974.

Fulignati, P., Sbrana, A., 1998. Presence of native gold and tellurium in the active high-sulfidation hydrothermal system of the La Fossa volcano, Vulcano, Italy. Journal ofVolcanology and Geothermal Research 86, 187–198.

Giggenbach, W.F., 1987. Redox processes governing the chemistry of fumarolic gasdischarges from White Island, New Zealand. Applied Geochemistry 2, 143–161.

Hastie, J.W., 1975. High Temperature Vapors. New York, Academic Press. 480 pp.

Heald, P., Foley, N.K., Hayba, D.O., 1987. Comparative anatomy of volcanic-hostedepithermal deposits: acid sulfate and adularia-sericite types. Economic Geology 82,1–26.

Hedenquist, J.W., 1995. The ascent of magmatic fluid: Discharge versus mineralization.Mineralogical Association of Canada Short Course 23, 263–289.

Hedenquist, J.W., Matsuhisa, Y., Izawa, E., White, N.C., Giggenbach, W.F., Aoki, M., 1994.Geology, geochemistry and origin of high-sulfidation Cu–Au mineralization in theNansatsu district, Japan. Economic Geology 89, 1–30.

Hedenquist, J.W., Arribas, A., Reynolds, T.J., 1998. Evolution of an intrusion-centeredhydrothermal system; Far Southeast-Lepanto porphyry and epithermal Cu–Audeposits, Philippines. Economic Geology 93, 373–404.

Heinrich, C.A., 2005. The physical and chemical evolution of low-salinity magmaticfluids at the porphyry to epithermal transition: a thermodynamic study.Mineralium Deposita 39, 864–889.

Heinrich, C.A., Driesner, T., Stefansson, A., Seward, T.M., 2004. Magmatic vaporcontraction and the transport of gold from the porphyry environment toepithermal ore deposits. Geology 32, 761–764.

Hemley, J.J., Hostetler, P.B., Gude, A.J., Mountjoy, W.T., 1969. Some stability relations ofalunite. Economic Geology 64, 599–612.

Hemley, J.J., Montoya, J.W., Marineko, J.W., Luce, R.W., 1980. Equilibria in the Systemand Some General Implications for Alteration/Mineralization Processes. EconomicGeology 75, 210–228.

Henley, R.W., Berger, B.R., 2000. Self-ordering and complexity in epizonal mineraldeposits. Annual Reviews of Earth and Planetary Sciences 28, 669–719.

Henley, R.W., Ellis, A.J., 1983. Geothermal systems ancient and modern: a geochemicalreview. Earth Science Reviews 19, 1–50.

Henley, R.W., McNabb, Alex, 1978. Magmatic vapor plumes and groundwaterinteraction in porphyry copper emplacement. Economic Geology 73, 1–20.

Huang, C.K., Chou, S.A., 1975. Parageneses of enargite and luzonite–famatinite in theChinkuashih gold–copper deposits, Taiwan. Acta Geologica Taiwanica 18, 26–35.

Huang, C.K., Tsai, H.L., 1984. Chemical variations of enargite and luzonite–famatinitefrom the Chinkuashih gold–copper deposits, Taiwan. Acta Geologica Taiwanica 22,118–130.

Husen, S., Smith, R.B., White, G.P., 2004. Evidence for gas andmagmatic sources beneaththe Yellowstone Volcanic Field from seismic tomographic imaging. Journal ofVolcanology and Geothermal Research 131, 397–410.

Imai, Akita, 1999. Pyrite Disease in Luzonite from the Lepanto Cu–Au Deposit,Mankayan, Philippines: further example of disease texture and its origin. ResourceGeology 49, 163–168.

Jannas, R.R., 1995. Reduced and oxidizing high-sulfidation deposits of the El Indiodistrict, Chile. Unpublished Ph.D. dissertation, Harvard University, Cambridge, 421pp.

Jannas, R.R., Bowers, T.S., Petersen, U., Beane, R.E., 1999. High-sulfidation deposit typesin the El Indio district, Chile. In: Skinner, B.J. (Ed.), Geology and Ore Deposits of theCentral Andes: Society of Economic Geologists Special Publication, 7, pp. 219–266.

Kodas, T., Hampden-Smith, M., 1994. The Chemistry of Metal CVD. WeinheimPublishers, New York. 530 pp.

