Mineralium Deposita

https://doi.org/10.1007/s00126-021-01040-9

ARTICLE

Reduction of oxidized sulfur in the formation of the Grasberg

porphyry copper-gold deposit, Papua, Indonesia
Adi Sulaksono 1,2 & Yasushi Watanabe 1 & Antonio Arribas 1,3 & Takuya Echigo 1 & Reza Al Furqan 1,4 & Clyde A. Leys 4,5

Received: 3 July 2019 / Accepted: 7 January 2021

# The Author(s), under exclusive licence to Springer-Verlag GmbH, DE part of Springer Nature 2021

Abstract
The reduction of oxidized sulfur is essential in porphyry copper deposits whose mineralization predominantly comprises copper
sulfides, whereas their source magmas are oxidized with most of their sulfur as SO42− and with exsolved fluids having
SO2>>H2S. To estimate the redox state of sulfur, we examined drill cores that intersect potassic and unaltered intrusive rocks
in the deeper levels of the Grasberg porphyry copper-gold deposit. Magmatic oxybarometers such as the amphibole-titanite-
magnetite-quartz assemblage, anhydrite, and amphibole consistently show that Grasberg ore-forming magmas were oxidized
with fO2 > FMQ+3. Initial hydrothermal events formed sulfide-free, anhydrite-rich K-feldspar, and biotite alteration, followed by
successive vein stages of (1) magnetite, (2) biotite, (3) quartz, (4) anhydrite-chalcopyrite, (5) chalcopyrite ± sericite selvages, and
(6) pyrite-chalcopyrite-quartz + sericite selvages. The δ34S values of sulfide-sulfate mineral pairs indicate SO2-derived SO42− and
H2S in SO42−/H2S molar proportions of ~ 4:1 to ~ 3:1 at > 550 °C. The hydrothermal fluid then likely followed a rock-buffered
trajectory and became more reduced at < 550 °C. Hydrothermal biotite that replaces igneous amphibole and biotite has a
phlogopitic composition, suggesting that Fe2+ was liberated from igneous mafic minerals and oxidized by reaction with SO42−
to form magnetite, resulting in sulfide formation by the simplified reaction 12FeO + SO42− + 2H+ → 4Fe3O4 + H2S.

Keywords Reduction of oxidized sulfur . Porphyry copper . SO2 . SO42 − . Rock buffer . Fe2+

Introduction (Fig. 1a). Grasberg is the world’s largest gold-rich porphyry cop-
per deposit and its mineral district also contains a number of large
The Grasberg porphyry copper-gold deposit is located in the mineralized copper-gold skarns (Cooke et al. 2005). At a cutoff
Ertsberg-Grasberg mineral district in the province of Papua, grade of 0.1% Cu, the Grasberg intrusive-related system, com-
Indonesia and lies within the New Guinea fold-and-thrust belt prising the Grasberg copper-gold porphyry and Kucing Liar
skarn deposits, contains 7.5 billion tons of in situ mineralization
with a grade of 0.7% Cu and 0.64 ppm Au (Leys et al. 2012).
Editorial handling: D. Banks Porphyry copper deposits form in subduction-related arcs
where an oceanic plate and overlying sediments subduct and
* Adi Sulaksono transfer their water and soluble components into the overlying
[email protected]
mantle wedge (Hildreth and Moorbath 1988; Leeman 1996;
Best and Christiansen 2001; Mungall 2002; Richards 2003,
1
Department of Earth Resource Science, Akita University, 1-1 2011a, b; Annen et al. 2006; Tosdal et al. 2009; Dreyer et al.
Tegatagakuenmachi, Akita 010-8502, Japan
2010; Wilkinson 2013). Hydrous mafic magmas from the
2
Department of Geological Engineering, UPN Veteran Yogyakarta, mantle wedge are then transferred into MASH (melting, as-
Jl. Padjajaran Condongcatur, Yogyakarta 55283, Indonesia
similation, storage, homogenization) zones at the base of the
3
Present address: Department of Geological Sciences, The University continental crust, followed by the ascent of magmas to shal-
of Texas at El Paso, 500 W University Ave., El Paso, TX 79968,
USA
low levels in the crust where they form upper crustal batho-
4
liths (Hildreth and Moorbath 1988; Richards 2003). These
PT Eksplorasi Nusa Jaya, Jl. Sultan Iskandar Muda Kav. V-TA,
Pondok Indah, Jakarta Selatan 12310, Indonesia
plutons that feed overlying porphyritic intrusions which may
5
host economic porphyry copper mineralization (Cloos 2001;
Freeport-McMoRan Copper and Gold, Inc., 333 N. Central Ave.,
Phoenix, AZ 85004, USA
Tosdal et al. 2009; Sillitoe 2010; Richards 2015).
Miner Deposita

Fig. 1 a Lithotectonic map of the island of New Guinea (modified from Grasberg porphyry Cu-Au deposit is hosted by multiple intrusions of
Sapiie and Cloos 2004). b Geological map of the Erstberg-Grasberg the GIC. Also shown are locations of skarn ore bodies: Kucing Liar,
district modified from Leys et al. (2012) after New (2006). The Ertsberg, Ertsberg East skarn system (EESS), Big Gosan, and Dom

Subduction-related arc magmas are characteristically Hoog et al. 2001; Wallace 2005; Jenner et al. 2010; Wallace
enriched in sulfur and are thought to undergo magmatic sul- and Edmonds 2011; Richards 2015), which may lead to mag-
fide saturation due to the onset of magnetite crystallization (de ma pre-enrichment in base and precious metals (Wilkinson
Miner Deposita

2013). Although sulfide saturation in silicate magmas may trench, following a convergence reversal event (Dewey and
trigger porphyry copper mineralization when the sulfide melt Bird 1970; Hamilton 1979). However, Cloos et al. (2005)
or its crystallized products are accessible to—and soluble in— proposed that the north-dipping slab of the oceanic end of
exsolved fluids (Keith et al. 1997; Halter et al. 2002; Sillitoe the Australian plate subducting beneath the Pacific plate broke
2008; Wilkinson 2013), most porphyry copper deposits are off and continued to sink after collision jammed the subduc-
genetically related to oxidized magmas (Ishihara 1977; tion zone, which began at approximately 8 Ma. Magma was
Hedenquist and Lowenstern 1994; Ballard et al. 2002; de generated from 7 to 4 Ma by decompression melting as the
Hoog et al. 2004; Sillitoe 2010; Richards 2015; Hattori asthenosphere upwelled into thinned lithospheric mantle
2018), in which the bulk of sulfur is present as SO42− (Cloos et al. 2005). The magma ponded in the lower crust to
(Carroll and Rutherford 1985, 1988; Métrich and Clocchiatti form a chamber several cubic kilometers in volume (Cloos
1996; Métrich et al. 2002). The solubility of sulfur in oxidized et al. 2005) that melted and assimilated lower crustal material,
melts, as SO42− at fO2 ≥ FMQ+2 (for comparison, FMQ + 2 ≈ as evidenced by its Pb, Sr, and Nd isotope compositions
NNO +1.3), is higher than that in reduced melts, in which (Housh and McMahon 2000).
sulfur dominates as sulfide (Jugo et al. 2005; Hattori 2018). Most of the porphyry-style mineralization in the Grasberg-
This solubility has the additional effect of preventing the loss Ertsberg district is associated with Pliocene igneous breccias
of metals from the melt because sulfide remains undersaturat- and intrusions of the Grasberg Igneous Complex (GIC)
ed during magma evolution prior to ore deposit formation (Paterson and Cloos 2005) that were emplaced into mainly
(Hattori and Keith 2001; Jugo et al. 2005; Hattori 2018). siliciclastic Cretaceous Kembelangan Group and carbonate-
The hydrothermal fluids exsolved from such oxidized dominated Tertiary New Guinea Limestone Group rocks
magmas are expected to have a high oxidation state, similar (Leys et al. 2012; Fig. 1b). The GIC consists of several igne-
to that of their parental magmas (Richards 2015). However, ous units (MacDonald and Arnold 1994; McMahon 1999) that
the bulk of porphyry copper mineralization occurs as copper are divided into the following units: pre-mineralization Dalam
sulfides (i.e., reduced sulfur species). Recent studies at Igneous Complex (DIC) minor subaerial volcanics, near-
Grasberg indicate that the disproportionation of SO2 gas by surface breccias, and deeper intrusions interpreted by
a Ca-bearing phase could have increased the concentration of MacDonald and Arnold (1994) to represent a maar diatreme
H2S in hydrothermal fluids that combined with dissolved base complex; slightly pre-mineralization Main Grasberg Intrusion
metals to precipitate sulfide mineralization (Henley et al. (MGI); syn-mineralization Early South Kali Dike (ESKD) and
2015; Mavrogenes and Blundy 2017; Henley and Seward the post-mineralization Late South Kali Dike (LSKD). Zircon
2018). Liang et al. (2009) proposed an alternative process at U-Pb geochronology shows that the pre-mineralization DIC
the Yulong porphyry copper-gold deposit in China where magmatism occurred between 3.60 and 3.30 Ma and was
large amounts of H2S required to precipitate abundant Cu- followed by emplacement of the MGI at 3.22 ± 0.04, ESKD
Fe-sulfides were formed by the reduction of aqueous SO42− at 3.20 ± 0.04, and LSKD at 3.09 ± 0.05 Ma respectively
via oxidation of Fe2+ in mafic minerals to Fe3+. (Wafforn 2017). Field relationships indicate that an earlier
To understand how oxidized sulfur is reduced to form por- porphyry copper-molybdenum mineralizing event predates
phyry copper mineralization, this study aims to measure the MGI emplacement, but porphyry copper-gold mineralization
evolution of the redox state of sulfur in Grasberg to determine closely post-dates the MGI, overlaps the ESKD, and predates
its role in deposit formation. We start by documenting the the LSKD (Leys et al. 2012).
Grasberg hydrothermal vein and wallrock alteration sequence
from petrography, mineral chemistry, magnetite content (from
magnetic susceptibility), and sulfur isotopes of drill core sam- Samples and analytical methods
ples spanning a ~ 2-km depth range through the deposit. We
then use the paragenetic sequences to investigate the source of For this study, 198 samples were collected from four drill
SO42− in the hydrothermal fluids and discuss the sulfur reduc- holes (IF3580-99, INF41-02, GRD39-01, and KL98-10-21)
tion processes that formed the deposit. which in combination intersect a ~ 2-km depth range through
the Grasberg porphyry copper deposit (from ~ 3600 to ~
1600 m elevation; Fig. 2). Most samples come from 3600 to
Geological background 2600 m elevation because there are few deep drill holes below
2600 m elevation, and no samples were taken between 3600 m
The Ertsberg-Grasberg district contains deformed Cretaceous elevation and the pre-mine surface at 4400 m because drill
and Tertiary sedimentary rocks within the Papuan fold-and- cores from there are strongly weathered. Samples from
thrust belt (Sapiie and Cloos 2004; Fig. 1) that are cut by IF3580-99, INF41-02, and GRD39-01 are weakly to perva-
Pliocene intrusions. Magmatism in the district was previously sively altered, whereas samples from KL98-10-21 (1900–
believed to be the result of subduction at the New Guinea 1600 m elevation) are generally unaltered. One mineralized
Miner Deposita

