Accepted Manuscript

Significance of high temperature fluids and melts in the Grasberg

porphyry copper'gold deposit

Terrence P. Mernagh, John Mavrogenes

PII: S0009-2541(18)30496-0
DOI: doi:10.1016/j.chemgeo.2018.09.040
Reference: CHEMGE 18934
To appear in: Chemical Geology
Received date: 1 December 2017
Revised date: 1 September 2018
Accepted date: 25 September 2018

Please cite this article as: Terrence P. Mernagh, John Mavrogenes , Significance of high
temperature fluids and melts in the Grasberg porphyry copper'gold deposit. Chemge
(2018), doi:10.1016/j.chemgeo.2018.09.040

This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.
 ACCEPTED MANUSCRIPT

Significance of High temperature fluids and melts in the Grasberg porphyry copper-gold
deposit.

Terrence P. Mernagh* and John Mavrogenes

Research School of Earth Sciences, Australian National University, Acton, ACT, 2601, Australia

*Corresponding author: Email Terry.Mernagh@anu.edu.au

Keywords: High temperature fluids, melt inclusions, fluid inclusions, porphyry copper deposits

PT
Abstract
A study of quartz-sulfide veins from the Grasberg Cu-Au deposit has shown that the quartz crystals initially grew
from the walls of the veins into open fractures. Cathodoluminescence (CL) imaging shows concentric growth zones in

RI
the larger crystals and changes in the orientation of the growth zones with occasional truncation of the zoning.
Towards the centre of veins the crystals become smaller and irregularly shaped but exhibit CL banding parallel to vein

SC
walls, which is normally associated with crack-seal processes. The vein quartz contains silicate and sulfide-rich melt
inclusions, virtually water-free salt-melt inclusions, and coexisting hypersaline and vapour-rich inclusions. Late stage,
NU
secondary aqueous inclusions are also present in some veins. Most fluid inclusions homogenised at temperatures
below 620 °C. However, hypersaline (B2) inclusions had a very complex behaviour and exhibited partial
homogenisation to salt melt and clear immiscible fluid at temperatures ranging from 820 °C to 1300 °C. These
MA

inclusions contain the highest Cu concentrations (up to 6.3 wt.%), suggesting they play a role in metal transport.

The contemporaneous trapping of the K-feldspar-rich melt inclusions, hypersaline and vapour-rich inclusions suggests
a common source for these fluids and a process involving heterogeneous entrapment of immiscible silicate melt,
D

hypersaline fluid and vapour as previously proposed for other porphyry copper deposits. Furthermore, the Grasberg
deposit is estimated to have formed at pressures below 400 bar, which may explain the presence of salt-melt inclusions
E

as the vapour + halite field of the H2O-NaCl system is dominant below 800 °C at these pressures.
PT

The unrealistically high temperatures reported herein may have resulted from entrapment of either immiscible silicate
melt, hypersaline fluid and vapour or cumulates of an evolving silicate melt saturated in K-spar, brine and vapour.
CE

Although these temperatures may not be directly useful, these inclusions give important clues to metal-enrichment
processes. These melt and fluid inclusions record cycles of transitory, high-temperature (>700 °C) hydrofracturing,
melt and fluid release, and vein formation in a cooler (500−600 °C) background host-rock thermal regime. High sulfur
AC

and Cu contents of sulfide-melt and hypersaline B2 inclusions are interpreted to be the fluid composition before the
main sulfide precipitation event, and, therefore, may be used as tracers of the magmatic body from which they
exsolved at depth.

Introduction
Highly saline fluid inclusions are a common feature of porphyry copper deposits. Over one hundred years ago,
Lindgren (1905) recognised halite-bearing fluid inclusions at the Morenci porphyry copper deposit in Arizona and
interpreted these as proof that mineralisation was associated with high salinity fluids of likely magmatic origin. Since
then, there have been countless studies of fluid inclusions associated with this style of mineralisation (see reviews in
Bodnar et al., 2014; De Vivo and Frezzotti, 1994; Kouzmanov and Pokrovski, 2012; Roedder, 1992). Roedder (1984)
reported, “the very high Th values found for both polyphase and simple liquid-vapour inclusions from Bingham were
1
 ACCEPTED MANUSCRIPT
several hundred degrees above what most would have expected; hence on several occasions suggestions were made
that these data were spurious (e.g. resulting from the trapping of mixtures of vapour and liquid, leakage or necking
down), but the nature and occurrence of the inclusions from Bingham that were measured effectively excluded these
other interpretations.”

As noted by Bodnar et al. (2014), despite the large number of studies of porphyry copper deposits, there is a paucity of
fluid inclusion homogenisation data above approximately 600 °C, which most likely reflects the temperature limit of
common fluid inclusion heating stages. In accord with these observations, previous fluid inclusion studies of the
Grasberg porphyry copper deposit in Papua, Indonesia (Harrison, 1999; Baline, 2007) only report fluid inclusion
homogenisation temperatures up to 700 °C (the temperature limit of their fluid inclusion stage). Therefore, the current
study focuses on the extremely high temperature, highly saline, multiphase inclusions that occur in the mineralised

PT
quartz-sulfide veins of the Grasberg porphyry copper deposit. Although poorly understood, these inclusions are a view
into the early evolutionary stages of the giant Grasberg porphyry system.

RI
High temperature hypersaline and melt inclusions in other deposits
High temperature, saline, multiphase inclusions and silicate melt inclusions are observed in many porphyry copper

SC
deposits. Roedder (1971, 1984) reported that the Bingham deposit had many multiphase inclusions with
homogenisation temperatures ranging from 645 °C – 725 °C. Wilson et al. (1980) found many multiphase inclusions
from the Granisle and Bell porphyry copper deposits with homogenisation temperatures >800 °C, including some that
NU
failed to homogenise by 1290 °C, the maximum temperature of the Leitz heating stage. Eastoe and Eadington (1986)
recorded homogenisation temperatures ranging from 940 °C to 1080 °C for multiphase inclusions containing halite,
sylvite, chalcopyrite and other salts from the Panguna porphyry copper deposit in New Guinea.
MA

Harris et al. (2003) observed polyphase hypersaline inclusions and silicate melt inclusions with brine globules from
porphyritic intrusions of the Bajo de la Alumbrera porphyry copper deposit in Argentina. Heating experiments on the
polyphase, hypersaline liquid inclusions revealed phase transformations similar to those observed in brine globules in
D

associated silicate-melt inclusions. Final salt dissolution occurred between 480 and 540 °C with complete
E

homogenisation by vapour disappearance between 745 °C and 845 °C. Harris et al. (2003) also reported silicate melt
inclusions quenched from 800 °C containing silicate glass extremely rich in K2O, with compositions close to K-
PT

feldspar.
CE

Campos et al. (2006) reported multiphase inclusions with total salinities between 63 and 74 wt% NaCl equiv. and
homogenisation temperatures up to 900 °C in magmatic quartz phenocrysts from the Zaldívar porphyry copper deposit
of northern Chile. Klemm et al. (2007) observed saline inclusions (up to 60 vol.% salts) with hematite and
AC

chalcopyrite crystals from the El Teniente porphyry Cu-Mo deposit in Chile, where salt inclusions with no observable
vapour bubbles were also reported. Li et al. (2011) reported high temperature (vapour disappearance between 620 and
1030 °C), high salinity (between 34 and 82 wt.% NaCl equiv.) fluid inclusions containing multiple solids including
halite, sylvite, hematite and an opaque solid from the Duobuza porphyry copper–gold deposit in northern Tibet.

Pintea (2014) reviewed the occurrence of high temperature brine and melt inclusions in a number of Romanian
porphyry copper-gold deposits. All contained silicate glass inclusions, hydrous salt melt inclusions, “liquid-free” salt
melt inclusions, and globular sulfide melt inclusions. The silicate glass and hydrous salt melt inclusions homogenised
between 970 and 1363 °C while the “liquid-free” salt melt inclusions homogenised between 1115 and 1426 °C.
Koděra et al. (2014) studied Biely Vrch, in the Western Carpathians, and reported inclusions with consistent
proportions of salt crystals coexisting with vapour-rich inclusions. These water-free, salt melt inclusions still
contained a vapour bubble at 850 °C. Laser ablation–inductively coupled plasma mass spectrometry (LA-ICPMS)
analyses indicated that these inclusions contained ~50 wt.% FeCl2, ~30 wt.% KCl, and ~20 wt.% NaCl.
2
 ACCEPTED MANUSCRIPT
In the Late Cretaceous Elatsite porphyry Cu-Au-(Mo-platinum group element) deposit in Bulgaria, Stefanova et al.
(2014) reported that some B1 multiphase inclusions containing ~30 vol.% vapour, a large halite crystal, sylvite, an
opaque phase ± hematite, homogenise at temperatures above 650 °C. B2 multiphase inclusions were similar to the B1
inclusions but typically contained anhydrite as well and did not homogenise at temperatures up to 570 °C. Stefanova
et al. (2014) also observed silicate melt inclusions with variable proportions of crystals and vapour in quartz veins.
They concluded that highly variable Na and metal concentrations in unheated silicate melt inclusions resulted from
heterogeneous entrapment of coexisting silicate melt and brine..

Rottier et al. (2016) reported the presence of heterogeneous silicate melt inclusions in shallow porphyry type quartz
veins from the Cerro de Pasco district, Peru. They suggest that during the first stage, inclusions formed by
heterogeneous entrapment of an evolved, hydrous rhyolitic melt mixed with a hypersaline fluid phase at low pressure

PT
(270 bar) and high temperature (>600 °C). The second stage is marked by the entrapment of metal and sulfur-rich
hypersaline liquid inclusions (≈ 70 wt% NaCl equiv.) that originated from the adiabatic ascent of magmatic
hypersaline fluids from deeper parts of the system.