Kouzmanov, K., Pettke, T., Heinrich, C.A., 2010. Direct analysis of ore-precipitatingfluids: combined IR microscopy and LA-ICP-MS study of fluid inclusions in opaqueore minerals. Economic Geology 105, 351–373.

Krauskopf, K.B., 1964. The possible role of volatile metal compounds in ore genesis.Economic Geology 59, 341–410.

Larocque, A.C.L., Stima, J.A., Siebe, C., Greengrass, K., Chapman, R., Mejia, S.R., 2008.Deposition of a high-sulfidation Au assemblage from a magmatic volatile phase,Volcán Popocatépetl, Mexico. Journal of Volcanology and Geothermal Research170, 51–60.

Lindgren, W., 1913. Mineral Deposits, 1st edition. McGraw-Hill Book Co, New York. 883pp.

Losada-Calderon, A., Bloom,M.S., 1990. Geology, paragenesis and fluid inclusion studiesof a high sulphidation epithermal/porphyry system, Nevados del Famatina district,La Rioja Province, Argentina. Australasian Institute of Mining and Metallurgy,Pacific Rim Congress 90, Proceedings 2, pp. 457–464.

Mancano, D.P., 1994. Microthermometry of enargite-hosted fluid inclusions from theLepanto, Philippines, high-sulfidation Cu–Au deposit using infrared microscopy.Unpublished M.Sc. thesis, NewMexico Institute of Mining and Metallurgy, Socorro,New Mexico, 74pp.

Mancano, D.P., Campbell, A.R., 1995. Microthermometry of enargite-hosted fluidinclusions from the Lepanto, Philippines, high-sulfidation Cu–Au deposit. Geochimicaet Cosmochimica Acta 59, 3909–3916.

Martin, R.S., Roberts, T.J., Mather, T.A., Pyle, D.M., 2009. The implications of H2S and H2

kinetic stability in high-T mixtures of magmatic and atmospheric gases for theproduction of oxidized trace species (e.g., BrO and NOx). Chemical Geology 263,143–150.

Mavrogenes, J., Henley, R.W., in prep, Sulfosalt melts and heavy metal(As–Sb–Bi)fractionation during volcanic gas expansion.

Mavrogenes, J., Henley, R.W., Reyes, A., Berger, B., 2010. Sulfosalt melts: evidence ofhigh-temperature vapor transport of metals in the formation of high-sulfidationlode gold deposits. Economic Geology 105, 257–262.

Meyer, C., Hemley, J.J., 1967. Wall rock alteration. In: Barnes, H.L. (Ed.), Geochemistry ofHydrothermal Ore Deposits. Holt, Rinehart and Winston Inc, pp. 166–235.

Montegrossi, G., Tassi, F., Minissale, A.A., Vaselli, O., Buccianti, A., 2008. Naturalfluctuations of sulfur species in volcanic fumaroles. Journal of Non-EquilibriumThermodynamics 33, 75–102.

Moore, J.N., Allis, R.G., Nemcok, M., Powell, T.S., Bruton, C.J., Wannamaker, P.E., Raharjo, I.B.,Norman,D.I., 2008. The evolution of volcano-hosted geothermal systemsbased ondeepwells from Karaha–Telaga Bodas, Indonesia. American Journal of Science 308, 1–48.

74 R.W. Henley, B.R. Berger / Ore Geology Reviews 39 (2011) 63–74

Morey, G.W., Hesselgesser, J.M., 1951. The solubility of quartz and some other substancesin superheated steam at high pressures. Economic Geology 46, 821–835.

Moritz, R., 2006. Fluid salinities obtained by infrared microthermometry of opaqueminerals: implications for ore deposit modeling—a note of caution. Journal ofGeochemical Exploration 89, 284–287.

Moritz, R., Benkhelfa, F., 2009. Fluids in high-sulfidation epithermal ore deposits:constraints from infrared microthermometry of enargite-hosted fluid inclusions.Extended abstract, 10th Biennial SGA Conference, Townsville, Australia, 16–20August, pp. 533–535.

Muntean, J.L., Einaudi, M.T., 2001. Porphyry-epithermal transition: Maricunga Belt,northern Chile. Economic Geology 96, 743–773.

Nuccio, P.M., Paonita, A., Sortino, F., 1999. Geochemical modeling of mixing betweenmagmatic and hydrothermal gases: the case of Vulcano Island, Italy. Earth andPlanetary Science Letters 167, 321–333.