mineral microanalyses and sulfur isotope analysis) are provid-

ed in the Electronic Supplementary Material 1.

Results

Igneous petrography and mineral chemistry

The main GIC rock units studied are the pre-mineralization

DIC diorite (Fig. 3a), MGI quartz monzodiorite (Fig. 3b),
ESKD monzodiorite (Fig. 3c), and post-mineralization
LSKD quartz diorite (Fig. 3d). The DIC diorite surrounds
the MGI body, and is moderately to pervasively altered. It
has a slightly porphyritic texture with approximately 50 to
80% altered 1–10-mm-long plagioclase and mafic mineral
phenocrysts.
The MGI porphyritic quartz monzodiorite is generally
strongly and pervasively altered and hosts high-grade copper
mineralization (Fig. 2). It comprises, biotite, and plagioclase
and distinctive needle-shaped amphibole phenocrysts up to
65–85 vol.%, generally from 0.5 to 4 mm (rarely 10 mm), in
a groundmass that consists of K-feldspar, quartz, minor anhy-
drite, and accessory magnetite, ilmenite, apatite, and zircon.
Amphibole phenocrysts from sample FrMGI-2 (2630 m ele-
vation) show a narrow compositional range with no chemical
zoning and are principally alkali-poor, low-Al Mg-hornblende
with (Na + K)A, Altot, and Mg# [Mg/(Mg + Fe2+)] values of
0.03–0.12, 0.44–0.61, and 0.82–0.87, respectively (Electronic
Supplementary Material 2). Primary igneous biotite pheno-
Fig. 2 General SW-NE cross section of the GIC with the drill holes crysts vary from 0.2 to 5 mm in length and comprise up to
IF3580-99, INF41-02, GRD39-01, and KL98-10-21 (red solid lines are 15 vol.% of the MGI, and are distinguishable from hydrother-
samples studied). All samples from the IF3580-99, INF41-02, and mal biotite by their sub- to euhedral shape and inclusions of
GRD39-01 are hydrothermally altered, whereas samples from the
KL98-10-21 are unaltered. Heavy sulfide zone contains > 20% pyrite, apatite, magnetite, anhydrite, and zircon. The Mg# [Mg/(Mg+
with covellite, galena, enargite, and sphalerite Fe)] of igneous biotite varies from 0.62 to 0.67 (Electronic
Supplementary Material 2).
Unaltered ESKD monzodiorite at 1850 m elevation
(sample FrESKD-1) is porphyritic with phenocrysts of amphi-
vein sample was also collected from the Grasberg open pit at bole, biotite, plagioclase, and K-feldspar. These phenocrysts
3100 m elevation. Magnetic susceptibilities used to determine make up ~ 75 vol.% of the ESKD and are up to 7 mm long,
magnetite abundance were measured using a SM30 MS meter. and the groundmass of similar composition also contains ac-
Vein and alteration mineral paragenesis was determined from cessory anhydrite (1.5 vol.%; Fig. 3e, f, g), magnetite, ilmen-
core logging observations and from thin section examinations ite, titanite, apatite, zircon, and trace sulfides. The sulfides
using a Nikon ECLIPSE LV100N POL microscope and a comprise very small (~ 10 μm) chalcopyrite and pyrite inclu-
JEOL® JSM-6610 LV scanning electron microscope sions within igneous magnetite. Anhydrite occurs as inclu-
equipped with an Oxford X-MaxN energy-dispersive X-ray sions within amphibole and biotite (Fig. 3e, f) and in the
detector at Akita University. Amphibole and biotite were an- groundmass together with quartz and K-feldspar (Fig. 3f, g).
alyzed for major elements in situ using a JEOL® JXA-8800 Amphibole phenocrysts, even those containing anhydrite in-
SuperProbe at Akita University. Sample preparation for sulfur clusions, show a narrow compositional range. This amphibole
isotope analyses followed the procedure of Yanagisawa and is principally alkali-poor, low-Al Mg-hornblende, with (Na +
Sakai (1983), and sulfur isotope compositions were deter- K)A, Altot, and Mg# values of 0.13–0.30, 0.58–1.04, and
mined using a Thermo Fisher Scientific Delta V Advantage 0.75–0.84, respectively, and is sometimes associated with
isotope ratio monitoring mass spectrometer at Akita titanite, magnetite, and quartz (Fig. 3h). Igneous biotite phe-
University. Details of the analytical methods (quantitative nocrysts have Mg# of ~ 0.57, are subhedral to euhedral, 0.2 to
Miner Deposita

Fig. 3 Representative polished

slabs of the GIC intrusive rocks of
a pre-mineralization DIC diorite,
sample BV-9, 3347 m elevation;
b syn-mineralization MGI quartz
monzodiorite, sample GRD-80,
2668 m elevation; c syn-
mineralization ESKD
monzodiorite, sample FrESKD-1,
1850 m elevation; d post-
mineralization LSKD quartz dio-
rite, sample FrLSKD-3, 1875 m
elevation. e Photomicrograph of
the ESKD monzodiorite, showing
anhydrite inclusions within am-
phibole phenocrysts. Sample
FrESKD-1. f Anhydrite inclusion
within biotite phenocryst. Sample
FrESKD-1. g Anhydrite in a
groundmass. Sample FrESKD-
1. h Backscattered electron image
(BEI) of the assemblage of Mg-
rich amphibole, titanite, magne-
tite, and quartz. Sample FrESKD-
1. i BEI of hydrothermal biotite
with chemical compositions more
enriched in phlogopite compo-
nent after igneous amphibole.
Sample INF-11, 3215 m eleva-
tion. j Hydrothermal phlogopite
after igneous biotite. Also shown
are associated hydrothermal
magnetite grains. Sample INF-11.
Amp, amphibole; Anh, anhydrite;
Ap, apatite; Bt, biotite; Bn, born-
ite; Ccp, chalcopyrite; Kfs, K-
feldspar; Mag, magnetite; Mg-
Amp, magnesium-rich amphi-
bole; Mol, molybdenite; Pl, pla-
gioclase; Phl, biotite with
phlogopitic composition; Py, py-
rite; Qz, quartz; Ser, sericite; Ttn,
titanite; Zrn, zircon