RI
Thus, inclusions with complex mixtures of magmatic and hydrothermal fluids that homogenise at unrealistically high

SC
temperatures are present in many (if not most) porphyry deposits. However, the formation and entrapment of these
complex melt and fluid inclusions remains poorly understood. In this contribution we present similar data for the
Grasberg porphyry Cu-Au deposit and speculate on their formation and the role of salt melts in porphyry Cu deposits.
NU
The Grasberg porphyry Cu-Au deposit
Geology
MA

The Late Pliocene Ertsberg-Grasberg porphyry-skarn Cu-Au-(Mo) district occupies an area of about 50 km2 within the
central highlands of the Papua province in eastern Indonesia. The district contains intrusive rocks generated during
two temporally and geochemically distinct magmatic events: an early (4.4–3.5 Ma) and volumetrically minor,
relatively low K suite of dykes and sills and a later (3.5–2.6 Ma), high K suite of more voluminous intrusions of which
D

the largest are the Grasberg and Ertsberg intrusions (Leys et al., 2012). The Grasberg deposit is hosted within the
E

Grasberg Igneous Complex (GIC). The complex has extensive hydrothermal alteration (e.g. MacDonald and Arnold,
1994; Pollard et al., 2005) and is composed of three main igneous events, from oldest to youngest, the Dalam igneous
PT

stage, the Main Grasberg Intrusion, and the Kali dykes (MacDonald and Arnold, 1994). A fourth intrusive stage
named the Gajah Tidur has also been identified under the southwestern contact of the GIC. The Gajah Tidur alteration
CE

system effectively removed pre-existing mineralisation from the Grasberg orebody (Leys et al., 2012). The destruction
of Main Grasberg Intrusion-stage and earlier mineralisation by the extensive Gajah Tidur alteration implies a
relatively young age.
AC

Mineralisation and alteration

The Grasberg Cu-Au deposit is a pipelike body approximately 950 m in diameter, which widens above 3,400 m,
reaching approximately 2.4 by 1.7 km at surface (Fig. 1). A quartz-magnetite-anhydrite with Au-bearing chalcopyrite
+bornite stockwork vein system, ~1,500 m in vertical extent, is centred on the axis of the Main Grasberg Intrusions.
Magmatic and hydrothermal anhydrite is abundant in the lower parts of the system (Sulaksono et al., 2017).
Chalcopyrite-bornite veins account for the bulk of the copper and gold in the Grasberg deposit. The spatial distribution
of the ore metals is shown in detail in Figure 4 of Leys et al. (2012). Pollard et al. (2005) group the major episode of
chalcopyrite-bornite mineralisation at Grasberg with the anhydrite-quartz veins and molybdenite veins rather than with
the heavy sulfide zone. The heavy sulfide zone (Fig. 2) occurs mainly near the margin of the GIC and consists of
zones of massive, fine-grained, replacement pyrite grading into veins with less well-developed alteration in peripheral
zones. Rare chalcopyrite-bornite veins that cut the heavy sulfide zone are considered as part of the mixed copper-

3
 ACCEPTED MANUSCRIPT
sulfide stage. The mixed copper-sulfide stage is composed largely of chalcopyrite, bornite, nukundamite, digenite-
chalcocite, covellite, and pyrite and commonly occurs within sericitic-argillic altered rocks (Pollard et al., 2005).

Similar to most porphyry deposits, alteration at Grasberg consists of a series of concentric alteration shells
surrounding the Main Grasberg Intrusion (Fig. 2). Paterson and Cloos (2005) defined the alteration types in the GIC
based on a series of distinct mineral assemblages from the upper levels of the GIC ranging in elevation from 3515 m
to 4129 m. The “potassic zone” is defined by the presence of secondary biotite, K-feldspar, magnetite and anhydrite.
Chalcopyrite is the dominant ore mineral and bornite prevalence increases with depth. The “phyllic zone” is marked
by the presence of sericite, pyrite and covellite. Propylitic alteration consists of epidote, chlorite, hematite and calcite.
The heavy sulfide zone occurs mainly near the margin of the GIC (Fig. 2) and consists of zones of massive, fine-
grained replacement pyrite grading into veins with less well-developed alteration in peripheral zones (Pollard et al.

PT
2005).

Skarn formation at Grasberg lies below 3,100-m elevation, suggesting that the causative intrusion did not extend

RI
above this level. Proximal skarn occurs around the entire circumference of the Dalam diorite, except where removed
by the South Kali dykes to the southeast. Distal skarn formation is related to favourable lithostratigraphic horizons in

SC
the lower Waripi and upper Ekmai formations (Leys et al. 2012). The large skarn on the left in Figure 2 is known as
the Kucing Liar Skarn, which is hosted within a heterogeneous package of rocks containing sandstone, calcareous
shale, and thinly bedded limestone. Ore-grade mineralisation extends for ~1.6 km along strike, ~800 m down-dip, and
NU
varies in thickness from 50 to >700 m. The skarn and mineralisation are in places contiguous with the Heavy Sulfide
Zone at the margin of the GIC (Leys et al. 2012).
MA

Quartz veins
Penniston-Dorland (1997, 2001) has carried out a detailed petrographic and cathodoluminescence study of the quartz
veining in the Grasberg Cu-Au deposit. Quartz veins in the GIC range in size from 1 mm to 10 cm wide, but most are
generally less than 1 cm wide. They are generally straight to slightly curved; some veins branch and others form a
D

three dimensional stockwork. Regions of dense quartz vein stockworks are shown in Figure 2. Crosscutting veins are
common, especially in the centre of the GIC. The veins dip steeply and have preferred orientations consistent with
E

emplacement into a pull apart region of a NW-trending, left-lateral strike-slip fault system. Quartz is commonly
PT

oriented perpendicular to the vein wall. Other important vein minerals include magnetite, chalcopyrite, and pyrite.
Biotite is found in veins post-dating the Late Kali Intrusion.
CE

Penniston-Dorland (1997, 2001) divided the quartz veins in the GIC into two distinct stages. Stage 1, the major ore-
forming event, created at least five generations of cross-cutting veins in the Main Grasberg Intrusion and Dalam
Igneous Complex. Stage 2 created at least five generations of cross-cutting veins observed in the Late Kali Intrusion
AC

and likely extending into the Dalam Igneous Complex and the Main Grasberg Intrusion. The quartz veins examined in
this study are hosted by the Dalam Igneous Complex and correspond with the Stage 1 c-d veins of Penniston-Dorland
(2001).

A composite CL map of a quartz vein from this study is shown in Figure 3. It can be seen from the vein symmetry
that the quartz crystals initially grew from the walls of the vein into an open fracture. The quartz crystals were initially
relatively small near vein walls but become larger as towards the centre of the vein. Concentric growth zones are
clearly observed in the larger crystals and there are changes in the orientation of the growth zones and also occasional
truncation of the zoning as noted by Penniston-Dorland (2001). Towards the centre of the vein the crystals become
smaller and irregularly shaped but exhibit CL colour banding parallel to vein walls, which is normally associated with
crack-seal processes. Finally, the remaining open space is infilled with sulfides (black colour) and almost pure white
quartz.

4
 ACCEPTED MANUSCRIPT
Previous fluid inclusion studies
Harrison (1999) studied fluid inclusions in quartz veins at Grasberg at elevations between approximately 3000 to 4000
m. Three types of aqueous fluid inclusions were identified in his study: (1) highly saline, multi-daughter inclusions,
(2) vapour-rich inclusions, and (3) halite-sylvite inclusions with liquid and vapour. The highly saline, multi-daughter
inclusions were reported to contain hematite, halite, sylvite, anhydrite, an opaque phase, Fe-chloride and an
unidentified daughter mineral in addition to liquid and vapour. Types 1 and 2 were reported to occur together in
abundance in the centre of the GIC. Inclusions associated with potassic alteration had the highest homogenisation
temperatures, with a range of 521 °C to >700 °C (the limit of their heating stage). Salinities ranged from 56 to 84
wt.% NaCl equiv. Fluid inclusions associated with phyllic alteration had homogenisation temperatures ranging from
345 °C to >700 °C and salinities ranging from 37 to 69 wt.% NaCl equiv.

PT
Baline (2007) measured fluid inclusion populations in 27 samples from the Dalam Intrusion, Main Grasberg Intrusion,
and Kali dykes and noted a similarity in fluid inclusion compositions, temperatures, textures, and chemical zonation in
quartz crystals in the veins in all three intrusive units. Five types of inclusions were identified: (1) liquid + vapour, (2)

RI
vapour-rich, (3) liquid + vapour + halite, (4) liquid + vapour + multi-daughter crystals, and (5) liquid + vapour +
opaque. Homogenisation temperatures for types 2, 3, and 4 were reported to be similar. Homogenisation temperatures

SC
ranged from 266 °C to >700 °C (the limit of their heating stage). Salinities ranged from 32.3 to 72.1 wt.% NaCl equiv.
Fluid inclusions in anhydrite from late-stage veins record significantly lower temperatures, approximately 250°C at a
salinity near 10 wt.% NaCl equiv.
NU
Methods
The samples for this study were vein quartz collected from Drill Hole GRS-37-33 at a depth of 241.5 m, which
MA

equates to an elevation of 3713.2m (Fig. 1). All samples are mineralised and are associated with proximal phyllic
alteration of the Dalam Fragmental Unit with pervasive silicification and chalcopyrite as the dominant sulfide mineral
in the veins.
D

Microthermometry was carried out on a Linkam THMSG600 heating/cooling stage and a Linkam TS1400XY heating
E

stage. The THMSG600 stage was calibrated with synthetic fluid inclusions and is accurate to ± 0.2 °C below 30 °C
and ± 2.0 °C above 30 °C. In the low temperature runs a heating rate of 20 °C per minute was used but was reduced to
PT

5 °C per minute over intervals where phase changes were observed. High temperature inclusions were studied on the
TS1400XY stage, which was calibrated at the melting points of NaCl (800.7 °C) and gold (1064.4 °C) and
CE

measurements were accurate to ± 5 °C. A heating rate of 20 °C per minute was used in the high temperature studies.
Student and Bodnar (1999) and Danyushevsky et al. (2002) have shown that fast heating rates can result in an
overestimate of homogenisation temperatures by up to 100 °C in some melt inclusions. To test for this effect a few
AC

inclusions were heated at a slower rate of 5 °C per minute above 800 °C but no significant change in homogenisation
temperature was observed. It was not possible to use slower heating rates due to problems with sample oxidation and
melting of the host quartz at such high temperatures, as well as potential damage to the microscope optics.

