Earth and Planetary Science Letters 183 (2000) 7^14

www.elsevier.com/locate/epsl

Di¡erent crustal sources for Au-rich and Au-poor ores of the

Grasberg Cu^Au porphyry deposit
Ryan Mathur a; *, Joaquin Ruiz a , Spencer Titley a , Stacie Gibbins a ,
Widodo Margotomo b
a
Department of Geosciences, University of Arizona, Tucson, AZ 85721, USA
b
PT Freeport Indonesia, P.O. Box 51777, New Orleans, LA 70151-1777, USA

Received 10 August 2000; received in revised form 29 August 2000; accepted 29 August 2000

Abstract

The Grasberg is a porphyry copper deposit that is crosscut by a second stage mineralization which is greatly enriched
in gold. Sulfides from the porphyry-type event yield a 2.9 þ 0.3 Ma Re^Os isochron that agrees with published
geochronology. The initial 187 Os/188 Os ratio of the isochron is 0.56 þ 0.02, and implies a significant crustal component
for the source of Os and by inference the other base metals. The samples from the crosscutting secondary event do not
form an isochron, but form a mixing line with the older porphyry style mineralization as one of the end members and a
very radiogenic end member possibly shale. The initial 187 Os/188 Os ratios from samples of the second event range from
0.81 to 1.26 and correlate with gold content. The most radiogenic samples have the highest gold. The Re^Os isotope
data indicate different crustal sources for the ore-forming elements at the Grasberg Cu^Au deposit and support a model
in which gold is derived from sedimentary protoliths, that may have been pre-enriched by sedimentary processes. The
gold derived from these sources is concentrated by magmatic/hydrothermal systems. The requirement of a sedimentary
source for gold to produce hydrothermal gold deposits has been debated for decades and this study provides strong
support for the model. ß 2000 Elsevier Science B.V. All rights reserved.

Keywords: Re/Os; porphyry copper; gold

1. Introduction lution and history, and second, the question of

whether metal-`source' terranes, characterized by
The source of gold in ores is a signi¢cant and pre-enrichment of certain metals, should be uti-
long-standing question that has been addressed lized in resource evaluation. Conventional
historically in qualitative ways. The importance thought has viewed metals as indigenous to the
of the source problem is two-fold, addressing ¢rst, igneous intrusions' source of magma generation.
questions of geochemical processes of crustal evo- In contrast to this perspective, other hypotheses
consider metal as derived from volcanic and sedi-
mentary rocks of the crust [1^3] in which the in-
trusion resides, remnants of deep metallogenic
* Corresponding author. Tel.: +1-520-621-2365; heterogeneity in the mantle [4], or the melted sub-
E-mail: rmathur@geo.arizona.edu ducted oceanic crust [5].

0012-821X / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 0 1 2 - 8 2 1 X ( 0 0 ) 0 0 2 5 6 - 9

