Prograde Sulfide Metamorphism in Blueschist and Eclogite, New Caledonia

JOURNAL OF PETROLOGY VOLUME 55 NUMBER 3 PAGES 643^670 2014 doi:10.1093/petrology/egu002 Prograde Sulfide Metamorphism in Blueschist and Eclogite...
Author: Brendan Holt
1 downloads 0 Views 2MB Size
JOURNAL OF PETROLOGY

VOLUME 55

NUMBER 3

PAGES 643^670

2014

doi:10.1093/petrology/egu002

Prograde Sulfide Metamorphism in Blueschist and Eclogite, New Caledonia JULIE L. BROWN1*, ANDREW G. CHRISTY2, DAVID J. ELLIS3 AND RICHARD J. ARCULUS3 1

CANADIAN NUCLEAR SAFETY COMMISSION, OTTAWA, ON K1P 5S9, CANADA

2

CENTRE FOR ADVANCED MICROSCOPY, AUSTRALIAN NATIONAL UNIVERSITY, CANBERRA, ACT 0200, AUSTRALIA

3

RESEARCH SCHOOL OF EARTH SCIENCES, AUSTRALIAN NATIONAL UNIVERSITY, CANBERRA, ACT 0200, AUSTRALIA

RECEIVED DECEMBER 22, 2012; ACCEPTED DECEMBER 31, 2013 ADVANCE ACCESS PUBLICATION JANUARY 29, 2014

In New Caledonia, blueschist and eclogite preserve, as inclusions in porphyroblastic minerals, a record of sulfide present during prograde subduction processes. Sulfide inclusions in prograde garnet and lawsonite became chemically isolated from the matrix whereas sulfide minerals in the matrix continued to equilibrate with matrix fluids, or grew later, during retrogression. Cu^Fe sulfide mineral inclusions have been found across metamorphic grade within silicate-defined metamorphic mineral zones spanning a crustal profile of 30 km. Bulk area scans of sulfide inclusions provide compositions that represent mixtures of the solid sulfide that were included as the host silicate minerals grew. In general, single sulfide inclusion compositions and aggregate sulfide assemblages are distinct from those of matrix phases. High Cu contents in sulfide inclusions are interpreted to be a consequence of Fe lost from sulfide to growing garnet, rather than the result of intrinsically high Cu in the bulk-rock. The distribution of sulfide inclusion compositions across metamorphic grade, considered together with the available thermodynamic data, suggests that covellite/nukundamite-bearing inclusions in lawsonite, high in both Cu and S, disappear at higher grades as these sulfide minerals are no longer stable. Similarly, clustering near the ratio Fe:Cu ¼ 1:1 may cease with increasing grade owing to the replacement of chalcopyrite/intermediate solid solution (iss) by the denser assemblage pyrite þ pyrrhotite þ bornite/digenite.

I N T RO D U C T I O N

New Caledonia; sulfide metamorphism; blueschist; eclogite; subduction

Most studies of regional metamorphism ignore sulfide mineralogy because sulfide minerals re-equilibrate much more quickly than silicate minerals (Barton, 1974), and therefore do not represent the conditions of peak metamorphism of the host-rock. Furthermore, although there are abundant experimental data concerning the stabilities of common low-pressure sulfide minerals of Cu and Fe over wide ranges in temperatures, it is currently difficult or impossible to incorporate such phases with confidence in thermodynamic models of high-pressure metamorphic reactions because little work has been done on the system Cu^Fe^S at pressures above 1atm. As a result, the geochemical behaviour of Cu in metamorphic rocks remains very poorly understood. However, Kawakami et al. (2006) showed that in amphibolite- and granulite-facies rocks, sulfide inclusions within porphyroblasts preserved compositions consistent with equilibration during prograde metamorphism, whereas matrix sulfide was retrogressed. The study of sulfide inclusions trapped under metamorphic conditions therefore affords a unique window into the high-pressure phase relations of natural Cu-bearing sulfide minerals. This study extends this approach by investigating whether sulfide inclusions in silicate minerals that formed during subduction show evidence for preserving high-pressure, low-temperature assemblages and compositions. By examining blueschist- and eclogite-facies rocks from

*Corresponding author. Telephone: þ1 (613) 944-1984. Fax: þ1 (613) 995-5086. E-mail: [email protected]

ß The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@ oup.com

KEY WORDS:

JOURNAL OF PETROLOGY

VOLUME 55

New Caledonia, and in particular sulfide inclusions within garnet and lawsonite porphyroblasts, we present evidence for sulfide processes characteristic of the Earth’s deep sulfur cycle. Sulfide inclusions in porphyroblasts should retain their overall bulk chemical composition from the time of enclosure, reflecting the solid sulfide phases that were stable under subduction conditions. In the absence of cracks or fractures in host silicate minerals, sulfide compositions remain isolated from, and unmodified by, subsequent retrograde fluid^mineral processes that characterize much of blueschist^eclogite excavation metamorphism. A sulfide included during prograde porphyroblast growth is nearly a closed system. Major chalcophile elements such as Ni, Cu and S are not accommodated in silicate phases above trace levels, and are trapped in the inclusion. The only element that can act as a major component in both sulfide inclusions and the silicate host, and thus may participate in exchange reactions between them, is Fe. Iron present in sulfide trapped at low pressure can change its behaviour from chalcophile to lithophile at high pressures, principally owing to garnet-producing reactions. Thus, sulfide trapped within garnet may become enriched in those elements that remain chalcophile (e.g. Cu) and do not partake in blueschist^eclogite silicate-forming reactions. At temperatures of 7008C, rates of intra-granular diffusion in garnet are high enough to eliminate zoning at a scale larger than a single crystal; at such temperatures, an entire porphyroblast may be in chemical equilibrium with the matrix (Yardley, 1977). At lower temperatures, such as the conditions in this study, diffusion would occur on a scale that is smaller than the grain size of the enclosing garnet host (Marmo et al., 2002). The range of processes that can be experienced by a trapped inclusion is therefore limited to the following. (1) The loss or gain of Fe from the surrounding silicate, with a concomitant change in the cation:sulfur ratio of the inclusion, and the ratio of Fe to non-Fe cations. The associated need to change the cation:oxygen ratio in the host, or to diffuse other components through it, may be difficult to accommodate, which would inhibit such processes. (2) Even in the absence of Fe exchange with the host, the sulfide assemblage will still re-equilibrate on both the heating and cooling legs of the P^T path, but such changes will be isochemical. Possible isochemical changes to sulfide include the following. (1) The polymorphic transformation of a phase, with no reaction between phases. (2) The change in composition of a solid-solution phase in response to P^T changes. If the overall inclusion

NUMBER 3

MARCH 2014

composition remains fixed, then at least one other phase must be exsolved or resorbed. (3) The homogenization of a multiphase assemblage to a single solid-solution phase. (4) The breakdown of a phase to a multiphase assemblage of the same overall mean composition. Based on these possible isochemical changes, we anticipate that the current phase assemblages occurring as inclusions may be different from either those present at peak metamorphism or those originally incorporated into the host porphyroblast, as discussed by Vernon et al. (2008). However, it is likely that the bulk composition of a given inclusion remains constant. At present, there is very little published work on highpressure phase equilibria in petrologically important sulfide systems, or on sulfide^silicate reactions, although Kullerud (1967) presented some early data for both. This study is a reconnaissance of sulfide minerals from a suite of rocks from the high-pressure belt of New Caledonia, focusing on inclusions trapped in prograde porphyroblasts such as garnet. We examined these inclusions for systematic differences from matrix sulfide minerals, which are vulnerable to retrogression, and also for evidence of the phases that were trapped during porphyroblast growth, and the processes that they have undergone.

The Fe^Cu^Ni^S system at 1 bar Many phases are known in the Fe^Cu^Ni^S system: their phase relations at 1bar and high temperature have been reviewed by Fleet (2006). The Fe^S subsystem contains, in addition to pyrite (FeS2) and pyrrhotite (Fe1^xS), the minerals greigite, smythite, mackinawite and marcasite, which are known either to never be stable or at best to be stable only at very low temperatures (Kullerud & Yoder, 1959; Lennie & Vaughan, 1996; Fleet, 2006). ‘Pyrrhotite’ is actually a group of closely related monoclinic, orthorhombic or hexagonal minerals with differently ordered cations and vacancies. Troilite, the stoichiometric hexagonal relative of pyrrhotite, is stable only below 1478C at 1bar (Fleet, 2006). A distinctive feature of the Fe^Cu^Ni^S system is that the large number of ordered, stoichiometric phases known at low temperature tend to coalesce into a small number of variable-composition phases with broad solid solution fields at higher temperature. The very large number of minerals such as yarrowite, anilite and djurleite that exist between Cu2S and CuS (Potter & Evans, 1976) no longer exist above 2008C, where the only phase between high-chalcocite (Cu2S) and covellite (CuS) is an increasingly broad digenite solid solution, which eventually replaces chalcocite as well, above 4358C (Barton, 1973). This solid solution spans the range Cu2S^Cu1·75S at 6008C, and also extends well into the Cu^Fe^S ternary system, beyond the ideal composition

644

BROWN et al.