LA-ICPMS analysis used an F2 157 nm Coherent LPF202 laser, which was coupled with a Varian 820-MS
inductively-coupled mass spectrometer at the Research School of Earth Sciences, Australian National University. The
157 nm laser was passed through an aperture to produce a square ablation pit, 50 x 50 microns in size. The size of the
ablation spot was increased for large melt inclusions to ensure that whole inclusions were ablated. The laser energy
density on the quartz surface was set to 5 J/cm2 with a repetition rate of 20 Hz. Masses of 7Li, 23Na, 24Mg, 27Al, 29Si,
34
S, 35Cl, 39K, 43Ca, 47Ti, 57Fe, 65Cu, 75As, 77Se, 85Rb, 88Sr, 89Y, 95Mo, 133Cs, 140Ce, 172Yb, and 208Pb were analysed with
dwell times of 10 ms and 197Au was analysed with a dwell time of 40 ms. As reported in Table 1 most FIAs only
contain one type of melt or fluid inclusion with the exception of B2 Hypersaline inclusions which also have vapour-

5
 ACCEPTED MANUSCRIPT
rich inclusions in their FIA. However, reliable LA-ICPMS analyses could only be obtained from a few vapour-rich
inclusions. The LA-ICPMS data for these FIAs and other individual analyses are reported in the electronic
supplementary material.

The SILLS software package (Guillong et al. 2008) was used for LA-ICPMS data reduction. All elements were
quantified against NIST SRM612 as the external standard. Absolute concentration values were calculated after LA-
ICPMS analysis by internal standardisation against Na concentrations determined by microthermometry on similar
inclusions in the same FIA. Most fluid types contained significant amounts of other cations as well, so that the NaCl
equivalent salinity (NaCl equiv.) is not identical to the true fluid salinity. However, true Na concentration for internal
standardisation can be reasonably well approximated by the mass balance procedure as described by Heinrich et al.
(2003). Limits of detection were calculated using Eq. (6) in Pettke et al. (2012), which yields concentration of an
element at the 95% confidence level. The method is implemented in the SILLS software package.

PT
A Renishaw inVia Reflex Spectrometer System equipped with a standard confocal microscope was used for Raman

RI
spectral analysis. A Renishaw diode-pumped solid state laser provided 532 nm laser excitation with 5 mW power at
the sample. A 2400 grooves/mm grating was used giving a spectral resolution of 0.5 cm-1. Single Raman spectra were

SC
obtained using a 1 second integration time with 10 accumulations and a 100x Leica microscope objective, which
focused the beam to a spot size of 0.8 µm. Wavenumber calibration was carried out using an internal silicon standard
and was performed as an automated procedure using the Wire version 4.2 software. Two-dimensional Raman mapping
NU
was carried out using the StreamLineHR™ mode. The maps were produced by rastering the region of interest in 1 µm
steps and recording a Raman spectrum at each step with a 1 second integration time and 1 accumulation. Maps of
different minerals were produced by plotting the intensity of definitive Raman bands for each mineral at each step.
MA

These were then displayed as coloured maps using the Wire software.

Melt and fluid inclusion petrography

D

A summary of the melt and fluid inclusion types observed in this study, and their associated fluid inclusion
E

assemblages (FIAs), is given in Table 1.

Primary and Pseudosecondary Inclusions
PT

Melt Inclusions
Isolated silicate melt inclusions and sulfide-rich melt inclusions coexist with fluid inclusions in the vein quartz from
CE

Grasberg. Figure 4 shows the Raman map of a silicate melt inclusion containing partially recrystallised K-feldspar,
small crystals of gypsum, calcite, and hematite and several vapour bubbles. The amount of vapour is greater than
normally observed in melt inclusions (Cannatelli et al., 2016) suggesting heterogeneous entrapment of these phases.
AC

Although rare, these silicate melt inclusions coexist with nearby hypersaline brine and vapour inclusions indicating the
contemporaneous trapping of these three inclusion types within the quartz veins.

A sulfide-rich melt inclusion with a large crystal of chalcopyrite and smaller crystals of hematite and anhydrite is
shown in Table 1. Sometimes pyrite is present rather than chalcopyrite. Such inclusions are common in the quartz
veins and the fact that they contain hematite and anhydrite daughter crystals suggests that the sulfides are not
accidentally trapped crystals.

Figure 5 shows an example of a rare type of FIA observed in some quartz veins, which contain coexisting salt- and
metal-rich, vapour-free inclusions and vapour-rich inclusions. Crystals of pyrite, chalcopyrite, hematite and anhydrite
are observed in both types of inclusion, as confirmed by Raman spectroscopy. The fact that the same minerals occur in
all inclusions in this FIA indicates they are not accidentally trapped solids but were most likely in a liquid state (i.e. a
6
 ACCEPTED MANUSCRIPT
salt melt) at the time of trapping. The presence of vapour-rich inclusions with a small amount of solids, indicates that
heterogeneous trapping of melt and vapour has occurred.

Aqueous fluid inclusions

This study only reports on samples from the early stage of mineralisation at Grasberg but to facilitate comparison with
studies of other porphyry Cu-Au deposits we have adopted the same fluid inclusion nomenclature as used in Klemm et
al. (2007). Figure 6(a) shows an example of subhedral quartz crystals in the veins examined in this study. The crystals
have corroded margins with cavities that are partially infilled by late sulfides. As shown in Figure 6(b), the majority of
inclusions are either primary or pseudosecondary and detailed FIA descriptions are given below. It is not possible to
further refine the timing of the FIAs as crosscutting relationships differ from crystal to crystal. Where present, trails of
secondary FIA appear to contain only two-phase aqueous inclusions (Table 1). As the samples in this study are from a

PT
relatively high level in the system (Fig. 1), intermediate density inclusions were not observed in these veins.

An FIA of B1 Hypersaline inclusions is shown in Table 1. These are multiphase inclusions containing halite ± sylvite

RI
and a small opaque crystal. Vapour bubbles typically fill ~20 to 30 vol.% of inclusions. The constant ratios of vapour,
halite and liquid indicated that these inclusions formed from a homogeneous, high salinity brine at high temperature.

SC
These FIA occur in proximity to B2 Hypersaline inclusions but crosscutting relationships could not be established.

FIAs of coexisting B2 Hypersaline inclusions and vapour-rich inclusions are shown in Figure 7. Halite in the B2
NU
Hypersaline inclusions was distinguished by its cubic habit while sylvite, rod-shaped anhydrite, hematite, and
chalcopyrite ± pyrite were identified by Raman spectroscopy. These inclusions may also contain other colourless
crystals, e.g. Fe-chloride, as identified by SEM-EDS analysis. The salts generally fill over 60 vol.% of the inclusion
MA

and only a small amount of liquid is present. The vapour bubble typically occupies about 10 - 40 vol.% and is
commonly squeezed between the inclusion wall and the daughter crystals. The vapour-rich inclusions (Table 1)
typically contain 100 vol.% vapour at room temperature although some with rims of aqueous fluid are occasionally
D

observed. Sometimes a thin layer of crystals is observed on the walls of the inclusions (Fig. 7a). Figure 7(b) shows a
healed fracture containing an FIA of dense, B2 Hypersaline inclusions adjacent to a separate array of vapour-rich
E

inclusions. The coexistence of these two types of inclusions in a single healed fracture suggests that immiscibility has
PT

occurred.

Secondary Inclusions
CE

Secondary trails of two-phase, aqueous inclusions are also observed in some veins (Table 1). These can be
distinguished from other inclusions by their irregular shapes, larger size and the cross-cutting trails. The vapour
bubble occupies 15 – 30 vol.% of the inclusion.
AC

Microthermometry
A histogram of the homogenisation and partial homogenisation temperatures of the fluid inclusions is shown in Figure
8 and details of the microthermometric data are given in electronic supplementary material. Although most fluid
inclusion assemblages (FIAs) also contained vapour-rich inclusions, their homogenisation temperatures were not
measured due to the very small amount of liquid they contain which makes total homogenisation difficult to observe.
Sterner (1992), reported that total homogenisation to the vapour phase (via liquid disappearance) in fluid inclusions
may yield erroneously low temperatures by up to several hundred degrees Celsius. No gases were detected during
Raman spectroscopic analysis of the vapour phase of inclusions in this study but a previous study has detected the
presence of CO2 in some inclusions from Grasberg (Fu et al., 2003).