EPSL 5610 7-11-00

8 R. Mathur et al. / Earth and Planetary Science Letters 183 (2000) 7^14

Gold has been of particular interest because of by the Re^Os systems in the porphyry deposit of
its occurrence in a wide variety of styles of ore Lihir [11], in the Bushveld layered ma¢c intrusion,
deposits and its time-enduring habits of occur- and in the ancient basin of the Witswatersrand
rence. Titley [6] reviewed the occurrence of gold [12] indicated a single source for metals that in-
in association with Phanerozoic sedimentary volved the mantle or partially metasomatized
rocks and proposed that there is pre-enrichment mantle as a predominant source for gold [11,12].
of gold in lower Paleozoic and mid to upper Mes- The use and potential of the Re^Os isotopic
ozoic strata, which may be regenerated and en- system to trace the source of metals and deter-
riched in younger deposits by a variety of crustal mine mineralization ages in ore deposits are start-
processes. This general notion is not new. Boyle ing to be realized [11^16]. 187 Re decays to 187 Os
[7] and Sangster [8] reviewed the occurrence of (with a half-life of 42.3 billion years) by beta
gold in turbidite successions and proposed that emission. Importantly, rhenium and osmium are
gold-bearing ores of many regions were derived concentrated in sul¢de phases and permit geo-
from pre-enriched source beds through a process chronologic and genetic information to be ob-
that involves the devolitization and transport of tained from the ore mineral rather then alteration
gold from the sedimentary rocks through meta- silicates. Crustal and mantle reservoirs within the
morphism. subduction zone environments develop distinct
187
Here we address the origin of gold in large in- Os/188 Os ratios through time because crustal
trusion-centered copper deposits, the `porphyry materials have higher Re/Os [17]. Construction
copper deposit' of common usage. These deposits of isochron diagrams for these sul¢des provide
form along active continental margins and on is- the age of mineralization, and the initial 187 Os/
188
land arcs. They are the result of the shallow Os value of the ore-forming £uids. The initial
187
(6 km) emplacement of small plutons or pluton Os/188 Os can be used to trace the source of Os,
complexes (V4 km2 area), commonly associated and by inference the other base metals (Cu) which
with volcanic centers. The metal deposits origi- should have a similar geochemical behavior. In
nate from the cracking of large volumes of rock regard to Os similarities to Au, they are both
and the development of convective £uid £ow as- noble metals, and believed to be soluble in similar
sociated with the thermal and mechanical e¡ects types of aqueous solutions [18]. Thus, the source
of emplacement and cooling of plutons at shallow of Os can be used as a proxy for the source of the
levels in the crust [9]. The sources of copper and metals in ore deposits.
gold present in deposits of this sort are conven- Grasberg is the largest known porphyry Cu^Au
tionally believed to be similar and acquired from deposit on earth containing about 2.1 billion tons
magmatic £uids. of Cu ore at 1.2% Cu and 1.2 g per ton Au [19].
We present the result of studies of the Grasberg The ore deposit formed during recent magmatic
deposit in West Irian Jaya where isotopic evidence events corresponding with the collision of the
[10] allows metallogenic interpretation of por- Australian and Indo^Paci¢c plates and occurs
phyry magmas that originated from melted Aus- within the central mobile fold belt [20,21] (Fig.
tralian continental crust. The magmas have cross- 1). In general, the magmatic rocks of this ore
cut and potentially involved a succession of deposit intruded highly deformed/metamorphosed
Proterozoic and lower Paleozoic strata of the Proterozoic and Paleozoic through Cenozoic sedi-
sort described here [6^8] at some great depth mentary rocks. The basement in this area is
( s 4 km) beneath the deposit. We have used the thought to be the Australian continental crust
Re^Os isotopic system to determine if strata of [21]. The intrusions consist of three porphyries
Phanerozoic age may have contributed to the (Dalam [oldest]^Main Grasberg Intrusion^Kali
high gold content of the Grasberg ores, thereby [youngest]) of diorite to quartz monzonite compo-
testing the notion that there are potentially di¡er- sition that intruded a succession of clastic and
ent sources for the base metals in certain types of carbonate strata (Fig. 1). Housh and McMahon
mineral deposits. The source of gold as revealed [10] studied the geochemistry of the volcanic rocks

EPSL 5610 7-11-00

 R. Mathur et al. / Earth and Planetary Science Letters 183 (2000) 7^14 9

event crosscuts and borders the intrusions, and

is characterized by massive, meter thick veins of
pyrite, chalcopyrite, and magnetite. In order to
study the sources of metals and relative timing
of these events, samples were selected from drill
core that intersected rocks carrying mineralization
of both episodes (Fig. 1).