PROGRADE SULFIDE METAMORPHISM

of bornite (Cu5FeS4) to about Cu4·1Fe1·4S4 (Cabri, 1973). This is the metal-rich ‘bornite^digenite solid solution’ (bss). Bornite in various structural states exists as a separate phase of narrow compositional range up to 2658C, where it is subsumed by this solid solution (Grguric et al., 2000). Although the nickel-rich part of the system is of only marginal relevance to the current study, in which Ni concentrations were found to be small, we note that pyrrhotite (Fe1^xS), an important phase in our rocks, forms a ‘monosulfide solid solution’ (‘mss’) towards Ni1^xS that becomes complete at a temperature between 300 and 4008C (Misra & Fleet, 1973). Solid solution in pyrrhotite towards CuS is much more restricted (0·3 at. % Cu at 3008C and still only 0·6% at 3508C; Sugaki et al., 1975). Covellite, with a different structure type from that of pyrrhotite, exhibits no significant solid solution, being close to pure CuS up to its incongruent melting at 5078C (Fleet, 2006). There are two known Cu-rich phases with 1:1 metal:sulfur ratios in the Cu^Fe^S system: idaite (‘orange bornite’, Cu3FeS4) and nukundamite (Cu3·38Fe0·62S4). Nukundamite is a rare fixed-composition phase occurring with covellite and sulfur, structurally related to covellite (Rice et al., 1979; Sugaki et al., 1981). Idaite is poorly characterized: the original report of a supergene alteration product of bornite by Frenzel (1959) assumed a formula identical to that of nukundamite, and the two phases have occasionally been confused in the literature, but the true composition of this mineral appears to be nearer Cu3FeS4 (Constantinou, 1975; Sugaki et al., 1981). This idaite may be metastable, and in any case decomposes to chalcopyrite þ nukundamite above 2608C (Wang, 1984), which itself appears to be a metastable assemblage. Seal et al. (2001) re-evaluated earlier thermodynamic data for nukundamite and concluded that its rarity is due to a stability field that spans a very narrow range (maximum 0·4 log units) of fS2, sulfidizing to covellite þ pyrite above that range, and desulfidizing to bornite þ pyrite below. It is not stable at all below 2248C, and is stable with excess sulfur only in the interval 434^5018C. The stability of pyrite þ bornite over that whole temperature range implies that nukundamite þ chalcopyrite is never stable (Seal et al., 2001; Inan & Einaudi, 2002). The third major solid solution is the ‘intermediate solid solution’ (iss), lying in the middle of the Cu^Fe^S triangle. At 6008C, this spans a range of Fe/(Fe þ Cu) ¼ 0·4^0·7, (Fe þ Cu)/S ¼1·00^1·15 (Cabri, 1973). The compositional field shrinks and breaks up at lower temperature, differentiating into ordered phases of restricted composition range such as chalcopyrite (CuFeS2), cubanite (CuFe2S3), and a range of low-temperature metal-rich phases such as mooihoekite, Cu9Fe9S16 (see Cabri & Hall, 1972). The mineral isocubanite is a rare instance of cubic iss solid solution persisting at low temperature in nature. It is formed by

quenching of black smoker fluids in near-freezing ocean water (Caye et al., 1988), but becomes the stable polymorph of CuFe2S3 only above 2008C. Chalcopyrite is thermally the most stable of the ordered iss derivatives: it persists as an ordered tetragonal phase with a distinct composition field until 5578C, when it breaks down to iss þ pyrite (Barton, 1973). The common association of covellite with native sulfur in nature and sulfur vapour in experiments suggests that CuS2 is unstable at near-ambient conditions, and indeed the natural Cu-dominant pyrite mineral, villamaninite, always contains substantial amounts of other components such as Fe, Co and Ni, and is probably metastable (Fleet, 2006). ‘Fukuchilite’, nominally Cu3FeS8, appears to be Fe-rich villamaninite, implying the occurrence of substantial solid solution towards pyrite (Bayliss, 1989). However, the high-P/high-T synthetic study of Munson (1966) showed that pure pyrite-structure CuS2 formed from covellite þ sulfur at pressures above 15 kbar. Many of the sulfide minerals discussed above could have been present in seafloor basalts and sediments, either as primary precipitates or secondarily through sulfidation/ desulfidation reactions. Some sulfides such as pyrite possess strong, well-directed bonding, as evidenced by their high hardness. Therefore, they are kinetically inert, and can persist metastably and retain minor element zonation through geological time. However, most sulfide minerals are considerably more reactive. The small copper cation often shows high diffusivity in sulfide, which allows fixedcomposition phases to merge into broad solid solutions through cation disorder at geologically low temperatures as described above, and is manifest to an extreme degree in Cu2S, where the Cu sublattice ‘melts’ above 1058C to form the solid-electrolyte high-chalcocite phase (Wang, 2012). Given the resulting fast kinetics of sulfide reactions, Cu-bearing phases would have re-equilibrated at an early stage of subduction to produce intermediate and bornite^ digenite solid solutions. Thus, it is unlikely that Cu-bearing sulfide incorporated into garnet growing at depth would reflect the earlier unsubducted seafloor mineralogy. Instead, some such inclusions would be expected to sample single-phase solid-solution phases that were stable under the P^Tconditions of incorporation. Given the paucity of experimental data for sulfide systems in situ at high pressure, it is uncertain at present how increasing pressure modifies the extent of solid solutions, or the compatibility between phases. However, inclusions give us access to the bulk sulfide compositions that were in existence at the time of their isolation in growing porphyroblasts. Some inclusions, such as those of single-phase pyrite, are stable over the whole P^T path inferred for New Caledonian subduction, and may have undergone little change. Other sulfide inclusions will have resorbed or exsolved phases owing to a change in composition range of a solid solution

645

JOURNAL OF PETROLOGY

VOLUME 55

or its complete breakdown, or will have undergone crossed tie-line reactions, at various points during subduction and exhumation. In all cases, we expect that high-P/high-T phases will have been replaced during uplift by those stable at or near ambient conditions by the time they reach the surface. However, because the inclusions are diffusionally isolated from the rock matrix, the bulk composition of each inclusion should remain unchanged. Bulk compositions should correspond to either single phases or assemblages that were stable at metamorphic conditions. Therefore, the analysis of populations of such compositions will give an indication of high-P phase relations. There are cases where sulfide inclusions within a porphyroblast are in fact connected to the matrix through fractures, and have been exposed to retrogressive fluids. These sulfide minerals were termed ‘pseudo-inclusions’ by Kawakami et al. (2006). Because we are primarily interested in inclusions that were trapped on the prograde path, we aim to distinguish genuine prograde inclusions from pseudo-inclusions, as well as from inclusions that are hosted in minerals that formed during retrograde metamorphism. Nevertheless, we record pseudo-inclusion and matrix sulfide, because systematic differences between the phases present in these microenvironments and in true inclusions provide evidence that the entrapment of inclusions protects them from reacting with fluids during retrogression. It should be noted that such fluids may have reintroduced sulfur into the system (e.g. Itaya et al., 1985), resulting in the precipitation of new sulfide minerals or the replacement of old ones. Alternatively, sulfur may have been leached from the system, resulting in the absence of matrix sulfide minerals. Any new sulfide mineral may have sourced Fe either from existing metamorphic minerals or from the fluid, and would have incorporated Cu that was present in the matrix at the time of fluid influx. Here, we examine the sulfide phases present in rocks sampled across a high-pressure, low-temperature metamorphic belt, and determine the following: (1) whether sulfide assemblages in the matrix of a host-rock are different from those trapped as inclusions in porphyroblasts; (2) whether there are systematic differences in these assemblages as a function of lithology or metamorphic grade; for example, pyrite is generally common in metamorphosed basalt, whereas pyrrhotite is common in metasedimentary rocks (Frost, 1991); (3) what the identities of single-phase sulfide inclusion assemblages and the mean compositions of multiphase sulfide assemblages reveal about the sulfide compositions that were trapped at high pressure. Area scan analyses of polymineralic inclusions were performed to estimate original sulfide compositions. The area scan technique and data collection method have been described by Brown (2007), and are described below.