7
 ACCEPTED MANUSCRIPT
The salinity of B1 Hypersaline inclusions was determined by final dissolution of halite and ranged from 39.3 wt.%
NaCl eq. to 54.0 wt.% NaCl eq. Total homogenisation or partial homogenisation (V+L+S → L+S) took place by
disappearance of the vapour bubble and occurred over a temperature range from 356 °C to 602 °C (Fig. 8).
Dissolution of opaque solids was not observed at temperatures below 650 °C. The B2 Hypersaline inclusions yielded
the highest salinities and extremely high homogenisation temperatures (Fig. 8). Salinity estimates based on the
dissolution temperature of halite and sylvite and the NaCl-KCl-H2O diagram and with the software package “Fluids”
(Bakker, 2012) yield values between 57.2 to 64.4 wt.% NaCl and 14.2 to 17.6 wt.% KCl and a total salinity between
65.0 wt.% NaCl equiv. and 74.7 wt.% NaCl equiv.

The behaviour of B2 Hypersaline inclusions during heating is very complex and some of the dissolution processes are
shown in Figure 9. At room temperature the bubbles in this FIA of inclusions are compressed against the walls of the

PT
inclusions by a mass of colourless salt crystals (Fig. 9a). A reddish hematite and opaque chalcopyrite crystal are also
observable in most inclusions. Upon heating, sylvite is the first crystal to dissolve between 147 °C and 292 °C. An
unidentified crystal is the next to dissolve between 318 °C and 439 °C. As heating progresses, the halite crystal begins

RI
to dissolve and become rounded in shape (Fig, 9b) and finally dissolves between 538 °C and 605 °C. Anhydrite
dissolves after halite at temperatures between 600 °C and 648 °C, leaving chalcopyrite and hematite as the only

SC
remaining solids (Fig. 9c). When observable, the chalcopyrite crystal dissolves between 738 °C and 989 °C. Hematite
is the last solid to dissolve typically between 976 °C and 990 °C, leaving a vapour bubble, a salt melt and a very small
rim of a second immiscible phase, thought to be silicate melt (Fig. 9d). Upon further heating, the salt melt becomes
NU
rounded and the amount of the clear immiscible fluid appears to increase (Fig. 9e). During this time the bubble may
also disappear leaving only the two immiscible fluids, which do not homogenise below 1400 °C (the limit of the
heating stage). In some inclusions a small sulfide globule and a vapour bubble remain within the salt melt above 1250
MA

°C (Fig. 9f).

Cooling of the inclusions from the high temperatures is characterised by the reappearance of the vapour bubble at 10
to 30°C below the bubble disappearance temperature. Continued cooling results in re-nucleation (in reverse order) of
D

all the dissolved solid phases: first the opaque hematite and chalcopyrite followed by anhydrite, halite, and finally
E

sylvite, which crystallises below 100° C, in the free space left by the earlier nucleated solids. At room temperature, all
the initial phases are present but appear as quenched, glassy phases rather than as the original crystals due to the rapid
PT

rate of cooling. Repeated heating of the same inclusions generally resulted in an increase of the bubble disappearance
temperature by 20 to 50 °C. The expansion of the clear immiscible phase in Figure 9(e) suggests that additional silica
CE

has been dissolved from the inclusion walls during heating of the inclusions, which increases the volume of the
inclusions and the subsequent dissolution temperatures.

Secondary aqueous inclusions were relatively rare in this study and exhibited eutectic melting temperatures between -
AC

31 °C and -26 °C with hydrohalite melting around -22.2 °C. Final ice melting occurred between -16.6 °C and -10.6 °C
with a mode at -10 °C equating to salinities between 14.1 and 19.9 wt.% NaCl equiv. Homogenisation to the liquid
phase took place over the range from 255 °C to 521 °C with a mode at 325 °C (Fig. 8).

Laser ablation-ICPMS microanalytical results

All LA-ICPMS analyses were carried out on unheated samples in order to avoid the problem of Cu and other element
diffusion into the inclusions as reported by Rottier et al. (2017). Determination of the internal standard Na
concentration by microthermometry was carried out on other inclusions in the same FIA after LA-ICPMS analysis as
discussed above. Only three vapour-rich inclusions yielded reliable LA-ICPMS data, which is shown along with the
analyses of the coexisting B2 Hypersaline inclusions from these FIAs in the electronic supplementary material. All
other LA-ICPMS data reported in the electronic supplementary material are analyses of single inclusions as analyses
of the coexisting vapour-rich inclusions could not be obtained.
8
 ACCEPTED MANUSCRIPT
Sodium, K and Fe were the major elements in B2 Hypersaline inclusions and these were measured with precisions of
1 – 2 wt.%. Lesser amounts of Li, Mg, Al, S, Ca, Ti, Cu, As, Se, Rb, Sr, Y, Mo, Cs, Ce, Yb, and Pb were also
recorded with precisions better than 300 µg/g and typically below 50 µg/g. The concentration ratios of Na, K and Fe
are shown in Figure 10 with Na typically between 20 – 60 %, K typically between 20 – 40 % (occasionally up to 70
%), and Fe generally between 20 – 50 %.

Figure 11 shows a plot of Cu versus Na+K+Mg concentration for selected inclusion types. Na+K+Mg were selected
because they are expected to be the major chloride salts in the inclusions. Ca was not included due to its poor
sensitivity in LA-ICPMS analyses. The B2 Hypersaline inclusions have the highest combined Na+K+Mg
concentrations, which range from 20 – 46 wt.%. These inclusions also show a wide range of Cu concentrations, which
are in accord with the fact that chalcopyrite is not readily observed in every B2 Hypersaline inclusion. However, they

PT
do record the highest observed Cu concentration of 6.3 wt.% (Fig. 11). The aqueous and vapour-rich inclusions record
lower Cu concentrations. Gold was generally below detection but concentrations of up to 2 µg/g were obtained in a
few B2 Hypersaline inclusions (see the electronic supplementary material).

RI
A number of sulfide-rich melt inclusions were also analysed by LA-ICPMS. As no reliable internal standard was

SC
available for these analyses, ratios of counts from each element against Ti (a minor element) are used. The resulting
data are presented in the electronic supplementary material. An inclusion containing chalcopyrite showed high ratios
of Fe and Cu as expected. Other inclusions with pyrite showed enhancement of the Fe ratio but only minor
NU
enhancement of the Cu ratio. The ratios of Na, K, Rb and Pb were also slightly elevated in the sulfide-rich inclusions
while a few inclusions showed a moderate enhancement in the Mg ratio. Unfortunately, reliable LA-ICPMS data
could not be obtained from the silicate and salt melt inclusions due to their very small size.
MA

Titanium-in-quartz geothermometer
The titanium-in-quartz geothermometer is based on the assumption that the Ti concentration in quartz is related to the
mineral formation temperature. A higher crystallisation temperature results in a higher Ti concentration in quartz
D

(Wagner and Pruß, 2002). Several authors have previously used the titanium-in-quartz geothermometer to estimate the
E

formation temperature of quartz in porphyry copper deposits (e.g. McInnes et al., 2005; Muller et al., 2010; Mercer
and Reed, 2013).
PT

The Ti values in the host vein quartz were obtained during the abovementioned LA-ICPMS analyses and ranged from
31 – 170 µg/g (see the electronic supplementary material). As shown in Figure 3 the oscillatory zoning in the quartz
CE

crystals is much finer than the laser spot (50 µm), and hence, the analyses give an average Ti concentration in the
crystal. McInnes et al. (2005) estimate that the Grasberg deposit formed at depths of up to 2 km below the
paleosurface, which equates to lithostatic pressures of around 400 bar. Rutile has not been observed in the vein quartz
AC

at Grasberg and we have therefore assumed an αTiO2 = 0.5, which is similar to that used for the Butte porphyry Cu
deposit (Mercer and Reed, 2013).

Crystallisation temperatures were calculated using the methods of Thomas et al. (2010; 2015) and Huang and Audétat
(2012) and are shown in Figure 12. Both methods produce similar results with the majority of temperatures lying
between 560 and 600 °C. These temperatures are significantly cooler than indicated from the fluid and melt inclusion
data (Fig. 8). Some possible reasons for this include: (1) the vein quartz porphyry was formed at pressures higher than
expected, (2) the porphyry ore deposit formed much deeper than has been proposed, (3) the Ti concentrations in quartz
phenocrysts changed significantly after crystallisation, (4) the Ti concentration did not equilibrate in quartz, (Note that
Huang and Audetat observed that Ti concentrations in quartz depend strongly on the quartz growth rate), and (5) the
thermobarometers used for these calculations do not apply at such low pressure. This latter point is important and we
suggest that the application of the Ti-in-quartz thermometer at low pressures is problematic and therefore temperatures
9
 ACCEPTED MANUSCRIPT
calculated using the Ti contents in quartz are only indicative at these conditions. Mercer and Reed (2013) also reached
similar conclusions in their study of the Butte porphyry Cu deposit. They suggested that the higher temperatures in
their study reflected the transient thermal conditions related to variable cooling of hot (~700°C) fluids upon contact
with cooler (~450°−500°C) wall rock as magmatic-hydrothermal fluids were injected into the crust along
hydrofractures.

Discussion
The presence of silicate melt inclusions and high temperature B2 Hypersaline inclusions both suggest that the quartz
veins trapped magmatic fluids as they became saturated in a volatile phase. Previous studies of coexisting melt and
fluid inclusions (Lowenstern, 1994); Thomas, 1994; Student and Bodnar, 1999) have found that homogenisation
temperatures obtained with a heating stage differed by as much as 75°C from values obtained using a tube

PT
furnace. However, even accounting for these errors, it is still apparent that the B2 Hypersaline inclusions
homogenise at temperatures much higher than the commonly accepted magmatic temperatures associated with
porphyry deposits (Kouzmanov and Pokrovski, 2012). The reasons for the existence of these extremely high

RI
temperature melt and hypersaline inclusions in porphyry copper systems remains poorly understood.