2. Techniques and results

Separation of sul¢des from the drill core for

Re^Os analysis entailed wrapping each sample
in paper (in order to reduce possible PGE con-
tamination), crushing the drill core with a ham-
mer, and hand picking the sul¢des. Concentra-
tions of Re^Os were determined by isotope
dilution. Typically 0.5^1.5-g samples were loaded
into a Carius tube and dissolved in a reverse
aquaregia solution. The Os was extracted from
the solution through a two-distillation process
[15] and Re through column chemistry [24]. Sam-
ples were analyzed on a negative thermal ioniza-
Fig. 1. Location and generalized section of the pro¢le tion mass spectrometer. The details of sample
sampled in the southwest quadrant of Grasberg intrusive preparation and running procedures are recorded
complex. The cross section illustrates the location of each in [25,26]. Procedural blanks for Os ranged from
sample with respect to alteration and host lithology. Solid
0.9 to 1.6 pg and Re 25^35 pg, which were cor-
circles indicate samples from the initial porphyry style miner-
alization, and X indicate samples taken from the younger rected for in this study. Gold analyses were con-
massive sul¢de event. ducted by ¢re assay.
All results are presented in Table 1. The con-
centrations of Re and Os for these samples range
in the region and interpreted variable Pb (206 Pb/ from 6 to 20 ppb (parts per billion, or 1039 g per
204
Pb 18.002^18.291, 207 Pb/204 Pb 15.535^15.616, g) and 0.005^0.071 ppb, respectively. On average,
and 208 Pb/204 Pb 38.395^38.744), elevated Sr
(87 Sr/86 Sr 0.70611^0.70707), and negative Nd
(ONd 10.3^15.3) values as evidence of ancient
(Proterozoic and Archean) crustal and depleted
mantle components of the magmas in the district.
The distribution of all metals in the deposit forms
a horseshoe-shaped aureole around the Kali intru-
sion, suggesting that this intrusion was responsi-
ble for the bulk of mineralization [22].
In general, two distinct mineralization events
are recognized in the district. The older event as-
sociated with the porphyritic rocks and surround-
ing sedimentary rocks deposited traditional stock- Fig. 2. Isochron plot of the data from the porphyry style
work style porphyry vein, disseminated, and mineralization event, ages and initial values were calculated
contact sul¢de mineralization [23]. The second from algorithms of York [40].

EPSL 5610 7-11-00

10 R. Mathur et al. / Earth and Planetary Science Letters 183 (2000) 7^14

Table 1
187
Sample name Mineralization Re Os Re/188 Os 187
Os/188 Os 187
Os/188 Osinitial Au
(ppt) (ppt) (ppm)
High sul¢des py massive sul¢de 125 759 1.28 17.66
High sul¢des py massive sul¢de 140 530 636 1 226 þ 11 1.30 þ 0.003 1.24 17.66
High sul¢des py massive sul¢de 138 214 1 024 751 þ 8 1.30 þ 0.003 1.26 17.66
High sul¢des py massive sul¢de 73 074 571 711 þ 8 1.31 þ 0.003 1.26 17.66
Grs194 mag massive sul¢de 11 530 53 1 185 þ 30 1.14 þ 0.02 1.08 14.5
Grs194 cpy massive sul¢de 30 368 38 4 495 þ 130 1.31 þ 0.02 1.09 14.5
Grs37194-706 py massive sul¢de 34 811 42 4 676 þ 185 1.29 þ 0.02 1.07 6.11
Grs37194-706 py massive sul¢de 37 473 60 3 475 þ 104 1.27 þ 0.02 1.07 6.11
Hsxc22-py massive sul¢de 32 740 36 5 030 þ 195 1.03 þ 0.03 0.79
Hsxc22-py massive sul¢de 33 083 32 5 653 þ 200 1.06 þ 0.03 0.79
Hsxc22-py massive sul¢de 46 664 47 5 523 þ 220 1.16 þ 0.03 0.81
GSP cov+py-1a cov.-pyrite 10 616 7 14 044 þ 1 100 1.22 þ 0.07 0.56 6.90
KL4-40 cpy+pya quartz-ser. 7 871 5 9 854 þ 900 1.00 þ 0.05 0.56 1.61
KL4-40 cpy+pya quartz-ser. 7 242 6 6 539 þ 530 0.93 þ 0.05 0.56 1.61
Kl40-05-423.8a Ter. sed. 20 148 56 1 869 þ 74 0.67 þ 0.02 0.56 4.20
Kl40-05-423.8a Ter. sed. 16 228 34 2 463 þ 170 0.64 þ 0.02 0.56 4.20
Kl40-05-423.8a Ter. sed. 15 103 71 1 121 þ 60 0.62 þ 0.01 0.56 4.20
Grs37194-08-470.5a biot.-ortho. 4 299 12 3 109 þ 242 0.71 þ 0.05 0.56 1.20
All numbers following or associated with samples re£ect drill core and depths of drill cores. Location of samples is labeled in the
mineralization column. Abbreviations represent the dominate phase in the sample analyzed: py = pyrite, cpy = chalopyrite, mag =
magnetite, Ter. sed. = tertiary sediments, biot.-ortho. = indicates sul¢des taken from the biotite^orthoclase alteration silicates,
quartz-ser. = indicates sul¢des taken from the quartz^sericite alteration silicates. All Au analysis done by ¢re assay techniques of
Jacob's Fire Assay, Tucson, AZ, USA.
a
Indicates values used to construct isochrons, error calculation is discussed in text, and isochron statistics were determined using
algorithms of York [40].