NUMBER 3

MARCH 2014

GEOLOGIC A L S ET T I NG New Caledonia is a 400 km long island in the SW Pacific that is the largest emergent portion of the Norfolk Ridge (Sdrolias et al., 2004). It is surrounded by submerged continental ridges, volcanic-arc ridges and both extinct and active back-arc basins. The Norfolk Ridge is considered to be a piece of continental crust rifted from the margin of Gondwana in the late Cretaceous (Sdrolias et al., 2004). New Caledonia is a mosaic of diverse terranes, including a Mesozoic basement complex overlain by a series of Cretaceous to Eocene sedimentary sequences and a variety of mafic and ultramafic ophiolitic nappes (Fig. 1). The basement rocks of New Caledonia comprise three Mesozoic, arc-derived terranes: the Permian to Jurassic Teremba island arc terrane (Paris, 1981); the schistose metasediments of the Boghen terrane (Aitchison et al., 1995); and the late Carboniferous, fore-arc related Koh ophiolite (Aitchison et al., 1998; Cluzel et al., 2001). These terranes were amalgamated to Australian Gondwana by the Early Cretaceous. This basement is unconformably overlain by a sequence of Late Cretaceous to Eocene sandstones, siltstones, shales and carbonates, the lowermost units marking the onset of rifting during the break-up of Gondwana (see Brothers, 1974; Paris, 1981; Aitchison et al., 1995; Cluzel et al., 2001). A more complete account of the New Caledonian terranes has been given by Brown (2007). The high-P/low-T metamorphic rocks, which are the focus of this study, are located in the northeastern portion of the island. This high-pressure metamorphic belt is one of the most extensive blueschist- to eclogite-facies terranes in the world, covering an area of more than 2000 km2 (Lillie, 1975). From SW to NE, there is a progressive increase in metamorphic grade (Black, 1977). In many places, true isograds have been removed by normal faulting related to late-stage exhumation and extension (Rawling & Lister, 2002), accounting for the rapid increase in metamorphic grade to the north. The high-P/low-T rocks have been divided into the Diahot and Poue¤bo terranes on the basis of differences in protolith lithology, geochemistry, metamorphic grade, and structural features (Brothers, 1974; Black & Brothers, 1977; Briggs et al., 1977; Brothers & Yokoyama, 1982; Cluzel et al., 1994; Clarke et al., 1997; Carson et al., 1999; Rawling & Lister, 2002). Early work established that the highest stratigraphic units retain the highest-pressure mineral assemblages. The Diahot terrane is thought to be tectonically underlain by the Poue¤bo terrane (Carson et al., 1999, 2000). It consists of a range of lawsonite- to epidote^omphacitebearing blueschists, whose dominant protoliths were Cretaceous sandstones and siltstones. It also contains subordinate lenses of basalt, rhyolite, chert, conglomerate and carbonate (Briggs et al., 1977). The petrology of the Diahot terrane metabasites between the villages of Pam and

646

BROWN et al.

PROGRADE SULFIDE METAMORPHISM

Fig. 1. Regional geology of New Caledonia after Aitchison et al. (1995) and Clarke et al. (1997). The box outlines the study area, depicted at a larger scale in Fig. 2.

Oue¤goa indicates two distinct prograde metamorphic events (Fitzherbert et al., 2003). M1 assemblages are patchily preserved; typically, metamorphic minerals (omphacite, chlorite, lawsonite, glaucophane) completely pseudomorph igneous plagioclase and augite. S2 deformation fabrics envelop M1 assemblages, increasing in intensity toward the NE, and preserve M2 transitional blueschist- to lower eclogite-facies assemblages. Peak metamorphic conditions indicated from the Diahot metabasites are P ¼1·7 GPa and T ¼ 6008C (Fitzherbert et al, 2003). The lithology of the Poue¤bo terrane is dominated by metamorphosed basaltic units, referred to as eclogite and garnet glaucophanite (Clarke et al., 1997; Carson et al., 1999). Black & Brothers (1977) suggested that the mafic eclogites of the Poue¤bo terrane represent a metamorphosed back-arc basin sequence. However, Spandler et al. (2005) pointed out that there is a wide variety of rock types in this terrane, including pelitic and ultramafic material, and referred to it as the Poue¤bo Eclogitic Me¤lange. Peak metamorphic conditions of P 1·9 GPa and T  6008C have been deduced for the eclogite (Carson et al., 1999). Semi-pervasive fluid influx during isothermal

decompression formed garnet glaucophanite at P 1·6 GPa (Carson et al., 2000). The entire belt experienced high-P subduction metamorphism in the Eocene, and has been interpreted as a metamorphic core complex (Aitchison et al., 1995; Clark et al., 1997). However, the structural studies of Rawling & Lister (1999, 2002) show that the eclogites form a sheet that lies along the northern limb of a regional antiform. A link has been proposed between the obduction of the New Caledonian Ophiolite Nappe and the eclogite-forming event (Cluzel et al., 2001; Rawling & Lister, 2002). However, Spandler et al. (2005) constrained the timing of peak metamorphism to 44 Ma, 10 Myr prior to the obduction of the New Caledonian Ophiolite Nappe, and suggested that the metamorphic peak was related to the large-scale reorganization of the plate-tectonic configuration in the SW Pacific at this time.

ST U D I E D SA M P L E S The high-pressure belt has been divided into metamorphic zones defined on the basis of silicate equilibria by

647

JOURNAL OF PETROLOGY

VOLUME 55

Black (1977), Carson et al. (1999) and Fitzherbert et al. (2003). In order of increasing pressure, they identified lawsonite, epidote, omphacite and hornblende zones, spanning a range of conditions from c. 0·7 GPa, 3508C up to 1·9 GPa, 6008C. A traverse across these zones thus allows us to access sulfide inclusions from a wide range of high-P conditions. Sample locations are shown in Fig. 2, and their lithologies and non-sulfide mineralogy are given in Table 1. P^T estimates for the zones (above) define similar trajectories to those shown in Fig. 3, but with rocks of the different zones reaching their metamorphic peak at different points along the path prior to exhumation. Inclusions were found in almost all zones. Assignment of a rock to a metamorphic zone was used to constrain the pressure and temperature conditions of the sulfide inclusions, given it is the growth of the distinctive silicate porphyroblasts of garnet and lawsonite that isolated sulfide inclusions with compositions reflecting the high-pressure sulfide mineralogy. No prograde sulfide inclusions were found in samples of the epidote zone, although numerous garnet porphyroblasts were studied from samples collected for this study.

NUMBER 3

MARCH 2014

We propose that this exception does not simply reflect insufficient sampling, although that remains a possibility. The composition and textural associations of matrix and inclusion sulfide within specified metamorphic zones are described below, in order of increasing P. P^T estimates from silicate assemblages are based on the previous studies of Clarke et al. (1997), Carson et al. (1999), Marmo et al. (2002) and Fitzherbert et al. (2003), and are summarized in Fig. 3. A full account of analytical methods and interpretation of analyses is presented in the Appendix.

S U L F I D E M I N E R A L C H E M I S T RY A N D P E T RO G R A P H Y Lawsonite-zone sulfide The lawsonite zone is characterized by the appearance of lawsonite in metamorphosed sedimentary and basalt units. P^T conditions for the lawsonite zone of 0·7^ 1·0 GPa and 350^4008C are based on the metabasalt silicate mineral assemblages (Fitzherbert et al., 2003). No Co

Fig. 2. Simplified geological map of study area, showing metamorphic isograds. Geology and isograds compiled from Black (1977), Maurizot et al. (1989), Rawling & Lister (1999) and Fitzherbert et al. (2003). The apparent epidote-in isograd is actually a fault. Samples from locality numbers are: (1) J36, (2) J15, (3) J35, (4) J16, (5) 23956, (6) 23903, (7) 23897, (8) 73004, (9) 23874, (10) J17, (11) E803, (12) J22, (13) 729xx series, (14) J40, (15) 96312k, (16) 72811, (17) NC2, (18) 99xx series, (19) J37, (20) 73101, (21) 73102, (22) J43, (23) J38, (24) 73112. The locations of the Balade and Me¤re¤trice mines are shown by crossed hammers; Balade is located with (17) in the omphacite zone whereas Me¤re¤trice lies further to the south along the edge of the high-pressure belt.

648

BROWN et al.

PROGRADE SULFIDE METAMORPHISM

Table 1: Lithologies and major silicate or oxide minerals present in rocks of this study Sample no.