SC
Heterogeneous entrapment of silicate melt and salt-rich fluid inclusions
Roedder (1992) stated that the great bulk of magmatic inclusions arise from one or more stages of immiscible fluid
separation. Roedder claimed, “immiscibility is involved in the formation of many or even most high-temperature
NU
hydrous highly saline inclusions, as well as in the formation of most of the “ordinary” low- to moderately-saline
aqueous (+CO2) inclusions present in igneous rocks and associated ore deposits.” Furthermore, Student and Bodnar
(2004) noted that melt inclusions are common in porphyry copper deposits, but relatively few data are available
MA

because such melt inclusions are not easily recognised.

Heterogeneous entrapment has already been proposed by several authors to explain the presence of silicate melt
inclusions in quartz from porphyry quartz veins (Harris et al., 2003; Pintea, 2014; Rottier et al., 2016). Muntean and
D

Einaudi (2001) originally proposed a model in which a supercritical, salt-rich aqueous fluid ascends along a quasi-
E

adiabatic path (Figure 13). This fluid encounters the two-phase field at depths of ~6 to 8 km and separates into a high-
density brine and a low-salinity vapour (Bodnar et al., 1985; Driesner and Heinrich, 2007). However, as shown in
PT

Figure 13, this process occurs at sub-solidus conditions (see Muntean and Einaudi, 2001) allowing low-density
vapour, high density, heterogeneous, salt-rich fluid and silicate melt to coexist under these conditions. This is
CE

supported by the contemporaneous trapping of three components within the inclusions of this study. Coexistence of
vapour (+ brine?) in K-feldspar melt inclusions (Fig. 4) and virtually water free, salt-rich and vapour-rich inclusions
(Fig. 5) supports this scenario. Very similar high temperature, multiphase inclusions from other porphyry copper
AC

deposits (Li et al., 2011; Pintea, 2014) also contain rutile and K-feldspar. Harris et al. (2003) reported the occurrence
of silicate and hypersaline inclusions in mineralised quartz veins from the Bajo de la Alumbrera porphyry (see Fig. 2
in Harris et al., 2003). The hypersaline liquid inclusions in the vein quartz were interpreted to be the most primitive
and copper-rich ore fluids that exsolved from the crystallising melt at ~100 MPa. They suggest there was
heterogeneous trapping of the silicate melt and hypersaline metal-enriched fluids, which is characteristic of the
magmatic to hydrothermal transition.

The presence of silicate melt inclusions consisting almost entirely of K-feldspar and vapour is most unusual. It is
possible that these may form by a fractional crystallisation process that occurs above the magma solidus temperature.
Chappell and Wyborn (2004) have used the term “cumulative rocks” to describe granites formed by the concentration
of crystals in a melt for which the bulk remains mobile or magmatic. The cumulative process has been used to

10
 ACCEPTED MANUSCRIPT
describe any process that fractionates melt from a mixture of melt, fluid and crystals. Hence, we suggest that the K-
feldspar dominant silicate melt inclusions observed at Grasberg may have formed from a cumulative process.

Furthermore, the phase changes observed during heating of the B2 Hypersaline inclusions in this study are almost
identical to the phase changes reported for the heterogeneous silicate melt inclusions in the study of Rottier et al.
(2016; Figure 7) resulting in chloride-rich melt containing a vapour bubble and an opaque phase all surrounded by a
clear glassy phase (see Figure 7(f) in Rottier et al., 2016).

The heterogeneous entrapment of a silicate phase and a hypersaline fluid has been explained in a number of ways.
Cloos (2001) proposed that residual melt and hypersaline fluids exist as a heterogeneous mixture of two immiscible
fluids that migrate toward shallow levels and are then trapped. He suggests that when the H2O-unsaturated magma is
emplaced into wall rock that is cool enough, a steep thermal gradient creates a narrow solidification front. Upwelling

PT
of saturated sidewall magma entrains deeper-seated, nearly saturated magma, which decompresses and saturates as it
rises. As the system cools, the depth of H2O saturation and sidewall upwelling increases. Bubbles of mineralising fluid

RI
are generated where the saturation front extends to depths of ~6 km or more.

Rottier et al. (2016) invoke the ascent of a hydrous, evolved melt as a single phase to shallow upper crustal levels.

SC
This single phase subsequently exsolves an aqueous fluid at P-T conditions below the miscibility field of a vapour
phase and a hypersaline fluid. The evolved melt and the newly formed hypersaline fluids are then trapped randomly
NU
creating heterogeneous silicate melt inclusions, while the low-density vapour phase escapes due to its low density. K-
feldspar in the fluid inclusions could represent entrapment of cumulate phases. Exsolution below the miscibility field
of vapour and hypersaline liquid requires anomalously high-T for a given pressure; such a hot environment is ideally
MA

provided by shallow porphyry intrusions.

Blundy et al. (2015) proposed a model whereby brines accumulated at shallow levels above arc systems react with the
sulfur-bearing magmatic gas from deeper primitive melts to form porphyry deposits. At Grasberg, coexisting silicate
D

and salt-rich inclusions may have formed as a result of mingling between magma and brines. It is possible that such a
process could lead to second boiling and the trapping of extremely hypersaline and vapour-rich inclusions.
E

Furthermore, Mavrogenes and Blundy (2017) demonstrated that the reaction of magmatic SO2 gas with calcite, in the
PT

presence of aqueous brines, produces anhydrite and sulfide. This process may have been responsible for the formation
of the abundant anhydrite-sulfide veining in the Grasberg deposit, which is hosted by carbonate rocks (Leys et al.,
2012).
CE

Post-entrapment modification of inclusions?

Post-entrapment modification of fluid and melt inclusions may also result in an increase in the homogenisation
AC

temperature (Bodnar, 2003). An important observation in this study is the presence of anhydrite in the B2 Hypersaline
inclusions. The coexistence of both sulfides (pyrite, chalcopyrite) and CaSO4 (anhydrite) in the same inclusion may
suggest that the inclusions have become partially oxidised after entrapment by the process of hydrogen loss
(Mavrogenes and Bodnar, 1992, 1994). The loss of hydrogen and possibly other elements would decrease the density
of the inclusions and result in a corresponding increase in the homogenisation temperature. To test this possibility, the
method of Mavrogenes and Bodnar (1992; 1994) was used to diffuse H2 into inclusions in a sample of the original,
unheated vein quartz. After H2 diffusion the inclusions appeared basically unchanged with the exception that the
hematite was converted to magnetite. However, no change was observed in the heating behaviour of the inclusions
during later microthermometric studies. The details of these experiments will be the subject of a future publication.

These results indicate that significant H2 loss has not occurred, and agrees with previous observations by Liang et al.
(2009) who reported that primary fluid inclusions from the Yulong porphyry Cu (Au) deposit contained mainly

11
 ACCEPTED MANUSCRIPT
anhydrite and halite as daughter minerals, whereas secondary inclusions also contained sulfide daughter minerals,
similar to those observed in the mineralised quartz veinlets. They argue that the fluids changed from being sulfate
dominant during magmatic processes to sulfide dominant during the main mineralisation processes, corresponding to a
decrease in fluid redox potential from the porphyry to the hydrothermal stage.

The oxidation state of the fluid may also affect other fluid properties as well. Newton and Manning (2005) conducted
experiments to show that CaSO4 solubility increases enormously with NaCl activity and at 800 °C and 10 kbar and
XNaCl of 0.3, CaSO4 molality is 200 times higher than with pure H2O. They propose that the very high solubility of
CaSO4 in concentrated NaCl solutions at near-magmatic temperatures (>750 °C) indicates that a large proportion of
solute sulfur in a mafic magma would partition into an exsolved saline fluid as sulfate, and hence, result in the
trapping of primary anhydrite in fluid inclusions at these temperatures. Although the magmatic fluids in the current

PT
study are at much lower pressures, the presence of anhydrite in the inclusions once again suggests the fluids were
sulfate dominant during the mineralisation process.

RI
The role of salt melts
The mineralised portion of the Grasberg deposit is estimated to have formed at less than 2 km below the paleosurface

SC
(McInnes et al., 2005) at pressures below 400 bar. As shown in Figure 13 a large part of the P-T field below 800 °C at
these pressures is occupied by the vapour + halite field. In this case, the system is essentially water-free and the fluid
would be a salt melt. In this way, the formation conditions at Grasberg can be considered to be similar to high
NU
temperature and low pressure formation conditions of the porphyry-style mineralisation at the Cerro de Pasco deposit
in Peru (Rottier et al., 2016).
MA

As the natural system is more complex than the basic H2O-NaCl system shown in Figure 13, it is useful to consider
the effect of other salts. An examination of the ternary NaCl-KCl-FeCl2 system (Fig. 14) shows that salt melts can
exist to temperatures as low as 309 °C and that the existence of salt melts in vapour-rich, water-poor volcanic
environments may be much more common than previously expected. Salt melt inclusions have also been reported
D

from the Biely Vrch deposit in the Western Carpathian magmatic arc (Kodera et al., 2014) and from primary FIAs in
ruby from Mogok and other deposits in central and southeast Asia (Giuliani et al., 2015).
E
PT

Figures 5(a) and (b) support the presence of molten salt in the veins as the same daughter crystals are present in both
the vapour-free (Fig 5a) and vapour-rich inclusion (Fig. 5b). This is analogous to the heterogeneous trapping of liquid
water and vapour in low salinity systems (Roedder, 1984). Furthermore, salt melts appear to be highly efficient at
CE

sequestering metals as is evident by the volume of chalcopyrite and iron minerals in the salt melt inclusions (Figures
5a and b). The laser ablation data also indicate that the salt-rich inclusions carry the highest concentrations of Cu. This
confirms the studies of fluid and melt inclusions in magmatic rocks and associated ores carried out by Borisenko et al.
AC

(2006). They reported that the high metal-bearing capacity of these fluids are due to their heterogeneous state, the high
extraction capacity of their components, the dynamics and high temperature of fluid separation from melts (>700 °C),
and high concentrations of ore elements in the original ore-bearing melts. Therefore, the formation of salt melts may
play an important, and previously unrecognised role in the metal enrichment processes in porphyry systems.