these samples contained 3.47 ppm (part per mil- are also higher in this zone averaging 12.65 ppm.
lion, or 1036 g per g) gold and are considered This is de¢ned as the gold-rich ore because there
samples of ore with the lower concentration is a relative increase in gold content without an
gold. Gold and copper usually follow a 1 to 1 accompanying increase in copper content. Di¡er-
ratio for this part of the mineralized ore body. ences between the concentrations of Au, Re, and
The data yield an isochron with an age of
2.9 þ 0.3 Ma (MSWD 1.2) which agrees with pre-
vious age determinations of 3 Ma on alteration
silicates [22]. The initial 187 Os/188 Os determined
from the isochron is 0.56 þ 0.02 (Fig. 2). Errors
for each analysis are calculated by changing the
blank concentration between 0.9 and 1.6 pg Os
since this is the greatest source of error in the
analysis. The 187 Os/188 Os isotopic composition of
the blank is constant and around 0.175. The fact
that the data lie on an isochron plot with the
correct age indicates that the ore-forming £uids
were isotopically homogeneous in Os. Fig. 3. Concentrations of element along tested drill core
GRS37194. Solid line indicates Os concentration, and X on
Concentrations of Re and Os for these samples the line indicates sample position, with the 187 Os/188 Os la-
are much greater, ranging from 36 to 140 ppb and beled. The dotted line represents the Cu concentration in %,
0.047 to 1 ppb, respectively. Gold concentrations and dashed line indicates the Au concentration in ppm.