Lithology

Lws

Czo/Ep

Om

Gln

Hbl

Phg

Chl

x

x

Prg

Grt (Alm)

Qtz

Ab

x

x

x

x

Others

Lawsonite zone J15a

mafic lawsonite blueschist

J15

metasedimentary schist

J16b

metasedimentary lawsonite schist

J16c

metasedimentary schist

J36

schist

x

x

x x

x

x Ttn, Stp

x

x

x

Sps

x

Ttn

x

x

x

Sps

x

Ttn

x

x

x

Epidote zone 73004a

glaucophanite

x

x

23874

mica schist

x

x

23897

mica schist

x

x

23903

quartzite

23956

mineralized metasedimentary schist

x x

x

x x

x

x

x

x

x

x

x

x

Ttn

x

Ilm x x

Ilm

x

x

Ilm

Omphacite zone 96312k

aluminous glaucophanite

x

x

9908

glaucophanite

x

x

x

x

9918

aluminous eclogite

x

x

x

9930b

eclogite boudin in schist

x

x

x

9931a

eclogite

x

9939

eclogite

x

x

9941b

glaucophanite

x

x

9946

eclogite

x

x

9949

eclogite

x

x

72811

glaucophanite

x

73101

schistose aluminous eclogite

x

x

x

x x

x

x

x

x

x

x

x

x

Rtl

Rtl

x

x

x

x

x

x

x

Rtl Rtl

x

x

x

Ttn, Rtl

x

x x

x x

x

x

x

x

x

Ttn Rtl

x

x

Ttn

x

Ttn

x

Ttn

Ttn

(metasedimentary?) J17

eclogite

x

J40

aluminous glaucophanite (metagabbro)

x

73102

felsic rock

x

9919a

metasedimentary schist

9919t

metasedimentary schist

9923

metasedimentary mylonite

x

x

x

x

x

x

x

9924a

metasedimentary schist

x

x

x

x

x

x

x

9931b

metasedimentary schist

x

x

9931c

metasedimentary schist

x

x

x

x

NC2

mineralized schist from Balade mine

x

x

x

x

x x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

Ttn Ttn

x

Gr, Mt

x

Rtl

Hornblende zone J22

glaucophanite

J37

amphibolite

J38

amphibolite

x

x

x

x x

x

Brs

x

x x

xSps core

Rtl, Ilm in Grt

and rim 72908

amphibolite

x

x

x

x

x

72909

glaucophanite

x

x

Brs

x

x

x

x

72915

amphibolite

x

x

x

x

x

x

x

73112

amphibolite

x

Brs

x

x

x

x

E803

glaucophanite

x

J37a

metasedimentary schist

x

J43

metasedimentary schist

x x x

x

x Act

x x

Rtl

Rtl Rtl Rtl

x

Rtl, Stp

x

Rtl

Ab, albite; Act, actinolite; Alm, almandine; Brs, barroisite; Chl, chlorite; Czo, clinozoisite; Ep, epidote; Gln, glaucophane; Gr, graphite; Grt, garnet group; Hbl, hornblende or other Ca-rich amphibole; Ilm, ilmenite; Lws, lawsonite; Om, omphacite; Qtz, quartz; Phg, phengite; Prg, paragonite; Rtl, rutile; Sps, spessartine; Stp, stilpnomelane; Ttn, titanite.

649

JOURNAL OF PETROLOGY

VOLUME 55

NUMBER 3

MARCH 2014

Fig. 3. P^T path for a New Caledonian eclogite reaching hornblende-zone conditions. Mineral assemblages and phase relationships are from Schmidt & Poli (1998) based on experiments for water-saturated mid-ocean ridge basalt. The lawsonite to anorthite transition is from Liou (1971). grt, garnet; ep, epidote; zo, zoisite; amph, amphibole. Other abbreviations are given in Table 2. P^Testimates are correlated to mineral zones based on studies by Clarke et al. (1997), Carson et al. (1999, 2000) and Fitzherbert et al. (2003). P^Tregions: Lws, lawsonite zone; Ep, epidote zone; Om, omphacite zone; Hbl, hornblende zone.

or Ni was detected in sulfide inclusions from the lawsonite zone. The phases observed or deduced by composition to occur in inclusions and pseudo-inclusions of prograde porphyroblasts are summarized in Table 2, distinct from those that were observed in the matrix or included in silicates that are believed to be retrograde (also given in Table 2).

Lawsonite-zone metamafic rocks The only lawsonite-zone mafic rock studied is metabasalt sample J15a. The metabasalt forms a boudinaged sheet intruding the metasediments from which sample J15 was taken (see below). It contains abundant, fracture-free lawsonite grains. Sulfide minerals were not observed in thin section, but grain mounts made from lawsonite mineral separates revealed rare sulfide inclusions. The dense mineral separates also contained admixed pyrite, which probably originated from the matrix. Most sulfide inclusions found in lawsonite were of the order of 1 mm in size, smaller than the microprobe interaction volume. Consequently, many compositions were extrapolated from area scans. The compositions obtained cluster around the chalcopyrite^iss region of the Cu^Fe^S triangle (Table 3, Fig. 4). Point analyses of single phases (with some cross-contamination) were possible only for a single much larger inclusion that contained

pyrite þ covellite þ chalcopyrite, and a nearby separate grain of digenite with an in situ weathering rim implying that it was a pseudo-inclusion (Table 3, Fig. 5a and b). Covellite was not observed to be in contact with chalcopyrite at the surface of the section through the three-phase inclusion; this is a case where three-dimensional (3D) mapping of the inclusion would be required to verify such a contact relationship. It should be noted that the above three-phase assemblage is not stable in the 1atm phase diagrams of Fig. 6 (see Discussion). Apart from the Cu-rich grain discussed above, a few pseudo-inclusions were found along the edges of lawsonite grains, which are easily recognized by the presence of Fe^O^OH weathering rims and/or by visible fractures connecting the sulfide to the edge of the host lawsonite. Pseudo-inclusion point analyses gave compositions close to the FeS^FeS2 line. Sulfide grains in the matrix are exclusively pyrite. These are generally 50^100 mm, much larger than sulfide inclusions, and comparable in size with the lawsonite and with matrix baryte grains. Baryte grains are found both as inclusions in lawsonite and as a matrix phase in this sample. Previously in the study area, baryte was reported only from the now-defunct Me¤re¤trice mine (Briggs et al., 1977), located in the lawsonite zone.

650

BROWN et al.

PROGRADE SULFIDE METAMORPHISM

Table 2: Sulfide mineral inclusions and pseudo-inclusions in prograde porpohyroblasts in lawsonite-zone rocks Sample no.

Host

Point analyses of inclusions

Observed

Phases deduced

contacts in

from area scans

Pseudo-

Matrix sulfide

inclusions

Contacts observed

inclusions

among matrix sulfide

J15a

mafic lawsonite

Lws

blueschist

Py( Cu), Cp, Cv(Fe), MP,

Py þ Cp,

Cp, MP, MR

Py þ Cv

metasedimentary

Alm

Py, Bn

Py(Cu)

Cc(Fe),

[Po þ Py] J15

Po, Py(Cu),

[Py þ Po] Py þ Cv

[Po þ Py],

schist

Po, [Py þ Po] Py

Py(  Cu), Cv(Fe), MP

J16b

metasedimentary

Sps

Py, Cp, Bn, MR

Py þ Cp

MP, MR, Py( Cu)

Py, Cp, Gln

lawsonite schist J16c

metasedimentary

Py þ Cp, Py þ Gln

incl. in Py Sps

[Po þ Py], Py(

schist

Py, Cp, MR,

Cu), Cp, MP

Py þ Cp, Py þ Gln

Gln incl. in Py

J36

schist

Ab

Po(Cu)R

Py

Also shown are sulfide inclusions in a retrograde silicate host (in this case albite; retrograde sulfide inclusion is denoted R) and matrix sulfide. Bn, bornite; Cp, chalcopyrite; Cv, covellite; MP, bulk composition similar to metal-poor iss with metal/ sulfur50·95; MR, bulk composition similar to metal-rich iss with metal/sulfur41·05; Po, pyrrhotite; Py, pyrite; Lws, lawsonite; Alm, almandine; Sps, spessartine; Ab, albite; Gln, galena. Element(s) in parentheses indicate that the phase contains significant quantities of the element in solid solution or unresolved intergrown phase. Square brackets indicate unresolved intergrowth of two phases.