1.0 mm long, have brown pleochroism, and occasionally con- hornblende with (Na + K)A, Altot, and Mg# values of 0.07–
tain apatite, magnetite, anhydrite, and zircon inclusions. 0.16, 0.48–0.71, and 0.81–0.85, respectively.
The youngest GIC unit studied is the LSKD quartz diorite
at elevations of 1940, 1890, and 1875 m. This equigranular Hydrothermal alteration-mineralization
unit is generally fresh-to-weakly altered and consists mainly
of biotite, plagioclase, quartz, amphibole, and K-feldspar, The Grasberg porphyry deposit is dominated by a sub-vertical,
with accessory magnetite, ilmenite, anhydrite, zircon, apatite, cylindrical zone of potassic alteration centered on the MGI
and rare sulfides. A LSKD amphibole from sample FrLSKD-2 that is surrounded by a selvage of sericite alteration that ex-
(1890 m elevation) has a narrow compositional range with no tends to the margins of the GIC. These alteration zones patch-
chemical zoning and is principally alkali-poor, low-Al Mg- ily overlap the margins of the ESKD but are truncated by the
Miner Deposita

LSKD (Leys et al. 2012). The highest copper grades (> 2% hydrothermal biotites have chemical compositions more
Cu) are associated with the strongest and most pervasive po- enriched in the phlogopite component than igneous biotite
tassic, and to a lesser extent sericite alteration zones. Almost (Fig. 5).
all of the drill core samples studied lie within the predomi-
nantly potassic-altered Grasberg > 0.5 wt.% Cu zone, with Magnetite veins
only a few from a lower-grade, deeper zone in drill hole
GRD39-01 (Fig. 2). In addition to replacing mafic minerals within the GIC potas-
Most samples studied are potassic altered that comprise K- sic alteration zone, magnetite is commonly found as veins
feldspar, biotite, magnetite, quartz, anhydrite, chalcopyrite, (Fig. 4f, g) in samples from drill holes IF3580-99 and
bornite, apatite, and rutile. Very few samples contain sericite INF41-02 and the upper part of GRD39-01. These magnetite
alteration that overprints potassic alteration and comprises veins may also contain quartz and biotite, are generally < 3-cm
white mica, quartz, chlorite, anhydrite, pyrite, chalcopyrite, thick, and lack selvages. Magnetite veins are particularly
bornite, apatite, and rutile. This alteration typically replaces abundant in a sample from drill hole INF41-02 where they
plagioclase and K-feldspar in centimeter-scale (< 5 cm wide) occur as sheeted structures (Fig. 4g).
alteration selvages to 1–10-mm-wide chalcopyrite and quartz-
pyrite-chalcopyrite veins. Sericite alteration is uncommon in Biotite veinlets
the samples studied because sampling was focused on MGI
and DIC intrusions within the central potassic alteration zone; Biotite veinlets are common within GIC potassic alteration
however, in DIC rocks outboard of this zone, sericite alter- and are typically < 1-mm thick, discontinuous, and irregular
ation is intense and pervasive (Leys et al. 2012). and in places associated with magnetite, K-feldspar, anhy-
Based on detailed mineralogical and textural relationships, drite, quartz, and minor chalcopyrite and bornite (Fig. 4f).
the potassic and sericite alteration zones are sub-divided into K-Feldspar selvages around biotite veinlets which cut plagio-
10 detailed alteration and vein stages (Electronic clase phenocrysts are often significantly wider than the vein-
Supplementary Material 2). lets. Vein biotites have chemical compositions enriched in the
phlogopite component (Fig. 5).
K-Feldspar alteration
Sugary quartz veins
K-Feldspar is one of the most abundant hydrothermal min-
Sugary quartz veins are widely seen in potassically altered
erals in potassically altered rocks, and typically replaces pla-
samples from drill hole INF41-02 and the upperpart of
gioclase phenocrysts and groundmass phases. K-Feldspar has
GRD39-01 as locally intense stockworks. The veins are wavy,
mostly irregular crystal shapes with serrate or wavy margins.
irregular, and discontinuous, generally < 5 cm thick, and con-
K-Feldspar also occurs as selvages around anhydrite veinlets
sist of fine-grained, polygonal quartz aggregates often with
(Fig. 4a) as well as within them (Fig. 4b). Locally, K-feldspar
magnetite, biotite, chalcopyrite, bornite, and molybdenite
and magnetite replace igneous biotite (Fig. 4c).
(Fig. 4f, h, l).

Replacement biotite Comb quartz veins

Hydrothermal biotite within the potassic alteration zone is Comb quartz veins are abundant in potassically altered sam-
common in samples from drill holes IF3580-99, INF41-02, ples from well-mineralized (1~4% Cu) sections of drill hole
and GRD39-01 as pseudomorphs after igneous amphibole IF3580-99 and INF41-02 and are characterized by subhedral
(Figs. 3i, and 4d, e), and less commonly replaces primary quartz crystals (Fig. 4h, i). These veins are typically 1 to 5 cm
biotite phenocrysts (Fig. 3j). Igneous amphibole is replaced thick, lack selvages, and often contain central zones rich in
by shreddy hydrothermal biotite, with or without magnetite, chalcopyrite and bornite especially in their wider segments
K-feldspar, quartz, rutile, apatite, anhydrite, and minor chal- (Fig. 4i).
copyrite and bornite. Igneous biotite is replaced by fine-
grained biotite aggregates which occasionally contain hydro- Planar quartz veins
thermal magnetite (Fig. 3j).
Hydrothermal biotite after amphibole (Fig. 3i) and biotite Planar quartz veins (Fig. 4i, k) are common in potassically
phenocrysts (Fig. 3j) is characterized by higher Mg# and low- altered samples from holes GRD39-01 and INF41-02 and in
er Fe, Ti, and Altot contents than igneous biotite (Electronic places contain minor amounts of biotite, chalcopyrite, and
Supplementary Material 2). In annite-siderophyllite- bornite. Veins vary from < 1 up to 5 cm and typically have
phlogopite-eastonite quadrilateral diagrams, most sharp planar walls with no selvages.
Miner Deposita

Fig. 4 Successive vein and

alteration stages observed in the
drill cores studied. a Anhydrite
veinlet with K-feldspar alteration
selvage. Sample GRD-35,
2887 m elevation. b
Discontinuous, wavy anhydrite-
dominated veinlet associated with
K-feldspar and biotite. Sample
GRD-35. c Biotite phenocryst
that is partially replaced by K-
feldspar and magnetite. Circles
indicate a replacement texture.
Sample BVG-24, 2703 m eleva-
tion. d, e Amphibole phenocryst
that is converted to hydrothermal
biotite, K-feldspar, anhydrite,
magnetite, and quartz. Sample: d
GRD-35; e FrESKD-3, 2594 m
elevation. f Cross-cutting vein re-
lationships: magnetite veins are
cut by biotite stage veins, and
both magnetite and biotite veins
are cut by a sugary quartz vein
and a chalcopyrite vein with
sericite selvage. Sample INF-22,
3296 m elevation. g Parallel,
sheeted vein structure of magne-
tite. Sample GMCINF-12, 3314
m elevation. h Sugary quartz
vein, cut by a comb quartz vein,
which contains a centerline of
chalcopyrite. A chalcopyrite seam
extending from the chalcopyrite
centerline also cuts the sugary
quartz vein. Sample INF-8,
3199 m elevation. i Comb quartz
vein that contains a chalcopyrite
centerline, cut by planar quartz
vein with minor biotite. Sample
INF-13, 3218 m elevation. j
Anhydrite-chalcopyrite veins
with some quartz and molybde-
nite. Sample GRD-45, 2838 m
elevation. k Planar quartz vein
with minor chalcopyrite and
bornite cut by an anhydrite-
chalcopyrite vein (Grasberg open
pit mine, 3100 m elevation). l
Pyrite-chalcopyrite-quartz vein
with sericite alteration selvage
cutting an earlier sugary quartz
vein. Sample INF-18, 3275 m el-
evation. See Fig. 3 caption for
mineral abbreviations

Anhydrite-chalcopyrite veins amounts of anhydrite relative to chalcopyrite with depth, and

are associated with quartz, biotite, bornite, and molybdenite.
Anhydrite-chalcopyrite veins are generally 1 to 5 cm thick,
although they can reach > 50 cm in places, and are seen in all Chalcopyrite veins ± sericite selvages
IF3580-99, INF41-02, and GRD39-01 potassically altered
samples, especially in zones containing high Cu grades up to Chalcopyrite veins are widely observed within all potassically
~ 4% Cu. Veins lack selvages (Fig. 4j, k), display increasing altered samples from IF3580-99, INF41-02, and GRD39-01
Miner Deposita

anhydrite (Fig. 6). Magnetic susceptibility measurements are

highest (> 100 × 10−3 SI) at 3300~3360 m elevation, where
the DIC diorite contains > 15 vol.% magnetite and 1~3%
Cu. In the upper part of drill hole GRD39-01 at
2900~3040 m elevation, the highest Cu values (2~4% Cu)
are associated with magnetic susceptibilities ranging from 10
to 100 × 10−3 SI, which corresponds to 2~15 vol.% magnetite
from visual estimates. In the middle part of hole GRD39-01,
the DIC diorite contains no magnetite but has grades of ≤ 2%
Cu due to sericite alteration that overprints the potassic alter-
ation and has removed its hydrothermal magnetite.