EPSL 5610 7-11-00

 R. Mathur et al. / Earth and Planetary Science Letters 183 (2000) 7^14 11

Sul¢des from di¡erent silicate alteration events

in the Dalam intrusion and the Tertiary sedimen-
tary rocks lie on the same isochron. This indicates
a similar source of metals for sul¢des regardless of
the associated silicate alteration assemblages or
the host rocks.
Inasmuch as the chondritic mantle 187 Os/188 Os
ratio is close to 0.13 [17,31], the initial 187 Os/188 Os
value of the sul¢des at about 0.56 indicates a sig-
ni¢cant crustal component for the Os, Cu, and
Au. Possible radiogenic sources in the subduction
environment are the descending oceanic slab (and
Fig. 4. Chemical mixing diagram, the solid square represents accompanying sediments) [27], the upper/lower
the average concentration and initial 187 Os/188 Os value of the continental crust [28], and a partially metasomat-
main porphyry style-mineralizing event, whereas the dark- ized mantle wedge [12,29]. Simple two component
ened circles represent the average concentration and initial
187
mixing models using chondritic [31] mantle with
Os/188 Os values from various samples in the massive sul-
¢de event. the mean lower crustal values from Proterozoic/
Archean Northern Australian continental crust
[30] that exists beneath Grasberg suggest at least
Os of the massive sul¢des and porphyry vein style 40% of the metals could have been derived from a
event are illustrated in Table 1 and Fig. 3. radiogenic ancient lower crust. The relative pro-
The ¢ve samples from di¡erent locations in the portion of crustal metals increases drastically if a
massive sul¢de ore body (Fig. 1) do not form an lower concentration source such as the oceanic
isochron ; instead they form a mixing line (Fig. 4). slab is considered. Regardless of the radiogenic
source hypothesized, the amount of crustal metals
present in the sul¢des in the Grasberg ore body is
3. Discussion signi¢cant. The radiogenic 187 Os/188 Osi isotopic
signature of the metals combined with the ex-
3.1. Older mineralizing event associated with the tremely negative ONd and elevated 87 Sr/86 Sr val-
intrusion of the Kali ues for the magmas [10] suggest a model for this
mineralizing event in which most of the metals
This event is characterized as the typical `por- could have been derived from a magma that had
phyry copper' style mineralization and alteration extensive interaction with the Proterozoic/Arche-
that evolved in a highly fractured stockwork. Sul- an lower crust.
¢de samples from the high temperature alteration
(biotite^orthoclase, quartz^sericite), the late stage 3.2. Younger massive sul¢de event
covellite^pyrite mineralization, and the hydrother-
mal pyrites occurring in the Tertiary sedimentary The massive sul¢de ore bodies and their rela-
rocks generate an isochron. The young age of the tionship to the porphyry mineralization remain
deposit and the analytical error obtained from an enigmatic. Models for the mineralization range
isochron constructed with sul¢des from the bio- from £uids derived from a hidden porphyry,
tite^orthoclase and quartz^sericite alteration which migrate along previous fractures, to a late
zones suggest that the duration of mineralization stage £uid associated with the porphyry intrusion
in the Grasberg ore body was less than 600 000 yr. [32]. Our Re^Os results indicate that the £uids
This implies that the various biotite^orthoclase, responsible for precipitating the sul¢de minerals
quart^sericitic, and late stage covellite^pyrite al- of the massive sul¢de had evolving 187 Os/188 Os
teration/mineralization phases occurred within isotopic ratios. Fig. 3 illustrates the di¡erence be-
this geologically short period. tween these two mineralized systems by a sharp