Lawsonite-zone metasedimentary rocks Samples J15, J16b and J16c are metasedimentary rocks from the lawsonite zone, where sample J15 was collected from the same outcrop as the lawsonite metabasalt J15a. Garnet is the host for inclusion sulfide in all cases, and was in general recovered from mineral separates as few grains were found in thin section. The garnets of sample J15 are almandine-rich with XFe ¼ 0·61^0·64 and XMn ¼ 0·02^0·09, in contrast to the garnets of J16b and J16c, which are spessartine with XFe ¼ 0·29^0·55 and XMn ¼ 0·21^0·61 (Brown, 2007). Spessartine garnet (samples J16b and J16c) has been documented from the upper portions of the lawsonite zone (Black, 1977), whereas Fe dominance had not been recorded in lawsonite-zone garnets prior to Brown (2007); it probably reflects the greater Fe content of the host-rock. Spessartine cores in samples J16b and J16c are often zoned to Fe-rich rims, and this accounts for the spread of analyses. Sulfide inclusions in these lawsonite-zone garnets are normally very small (1^3 mm), with few exceptions. Larger 15^20 mm inclusions appear to consist of homogeneous single phases (Fig. 5c). Most sulfide inclusions are found in garnet rims that are generally enriched in the almandine component relative to garnet cores. Pseudoinclusions lie along narrow quartz-filled fractures within garnets. Matrix sulfide is abundant, comprising much larger euhedral pyrite crystals (up to 1mm), which

commonly overgrow silicate deformation fabrics evidenced by corresponding inclusion trails. Sulfide inclusion backscattered electron (BSE) images are shown in Fig. 5c^e. Representative analyses of inclusion sulfide are reported in Table 3, and the overall distribution of compositions for inclusion and matrix sulfide in Fig. 4. Sulfide inclusion compositions from the almandine-bearing metasediment (J15) differ from those in the lawsonite metabasalt (J15a) at the same location. Whereas mafic rock inclusions tend to be Cu-rich,18 out of 20 sulfide inclusions in metasediment J15 are pyrite (Fig. 5c), with only two Cu-bearing inclusions. One is an isolated, homogeneous bornite inclusion. The bulk area scan of the other plots along the covellite^chalcopyrite line (Fig. 4), implying intergrowth of covellite with Fe-bearing phases, probably nukundamite or pyrite þ bornite. Pyrite and pyrrhotite pseudo-inclusions are hosted in fractures in almandine; the pseudo-inclusions range in average composition from Fe0·75S to Fe4S as a result of oxidation. Samples J16b and J16c were collected from the same location (Fig. 2). Recent blasting in the area provided particularly fresh samples from locality J16, including metasedimentary and ultramafic rock types, although the ultramafic rocks contain no sulfide minerals. The mineral assemblage in the metasediments (spessartine^chlorite^ paragonite^phengite^quartz) corresponds to the upper

651

JOURNAL OF PETROLOGY

VOLUME 55

NUMBER 3

MARCH 2014

Table 3: Representative lawsonite-zone sulfide mineral compositions extrapolated from area scans and from point analyses Extrapolated from area scans (normalized to 100%) for lawsonite-zone sulfide minerals Sample:

J15a

J15a

J15a

J15

J15

J15

J16b

J16b

J16b

J16c

J16c

J16c

Location:

J15a-a2

J15a-a5

J15a-a6

J15-a1

J15-a2

J15-a3

J16b-a9

J16b-a10

J16b-a11

J16c-i5

J16c-a2

J6c-i3

Silicate host:

Lws

Lws

Lws

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

at. % Fe

22·22

24·81

25·67

37·12

33·43

13·55

30·05

31·78

25·39

24·05

38·27

4·87

at. % Co

























at. % Ni

























at. % Cu

25·77

30·48

25·16

45·01

0·00

0·00

41·31

11·37

1·15

23·94

23·42

0·00

at. % S

52·01

44·71

49·17

50·12

62·88

66·57

45·14

58·58

67·07

50·67

52·52

61·73

Fe/M

0·46

0·45

0·50

0·10

1·00

1·00

0·25

0·73

0·97

0·51

0·51

1·00

Co/M

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

Ni/M

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

Cu/M

0·54

0·55

0·50

0·90

0·00

0·00

0·75

0·27

0·03

0·49

0·49

0·00

M/S

0·92

1·24

1·03

1·00

0·59

0·50

1·22

0·71

0·49

0·97

0·90

0·62

MP

MR

Cp

Cv(Fe)

Py

MR

Cp

MP

Identity

[Py þ Po]

MP

Py(Cu)

[Py þ Po]

Representative point analyses for lawsonite-zone sulfide minerals Sample:

J15a

J15a

J15a

J15a

J15a

J15a

J15

J15

J16b

J16b

J16b

J16b

Location:

J15a-i10

J15a-i11

J15a-i13

J15a-i12

J15a-i16

J15a-i17

J15-i1

J15-i2

J16b-i1

J16b-i2

J16b-i5

J16b-i6

Silicate host:

Lws

Lws

Lws

Lws*

Lws

Lws

Grt

Grt

Grt

Grt

Grt

Grt

wt % Fe

46·40

30·70

0·33

1·40

29·17

30·48

14·10

46·04

15·09

12·60

32·25

30·64

wt % Cu

0·23

34·62

69·72

77·75

33·59

30·75

62·03

0·00

56·51

61·58

35·14

34·81

wt % S

52·99

34·65

30·04

21·04

36·84

38·99

26·35

52·94

26·59

25·12

32·75

34·82

Total

99·62

99·97

100·09

100·19

99·59

100·23

102·48

98·98

98·19

99·30

100·14

100·27

at. % Fe

33·25

25·29

23·55

24·34

12·32

33·23

11·41

11·28

26·84

25·14

wt % Co wt % Ni

0·30

2·80

at. % Co























at. % Ni























at. % Cu

0·15

25·04

53·86

63·42

24·05

21·44

47·60

0·00

48·99

48·71

25·70

25·10

at. % S

– –

66·59

49·67

45·84

33·78

52·40

54·21

40·08

66·57

39·60

40·01

47·46

49·75

Fe/M

1·00

0·50

0·01

0·04

0·49

0·53

0·21

1·00

0·19

0·19

0·51

0·50

Co/M

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

Ni/M

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

Cu/M

0·00

0·50

0·99

0·96

0·51

0·47

0·79

0·00

0·81

0·81

0·49

0·50

M/S

0·50

1·01

1·18

1·96

0·91

0·84

1·50

0·50

1·52

1·50

1·11

1·01

Py

Cp

Cv(Fe)

Dg(Fe)

MP

MP

Bn

Py

Bn

Bn

MR

Cp

Identity

*Pseudo-inclusion.

portion of the lawsonite zone (1·0 GPa, 400^4508C; Fitzherbert et al., 2003), with no glaucophane present. J16b inclusion compositions are almost all chalcopyrite (Fig. 5d), pyrite (Fig. 5e), bornite or metal-rich ‘MR’

compositions (Table 3). Inclusions appear homogeneous at the resolution of analysis, implying that any intergrowth is on a very fine (submicron) scale. Although the majority of area scan compositions for J16b inclusions lie between

652

BROWN et al.

PROGRADE SULFIDE METAMORPHISM

Fig. 4. Composition distribution for lawsonite-zone sulfide, distinguished by host lithology. Continuous tie lines indicate observed contacts.

pyrite and chalcopyrite, some approach bornite (Table 3, Fig. 4). Area scans for J16c inclusions indicate chalcopyrite, pyrite and, in one case, a mixture of pyrite þ pyrrhotite (Table 3). Pseudo-inclusions of pyrite in garnet occur along quartz-filled fractures. Matrix sulfide in both J16b and J16c is dominated by 200^500 mm euhedral pyrite and 50^100 mm subhedral chalcopyrite aggregates. Matrix pyrite grains contain silicate inclusion trails contiguous with the deformation fabric in the silicate matrix, suggesting that the pyrite is post-tectonic. Matrix pyrite grains in sample J16c show Co-rich zones in X-ray maps (Brown, 2007) and contain small galena inclusions, which form elongate blebs at a high angle to folded silicate inclusion trails (Brown, 2007). The chalcopyrite is free of such inclusions. Pyrrhotite was found hosted within albite in other metasedimentary rocks in the lawsonite zone (sample J36, Table 2). Albite is interpreted to have grown subsequent to peak metamorphism, as suggested by Fitzherbert et al. (2005), which would imply that the included sulfide is also retrograde. In summary, the lawsonite-zone rocks demonstrate differences between matrix and inclusion assemblages, and also between the inclusion assemblages of different host lithologies. Inclusion hosts also differ between lithologies, being lawsonite in metabasalt but garnet in the

metasediments. Inclusions contain a variety of Fe^Cu^S phases, and are relatively Cu-rich in the metabasalt and spessartine-bearing metasediments in comparison with the almandine-bearing metasediments. Matrix sulfide minerals are dominated by pyrite, with textural evidence of late development in at least J16c. The latter sample also shows evidence of incursion by late mineralizing fluids, in the form of matrix chalcopyrite, the high Co content of pyrite, and galena inclusions in pyrite. Pyrrhotite grains are infrequently observed as visibly altered pseudo-inclusions, and as two inclusions in albite, which is likely to be retrograde.