Sulfur isotope ratios and temperature estimates

Fig. 5 Composition of hydrothermal and igneous biotite crystals from the The sulfur isotope ratio of sulfur-bearing minerals in sixty
DIC diorite and MGI quartz monzodiorite expressed in the Fe/(Fe+Mg) altered and one unaltered igneous samples was measured
versus Altot diagram. Most hydrothermal biotite crystals, including those
replacing igneous amphibole and biotite minerals, have chemical
(Table 1, Fig. 7). Sulfur-bearing minerals analyzed include
compositions more enriched in phlogopite component than igneous those contained in (a) comb quartz veins with chalcopyrite
biotite crystals centerlines, (b) anhydrite-chalcopyrite veins, (c) chalcopyrite
veins with sericite selvages, and (d) anhydrite-bearing pyrite-
drill holes and are characteristically narrow (~ 1 to 5 mm) but chalcopyrite-quartz veins with sericite selvages.
may reach 3 cm and are mostly planar. The veins locally The δ34S values of anhydrite are invariably higher than
contain bornite, anhydrite, quartz, and apatite, and have those of coexisting sulfides. The ranges of δ34S values of
sericite selvages (Fig. 4f). The selvages contain white mica sulfur-bearing minerals hosted in the anhydrite-chalcopyrite
with or without chlorite, anhydrite, quartz, pyrite, chalcopy- veins are 1.2–4.2‰ for chalcopyrite (n = 24; avg. 2.3‰), 1.1
rite, bornite, apatite, and rutile. This vein type typically occurs and 2.1‰ for bornite (n = 2), and 10.5–13.8‰ for anhydrite
in the well-mineralized transition zone from the GIC potassic (n = 26; avg. 12.2‰). Two chalcopyrite samples that occur
to the sericite alteration zone where grades can locally reach along the centerlines of comb-textured quartz veins have
up to ~ 3% Cu. δ34S values of 2.4 and 2.5‰ (n = 2). One chalcopyrite vein
that is accompanied by sericite alteration has δ34S = 2.1‰.
Pyrite-chalcopyrite-quartz veins with sericite selvages The δ34S values for pyrite in a veinlet with sericite selvage
are 2.8 to 4.8‰ (n = 3; avg. 3.7‰), and those of anhydrite are
Pyrite-chalcopyrite-quartz (± anhydrite) veins are common in 14.6–15.2‰ (n = 2). A sample of the syn-mineralization
samples from INF41-02 and GRD39-01, are mostly 1 to ESKD monzodiorite yields a δ34S of 9.4‰.
10 mm thick, and have associated sericite selvages that are The δ34S values of chalcopyrite in comb quartz veins and
mainly < 5 cm thick (Fig. 4l). The selvages comprise white anhydrite-chalcopyrite veins are similar, whereas pyrite from
mica that replaces plagioclase and K-feldspar that can have pyrite-chalcopyrite-quartz veins with sericite selvages has
associated chlorite, anhydrite, quartz, pyrite, chalcopyrite, higher average δ 34 S than chalcopyrite in anhydrite-
bornite, apatite, and rutile. Veins of this type are abundant chalcopyrite veins (Fig. 7). Likewise, anhydrite from pyrite-
within GIC sericite alteration zones and where veining is most chalcopyrite-quartz veins with sericite selvages has a heavier
intense all wallrock minerals are completely replaced by isotopic composition than anhydrite in anhydrite-chalcopyrite
sericite. veins.
Assuming equilibrium and using fractionation calculations
Hydrothermal anhydrite and magnetite abundance in from Ohmoto and Rye (1979) and Ohmoto and Lasaga
relation to copper grade (1982), the Δ values of coexisting sulfate and sulfide minerals
in anhydrite-chalcopyrite-bearing veins within high-grade (1–
To study the relationship between hydrothermal anhydrite and 4% Cu) core intervals indicate temperatures of 645–475 °C
magnetite with copper grade, the abundance of these minerals (Fig. 7). By contrast, temperatures estimated with sulfate-
was estimated both visually and in thin section. Magnetite sulfide pairs from anhydrite-bearing pyrite-chalcopyrite-
estimation was supported by magnetic susceptibility measure- quartz veins with sericite selvages indicate formation temper-
ments, since little to no relict igneous magnetite occurs within atures of 465–460 °C. This matches observations from many
GIC altered units. Between 3600 and 2600 m elevation, hy- other porphyry deposits whose potassic alteration is believed
drothermal magnetite content correlates negatively with to be the product of high-temperature hydrothermal fluids that
Miner Deposita

Fig. 6 Relationships among rock

unit, alteration, hydrothermal
magnetite abundance, magnetic
susceptibility, hydrothermal
anhydrite abundance, and copper
concentration. Potassic alteration
zone contains centimeter-scale
sericite alteration of ~1 vol.%.
Copper mineralization has a pos-
itive correlation with the abun-
dance of hydrothermal magnetite
and magnetite susceptibility,
which is in inverse proportion to
the abundance of hydrothermal
anhydrite. Com., common; Abd.,
abundant

cool as they ascend and radiate outwards, forming a sericite et al. 2008; Hutchinson and Dilles 2019). Large amounts of
alteration halo around a core of potassic alteration (Seedorff sulfur in low-temperature, intermediate-composition magma
et al. 2005). are stored in anhydrite (1 vol.% anhydrite is ~ 2300 ppm S)
that may break down to provide sulfur to mineralizing fluids
(Hutchinson and Dilles 2019). At Grasberg, the syn-
mineralization MGI and ESKD and post-mineralization
Discussion
LSKD contain igneous anhydrite in the groundmass (Fig.
3f, g) and as inclusions within amphibole and biotite pheno-
Magma oxidation state
crysts (Fig. 3e, f). Anhydrite in the groundmass indicates that
the melt remained oxidized after phenocryst crystallization.
Porphyry copper deposits are genetically related to oxidized
High magma fO2 is also shown by Mg-rich amphibole rims
magmas (Ishihara 1977; Hedenquist and Lowenstern 1994;
(Mg# = 0.80–0.86; see Wones 1989) that contain associated
Ballard et al. 2002; de Hoog et al. 2004; Sillitoe 2010;
igneous titanite-magnetite-quartz assemblages that are well
Richards 2015; Hattori 2018) whose high oxidation state is
preserved in the ESKD monzodiorite (Fig. 3h). Although the
reflected by the presence of igneous anhydrite (Chambefort
Miner Deposita

Table 1 Sulfur isotope data

sorted by elevation No. Sample Elevation (m) Vein mineralogy/rock sample δ34S (‰)

Anh Ccp Py Bn Bulk

1 IF-1 3549.2 Qz-Bt+Ccp-Bn centerline 2.5

2 IF-2 3509.8 Qz-Bt+Ccp-Bn centerline 2.4
3 IF-3 3462.5 Anh-Qz-Ccp-Bn 11.9 2.6
4 INF-28 3317.4 Anh-Qz-Ccp 12.3 3.2
5 INF-27 3313.5 Anh-Qz-Ccp 12.4 2.7
6 INF-18 3275.4 Py-Ccp-Anh-Qz+Ser selvages 14.6 2.8
7 INF-16 3272.5 Py-Ccp-Anh-Qz+Ser selvages 15.2 3.6
8 INF-12 3216.5 Anh-Qz-Ccp 11.2 2.9
9 INF-9 3206.0 Anh-Qz-Ccp-Bt 13.3 2.6
10 INF-6 3194.2 Anh-Ccp 11.8 3.3
11 INF-4 3180.3 Anh-Ccp 12.3 2.5
12 GOP-1 3100.0 Anh-Qz-Ccp-Bn 12.6 2.0 2.1
13 GRD-12 3007.3 Bt-Anh-Qz-Ccp-Bn 12.2 2.1
14 GRD-15 2996.8 Anh-Qz-Ccp-Bn 12.1 2.9
15 GRD-20 2981.4 Anh-Qz-Ccp 11.0 2.4
16 GRD-21 2975.8 Ccp+Ser selvages 2.1
17 GRD-27 2917.1 Anh-Qz-Ccp-Mol 13.0
18 GRD-30 2911.2 Anh-Qz-Ccp-Bn 10.5 2.1
19 GRD-33 2898.9 Anh-Qz-Ccp 13.6 2.3
20 GRD-35 2887.0 Anh-Ccp 13.2 2.7
21 GRD-36 2885.3 Anh-Qz-Ccp 12.9 2.5
22 GRD-40 2861.2 Anh-Qz-Ccp 12.9 1.5
23 GRD-45 2838.2 Anh-Qz-Ccp-Mol 13.8 1.6
24 GRD-50 2807.8 Anh-Qz-Ccp 11.7 2.1
25 GRD-53 2794.2 Anh-Qz-Ccp-Bn 11.0 1.1
26 GRD-55 2780.4 Anh-Qz-Ccp-Bn 12.9 4.2
27 GRD-60 2750.3 Py-Qz+Ser selvages 4.8
28 GRD-63 2738.8 Anh-Qz-Ccp 11.8 1.3
29 GRD-66 2720.9 Anh-Qz-Ccp 12.8 1.2
30 GRD-71 2700.9 Anh-Qz-Ccp-Bn-Bt 11.3 1.3
31 GRD-77 2684.2 Anh-Ccp 11.8 2.4
32 GRD-78 2679.5 Anh-Ccp 11.3 1.3
33 FrESKD-2 1703.0 Monzodiorite (ESKD) 9.4