EPSL 5610 7-11-00

12 R. Mathur et al. / Earth and Planetary Science Letters 183 (2000) 7^14

isotopic and metal grade boundary between the concentration Os and Au end members. The pro-
massive sul¢de and the main mineralizing event. portion of the black shale component in the ores
Gold contents in this zone increase signi¢cantly increases signi¢cantly when attempting to resolve
without appreciable changes in copper abun- the higher concentration radiogenic end members.
dance. The data (Fig. 3) indicate that the source These data suggest that the £uid precipitating
of metals responsible for enriching the outer sur- the massive sul¢de has a component of the orig-
rounding rim of mineralization in the ore deposit inal porphyry and £uids that had possibly leached
contained more gold, and Os. metals from the black shales in the lower crust.
The 1/Os versus Os concentration (Fig. 4) trend The Re^Os data support a model in which the
suggests a relationship between the £uids that pre- porphyry intruded with its own Os isotopic signa-
cipitated the main stockwork mineralization and ture, and as the system cooled £uids interacted
the massive sul¢de. In order to access relative with, and acquired metals from, metalliferous
proportions of crustal materials, a radiogenic rich metamorphosed black shales present in deep-
source rock with a high concentration of common er portions of the Phanerozoic middle crust. Per-
Os must have been accessed by the £uids or in- haps a change in the £uid pH and temperature
corporated into the melt. There are few reservoirs during metallogenesis invoked di¡erent chemical
in the geologic rock record that have appreciable reactions with the host rock, thereby making gold
concentrations of radiogenic Os. Brooks and more soluble and allowing it to enter the miner-
others [33] noticed a trend that higher concentra- alizing system.
tion materials such as the mantle (1^10 ppb) have This is the ¢rst quantitative evidence of a pro-
lower Re/Os ratios (less then 1) and unradiogenic cess that extracts gold from surrounding sedimen-
Os signatures. Lower concentration materials tary rocks during the genesis of an ore deposit.
(0.5^0.05 ppb) have higher Re/Os ratios ( s 10) Inasmuch as the £uids that mineralized the mas-
and radiogenic Os signatures. None of the reser- sive sul¢de were distinctly heterogeneous, and not
voirs reported in these studies provides an end analogous to the £uids associated with the main
member with enough Os and a radiogenic signa- intrusion, models ascribing the massive sul¢de to
ture to resolve the source of the high-end member a hidden porphyry at depth are unlikely.
on the mixing array present in the Grasberg sul-
¢des.
Recent work on black shales [34^37] indicates 4. Conclusions
that concentrations of Re^Os range from 20 to
300 ppb and 0.1 to 13 ppb, respectively. The con- This study provides geochemical evidence for
centration of these elements in black shales varies the presence of crustal, and more importantly,
signi¢cantly compared with other geologic mate- sedimentary rocks as an essential source that en-
rials. However, two studies have found the higher riches some of the metals in a porphyry copper
end concentrations of Re^Os in black shales of ore deposit. Titley [6] observed that formation of
Precambrian and Devonian age sedimentary units many gold deposits might be closely linked with
[36,37]. Van U¡ort [38] noted metamorphosed crustal units that contain turbiditic sediments of
black shale equivalents in the sedimentary succes- Paleozoic and Cenozoic age. The Re^Os data
sion present in the Phanerozoic crust in which the from the Grasberg suggest that these kinds of
Grasberg porphyries intruded. As Os concentra- rocks could be the source of metals for this de-
tions and 187 Os/188 Os in modern shale [39] are posit. Therefore, the relevance of this study is that
relatively uniform, these metamorphosed Precam- it numerically illustrates the in£uence of the lower
brian and Devonian shales provide a plausible Os and middle continental crust during porphyry
source at Grasberg. Simple mixing calculations copper genesis. Speci¢cally, it highlights two sour-
between the initial radiogenic porphyry signature ces for the bulk of copper and gold in porphyry
and the Devonian or Precambrian shales indicate copper hydrothermal environments, and suggests
at least 30% black shale component for the low that the interaction of the hydrothermal system