Epidote-zone sulfide A large regional fault divides the lawsonite from the epidote zone, cutting out the isograd corresponding to the appearance of epidote and disappearance of lawsonite (Fig. 2). Fitzherbert et al. (2003) constrained P^T conditions within the epidote zone to 1·4^1·5 GPa, 450^5008C (Fig. 3). The epidote zone is marked by the occurrence of almandine garnet and absence of spessartine garnet in all rock types. No sulfide inclusions were found within garnet from this zone. However, sulfide minerals were observed in the matrix, and as inclusions in albite and quartz in metasedimentary rock types. Pyrite and chalcopyrite were found

653

JOURNAL OF PETROLOGY

VOLUME 55

NUMBER 3

MARCH 2014

Fig. 5. Lawsonite-zone sulfide. (a) Reflected light micrograph of lawsonite-zone metabasalt J15a (lawsonite grain mount); the box indicates the area shown in (b). (b) Backscattered electron (BSE) image showing covellite, pyrite, chalcopyrite, and digenite inclusions in lawsonite. The bright edge effect around part of the pyrite grain boundary extending from the left of the triangular-shaped covellite grain should be noted. The darker gap between part of the covellite and the host pyrite is a void caused by plucking during polishing. (c) BSE image of lawsonite-zone metasedimentary sample J15, from the same outcrop as J15a, showing pyrite inclusions in almandine. (d, e) Garnet grain mounts from lawsonite-zone metasedimentary sample J16 showing pyrite and chalcopyrite inclusions in spessartine.

included within quartz in quartzite sample 23903, whereas chalcopyrite was found in quartz, and pyrrhotite was found in albite in 23897, and both chalcopyrite and pyrrhotite were found in albite in 23874. The matrix contains pyrite þ chalcopyrite in those host-rocks, schist 29356 and also quartzite 23903, where sulfide is both entrained within and apparently replacing quartz that may itself have been retrograde. Albite formed subsequent to peak metamorphism, during decompression from 1·5 GPa to 0·8 GPa. Therefore, the albite inclusions are interpreted to be retrograde. The same is likely to be true of the quartz, which recrystallizes readily. In glaucophanite sample 73004a, matrix sulfide minerals are pyrite, with minor chalcopyrite and pyrrhotite. These minerals are texturally different: pyrite forms large and commonly euhedral crystals, whereas chalcopyrite forms much smaller, sub- to anhedral grain masses. Pyrrhotite contains inclusions of the other two sulfide minerals. The epidote-zone sulfide minerals found in the matrix and hosted by albite or quartz are summarized in Table 4, and their compositions are plotted in Fig. 7.

Omphacite-zone sulfide Similar to the lawsonite^epidote transition, the boundary between the epidote and omphacite zones in New Caledonia is also defined by a fault that was probably

modified by late extension (Rawling & Lister, 2002). The omphacite mineral zone is characterized by the appearance of omphacite in metasedimentary rocks (Black, 1977). Omphacite-zone rocks formed at P^T conditions of 1·4^1·6 GPa, 550^6008C (Fitzherbert et al., 2003). The rocks collected for this study are likely to represent a spread of temperatures, as 96312k was found close to the epidote zone whereas J17 and J40 were obtained near the hornblende zone, with the other omphacite-zone samples structurally between them. Host-rocks studied are summarized in Table 1, and include metabasaltic eclogites (‘type I’ of Clarke et al., 1997), and more aluminous metagabbros with abundant large clinozoisite porphyroblasts (‘type II’ of Clarke et al., 1997). These are relatively fresh rocks occurring as boudins or sheets within omphacitebearing schists. Glaucophanites and the felsic clinozoisite^ phengite^garnet rock 73102 show fracturing of garnet and abundant chlorite alteration. The glaucophanites are mainly inferred to be metabasaltic, but two correspond to metagabbroic cumulates on the basis of their clinozoisiterich petrology and resulting Al-rich composition (and relict gabbroic texture in the case of J40). Some metasedimentary schists were also collected. Sulfide inclusions are found almost exclusively in garnet, although a few were also found in clinozoisite and paragonite. Even in highly altered glaucophanite garnets, where the garnets were fractured and extensively replaced by

654

BROWN et al.

PROGRADE SULFIDE METAMORPHISM

Fig. 6. Experimental phase relationships in the Cu^Fe^S system at atmospheric pressure from Craig & Scott (1974), modified to include the nukundamite stability data of Seal et al. (2001), shown for temperatures relevant to the New Caledonian belt. Tie lines intersecting the bases of the diagrams point to metallic Cu and Fe. These diagrams illustrate the extent of ternary solid solution as a function of temperature. Pure end-member compositions such as that of Cc are shown as reference marks, even though the phases may be subsumed into growing solid solution fields (for instance, Cc is subsumed into the bss solid solution above 4358C).

chlorite during retrogression, sulfide mineral inclusions are preserved. Overall, sulfides enclosed within garnet are much larger in this metamorphic zone (410 mm on average) than sulfide inclusions in lawsonite and garnet at lower pressures. Most of the samples with sulfide inclusions are located on the south side of the Pam Peninsula (Fig. 2), in well-preserved, interlayered metamafic and medasedimentary rocks. Sulfide phases observed or deduced to be present in omphacite-zone rocks are summarized in Table 5. Representative analyses of inclusions are given in Table 6.

Omphacite-zone metamafic rocks Sample 96312k was collected near the boundary with the epidote zone (Fig. 2). Although it is dominated by almandine, clinozoisite and glaucophane and contains no

omphacite, its silicate compositions are consistent with omphacite-zone conditions. Figure 8 includes a BSE image illustrating the variety of sulfide inclusions found in one garnet from this rock. The 5 mm pyrrhotite inclusion appears to be single-phase; however, whereas a point analysis contained no Cu, an area scan shows that it contains almost 7 wt % (5 at. %) Cu (Table 6). Sample 96312k marks the first appearance of unequivocal single-phase pyrrhotite inclusions trapped in garnet. The two larger inclusions in Fig. 8 are more complex. The lower of the two contains a phase that is close to covellite in composition but contains 1wt % Fe, possibly owing to unresolved intergrown species, coexisting with pyrite (Table 6). The uppermost inclusion in Fig. 8 contains pyrite and digenite. Some point analyses also indicate intimate mixtures of these minerals  covellite. It should be noted that the

655

JOURNAL OF PETROLOGY

VOLUME 55

NUMBER 3

MARCH 2014

Table 4: Sulfide minerals present as inclusions in retrograde silicates and in the matrix of epidote-zone rocks Sample no.

Host

Inclusions

Matrix sulfide

Contacts observed among matrix sulfide

73004a

glaucophanite

23874

mica schist

Ab

Po(Ni), Py( Co, Ni, Cu), Cp

Po þ Py þ Cp

Po, Cp

Py, Cp

Py þ Cp

23897

mica schist

Ab

Po, Cp; in contact

[Py þ Cp]

23903

quartzite

Qtz

Py, Cp; in contact

Py, Cp

Py þ Cp

23956

mineralized schist

Py(Ni), Cp

Py þ Cp

Fig. 7. Composition distribution for epidote-zone sulfide, distinguished by host lithology.

stability of pyrite^digenite^covellite is inconsistent with the tie lines in Fig. 6; the possibility that this is a disequilibrium assemblage is discussed below. An area scan of the upper inclusion yielded an average composition that is S-poor relative to the pyrite^digenite tie line (Table 6), consistent with the presence of a metal-rich third phase such as pyrrhotite or chalcopyrite. Although the data for each inclusion were obtained only from a 2D section through a 3D object, which results in bulk compositions scattered along a tie line for genuinely two-phase inclusions, departure of the average composition from that tie line cannot occur without the presence of a third phase.

In all of the low-P experimental data, pyrrhotite never coexists with digenite (Fig. 6), so it is likely that the S-poor phase is chalcopyrite. This is the highest-pressure rock in which covellite was observed as an inclusion phase. No matrix sulfide was observed in 96312k. Pyrrhotite is an inclusion phase in garnet in every other omphacite-zone metamafic rock in which sulfide inclusions were observed (9918, 9930b, 9931a, 9939, 9941b, 9946, 9949, 73102, J17, J40). Figure 9 shows examples of pyrrhotite inclusions, some of which are in contact with chalcopyrite. The range of sulfide analyses is illustrated in Fig. 10 for rocks other than the atypical sample 96312k (discussed

656

schistose aluminous eclogite

73101

657

metasedimentary schist

metasedimentary mylonite

metasedimentary schist

metasedimentary schist

glaucophanite

glaucophanite

glaucophanite

schistose aluminous eclogite

9919t

9923a

9924a

9931c

9908

9941b

72811

73101

Czo

Alm

Alm

Alm

Alm

Alm

Alm

Alm

Alm

Alm

Alm, Czo

Alm, Prg

Alm

Alm

Alm

Alm

PoR

Po

Po( Ni)

Po( Ni)

Po( Ni)

Po

MP, MM, MR

Po(Ni), Py( Ni), Cp,

Po( Co, Ni), Tsp

MP(Ni), PoR

Po þ Cp

Py þ Cp

Py(Ni), Cp, Po(Ni)R Po( Co, Ni, Cu),

Po þ MP

Po( Co, Ni, Cu)