Conclusions
Detailed petrography of quartz veins from the Grasberg Cu-Au deposit has shown the existence of silicate and sulfide-
rich melt inclusions, virtually water-free, salt-melt inclusions, and coexisting hypersaline inclusions and vapour-rich
inclusions. Late stage, secondary aqueous inclusions are also present in some veins. The copper (and sulfur) contents
of sulfide-rich melt inclusions and hypersaline liquid inclusions, determined by LA-ICPMS, are thought to represent
the fluid composition prior to the main stage of ore precipitation.

12
 ACCEPTED MANUSCRIPT
Microthermometric analyses show that most inclusions homogenised at temperatures below 620 °C. However, the B2
Hypersaline inclusions had a very complex behaviour and exhibited partial homogenisation to a salt melt and a clear
immiscible fluid at temperatures ranging from 820 °C to 1300 °C. A cumulative process may have formed the K-
feldspar-rich inclusions and the contemporaneous trapping of the K-feldspar-rich melt inclusions, hypersaline
inclusions and vapour-rich inclusions suggests a common source for these fluids and a process involving
heterogeneous entrapment of immiscible silicate melt, hypersaline fluid and vapour as previously proposed for other
porphyry copper deposits (Harris et al., 2003; Pintea, 2014; Roedder, 1992; Rottier et al., 2016).

The presence of salt melts, observed in some inclusions, is due to the dominance of the vapour + halite field of the
H2O-NaCl system below 800 °C at the estimated formation pressures for Grasberg (≤ 400 bar). Furthermore, the
presence of K-, Fe-, Ca- and other salts may reduce the eutectic melting point in such systems to as low as ~309 °C.

PT
The rapid expansion and cooling of a two-phase fluid in a low-pressure (fumarolic) environment would lead to the
formation of an anhydrous Na-K-Fe–rich salt melt. Salt melts have a high metal extraction capacity, and hence, may
play an important, and previously unrecognised role in the high temperature, metal enrichment processes in porphyry

RI
systems.

SC
The Grasberg porphyry Cu-Au deposit is thought to have formed from magmatic and hypersaline liquids, and vapours,
that exsolve from large, magmatic intrusions assembled in the shallow crust (Blundy et al., 2015). Heterogeneous
entrapment of immiscible silicate melt, hypersaline fluid and vapour results in the observation of unrealistically high
NU
homogenisation temperatures but gives important clues to the metal-enrichment processes that occur. We suggest that
the melt and fluid inclusions record cycles of transitory, high-temperature (>700°C) hydrofracturing, melt and fluid
release, and vein formation in a cooler (500−600 °C) background host-rock thermal regime. High sulfur and Cu
MA

contents of sulfide-melt and B2 Hypersaline inclusions are interpreted to indicate the fluid composition before the
main sulfide precipitation event, and, therefore, they may be used as tracers of the magmatic body from which they are
exsolved at depth.
D

Acknowledgements
E

The authors thank Tom Weiskopf, Paul Warren and Clyde Leys (Freeport-McMoRan Inc.) for their assistance and the
opportunity to collect samples from the Grasberg deposit. We would also like to thank Leslie Kinsley for assistance
PT

with the laser ablation ICPMS analyses. We also thank the Australian National Fabrication Facility (ANFF) for access
to the Raman microprobe. The anonymous reviewers are thanked for their helpful comments and suggestions.
CE

References
Audétat, A. Simon, A.C., 2012. Magmatic controls on porphyry Cu genesis. In: Econ. Geol. Monograph in honor of
Richard Sillitoe. Hedenquist, J.W., Harris, Camus, M.F. (Eds) Society of Economic Geologists Special
AC

Publication Number 16, 553-572.

Bakker, R.J., 2012. Package FLUIDS. Part 4: thermodynamic modelling and purely empirical equations for H2O-
NaCl-KCl solutions. Mineral. Petrol. 105, 1-29.
Baline, L.M., 2007. Hydrothermal Fluids and Cu-Au Mineralization of the Deep Grasberg Porphyry Deposit, Papua,
Indonesia. MSc. Thesis, The University of Texas at Austin, Austin, p. 268.
Blundy, J., Mavrogenes, J., Tattitch, B., Sparks, S., Gilmer, A., 2015. Generation of porphyry copper deposits by gas–
brine reaction in volcanic arcs. Nature Geosci., 8: 235-240.
Bodnar, R.J., 2003. Reequilibration of fluid inclusions., in: Samson, I., Anderson, A., Marshall, D. (Eds.), Fluid
Inclusions: Analysis and Interpretation. Mineral. Assoc. Canada, Ontario, pp. 213-230.
Bodnar, R.J., Burnham, C.W. and Sterner, S.M., 1985. Synthetic fluid inclusions in natural quartz: III. Determination
of phase equilibrium in the system H2O-NaCl to 1000 C and 1500 bars. Geochim. Cosmochim. Acta 49, 1861-
1874.
Bodnar, R.J., Lecumberri-Sanchez, P., Moncada, D. and Steele-MacInnis, M., 2014. Fluid inclusions in hydrothermal
ore deposits. , in: Holland, H.D., Turekian, K.K. (Eds.), Treatise on Geochemistry, Second Ed. Oxford:
Elsevier, pp. 119-142.
13
 ACCEPTED MANUSCRIPT
Borisenko, A.S., Borovikov, A.A., Zhitova, L.M. and Pavlova, G.G., 2006. Composition of magmatogene fluids and
factors of their geochemical specialization and metal-bearing capacity. Russian Geol. Geophysics 47, 1308-
1325.
Campos, E.A., Touret, J.L.R. and Nikogosian, I., 2006. Magmatic Fluid Inclusions from the Zaldívar Deposit,
Northern Chile: The Role of Early Metal-bearing Fluids in a Porphyry Copper System. Resource Geol. 56, 1–8.
Cannatelli, C., Doherty, A.L., Esposito, R., Lima, A. and De Vivo, B., 2016. Understanding a volcano through a
droplet: A melt inclusion approach. J Geochem. Explor. 171, 4-19.
Chappell, B.W., Wyborn, D., 2004. Cumulate and Cumulative Granites and Associated Rocks. Resource Geol., 54(3):
227–240.
Cloos, M., 2001. Bubbling Magma Chambers, Cupolas, and Porphyry Copper Deposits. Internat. Geol. Rev. 43, 285-
311.
De Vivo, B. and Frezzotti, M.L., 1994. Evidence for magmatic immiscibility in Italian subvolcanic systems., in: De
Vivo, B., Frezzotti, M.L. (Eds.), Fluid Inclusions in Minerals: Method and Applications. Short Course Working
Group International Mineralogical Association. International Mineralogical Association, Pontignamno-Siena,

PT
pp. 345–362.
Driesner, T. and Heinrich, C.A., 2007. The system H2O–NaCl. Part I: Correlation formulae for phase relations in
temperature–pressure–composition space from 0 to 1000 oC, 0 to 5000 bar, and 0 to 1 XNaCl. Geochim.

RI
Cosmochim. Acta 71, 4880–4901.
Eastoe, C.J. and Eadington, P.J., 1986. High-temperature fluid inclusions and the role of the biotite granodiorite in
mineralization at the Panguna porphyry copper deposit, Bougainville, Papua New Guinea. Econ. Geol. 81, 478-

SC
483.
Fu, B., Baker, T., Margotomo, W., Mernagh, T.P., Pollard, P.J., Ryan, C.G., Ulrich, T. and Williams, P.J., 2003.
Copper carried in CO2-bearing vapour? Microanalytical characterization of fluid inclusions from the Grasberg
NU
porphyry copper-gold deposit, Irian Jaya, Indonesia Geochim. Cosmochim. Acta 67, A105.
Giuliani, G., Dubessy, J., Banks, D.A., L’Homme, T. and Ohnenstetter, D., 2015. Fluid inclusions in ruby from Asian
marble deposits: genetic implications. Eur. J. Mineral. 27, 393–404.
Guillong, M., Meier, D.L., Allan, M.M., Heinrich, C.A. and Yardley, B.W.D., 2008. Appendix A6: SILLS: A
MA

Matlab-based program for the reduction of laser ablation ICP–MS data of homogeneous materials and
inclusions., in: Sylvester, P. (Ed.), Laser Ablation ICP–MS in the Earth Sciences: Current Practices and
Outstanding Issues. Mineral. Assoc. Canada, Vancouver, B.C., pp. 328–333.
Harris, A.C., Kamenetsky, V.S., White, N.C., van Achterbergh, E. and Ryan, C.G., 2003. Melt inclusions in veins:
Linking magmas and porphyry Cu deposits. Science 302, 2109-2111.
D

Harrison, J.S., 1999. Hydrothermal alteration and fluid evolution of the Grasberg porphyry Cu-Au deposit, Irian Jaya,
Indonesia. MSc. Thesis. University of Texas at Austin, Austin, p. 205.
E

Heinrich, C.A., Petke, T., Halter, W.E., Aigner-Torres, M., Audetat, A., Gunther, D., Hattendorf, B., Bleiner, D.,
PT

Guillong, M., Horn, I., 2003. Quantitative multi-element analysis of minerals, fluid and melt inclusions by
laser-ablation inductively-coupled-plasma mass spectrometry. Geochim. Cosmochim. Acta 67,3473-3496
Huang, R. and Audétat, A., 2012. The titanium-in-quartz (TitaniQ) thermobarometer: A critical examination and re-
calibration. Geochim. Cosmochim. Acta 84:75–89.
CE

Klemm, L.M., Pettke, T., Heinrich, C.A. and Campos, E., 2007. Hydrothermal evolution of the El Teniente deposit,
Chile: Porphyry Cu-Mo ore deposition from low-salinity magmatic fluids. Econ. Geol. 102, 1021–1045.
Koděra, P., Heinrich, C.A., Wälle, M. and Lexa, J., 2014. Magmatic salt melt and vapor: Extreme fluids forming
AC

porphyry gold deposits in shallow subvolcanic settings. Geology 42, 495–498.