EPSL 5610 7-11-00

 R. Mathur et al. / Earth and Planetary Science Letters 183 (2000) 7^14 13

with speci¢c units in the crust produces gold-rich posits, Southwestern North America: Tucson, University
of Arizona Press, 1982, pp. 59^72.
ore.
[10] T. Housh, T. McMahon, Ancient isotopic characteristics
of Neogene potassic magmatism in western New Guinea
(Irian Jaya, Indonesia), Lithos 50 (2000) 217^239.
Acknowledgements [11] S.R. Hart, E.D. Kinloch, Hart, Osmium Isotope system-
atics in Witwatersrand and Bushveld Ore Deposits, Econ.
The studies reported here are part of the grad- Geol. 84 (1989) 1651^1655.
[12] B.I. McInnes, J.S. McBride, N.T. Evan, D.D. Lambert,
uate research programs of Mathur and Gibbins. A.S. Andrew, Osmium isotope and noble metal recycling
Analytical work was supported by the National in subduction zones: Implication for metallogenesis of
Science Foundation Grants EAR 9708361, and porphyry-epithermal Cu^Au deposits, Science 286 (1999)
EAR 9628150, and a instrumentation grant from 512^516.
[13] T. McCandless, J. Ruiz, Rhenium^osmium evidence for
the W.M. Keck Foundation. We thank Freeport
regionally timed mineralization events in southwestern
McMoRan Copper and Gold, Inc. for providing North America, Science 261 (1993) 1282^1286.
assistance and support of the ¢eldwork, and per- [14] H.J. Stein, R.J. Markey, J.W. Morgan, A. Du, Y. Sun,
mission to publish this information. The authors Highly precise and accurate Re^Os ages for molybdenite
also would like to thank Jon Patchett, Chuck from East Quling molybdenite belt, Shaanxi Province
China, Econ. Geol. 92 (1997) 827^835.
Brannon, Rick Carlson and an anonymous re-
[15] R. Frei, Re^Os, Sm^Nd, U^Pb and stepwise lead leaching
viewer for their constructive comments on earlier isotope systemactics in shear zone hosted gold mineraliza-
drafts of this paper.[RV] tion: Genetic tracing and age constraints of crustal hydro-
thermal activity, Geochim. Cosmochim. Acta 62 (1998)
1925^1936.
References [16] R. Mathur, J. Ruiz, F. Munizaga, Relationship between
copper tonnage of Chilean base-metal porphyry deposits
[1] K.B. Krauskopf, Source rocks for metal bearing £uids, and Os isotopes, Geology 28 (2000) 555^558.
in: H.L. Barnes (Ed.), Geochemistry of Hydrothermal [17] R.J. Walker, R.W. Carlson, S.B. Shirey, F.R. Boyd, Os,
Ore Deposits, Holt, Rinehart and Winston, Inc., 1967, Sr, Nd, and Pb isotope systematics of Southern Africa
pp. 1^33. peridotite xenoliths; implication for the chemical evolu-
[2] J.A. Noble, Metal provinces of the western United States, tion of subcontinental mantle, Geochim. Cosmochim.
Geol. Soc. Am. Bull. 81 (1970) 1607^1624. Acta 53 (1989) 1583^1595.
[3] C.W. Burnham, Metallogenic Provinces of the Southwest- [18] Y. Xiong, S.A. Wood, Experimental quanti¢cation of hy-
ern United States and Northern Mexico, New Mexico drothermal solubility of platinum-group elements with
State Bureau of Mines and Mineral Resources Bulletin special reference to porphyry copper environments, Min-
65, 1959. eral. Petrol. 68 (2000) 1^28.
[4] R.H. Sillitoe, Relation of metal provinces in western [19] Freeport mining report, 1997.
America to subduction of oceanic lithosphere, Geol. [20] S.D. Van Nort et al., Geology and mineralization of
Soc. Am. Bull. 83 (1972) 813^818. the Grasberg porphyry copper^gold deposit, Irian Jaya,
[5] R.E. Zartman, Lead isotopic provinces in the cordillera of Indonesia, Mining Engineering March, 1991, pp. 300^
the western United States and their geologic signi¢cance, 304.
Econ. Geol. 69 (1974) 792^805. [21] D. Mc Donald, L.C. Arnold, R.J. Kyle, Factors respon-
[6] S.R. Titley, Phanerozoic ocean cycles and sedimentary- sible for extreme concentration of Cu and Au in the Gras-
rock hosted gold ores, Geology 19 (1991) 645. berg deposit, in: A.H. Clark (Ed.), Giant Ore Deposits II,
[7] R.W. Boyle, Gold deposits in turbidite sequences: their Kingston, Ont., 1995, pp. 101^113.
geology, geochemistry, and history of the theories of their [22] Freeport mining report, 1998.
origin, in: J.D. Keppie, R.W. Boyle, S.J. Haynes (Eds.), [23] S.R. Titley, The style and progress of mineralization and
Turbidite Hosted Gold Deposits, Geological Association alteration in porphyry copper systems, in: S.R. Titley
of Canada Special Paper, 1986, pp. 1^13. (Ed.), Advances in The Geology of the Porphyry Copper
[8] A.L. Sangster, Light stable isotope evidence for a meta- Deposits, Southwestern North America, University of
morphogenic origin for bedding-parallel, gold-bearing Arizona, Tucson, AZ, 1982, pp. 93^116.
veins in Cambrian £ysch, Meguma Group, Nova Scotia, [24] S. Shirey, R. Walker, Carius Tube digestion for low blank
Explor. Min. Geol. 1 (1992) 69. rhenium^osmium analysis, Anal. Chem. 67 (1995) 2136^
[9] D.L. Norton, Fluid and heat transport phenomena typical 2141.
of copper-bearing pluton environments, in: S.R. Titley [25] R.A. Creaser, D.A. Papanastassiou, G.J. Wasserburg,
(Ed.), Advances in Geology of the Porphyry Copper De- Negative thermal ion mass spectrometry of osmium, rhe-