Okamoto, A., Saishu, H., Hirano, N., Tsuchiya, N., 2010. Mineralogical and textural variationof silica minerals in hydrothermal flow-through experiments: implications for quartzvein formation. Geochimica et Cosmochimica Acta 74, 3692–3706.

Pokrovski, G.S., Zakirov, I.V., Roux, J., Testemale, D., Hazemann, J.-L., Bychkov, A.Y.,Golikova, G.V., 2005. Experimental study of arsenic speciation in vapor phase to500 °C: implications for As transport and fractionation in low-density crustal fluidsand volcanic gases. Geochimica et Cosmochimica Acta 66, 3453–3480.

Pokrovski, G.B., Borisova, A.Y., Harrichoury, J.-C., 2008. The effect of sulfur on vapor–liquid fractionation of metals in hydrothermal systems. Earth and Planetary ScienceLetters 266, 345–362.

Posfai, M., Buseck, P.R., 1998. Relationships between microstructures and compositionin enargite and luzonite. American Mineralogist 83, 373–382.

Posfai, M., Sundberg, M., 1998. Stacking disorder and polytypism in enargite andluzonite. American Mineralogist 83, 365–372.

Pudack, C., Halter, W.E., Heinrich, C.A., Pettke, T., 2009. Evolution of magmatic vapor togold-rich liquid: the porphyry to epithermal transition at Nevados de Famatina,northwest Argentina. Economic Geology 104, 449–477.

Ransome, F.L., 1909. The geology and ore deposits of Goldfield, Nevada. U.S. GeologicalSurvey Professional Paper, 66. 258 pp.

Ransome, F.L., 1910. Geology and ore deposits of the Goldfield district, Nevada.Economic Geology 5, 301–311.

Roberge, J., Delgado-Granados, H., Wallace, P.J., 2009. Mafic magma recharge supplieshigh CO2 and SO2 gas fluxes from Popocatepetl Volcano, Mexico. Geology 37,107–110.

Ruaya, J.R., Ramos, M.N., Gonfiantini, R., 1992. Assessment of magmatic components ofthe fluids at Mt. Pinatubo volcanic-geothermal system, Philippines from chemicaland isotopic data. Report of the Geological Survey of Japan, 279, pp. 141–151.

Ruggieri, G., Lattanzi, P., Luxoro, S.S., Dessi, R., Benvenuti, M., Tanelli, G., 1997. Geology,mineralogy, and fluid inclusion data of the Furtei high-sulfidation gold deposit,Sardinia, Italy. Economic Geology 92, 1–19.

Rye, R.O., 1993. The evolution of magmatic fluids in the epithermal environment: thestable isotope perspective. Economic Geology 88, 733–752.

Rye, R.O., 2005. A review of the stable-isotope geochemistry of sulfate minerals inselected igneous environments and related hydrothermal systems. ChemicalGeology 215, 5–36.

Rye, R.O., Bethke, P.M., Wasserman, M.D., 1992. The stable isotope geochemistry ofsilica–alunite alteration. Economic Geology 87, 225–262.

Sack, R.O., Goodell, P.C., 2002. Retrograde reactions involving galena and Ag-sulfosaltsin a zoned ore deposit, Julcani, Peru. Mineralogical Magazine 66, 1043–1062.

Sakazaki, H., Ohtagaki, T., Chin, W., 1964. Geology and ore deposits of the Chin-qua-shihMine District (2). Mining Geology 14, 274–285.

Sillitoe, R.H., Hedenquist, J.W., 2003. Linkages between volcanotectonic settings, orefluid compositions and epithermal precious metal deposits. In: Simmons, S.F.,

Graham, I. (Eds.), Volcanic, Geothermal and Ore-forming Fluids: Rulers andWitnesses of Processes within the Earth: Society of Economic Geologists SpecialPublication, 10, pp. 315–343.

Simon, A.C., Pettke, T., Candela, P.A., Piccoli, P.M., Heinrich, C., 2008. The partitioningbehavior of silver in a vapor–brine–rhyolite melt assemblage. Geochimica etCosmochimica Acta 72, 1638–1659.

Skinner, B.F., 1960. Assemblage enargite–famatinite, a possible geologic thermometer(abs.). Geological Society of America Bulletin 71, 1975.