Po(Ni)

Po

Po( Ni,Cu)

Po(Ni)

Cv(Fe), MP

Py þ Cv, Py þ Dg

in inclusions

Po, Py, Cp, Dg,

Observed contacts

Point analyses of inclusions

Po( Cu)

[Po þ Py]

Po( Ni, Cu), Py,

Po(Cu)

Po, [Po þ Py]

Po(Cu)

Po( Ni, Cu)

MP( Co, Ni, Cu)

[Po þ Py], Po( Ni, Cu),

MP( Ni), MM, MR

Po( Co, Ni, Cu),

[Po þ Py], Tsp

Po( Co, Ni, Cu),

Po( Ni, Cu), MP(Ni)

Po( Cu), MP

Po, MR

Po( Ni, Cu)

Po( Cu)

[Dg þ Cv](Fe)

Po(Cu), MP,

from area scans

Phases deduced from

Po, Py, MR

Po, Py

Py

Py, Po( Cu)

Pseudo-inclusions

Cp, Py, Po

Po( Co, Ni)

Po, Cp

[Py þ Cp]

Po, Py, Cp

Cp

Po( Ni), Py, Cp

Po, Py, Cp

Po( Co)

Po(Ni), Cp, rare Py

Po(Ni)

Matrix sulfide

Po þ Py þ Cp

Po þ Cp

Po þ Py þ Cp

Po þ Py þ Cp

Po þ Py þ Cp

Po þ Cp, Py þ Cp

among matrix sulfide

Contacts observed

in glaucophanites 9908, 9941b, 72811, aluminous eclogite 73101, metasedimentary schist 9931b, mineralized schist NC2. No matrix 9931a, 9939, J40, 73102, 9923 or 9931c. documented in prograde (Alm) and retrograde (Prg, Czo) silicate mineral hosts. Abbreviations as for Table 2, plus Tsp, thiospinel MM, bulk composition close to iss with metal/sulfur ¼ 0·95–1·05 and Fe4Cu.

mineralized schist, Balade mine

NC2

*No prograde inclusion sulfides sulfide in 96312k, 9918, 9930b, Sulfide mineral inclusions were (polydymite–linnaeite–violarite);

metasedimentary schist

9931b

(metasedimentary?)

Felsic

metasedimentary schist

73102

9919a

aluminous glaucophanite

Eclogite

J40

J17

Eclogite

9949

(metasedimentary?)

Glaucophanite

Eclogite

9941b

Eclogite

9939

9946

eclogite boudin in schist

Eclogite

9930b

aluminous glaucophanite

9918

9931a

Alm

aluminous glaucophanite

96312k

Alm

Host

Sample no.

Table 5: Sulfide mineral inclusions, pseudo-inclusions and matrix sulfide in omphacite-zone rocks*

BROWN et al. PROGRADE SULFIDE METAMORPHISM

JOURNAL OF PETROLOGY

VOLUME 55

NUMBER 3

MARCH 2014

Table 6: Representative analyses for omphacite-zone sulfide mineral compositions extrapolated from area scans and from point analyses Extrapolated from area scans (normalized to 100%) for omphacite-zone sulfide minerals Sample:

96312k

96312k

9930b

9941b

9939

J40

J17

9949

9919t

9924a

9924a

Location:

S1

S2

S1

S9

S6

S3

S17

S6

S2

S18

S11

S3

Silicate host:

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

at. % Fe

43·55

24·14

41·81

42·65

38·13

34·36

44·22

44·61

45·38

46·80

38·85

29·14

at. % Co

0·00

0·00

0·00

0·00

0·00

1·38

2·20

0·26

0·00

0·00

0·00

2·69

at. % Ni

0·00

0·88

0·00

0·00

0·00

3·16

2·69

1·32

0·00

0·00

0·00

8·85

at. % Cu

4·47

20·90

3·80

3·06

15·99

7·86

0·00

1·02

0·00

0·00

0·00

0·99

at. % S

51·98

54·09

54·39

54·29

45·88

53·24

50·89

52·79

54·62

53·20

61·15

58·33

Fe/M

0·91

0·53

0·92

0·93

0·70

0·73

0·90

0·94

1·00

1·00

1·00

0·70

Co/M

0·00

0·00

0·00

0·00

0·00

0·03

0·04

0·01

0·00

0·00

0·00

0·06

Ni/M

0·00

0·02

0·00

0·00

0·00

0·07

0·05

0·03

0·00

0·00

0·00

0·21

Cu/M

0·09

0·46

0·08

0·07

0·30

0·17

0·00

0·02

0·00

0·00

0·00

0·02

M/S

0·92

0·85

0·84

0·84

1·18

0·88

0·97

0·89

0·83

0·88

0·64

0·71

Po(Cu)

MP

Po(Cu)

Po(Cu)

MR

Po

Po

Identity

MP(CoNiCu)

Po(CoNi)

Po(NiCu)

73101

Po þ Py

Tsp

Representative point analyses for omphacite-zone sulfide inclusions Sample:

96312k

96312k

96312k

9946

9949

J17

9941b

9919t

9924a

73101

73101

Location:

P4

P5

P10

P5

P8

P1

P4

P4

P12

P4

P7

73101 P8

Silicate host:

Grt

Grt

Grt

Grt

Grt

Grt

Czo

Grt

Grt

Grt

Grt

Grt

wt % Fe

46·11

2·22

1·14

41·92

60·74

59·15

56·26

60·72

60·43

3·08

13·25

60·19

wt % Co

0·32

0·00

0·00

0·23

0·00

0·00

0·82

0·00

0·00

17·76

7·46

0·49

wt % Ni

0·82

0·00

0·00

4·09

0·46

1·32

3·59

0·27

0·68

33·87

28·30

0·80

wt % Cu

0·86

70·03

66·93

0·00

0·00

0·29

0·35

0·00

0·00

2·37

4·21

0·00

wt % S

53·02

29·99

30·95

52·14

39·48

39·72

37·42

39·90

38·75

40·58

36·56

40·11

101·13

102·24

99·01

98·38

100·68

100·48

98·44

100·89

99·86

97·66

89·78

101·59

at. % Fe

32·87

1·91

1·00

30·63

46·75

45·56

44·67

46·54

47·00

2·47

11·56

45·85

at. % Co

0·22

0·00

0·00

0·16

0·00

0·00

0·62

0·00

0·00

13·47

6·17

0·35

at. % Ni

0·56

0·00

0·00

2·84

0·34

0·97

2·71

0·20

0·50

25·80

23·49

0·58

at. % Cu

0·54

53·06

47·34

0·00

0·00

0·20

0·24

0·00

0·00

1·67

3·23

0·00

at. % S

Total

65·82

45·03

51·66

66·36

52·92

53·28

51·75

53·26

52·49

56·59

55·55

53·22

Fe/M

0·96

0·03

0·02

0·91

0·99

0·98

0·93

1·00

0·99

0·06

0·26

0·98

Co/M

0·01

0·00

0·00

0·00

0·00

0·00

0·01

0·00

0·00

0·31

0·14

0·01

Ni/M

0·02

0·00

0·00

0·08

0·01

0·02

0·06

0·00

0·01

0·59

0·53

0·01

Cu/M

0·02

0·97

0·98

0·00

0·00

0·00

0·01

0·00

0·00

0·04

0·07

0·00

M/S

0·52

1·22

0·94

0·51

0·89

0·88

0·93

0·88

0·90

0·77

0·80

0·88

Identity

Cp

Cv(Fe)

Py(Ni)

Po(Ni)

Po(Ni)

Po(Ni)

Tsp

Tsp

Po(CoNi)

[Dg þ Cv] (Fe)

658

Po(CoNiCu)

Po

BROWN et al.

PROGRADE SULFIDE METAMORPHISM

Fig. 8. Compositions of inclusion sulfide for omphacite-zone sample 96312k. Arrows show correspondence between analysis points and the BSE image (right); continuous tie lines show observed contacts.

Fig. 9. Omphacite-zone photomicrographs. (a) BSE image of metabasaltic eclogite 9931a showing pyrrhotite and plagioclase in almandine. (b) Sample J17; BSE image of intergrown pyrrhotite þ chalcopyrite inclusion in almandine. Bright area is Cu-rich. (c) Reflected light photomicrograph of pyrrhotite and chalcopyrite in garnet in sample J17. (d) Reflected light photomicrograph of pyrrhotite þ chalcopyrite in the matrix of J17.