ESKD monzodiorite contains trace amounts of fine-grained and their contact or textural equilibrium with—and inclusions
(~ 10 μm) pyrite and chalcopyrite in magnetite grains, it and of—plagioclase and other minerals such as biotite, K-feldspar,
other intrusive rocks are generally sulfide-free. Therefore, the quartz, zircon, magnetite, titanite, and apatite. The Al-in-
contribution to magma sulfur content from sulfide at anhydrite amphibole geobarometer of Mutch et al. (2016) suggests that
saturation is insignificant (Jugo 2009), and its fO 2 is these amphibole rims crystallized in the upper crust at 0.9–-
interpreted to have been above the sulfide-sulfur oxide buffer. 1.3 kbar (~ 3–5 km depth).
Amphibole rim compositions also indicate strongly oxidized
magma with ΔFMQ = 3.4 ± 0.09 (n = 5), 3.1 ± 0.10 (n = 3), High-temperature SO2 disproportionation
and 3.4 ± 0.10 (n = 5; 1 s.d.) for the MGI, ESKD, and
LSKD, respectively (see Ridolfi et al. 2010 for fO2 and Sulfur dioxide is the principal sulfur species in gases evolved
temperature estimates). Such amphibole rims are inferred to from GIC magmas in which SO42− greatly exceeds H2S
have formed under near-solidus conditions, as indicated by (Henley et al. 2015; Mavrogenes and Blundy 2017; Henley
relatively low crystallization temperatures (~ 700–750 °C) and Seward 2018; Mernagh et al. 2020). The decrease in SO2/
Miner Deposita

Fig. 7 The δ34S values projected

for magmatic and hydrothermal
stages. *Some anhydrite-
chalcopyrite veins are interpreted
to be formed at the earliest stage
of potassic alteration

H2S during hydrothermal processes may be related to dispro- observed in DIC and MGI rocks (Paterson and Cloos 2005)
portionation of magmatic SO2 (Whitney 1988; Holland 1965; and other alteration minerals such as biotite, muscovite, or K-
Kusakabe et al. 2000): feldspar are ubiquitous in the GIC. These alteration minerals
may incorporate Al and Si generated by Reaction (2) if they
4SO2ðgÞ þ 4H2 OðaqÞ →H2 SðaqÞ þ 3H2 SO4ðaqÞ ð1Þ
react with alkalis, iron, and magnesium (Henley and Seward
2018), and andalusite that accommodates Al from Reaction (2)
This reaction occurs progressively as magmatic-
may have been altered to diaspore and pyrophyllite during ret-
hydrothermal fluids cool below 400–350 °C (Holland 1965;
rograde cooling and hydrolysis (Henley and Seward 2018).
Henley and McNabb 1978; Kusakabe et al. 2000). However,
Sulfur isotopes from anhydrite and sulfide minerals formed at
our sulfur isotope data indicate that coexisting sulfate and
> 620 °C show a significant increase in their δ34S ratios (Fig. 8d)
sulfide minerals formed at temperatures of 645 to 460 °C.
that may be due to enrichment of 34S in the SO42−- and H2S-
Alternatively, the decrease in SO42−/H2S (e.g., Ertsberg
bearing liquid (Field et al. 2005) after unmixing of a single-phase
skarn system; Gibbins 2006) within hydrothermal fluids may
magmatic SO2 gas into liquid and vapor. Hereafter, we assume
be due to addition of H2S degassed from mafic magma at
that SO2 gas was the main sulfur species exsolved from GIC
depth (Ohmoto and Rye 1979) and/or by the leaching and
magmas (Henley et al. 2015; Mavrogenes and Blundy 2017;
reprecipitation of sulfides (Ohmoto and Goldhaber 1997).
Henley and Seward 2018; Mernagh et al. 2020). When SO2
Another mechanism for the evolution from oxidized sulfur-
disproportionates, H2S is partitioned as both aqueous and gas-
dominant (S6+ and/or S4+) to reduced sulfur-dominant (S2−)
eous species into the liquid and low-density vapor, respectively,
fluid is by reaction between the hydrothermal fluid and igne-
whereas SO42− remains in the liquid as an aqueous species since
ous wallrocks (Carmichael and Ghiorso 1986; Rye 1993;
the concentration of H2SO4 gas is negligible. Although the par-
Ohmoto and Goldhaber 1997; Liang et al. 2009; Henley
tition coefficient for H2S between liquid and vapor is unknown,
et al. 2015; Mavrogenes and Blundy 2017; Henley and
removal of the H2S-bearing vapor would cause the remaining
Seward 2018).
liquid to become enriched in 34S because H2S is depleted in 34S
Recent studies provide evidence for SO2 disproportion-
relative to SO42−.
ation at high temperature and pressure by reaction with high-
Plots of δ34S values for associated sulfate and sulfide min-
anorthite plagioclase within igneous wallrocks (Henley et al.
erals versus the Δ values of the mineral pairs provide an
2015; Mavrogenes and Blundy 2017; Watanabe et al. 2018;
estimation of the molar proportion of SO42−/H2S (Field and
Henley and Seward 2018). This process may be written syn-
Gustafson 1976; Kusakabe et al. 1984; Imai 2000; Field et al.
optically as follows (Henley and Seward 2018):
2005; Hedenquist et al. 2017). The suites of coexisting sulfate-
3CaAl2 Si2 O8 þ 4SO2ðgÞ þ 22H2 OðgÞ →3CaSO4 sulfide mineral pairs in anhydrite-chalcopyrite veins suggest
that in the temperature range of 645 to 475 °C, the bulk sulfur
þ 6AlðOHÞ3ðgÞ þ 6H4 SiOðgÞ þ 3SiO2 þ H2 S ð2Þ isotope composition (δ34SΣS) was 6.4‰ and the SO42−/H2S
molar proportion was ~ 2:3 (XSO42− = 0.42; XH2S = 0.58), as-
Although our microscopic observations found no clear evi- suming that their proportion remained constant as temperature
dence for the conversion of plagioclase to anhydrite ± andalu- decreased (Fig. 8a). These results are not consistent with
site ± sulfide minerals or the growth of anhydrite on the surfaces Reaction (2), from which a molar SO42−/H2S of 3:1 is predict-
of plagioclase crystals, hydrothermal andalusite has been ed, and the calculated δ34SΣS of 6.4‰ is also much lower than
Miner Deposita