Sourirajan, S., Kennedy, G.C., 1962. The system H2O–NaCl at elevated temperatures andpressures. American Journal of Science 260, 115–141.

Sriwana, T., van Bergena, M.J., Varekampc, J.C., Sumartib, S., Takanod, B., van Ose, B.J.H.,Lengf, M.J., 2000. Geochemistry of the acid Kawah Putih lake, Patuha Volcano,WestJava, Indonesia. Journal of Volcanology and Geothermal Research 97, 77–104.

Stoffregen, R.E., 1987. Genesis of acid–sulfate alteration and Au–Cu–Ag mineralizationat Summitville, Colorado. Economic Geology 82, 1575–1591.

Stoffregen, R.E., Rye, R.O., Wasserman, M.D., 1994. Experimental studies of alunite: I.18O–16O and D–H fractionation factors between alunite and water at 250–450 °C.Geochimica et Cosmochimica Acta 58, 903–916.

Sung, J., 2005. Mineralogy and geochemistry of gold mineralization in the Goldfielddistrict, Nevada. unpublished MSc thesis, University of Utah, Salt Lake City, 155 pp.

Taran, Y.A., Bernard, A., Gavilanes, J.-C., Africano, F., 2000. Native gold in mineralprecipitants from high temperature volcanic gases of Colima volcano, Mexico.Applied Geochemistry 15, 337–346.

Taran, Y.A., Bernard, A., Gavilanes, J.C., Lunezheva, E., Cortes, A., Armienta, M.A., 2001.Chemistry and mineralogy of high temperature gas discharges from Colimavolcano, Mexico: implications for magmatic gas–atmosphere interaction. Journal ofVolcanology and Geothermal Research 108, 245–264.

Tomkins, A.G., 2010. Wetting facilitates late-stage segregation of precious metal-enriched sulfosalt melt in magmatic sulfide systems. Geology 38, 951–954.

Toulmin III, P., Clark Jr., S.P., 1967. Thermal aspects of ore formation. In: Barnes, H.L.(Ed.), Geochemistry of Hydrothermal Ore Deposits. Holt, Rinehart, and Winston,Inc, New York, pp. 437–464.

Ullrich, T., Gunther, D., Heinrich, C.A., 1999. Gold concentrations of magmatic brinesand the metal budget of porphyry copper deposits. Nature 399, 676–679.

Vennemann, T.W., Muntean, J.L., Kesler, S.E., O'Neil, J.R., 1993. Stable isotope evidencefor magmatic fluids in the Pueblo Viejo epithermal silica–alunite Au–Ag deposit,Dominican Republic. Economic Geology 88, 55–71.

Verhoogen, J., 1949. Thermodynamics of a magmatic gas phase. University of CaliforniaPublications in the Geological Sciences 28, 91–135.

Wang, Y., 2010. Physical and chemical characteristics of the host rocks in controlling themineralization of the Chinkuashih high-sulfidation gold–copper deposits, north-eastern Taiwan. Journal of Geochemical Exploration 104, 61–68.

Wang, Y., Sasaki, M., Sasada, M., Chen, C.-H., 1999. Fluid inclusion studies of theChinkuashih high-sulfidation gold–copper deposits in Taiwan. Chemical Geology154, 155–167.

Williams-Jones, A.E., Heinrich, C.A., 2005. Vapor transport of metals and the formationof magmatic-hydrothermal ore deposits. Economic Geology 100, 1287–1312.

Williams-Jones, A.E., Artashes, A.M., Sandy, M.A., Zhifeng, X., 2002. Vapor-transport ofore metals: in Hellmann, R., Wood, S.A., eds., Water–rock interactions, ore depositsand environmental geochemistry: a tribute to David A. Crerar. Geochemical SocietySpecial Publication 7, 279–305.

Wykes, J.L., Mavrogenes, J.A., 2005. Hydrous sulfide melting: experimental evidence forthe solubility of H2O in sulfide melts. Economic Geology 100, 157–164.

Yudovskaya, M.A., Distler, V.V., Chaplyygin, I.V., Mokhov, A.V., Trubkin, N.V.,Gorbacheva, S.A., 2006. Gaseous transport and deposition of gold in magmaticfluid: evidence from the active Kudryavy volcano, Kurile Islands. MineraliumDeposita 40, 828–848.