Kouzmanov, K. and Pokrovski, G.S., 2012. Hydrothermal Controls on Metal Distribution in Porphyry Cu (-Mo-Au)
Systems. Society of Economic Geologists, Inc. Special Publication 16, 573–618.
Leys, C.A., Cloos, M., New, B.T.E. and MacDonald, G.D., 2012. Copper-Gold ± Molybdenum Deposits of the
Ertsberg-Grasberg District, Papua, Indonesia. Econ. Geol. Special Publication 16, 215–235.
Li, J.X., Li, G.M., Qin, K.Z. and Xiao, B., 2011. High-temperature magmatic fluid exsolved from magma at the
Duobuza porphyry copper–gold deposit, Northern Tibet. Geofluids 11, 134–143.
Liang, H.Y., Sun, W., Su, W.C., Zartman, R.E., 2009. Porphyry Copper-Gold Mineralization at Yulong, China,
Promoted by Decreasing Redox Potential during Magnetite Alteration. Econ. Geol., 104(4): 587-596.
Lindgren, W., 1905. The copper deposits of the Clifton-Morenci district, Arizona. US. Geol. Survey Professional
Paper 43, 375.
Lowenstern, J.B., 1994. Chlorine, fluid immiscibility, and degassing in peralkaline magmas from Pantelleria, Italy.
Am. Mineral. 79:353–369.
MacDonald, G.D. and Arnold, L.C., 1994. Geological and geochemical zoning of the Grasberg Igneous Complex,
Irian Jaya, Indonesia. J. Geochem. Explor. 50, 143–178.

14
 ACCEPTED MANUSCRIPT
Mavrogenes, J., Blundy, J., 2017. Crustal sequestration of magmatic sulfur dioxide. Geology, 45(3): 211-214.
Mavrogenes, J.A. and Bodnar, R.J., 1992. Experimentally-induced dissolution of chalcopyrite daughter minerals in
natural fluid inclusions., PACROFI IV, Lake Arrowhead, California, USA,, p. 58.
Mavrogenes, J.A. and Bodnar, R.J., 1994. Hydrogen movement into and out of fluid inclusions in quartz:
Experimental evidence and geologic implications. Geochim. Cosmochim. Acta 58, 141-148.
McInnes, B.I.A., Evans, N.J., Fu, F.Q., Garwin, S., Belousova, E., Griffin, W.L., Bertens, A., Sukarna, D.,
Permanadewi, S., Andrew, R.L. and Deckart, K., 2005. Thermal History Analysis of Selected Chilean,
Indonesian and Iranian Porphyry Cu-Mo-Au Deposits., in: Porter, T.M. (Ed.), Super Porphyry Copper & Gold
Deposits: A Global Perspective. PGC Publishing, Adelaide, Australia, pp. 27-42.
Mercer, C.N. and Reed, M.H., 2013. Porphyry Cu-Mo Stockwork Formation by Dynamic, Transient Hydrothermal
Pulses: Mineralogic Insights from the Deposit at Butte, Montana. Econ. Geol. 108:1347–1377.
Muntean, J.L. and Einaudi, M.T., 2001. Porphyry-Epithermal Transition: Maricunga Belt, Northern Chile. Econ. Geol.
96, 743–772.
Newton, R.C. and Manning, C.E., 2005. Solubility of Anhydrite, CaSO4, in NaCl–H2O Solutions at High Pressures

PT
and Temperatures: Applications to Fluid–Rock Interaction. J. Petrol. 46 701-716.
Paterson, J.T., and Cloos, M., 2005, Grasberg porphyry Cu-Au deposit, Papua, Indonesia: 2. Pervasive hydrothermal
alteration., in: Porter, T.M. (Ed.), Super Porphyry Copper & Gold Deposits: A Global Perspective: Adelaide,

RI
PGC Publishing, p. 331-355.
Penniston-Dorland, S.C., 1997 Veins and Alteration Envelopes in the Grasberg Igneous Complex, Gunung Bijih
(Ertsberg) District, Irian Jaya, Indonesia, 402 p. MSc. Thesis. The University of Texas at Austin.

SC
Penniston-Dorland, S.C., 2001. Illumination of vein quartz textures in a porphyry copper ore deposit using scanned
cathodoluminescence. Am. Mineral., 86: 652–666.
Pintea, I., 2014. The magmatic immiscibility between silicate-, brine-, and Fe-S-O melts from the porphyry (Cu-Au-
NU
Mo) deposits in the Carpathians (Romania): a review. Romanian J. Earth Sci. 87, 79-82.
Pollard, P.J., Taylor, R.G. and Peters, L., 2005. Ages of Intrusion, Alteration, and Mineralization at the Grasberg Cu-
Au Deposit, Papua, Indonesia. Econ. Geol. 100, 1005–1020.
Robelin, C., Chartrand, P., Pelton, A.D., 2004. Thermodynamic evaluation and optimization of the (NaCl + KCl +
MA

MgCl2 + CaCl2 + MnCl2 + FeCl2 + CoCl2 + NiCl2) system. J. Chem. Thermo. 36, 809–828
Roedder, E., 1971. Fluid inclusion studies on the porphyry-type ore deposits at Bingham, Utah, Butte, Montana, and
Climax, Colorado. Econ. Geol. 66, 98-120.
Roedder, E., 1984. Fluid Inclusions. Mineralogical Society of America, Chelsea, Michigan.
Roedder, E., 1992. Fluid inclusion evidence for immiscibility in magmatic differentiation. Geochim. Cosmochim.
D

Acta 56, 5-20.

Rottier, B., Kouzmanov, K., Bouvier, A.-S., Baumgartner, L.P., Wälle, M., Rezeau, H., Bendezúd, R. and Fontboté,
E

L., 2016. Heterogeneous melt and hypersaline liquid inclusions in shallow porphyry type mineralization as
PT

markers of the magmatic-hydrothermal transition (Cerro de Pasco district, Peru). Chem. Geol. 447, 93–116.
Rottier, B., Rezeau, H., Casanova1, V., Kouzmanov, K., Moritz, R., Schlöglova, K., Wälle, M. and Fontboté, L.,
2017. Trace element difusion and incorporation in quartz during heating experiments. Contrib. Mineral. Petrol.
172:23
CE

Stefanova E., Driesner, T., Zajacz, Z., Heinrich, C.A., Petrov, P., Vasilev, Z., 2014. Melt and Fluid Inclusions in
Hydrothermal Veins. The Magmatic to Hydrothermal Evolution of the Elatsite Porphyry Cu-Au Deposit,
Bulgaria. Econ. Geol. 109:1359–1381.
AC

Sterner, S.M., 1992. Homogenization of fluid inclusions to the vapor phase: The apparent homogenization
phenomenon. Econ. Geol. 87, 1616-1623.
Student, J.J. and Bodnar, R.J., 1999. Synthetic Fluid Inclusions XIV: Coexisting Silicate Melt and Aqueous Fluid
Inclusions in the Haplogranite–H2O–NaCl–KCl System. J. Petrol., 40: 1509–1525.
Student, J.J. and Bodnar, R.J., 2004. Silicate melt inclusions in porphyry copper deposits: identification and
homogenization behavior. Can. Mineral. 42, 1583-1599.
Sulaksono, A., Echigo, T., Watanabe, Y., 2017. Oxidation State of Hydrothermal Fluids in the Grasberg Porphyry Cu
-Au ± Mo Deposit , Indonesia ,, SEG 2017, Ore Deposits of Asia: China and Beyond. Society of Economic
Geologists, Beijing. 17-20 September.
Thomas, J.B., Watson, E.B., Spear, F.S., Shemella, P.T., Nayak, S.K., Lanzirotti, A., 2010. TitaniQ under pressure:
the effect of pressure and temperature on the solubility of Ti in quartz. Contrib. Mineral. Petrol. 160:743–759
Thomas, J.B., Watson, E.B., Spear, F.S., Wark, D.A., 2015. TitaniQ recrystallized: experimental confirmation of the
original Ti‑in‑quartz calibrations. Contrib. Mineral. Petrol. 69:27
Thomas, R., 1994. Fluid evolution in relation to the emplacement of the Variscan granites in the Erzgebirge region: a
review of the melt and fluid inclusion evidence. In: Seltmann, R., Kampf, H. & Moller, P. (eds), Metallogeny of

15
 ACCEPTED MANUSCRIPT
Collisional Orogens Focused on the Erzgebirge and Comparable Metallogenetic Settings. Prague, Czech
Geological Survey, pp. 70–81.
Wagner, W. and Pruß, A., 2002. The IAPWS formulation 1995 for the thermodynamic properties of ordinary water
substance for general and scientific use. J. Phys. Chem. Ref. Data 31, 387-535.
Wilson, J.W.J., Kesler, S.E., Cloke, P.L. and Kelly, W.C., 1980. Fluid inclusion geochemistry of the Granisle and Bell
porphyry copper deposits, British Columbia. Econ. Geol. 75, 45-61.