Fig. 8 a Plot of δ34S values for associated sulfate and sulfides of veins 1:1 (more reduced) at a temperature below 550 °C, with δ34SΣS ≈ 7.4‰. d
formed during potassic episode vs. the Δ value of the pair. Assuming a Evolution of the oxidation state of hydrothermal fluids based on sulfur
constant proportion of SO42−/H2S, the convergence and slopes of the two isotope as a function of temperature. The fluid at > 550 °C record the
regression lines offer an approximation of the bulk sulfur isotope SO42− and H2S end-products of SO2 disproportionation which yields
composition (δ34SΣS) ≈ 6.4‰ and the proportion of oxidized to reduced SO42− and H2S in the molar proportion of ~ 4:1 to ~ 3:1, and δ34SΣS of
sulfur (~ 2:3 SO42−/H2S) in the hydrothermal system (adapted from Field ~ 9.4‰. The decreases in δ34 S ΣS and SO 42− /H 2S proportion are
et al. 2005). b Assuming that the proportion of SO42−/H2S was not con- interpreted to be due to early anhydrite precipitation and reduction of
stant, the SO42−/H2S proportion was ~ 4:1 (oxidized) at a temperature SO42− to H2S via a coupled oxidation of Fe2+ to Fe3+ (see text)
above 550 °C, with δ34SΣS ≈ 10.2‰. c The SO42−/H2S proportion was

the 9.4‰ observed in syn-mineralization E SKD 0.83; XH2S = 0.17) (Fig. 8b) and more reduced at 1:1 below
monzodiorite. 550 °C (Fig. 8c). The ~ 4:1 proportion of SO42−/H2S is slight-
An alternative interpretation is that the proportion of SO42 ly higher than the expected value of 3:1, perhaps indicating
−
/H2S was not constant during cooling of hydrothermal fluids, removal of H2S within magmatic vapor. Sulfur isotope results
as indicated by the sulfur isotope data that vary with temper- at > 550 °C show values of δ34SΣS = 10.2‰, which, although
ature. Above 550 °C, the SO42−/H2S was ~ 4:1 (XSO42− = reasonable, are much higher than the expected value of ~ 7‰
Miner Deposita

for an oxidized magma (e.g., Ohmoto and Rye 1979). The igneous wallrocks could have reduced SO42− to H2S via a
elevated δ34SΣS could be due to incorporation of wallrock coupled oxidation of Fe2+ to Fe3+ (Carmichael and Ghiorso
sulfate-bearing sediments into the magma (e.g., Butte district, 1986; Sun et al. 2004; Liang et al. 2009) and thus decreased
Montana: Field et al. 2005; Ertsberg skarn system: Gibbins the XSO42− in the hydrothermal fluid.
2006). A potential source of evaporite exists in the New The earliest K-feldspar and associated magnetite replace
Guinea Limestone Group, in which anhydrite nodules have biotite phenocrysts, where magnetite formation was likely
δ34S values between 20 and 22‰ (Claypool et al. 1980). due to oxidation of Fe2+ in biotite by aqueous SO42−. The
Reaction (2) ideally yields SO42− and H2S in the molar process is believed to have followed the reaction path predict-
proportion of 3:1 (XSO42− = 0.75; XH2S = 0.25). Assuming ed by Liang et al. (2009):
that the SO42− and H2S formed by such a reaction are subse-
quently incorporated into anhydrite and sulfides formed at > 8KFe3 AlSi3 O10 ðOHÞ2 þ 2H2 SO4ðaqÞ →8KAlSi3 O8
550 °C, we can obtain the δ34SΣS from: þ 8Fe3 O4 þ 8H2 O þ 2H2 S ð4Þ

δ34 SΣS ¼ 0:75 δ34 Ssulfates þ 0:25 δ34 Ssulfides ð3Þ Replacement of igneous amphibole and biotite by hydro-
by substituting the mean δ S values for anhydrite and sulfide
34 thermal biotite is observed in many porphyry deposits around
minerals (Field et al. 2005). The result of this calculation gives the world (Bingham, Utah: Moore and Czamanske 1973;
a δ34SΣS of 9.4‰ that equals the measured δ34S value of syn- Butte, Montana: Roberts 1975; Santa Rita, New Mexico:
mineralization ESKD monzodiorite, suggesting that this intru- Jacobs and Parry 1979; Casino, Canada: Selby and Nesbitt
sion directly represents the δ34SΣS of the system. 2000). Brimhall et al. (1985) also report replacement of
edenite (amphibole endmember) by an anhydrite-bearing al-
teration assemblage:
Reduction of SO2-derived aqueous sulfate
NaCa2 Mg5 ðAlSi7 O22 ÞðOHÞ2 þ 0:5K2 SO4ðaqÞ
Hydrothermal anhydrite with a negligible amount of sulfide þ 4H2 SO4ðaqÞ →KMg3 ðAlSi3 O10 ÞðOHÞ2 þ 4SiO2
minerals is observed in the earliest K-feldspar and biotite al-
teration (Fig. 4a–e). Because SO42− is enriched in 34S relative þ 2CaSO4 þ 0:5Na2 SO4ðaqÞ þ 2 MgSO4ðaqÞ þ 4H2 O ð5Þ
to the δ34SΣS of SO42−-bearing liquid, the precipitation of
anhydrite removes 34S-enriched SO42− so that the remaining This reaction applies to early alteration produced by SO42−-
liquid becomes depleted in 34S, assuming that aqueous SO42− bearing hydrothermal fluid, but note that there is no magnetite
and H2S in liquid contributed sulfur to the precipitation of precipitation and sulfur remains as sulfate. Most amphibole phe-
anhydrite ± sulfide vein/alteration. The ~ 3:1 proportion of nocrysts in the DIC diorite and MGI quartz monzodiorite have
SO42−/H2S at > 550 °C indicates a high activity of SO42− in been replaced by biotite (Figs. 3i, 4d), and biotite phenocrysts have
high-temperature magmatic-hydrothermal fluids (Brimhall re-equilibrated into hydrothermal biotite with lower Fe content
et al. 1985; Liang et al. 2009). The SO42−/H2S proportion (Fig. 3j). Hydrothermal biotite is commonly associated with hy-
was not constant during cooling of fluids as described above, drothermal magnetite, indicating that Fe2+ was liberated and oxi-
such that SO42−/H2S became more reduced (1:1) at < 550 °C. dized to form magnetite, resulting in the reduction of sulfur
Our observations suggest that Grasberg potassic alteration (Carmichael and Ghiorso 1986; Sun et al. 2004; Liang et al. 2009):
evolved through several alteration and vein stages (Fig. 4) from
12FeOðrockÞ þ SO4 2− þ 2Hþ →4Fe3 O4 þ H2 S ð6Þ
initial anhydrite-bearing K-feldspar and biotite alteration (Henley
and Seward 2018), followed by magnetite, biotite, quartz, and The role of this reaction in Grasberg porphyry copper min-
anhydrite-chalcopyrite veining (Pollard and Taylor 2002). Early eralization is illustrated in Fig. 9 that explains the increase in
K-feldspar and biotite alteration is frequently associated with hydrothermal fluid H2S concentration required to precipitate
hydrothermal magnetite that replaces igneous mafic minerals the abundant copper sulfides (Kouzmanov and Pokrovski
(Figs. 3j, 4e). Magnetite veins (Fig. 4f, g) formed during the 2012). The Grasberg-related system contains 52.5 Mt of Cu
mid-stage of potassic alteration, whereas end-stage potassic al- (7.5 Gt of ore at 0.7% Cu) with copper primarily contained in
teration formed anhydrite-chalcopyrite and chalcopyrite veins chalcopyrite (CuFeS2) and bornite (Cu5FeS4) with a relative
(Fig. 4h–k). Hydrothermal magnetite is interpreted to have abundance of 75% and 25%, respectively (Leys et al. 2012).
formed coincident with the reduction of SO42− during the potas- Thus, approximately 45 Mt of S2− is required to precipitate
sic episode. This process is consistent with drill core analysis, in 52.5 Mt of Cu. According to Reaction (6), the formation of
which copper mineralization has a positive correlation with the approximately 1.3 Gt of hydrothermal magnetite could have
abundance of hydrothermal magnetite, which is inversely corre- reduced approximately 45 Mt of S2−. The total amount of
lated with the abundance of hydrothermal anhydrite (Fig. 6). We hydrothermal magnetite of 1.9 Gt in the Grasberg deposit
infer that, in addition to early anhydrite precipitation, Fe in
Miner Deposita

sulfur species and oxidation states in the formation of porphyry

copper deposits. Recent studies at Grasberg indicate that SO2 gas
was the main sulfur species exsolved from GIC magmas, and is
believed to have disproportionated to form SO42− and H2S by
reaction with plagioclase at high temperature and pressure. Phase
separation of magmatic SO2 into liquid (aqueous SO42− and
H2S) and vapor (gaseous H2S) led to the enrichment in the 34S
of SO42−- and H2S-bearing liquid, as indicated by sulfur isotope
data. The high proportion of SO42−/H2S derived from SO2 gas
(e.g., ~ 4:1 to ~ 3:1 at Grasberg) is interpreted to have further
decreased due not only to early anhydrite precipitation but addi-
tionally by the reduction of aqueous SO42− to H2S via a coupled
oxidation of Fe2+ in igneous wallrocks and/or fluids to Fe3+. At
Grasberg, SO42−-rock interaction is suggested by the replacement
of amphibole by hydrothermal biotite and of igneous biotite by
phlogopitic hydrothermal biotite. Ferrous iron was liberated from
igneous amphibole and biotite and oxidized to form magnetite.
The S2− reacted with copper in hydrothermal fluids to precipitate
the abundant Grasberg copper sulfide mineralization. The fre-
quent association of copper sulfides with hydrothermal magnetite
in porphyry systems suggests that this may be a common process
in such deposits.

Supplementary Information The online version contains supplementary

material available at https://doi.org/10.1007/s00126-021-01040-9.