EPSL 5610 7-11-00

14 R. Mathur et al. / Earth and Planetary Science Letters 183 (2000) 7^14

nium, and iridium, Geochim. Cosmochim. Acta 55 (1994) ritic and basaltic magmatism of East Greenland: con-
397^401. straints on plume^lithosphere interactions and the genesis
[26] J. Ruiz, C. Freydier, T. McCandless, J.T. Chesley, F. of the Platinova reef, Skaergaard intrusion, Lithos 47
Munizaga, Re^Os isotope systematics from base metal (1999) 107^126.
porphyry and manto-type mineralization in Chile, Int. [34] G.E. Ravizza, K.K. Turekian, Application of the 187 Re^
187
Geol. Rev. 39 (1997) 317^324. Os system to black shale geochronometry, Geochim.
[27] M. Roy-Barman, G.J. Wasserberg, D.A. Papanastassiou, Cosmochim. Acta 53 (1989) 3257^3262.
M. Chaussidon, Osmium isotopic composition and Re^Os [35] G.E. Ravizza, K.K. Turekian, B. Hay, The geochemistry
concentration in sul¢de globules from basaltic glass, of rhenium and osmium in Recent sediments from the
Earth Planet. Sci. Lett. 154 (1998) 331^347. Black Sea, Geochim. Cosmochim. Acta 55 (1991) 3741^
[28] B.K. Esser, K.K. Turekian, The osmium isotopic compo- 3752.
sition of the continental crust, Geochim. Cosmochim. [36] M. Horan, J.W. Morgan, R. Coveney, J. Murowchick, L.
Acta 57 (1993) 3093^3104. Hulbert, Rhenium and osmium isotopes in black shales
[29] A.D. Brandon, R.A. Creaser, S.B. Shirey, R.W. Carlson, and Ni^Mo^PGE rich sul¢de layers, Yukon territory,
Osmium recycling in subduction zones, Science 272 (1996) Canada and Hunan and Guiznoun province China, Geo-
861^864. chim. Cosmochim. Acta 58 (1994) 257^265.
[30] A.E. Saal, R.L. Rudnick, G.E. Ravizza, Re^Os isotope [37] S. Sign, J. Trivedi, S. Krishumsumji, The Re^Os isotope
evidence for composition, formation, and age of the lower systematics in black shales from the Lesser Himalaya;
continental crust, Nature 393 (1998) 58^61. their chronology and role in the 187 Os/188 Os of seawater,
[31] T. Meisel, R.J. Walker, J.W. Morgan, The osmium iso- Geochim. Cosmochim. Acta 63 (1999) 2381^2392.
topic composition of the Earth's primitive upper mantle, [38] J. Van U¡ort, Ph.D. Thesis, The University of Texas,
Nature 383 (1996) 517^520. Austin, TX, 1996.
[32] S.R. Titley, Characteristics of high temperature Carbo- [39] G.E. Ravizza, K.K. Turekian, The osmium isotopic com-
nate-hosted replacement ores and some comparisons position of organic rich marine sediments, Earth Planet.
with Mississippi Valley ores, Soc. Econ. Geol. Spec. Sci. Lett. 110 (1992) 110^116.
Pub. 4 (1996) 244^254. [40] D. York, Least square ¢ttings of a straight line with cor-
[33] C.K. Brooks, R.R. Keays, D.D. Lambert, L.R. Frick, related errors, Earth Planet. Sci. Lett. 5 (1969) 320^324.
T.F. Nielsen, Re^Os isotope geochemistry of Tertiary pic-

EPSL 5610 7-11-00