PT
RI
SC
NU
MA
E D
PT
CE
AC

16
 ACCEPTED MANUSCRIPT
Figure Captions
Figure 1. Maps showing the location of the Grasberg Cu-Au deposit and (a) geological level plan at 3800 m, and (b)
geological cross section along GRS-37-33. The drill hole GRS-37-33 is also shown and the location of samples used
in this study is indicated by the star.
Figure 2. Cross section of the Grasberg deposit at the 3,100-m-elevation level showing major types of alteration
(modified from Leys et al. 2012).
Figure 3. Cathodoluminescence photomosaic of a Stage 1 c-d vein from the Grasberg Cu-Au deposit. The white
dashed lines delineate the margins of the vein walls. See text for details.
Figure 4. Raman band intensity map of the same inclusion. Green corresponds with the 513 cm-1 band of K-Feldspar,
magenta corresponds to the 1008 cm-1 band of gypsum, blue corresponds to the 1086 cm-1 band of calcite and red
corresponds to the 1317 cm-1 band of hematite.

PT
Figure 5. Examples of coexisting, (a) anhydrous salt-rich inclusions, and (b) vapour-rich inclusions with the same
daughter crystals of anhydrite, hematite, pyrite and chalcopyrite (in transmitted light and as identified by Raman
spectroscopy).

RI
Figure 6. (a) Transmitted light photograph showing vein quartz crystals with primary, pseudosecondary and
secondary trails of FIA, and (b) enlargement of the red square in (a) showing primary and pseudosecondary FIA

SC
containing coexisting B2 hypersaline inclusions and vapour-rich inclusions.
Figure 7. Photomicrographs of (a) an FIA containing vapour-rich inclusions coexisting with B2 Hypersaline
inclusions (the arrow denotes a fringe of precipitated crystals in a vapour-rich inclusion) and (b) a healed fracture
NU
containing an FIA of dense, B2 Hypersaline inclusions coexisting with vapour-rich inclusions.
Figure 8. Histogram of total homogenisation temperatures for the various types of fluid inclusions observed in this
study excluding the vapour-rich inclusions (see text for details).
MA

Figure 9. Phase changes during heating of B2 inclusions: (a) at 25 ºC the inclusions contain a vapour bubble and
multiple salt crystals, hematite and chalcopyrite, (b) at 500 ºC they contain a vapour bubble, a rounded halite crystal,
anhydrite, hematite and chalcopyrite, (c) at 777 ºC they contain a vapour bubble, hematite and chalcopyrite, (d) at
D

1050 ºC they contain two immiscible liquids and a vapour bubble, (e) at 1200 ºC the greenish liquid (salt melt) has
become rounded and the clear liquid is clearly visible and only a few inclusions still contain a vapour bubble, and (f)
E

shows a different inclusion at 1287 ºC which contains two immiscible fluids, a small opaque bubble of molten sulfide
PT

and a larger vapour bubble.

Figure 10. Triangular plot showing the ratios of Na, K and Fe determined by LA-ICPMS analysis of fluid inclusions
from Grasberg.
CE

Figure 11. A plot of Cu vs Na+K+Mg for the three types of inclusions shown in the legend.
Figure 12. Histograms of the crystallisation temperature of quartz calculated from the titanium-in-quartz
geothermometry method of Thomas et al. (2010) (shown in orange) and the method of Huang and Audetat (2012)
AC

(shown in blue).
Figure 13. Phase relations in the NaCl-H2O system modified from Muntean and Einaudi (2001). Black dots represent
fluid inclusion data from Harrison (1999), Baline (2007) and this study. Depths assuming a 1g/cm3 hydrostatic load
and a 2.5g/cm3 lithostatic load are also shown. Isopleths of NaCl in liquid in the two-phase vapour + liquid field are
shown by the short dashed lines. The shaded region depicts the vapour + halite field. Dashed curve A is the liquid
saturation curve for the NaCl-KCl-H2O system where Na/K ratios in the solution are fixed by equilibration with albite
and K feldspar at the indicated temperatures. The heavy dotted line is the H2O-saturated granite solidus from Audetat
and Simon (2012). The heavy dashed arrows show the proposed fluid path based on the fluid inclusion data. V =
vapour, L = liquid, H = halite.
Figure 14. The ternary salt melt system NaCl-KCl-FeCl2. Black dots show the low temperature ternary eutectics in
this system. The insert show a cross section along the NaCl-KFeCl3 join and the red arrows indicate a possible cooling
path for salt melt liquid. (adapted from Robelin et al., 2004). ss = solid solution.

17
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
DE
PT
CE
AC

18
 ACCEPTED MANUSCRIPT
Table 1 Types of fluid inclusions found at the Grasberg Cu-Au Deposit.

Melt and Fluid Inclusion Types Description Homogenisation Salinity

Silicate melt inclusions
consisting mainly of
recrystallised K-feldspar
(identified by Raman
mapping). They also contain
one or more vapour bubbles
and small crystals of other
minerals such as gypsum,
Temperature (Wt.% NaCl Eq.)
These inclusions An aqueous phase
partially was not observed in
homogenised to these inclusions
liquid silicate + although Harris et al.
vapour at (2003) reported
temperatures around small amounts of
1150 °C. brine in similar
inclusions from the

PT
calcite, and hematite. They Bajo de la
occur as isolated inclusions Alumbrera porphyry
and are quite rare. deposit.

RI
Sulfide-rich melt inclusions These inclusions did No aqueous phase
contain relatively large not homogenise could be detected in

SC
crystal of chalcopyrite (or below 1400 °C but these inclusions as
pyrite) and other daughter the sulphide they did not freeze at
minerals such as hematite and typically melted at temperatures down
anhydrite in a glassy matrix. temperatures from to -180 °C.
NU
Some inclusions also 850 – 1000 °C.
contained a vapour bubble
occupying up to 30 vol.%.
MA

These occur as isolated

inclusions and are not very
common.

Salt Melt inclusions These inclusions No aqueous phase

D

typically lack a vapour typically melted at could be detected in

E

bubble and contain several temperatures from these inclusions.

clear salts, hematite and 850 – 1000 °C.
PT

opaque sulfides with no

evidence of an aqueous
phase. These occur in small
CE

FIAs sometimes with

10 µm
coexisting vapour-rich
inclusions but are not very
AC

common.

B1 Hypersaline inclusions. These inclusions

The figure on the left shows homogenised to
39.3 – 53.4
an FIA of these inclusions liquid from 356 to
which contain a large halite 602 °C.
crystal ± sylvite and a small
opaque crystal. Vapour
bubbles typically occupy 20 –
30 vol.%. These FIAs were
not very common.

19
 ACCEPTED MANUSCRIPT

Table 1 (continued)

Melt and Fluid Inclusion Types Description Homogenisation Salinity

Temperature (Wt.% NaCl Eq.)
B2 Hypersaline inclusions These inclusions
contain halite, sylvite, rod- exhibit partial
63.5 – 74.4
shaped anhydrite, hematite, homogenisation to a
chalcopyrite and occasionally salt melt and a liquid
pyrite. Other crystals, e.g. Fe- silicate? phase over

PT
chloride, were identified by a range from 811 to
SEM-EDS analysis. The 1293 °C.
vapour bubble typically
occupies about 10 – 40

RI
vol.%. FIAs of these
inclusions also contain 100

SC
vol.% vapour inclusions (see
Fig. 8) and are very common.

Vapour-rich inclusions Homogenisation

NU
typically contain 80 – 100 data could not be 6.4 – 7.7
vol.% vapour. Vapour-rich reliably obtained
inclusions are very common from the vapour-rich
MA

and typically coexist with B2 inclusions.

Hypersaline inclusions in
FIAs as shown in the photo.
E D
PT

Aqueous inclusions An FIA These inclusions

of aqueous inclusions homogenise to liquid 14.1 – 19.9
containing 15 – 30 vol.% over the range 234 –
CE

vapour is shown in the photo. 458 °C.

These are late-stage,
secondary fluid inclusions but
AC

these FIAs were not very

common.

20
 ACCEPTED MANUSCRIPT
Significance of High temperature fluids and melts in the Grasberg porphyry copper-gold
deposit.

Terrence P. Mernagh and John Mavrogenes

Research School of Earth Sciences, Australian National University, Acton, ACT, 2601, Australia

Highlights
 Hypersaline fluid and melt inclusions in the Grasberg porphyry Cu-Au deposit exhibit unrealistically high
homogenisation temperatures up to 1300 °C.


PT
The high temperatures resulted from entrapment of either immiscible silicate melt, hypersaline fluid and
vapour or cumulates of an evolving silicate system saturated in K-feldspar, brine and vapour.
 The hypersaline inclusions contain the highest Cu concentrations (up to 6.3 wt.%), suggesting they play a role

RI
in metal transport.
 The melt and hypersaline inclusions are interpreted to be components of the fluid before the main ore

SC
precipitation event, and may be used as tracers of the magmatic body from which they exsolved at depth.
NU
MA
E D
PT
CE
AC

21
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10
Figure 11
Figure 12
Figure 13
Figure 14