Acknowledgments This project was completed with the help of many

geologists and support staff in PT Freeport Indonesia and PT Eksplorasi
Nusa Jaya. We thank Dr. Ryohei Takahashi and Dr. Hinako Sato of Akita
University for their assistance during sulfur isotope analysis. The authors
also acknowledge Georges Beaudoin and David Banks for their editorial
handling, and P. Voudrodis and three anonymous referees are also
thanked for their constructive comments.

Funding This project was funded by the Leading Program of Akita

Fig. 9 Schematic model showing the evolutions of sulfur species and University, and the fieldwork component was supported by PT Freeport
oxidation state of hydrothermal fluid at Grasberg with respect to Indonesia and PT Eksplorasi Nusa Jaya.
mineral assemblage and copper mineralization. The ~ 4:1 to ~ 3:1
proportion of SO 42−/H 2S produced by SO 2 disproportionation is
associated with magnetite-free, anhydrite-rich zone with ~ 0–2% Cu.
Higher concentrations of copper (~ 1–4% Cu) in the magnetite-bearing
zone indicate that mineralization is related to the increase in H2S by the
reduction of aqueous SO42− (see text)

(internal report by PT Freeport Indonesia) is more than References

enough to generate the observed 52.5 Mt of Cu. Therefore,
not all S2− was incorporated in copper minerals, some formed Annen C, Blundy JD, Sparks RSJ (2006) The genesis of intermediate and
silicic magmas in deep crustal hot zones. J Petrol 47:505–539
pyrite as veins, disseminations, or as massive bodies at the
Ballard JR, Palin JM, Campbell IH (2002) Relative oxidation states of
GIC periphery such as the heavy sulfide zone (Fig. 2) and magmas inferred from Ce (IV)/Ce (III) in zircon: application to
the rest was outgassed. porphyry copper deposits of northern Chile. Contrib Mineral
Petrol 144:347–364
Best MG, Christiansen EH (2001) Igneous petrology. Blackwell Science,
Edinburgh 458p
Conclusion Brimhall GH, Agee C, Stoffregen R (1985) The hydrothermal conversion
of hornblende to biotite. Can Mineral 23:369–379
Carmichael ISE, Ghiorso MS (1986) Oxidation-reduced relations in basic
Petrography, mineral chemistry, and sulfur isotope data com- magma: a case for homogenous equilibria. Earth Planet Sci Lett 78:
bined with alteration and vein parageneses reveal the roles of 200–210
Miner Deposita

Carroll MR, Rutherford MJ (1985) Sulfide and sulfate saturation in hy- volcanic sulphur dioxide flux and chemisorption. Nat Geosci 8:
drous silicate melts. Proceedings of the Fifteenth Lunar and 210–215
Planetary Science Conference Part 2. J Geophys Res 90:C601–C612 Hildreth W, Moorbath S (1988) Crustal contribution to arc magmatism in
Carroll MR, Rutherford MJ (1988) Sulfur speciation in hydrous experi- the Andes of Central Chile. Contrib Mineral Petrol 98:455–489
mental glasses of varying oxidation state: results from measured Holland HD (1965) Some applications of thermochemical data to prob-
wavelength shifts of sulfur X-rays. Am Mineral 73:845–849 lems of ore deposits II. Mineral assemblages and the composition of
Chambefort I, Dilles JH, Kent AJR (2008) Anhydrite-bearing andesite ore forming fluids. Econ Geol 60:1101–1166
and dacite as a source for sulfur in magmatic-hydrothermal mineral Housh T, McMahon TP (2000) Ancient isotopic characteristics of
deposits. Geology 36:719–722 Neogene potassic magmatism in western New Guinea (Irian Jaya,
Claypool GH, Holser WT, Kaplan IR, Sakai H, Zak I (1980) The age Indonesia). Lithos 50:217–239
curves of sulfur and oxygen isotopes in marine sulfate and their Hutchinson MC, Dilles JH (2019) Evidence for magmatic anhydrite in
mutual interpretation. Chem Geol 28:199–260 porphyry copper intrusions. Econ Geol 114:14–152
Cloos M (2001) Bubbling magma chambers, cupolas, and porphyry cop- Imai A (2000) Mineral paragenesis, fluid inclusions and sulfur isotope
per deposits. Int Geol Rev 43:285–311 systematics of the Lepanto Far Southeast Porphyry Cu-Au deposit,
Cloos M, Sapiie B, Quarles van Ufford A, Weiland RJ, Warren PQ, Mankayan, Philippines. Resour Geol 50:151–168
McMahon TP (2005) Collisional delamination in New Guinea: the Ishihara S (1977) The magnetite-series and ilmenite-series granitic rocks.
geotectonics of subducting slab breakoff. Geol Soc Am Spec Pap Min Geol 27:293–305
400:51p Jacobs DC, Parry WT (1979) Geochemistry of biotite in the Santa Rita
Cooke DR, Hollings P, Walshe JL (2005) Giant porphyry deposits: char- porphyry copper deposit, New Mexico. Econ Geol 74:860–887
acteristics, distribution, and tectonic controls. Econ Geol 100:801– Jenner FE, O’Neill HSC, Arculus RJ, Mavrogenes JA (2010) The mag-
818 netite crisis in the evolution of arc-related magmas and the initial
de Hoog JCM, Mason PRD, van Bergen MJ (2001) Sulfur and concentration of Au, Ag and Cu. J Petrol 51:2445–2464
chalcophile elements in subduction zones: constraints from a laser Jugo PJ (2009) Sulfur content at sulfide saturation in oxidized magmas.
ablation ICP-MS study of melt inclusions from Galunggung Geology 37:415–418
Volcano, Indonesia. Geochim Cosmochim Acta 65:3147–3164 Jugo PJ, Luth RW, Richards JP (2005) An experimental study of the
de Hoog JCM, Hattori K, Hoblitt RP (2004) Oxidized sulfur-rich mafic sulfur content in basaltic melts saturated with immiscible sulfide or
magma at Mount Pinatubo, Philippines. Mineral Deposita 146:750– sulfate liquids at 1300°C and 1·0 GPa. J Petrol 46:783–798
761 Keith JD, Whitney JA, Hattori K, Ballantyne GH, Christiansen EH, Barr
Dewey JF, Bird JM (1970) Mountain belts and the new global tectonics. J DL, Cannan TM, Hook CJ (1997) The role of magmatic sulfides and
Geophys Res 75:2625–2647 mafic alkaline magmas in the Bingham and Tintic mining districts,
Dreyer BM, Morris JD, Gill B (2010) Incorporation of subducted slab- Utah. J Petrol 38:1679–1690
derived sediment and fluid in arc magmas: B-Be-10Be-ɛNd system- Kouzmanov K, Pokrovski GS (2012) Hydrothermal controls on metal
atics of the Kurile Convergent Margin, Russia. J Petrol 51:1761– distribution in porphyry Cu (-Mo-Au) systems. Soc Econ Geol
1782 Spec Publ 16:573–618
Field CW, Gustafson LB (1976) Sulfur isotopes in the porphyry copper Kusakabe M, Nakagawa S, Hori S, Matsuhisa Y, Ojeda JM, Serrano L
deposit at El Salvador, Chile. Econ Geol 71:1533–1548 (1984) Oxygen and sulfur isotopic compositions of quartz, anhy-
Field CW, Zhang L, Dilles JH, Rye RO, Reed MH (2005) Sulfur and drite, and sulfide minerals from the El Teniente and Rio Blanco
oxygen isotopic record in sulfate and sulfide minerals of early, deep, porphyry copper deposits, Chile. Bull Geol Surv Jpn 35:583–614
pre-main stage porphyry Cu-Mo and late main stage base-metal Kusakabe M, Komoda Y, Takano B, Abiko T (2000) Sulfur isotopic
mineral deposits, Butte district, Montana. Chem Geol 215:61–93 effects in the disproportionation reaction of sulfur dioxide in hydro-
Gibbins SL (2006) The magmatic and hydrothermal evolution of the thermal fluids: implications for the δ34S variations of dissolved
Ertsberg intrusion in the Gunung Bijih (Ertsberg) mining district, bisulfate and elemental sulfur from active crater lakes. J Volcanol
West Papua, Indonesia. Dissertation, The University of Arizona Geotherm Res 97:287–307
Halter WE, Pettke T, Heinrich T, Heinrich CA (2002) The origin of Cu/ Leeman WP (1996) Boron and other fluid-mobile elements in volcanic
Au ratios in porphyry-type ore deposits. Science 296:1844–1846 arc lavas: implications for subduction processes. In: Bebout GE,
Hamilton WB (1979) Tectonics of the Indonesian region. US Geological Scholl DW, Kirby SH, Platt JP (ed) Subduction: top to bottom.
Survey Professional Paper 1078:345p Am Geophys Union Monogr, pp. 269–276
Hattori KH (2018) Porphyry copper potential in Japan based on magmat- Leys CA, Cloos M, New BTE, MacDonald GD (2012) Copper-gold±mo-
ic oxidation state. Resour Geol 68:126–137 lybdenum deposits of the Ertsberg-Grasberg district, Papua,
Hattori KH, Keith JD (2001) Contribution of mafic melts to porphyry Indonesia. Soc Econ Geol Spec Publ 16:215–236
copper mineralization: evidence from Mount Pinatubo, Philippines, Liang H, Sun W, Su W, Zartman RE (2009) Porphyry copper-gold min-
and Bingham Canyon, Utah, USA. Mineral Deposita 36:799–806 eralization at Yulong, China, promoted by decreasing redox poten-
Hedenquist JW, Lowenstern JB (1994) The role of magmas in the forma- tial during magnetite alteration. Econ Geol 104:587–596
tion of hydrothermal ore deposits. Nature 370:519–527 MacDonald GD, Arnold LC (1994) Geological and geochemical zoning
Hedenquist JW, Arribas A, Aoki M (2017) Zonation of sulfate and sulfide of the Grasberg Igneous Complex, Irian Jaya, Indonesia. J Geochem
minerals and isotopic composition in the Far Southeast porphyry Explor 50:143–178
and Lepanto epithermal Cu-Au deposits, Philippines. Resour Geol Mavrogenes J, Blundy J (2017) Crustal sequestration of magmatic sulfur
67:174–196 dioxide. Geology 45:211–214
Henley RW, McNabb A (1978) Magmatic vapor plumes and ground- McMahon T (1999) The Ertsberg intrusion and the Grasberg complex:
water interaction in porphyry copper emplacement. Econ Geol 73: contrasting styles of magmatic evolution and Cu-Au mineralization
1–20 in the Gunung Bijih (Ertsberg) mining district, Irian Jaya, Indonesia.
Henley RW, Seward TM (2018) Gas-solid reactions in arc volcanoes: Bul Geol 31:123–132
ancient and modern. Rev Mineral Geochem 84:309–349 Mernagh TP, Leys CA, Henley RW (2020) Fluid inclusion systematics in
Henley RW, King PL, Wykes JL, Renggli CJ, Brink FJ, Clark DA, porphyry copper deposits: the super-giant Grasberg deposit,
Troitzsch U (2015) Porphyry copper deposit formation by sub- Indonesia, as a case study. Ore Geol Rev 123:103570
Miner Deposita

Métrich N, Clocchiatti R (1996) Sulfur abundance and its speciation in Sapiie B, Cloos M (2004) Strike-slip faulting in the core of the central
oxidized alkaline melts. Geochim Cosmochim Acta 60:4151–4160 range of West New Guinea: Ertsberg mining district, Indonesia.
Métrich N, Bonnin-Mosbah M, Susini J, Menez B, Galoisy L (2002) Geol Soc Am Bull 116:277–293
Presence of sulfite (SIV) in arc magmas: implications for volcanic Seedorff E, Dilles JH, Proffett JM, Jr Einaudi MT, Zurcher L, Stavast
sulfur emissions. Geophys Res Lett 29:33–1–33-4 WJA, Johnson DA, Barton MD (2005) Porphyry deposits: charac-
Moore WJ, Czamanske GK (1973) Compositions of biotites from unal- teristics and origin of hypogene features. Econ Geol 100 th
tered and altered monzonite rocks in the Bingham Mining District, Anniversary Volume:251–298
Utah. Econ Geol 68:269–274 Selby D, Nesbitt BE (2000) Chemical composition of biotite from the
Mungall JE (2002) Roasting the mantle: slab melting and the genesis of Casino porphyry Cu-Au-Mo mineralization, Yukon, Canada: eval-
major Au and Au-rich Cu deposits. Geology 30:915–918 uation of magmatic and hydrothermal fluid chemistry. Chem Geol
Mutch EJF, Blundy JD, Tattitch BC, Cooper FJ, Brooker RA (2016) An 171:77–93
experimental study of amphibole stability in low-pressure granitic Sillitoe RH (2008) Major gold deposits and belts of the North and South
magmas and a revised Al-in-hornblende geobarometer. Contrib American Cordillera: distribution, tectonomagmatic settings, and
Mineral Petrol 171:85 metallogenic considerations. Econ Geol 103:663–687
New BTE (2006) Controls of copper and gold distribution in the Kucing Sillitoe RH (2010) Porphyry copper systems. Econ Geol 105:3–41
Liar deposit, Ertsberg mining district, West Papua, Indonesia.
Sun WD, Arculus RJ, Kamenetsky VS, Binns RA (2004) Release of
Dissertation, James Cook University
gold-bearing fluids in convergent margin magmas prompted by
Ohmoto H, Goldhaber MB (1997) Sulfur and carbon isotopes. In: Barnes
magnetite crystallization. Nature 431:975–978
HL (ed) Geochemistry of hydrothermal ore deposits, 3rd edn. Wiley,
New York, pp 517–611 Tosdal RM, Dilles JH, Cooke DR (2009) From source to sinks in aurif-
Ohmoto H, Lasaga AC (1982) Kinetics of reactions between aqueous erous magmatic-hydrothermal porphyry and epithermal deposits.
sulfates and sulfides in hydrothermal systems. Geochim Elements 5:289–229
Cosmochim Acta 46:1727–1745 Wafforn S (2017) Geo- and thermochronology of the Ertsberg-Grasberg
Ohmoto H, Rye RO (1979) Isotopes of sulfur and carbon. In: Barnes HL Cu-Au mining district, West New Guinea, Indonesia. Dissertation,
(ed) Geochemistry of hydrothermal ore deposits, 2nd edn. Wiley, The University Texas at Austin
New York, pp 509–567 Wallace PJ (2005) Volatiles in subduction zone magmas: concentrations
Paterson JT, Cloos M (2005) Grasberg porphyry Cu-Au deposit, Papua, and fluxes based on melt inclusion and volcanic gas data. J Volcanol
Indonesia: 1. Magmatic history. In: Porter TM (ed) Super porphyry Geotherm Res 140:217–240
copper and gold deposits: a global perspective. PGC Publishing, Wallace PJ, Edmonds M (2011) The sulfur budget in magmas: evidence
Adelaide, pp 313–329 from melt inclusions, submarine glasses, and volcanic gas emis-
Pollard PJ, Taylor RG (2002) Paragenesis of the Grasberg Cu-Au deposit, sions. Rev Mineral Geochem 73:215–246
Irian Jaya, Indonesia: results from logging section 13. Mineral Watanabe Y, Sato R, Sulaksono A (2018) Role of potassic alteration for
Deposita 37:117–136 porphyry Cu mineralization: implication for the absence of porphyry
Richards JP (2003) Tectono-magmatic precursors for porphyry Cu-(Mo- Cu deposits in Japan. Resour Geol 68:195–207
Au) deposit formation. Econ Geol 98:1515–1533 Whitney JA (1988) Composition and activity of sulfurous species in
Richards JP (2011a) High high Sr/Y arc magmas and porphyry Cu ± Mo quenched magmatic gases associated with pyrrhotite-bearing silicic
± Au deposits: just add water. Econ Geol 106:1075–1081 systems. Econ Geol 83:86–92
Richards JP (2011b) Magmatic to hydrothermal metal fluxes in conver- Wilkinson JJ (2013) Triggers for the formation of porphyry ore deposits
gent and collided margins. Ore Geol Rev 40:1–26 in magmatic arcs. Nat Geosci 6:917–925
Richards JP (2015) The oxidation state, and sulfur and Cu contents of arc Wones DR (1989) Significance of the assemblage titanite + magnetite +
magmas: implications for metallogeny. Lithos 233:27–45 quartz in granitic rocks. Am Mineral 74:744–749
Ridolfi F, Renzulli A, Puerini M (2010) Stability and chemical equilibri- Yanagisawa F, Sakai H (1983) Thermal decomposition of barium sulfate-
um of amphibole in calc-alkaline magmas: an overview, new vanadium pentaoxide-silica glass mixtures for preparation of sulfur
thermobarometric formulations and application to subduction- dioxide in sulfur isotope ratio measurements. Anal Chem 55:985–
related volcanoes. Contrib Mineral Petrol 160:45–66 987
Roberts SA (1975) Early hydrothermal alteration and mineralization in
Butte district, Montana. Thesis, Harvard University Cambridge
Rye RO (1993) The evolution of magmatic fluids in the epithermal envi- Publisher’s note Springer Nature remains neutral with regard to jurisdic-
ronment: the stable isotope perspective. Econ Geol 88:733–752 tional claims in published maps and institutional affiliations.