659

JOURNAL OF PETROLOGY

VOLUME 55

NUMBER 3

MARCH 2014

Fig. 10. Compositions of inclusion sulfide from the omphacite zone, excluding atypical samples 96312k, 73101 and NC2. (a) Eclogites 9918, 9931a, 9930b, 9939, 9946 and 9949. (b) Relatively heavily altered glaucophanites 9908, 9941b, 72811 and felsic rock 73102. (c) Metabasalt J17 and metagabbro J40 from near the hornblende zone. (d) Metasediments 9919a, 9919t, 9923, 9924a and 9931c. It should be noted that average compositions are confined below the Py^Dg line (dashed).

above; see Fig. 8), thiospinel-bearing 73101 and the heavily mineralized, but sulfide inclusion free, Balade Mine sample NC2 (discussed below). Cu was commonly detected in apparently homogeneous pyrrhotite inclusions, which may imply intergrowth with an unresolved Cu sulfide, probably chalcopyrite. Such compositions were observed in 9930b, 9941b, 9946 and J17 (Table 6). A metal-rich area analysis in 9939 can be interpreted as pyrrhotite þ chalcopyrite þ bornite/digenite. Pyrrhotite ( chalcopyrite) inclusions in three omphacitezone metamafic samples, 9946, J17 and J40, contain a small amount of Ni (up to 3 at. % in J40); the most Nirich inclusion in J40 also has 1·4 at. % Co. One point analysis from sample 9946 also corresponds to Ni-rich pyrite, and one pyrite analysis with no detectable Ni was obtained from J17 (Table 6). Pyrrhotite inclusions were observed not just in garnet but also in paragonite (9946) and clinozoisite (9941b, 9949). Nickel was detected in inclusions in paragonite and clinozoisite from 9941b (good analyses were not obtained for those from 9949). It was also detected in matrix

pyrrhotite for 9946, but matrix chalcopyrite and pyrrhotite of 9941b were Ni-free, and there was no matrix sulfide in 9949. Therefore, it seems likely that clinozoisite, like garnet, is effective at isolating inclusions from the matrix, whereas paragonite may not be. Definite pyrite and mixed pyrite^pyrrhotite pseudo-inclusions in J17 and 9931a are associated with chloritic alteration in garnet. Matrix sulfide is diverse and abundant in glaucophanites; the phases observed are chalcopyrite (9908), pyrrhotite (9946), pyrite þ chalcopyrite (72811) or all three phases (9941b). Some matrix pyrrhotite was also observed in metabasaltic eclogite 9946, whereas matrix sulfide in J17 is almost exclusively pyrrhotite þ chalcopyrite, except for one pyrite grain that contains patches of chalcopyrite aggregate. No matrix sulfide was seen in other eclogite samples 9918, 9930b, 9931a, 9939, 9949, 73102, and J40.

Omphacite-zone metasedimentary rocks These are well-layered quartz þ albite þ phengite þ almandine schists, usually preserving omphacite (Fitzherbert et al., 2003; Brown, 2007). Metasedimentary rock samples

660

BROWN et al.

PROGRADE SULFIDE METAMORPHISM

from the omphacite zone along the south side of the Pam Peninsula contained abundant sulfide inclusions in garnet. Bulk area scans yielded compositions that trend from close to troilite in composition (M:S ¼1:1) to pyrite (M:S ¼1:2). However, point analyses indicated that pyrrhotite was usually near Fe0·88S in composition (Table 6). The trend towards low M:S ratios probably indicates intergrowth of pyrite and pyrrhotite: some rims of pyrite around pyrrhotite were observed, even when inclusions were not obviously connected to the matrix by fractures. Variation towards high M:S may indicate the occurrence of a range of pyrrhotite compositions, or, when associated with high Co or Ni, intergrowth with a metal-rich phase that concentrates those elements, such as pentlandite. Matrix sulfide in metasedimentary rocks is usually pyrrhotite, and is abundant in some samples (9924a, 9931b, 9919). Chalcopyrite and pyrite are occasionally present. Nickel was variably detectable in matrix pyrrhotite in 9924a. Sample 73101 is interpreted as a strongly foliated eclogite, which may be of sedimentary origin. It is discussed separately because of the large amounts of Co and Ni in the sulfide inclusions. In addition to pyrrhotite and pyrite, which usually contain these elements above the detection limit, there are some inclusions whose compositions correspond to violarite^polydymite^greigite thiospinels (Fe,Co,Ni)3S4 with Ni up to 25·8 wt % and Co up to 13·5 wt % (Table 6). The matrix in this rock contains abundant pyrrhotite, usually with detectable Ni (up to 1·1wt %) and minor Ni-free chalcopyrite, although one ‘MP’ point analysis was both Ni- and Cu-rich, corresponding to (Fe0·48Co0·01Ni0·10Cu0·37)¼0·96S, which may represent a pyrrhotite^pentlandite^chalcopyrite intergrowth.

Balade mine The Balade Mine is located within the Diahot terrane, hosted by the metasedimentary rocks within the omphacite zone. The dominant ore metal is Cu, with lesser Pb and Zn. The host-rocks are variably altered, with abundant chloritoid and chlorite. According to Briggs et al. (1977), the Balade sulfide deposits were sourced from volcanic intrusions, cogenetic with basin sedimentation prior to subduction, and are hosted along stratigraphic horizons within the Diahot terrane. Briggs et al. (1977) stated that although these chalcopyrite-rich deposits did experience subduction, they do not preserve a record of subduction metamorphism. The massive sulfide sample NC2 from Balade Mine has 470% chalcopyrite, with pyrite more common than pyrrhotite. Pyrrhotite is rare, present as small, discrete grains within the chalcopyrite mass. The massive sulfide is interstitial to large quartz crystals. Graphite, where present, is rimmed by magnetite, and is never in direct contact with chalcopyrite. The presence of magnetite, the nearby chloritoid-bearing alteration zone and related normal faulting, and the fact that none of the sulfide is armoured against

any changes in external fluid composition during decompression and exhumation, all imply that these rocks do not record prograde sulfide metamorphism. Sulfide minerals will in general have been re-equilibrated or are the products of retrogression; Rawling & Lister (2002) reported the existence of normal faults nearby that would have facilitated fluid influx and sulfide retrogression. We emphasize that prograde relationships can be examined only where sulfide has been protected from retrogression and fluid influx.

Hornblende-zone sulfide Previous studies have hypothesized a gap in P^T conditions between the Diahot and Poue¤bo terranes, given the faulted contact between them. However, in this study, we assume that both terranes were metamorphosed together to high pressure, as the faulted contact predates subduction metamorphism. This zone includes rocks corresponding to zone 4 of Fitzherbert et al. (2003) within the Diahot terrane, and all of the Poue¤bo terrane. Peak P^T estimates for this zone are approximately 1·7^1·9 GPa, 600^6508C (Carson et al., 1999; Fitzherbert et al., 2003). All rocks collected for this study are from the Diahot terrane, except for 73112 and J38, which are from the Poue¤bo terrane (Fig. 2). The metamafic rocks of this zone are referred to as ‘amphibolite’ in Table 1, as omphacite is very rare and largely confined to inclusions in garnet (Brown, 2007). The omphacite þ garnet þ clinozoisite of unaltered eclogite is represented by a hydrous equivalent, barroisitic hornblende. Sulfide minerals observed in hornblende-zone rocks are summarized in Table 7, and representative analyses shown in Table 8. The distribution of point and area scan analyses are shown in Fig. 11.

Metamafic amphibolites Sulfide inclusions occur in almandine. In contrast to the omphacite-zone rocks, inclusions in garnet of the hornblende zone all contain pyrite as the dominant Fe-sulfide, except for 73112 (Tables 7 and 8, Fig. 11). Inclusions are typically single-phase pyrite. The Poue¤bo terrane samples show distinct differences from those of the Diahot terrane. Sample J38 point analyses indicate chalcopyrite, bornite and digenite inclusions, and contacts between pyrite and bornite were observed (Tables 7 and 8). Polyphase inclusions in sample 73112 gave analyses corresponding to pyrrhotite, chalcopyrite and MR intergrowths but not pyrite. One pyrrhotite analysis showed Ni above detection limit. Area analyses were consistent with expectations from the point analyses, corresponding to pyrite and bornite compositions (J38), or Fe^Cu sulfide intergrowths with (Fe þ Cu):S greater than, equal to or less than 1:1 (J38 and 73112). Matrix sulfide minerals observed are pyrite in J22 and J37 and chalcopyrite in E803. Samples 72908, 72909 and 72915,

661

JOURNAL OF PETROLOGY

VOLUME 55

NUMBER 3

MARCH 2014

Table 7: Sulfide mineral inclusions in prograde porphyroblasts and matrix sulfide in hornblende-zone rocks* Sample no.

Host

Point analyses

Observed

Phases deduced

of inclusions

contacts in

from area scans

Matrix sulfide

among matrix sulfide

inclusions

J22

glaucophanite

Alm

Py(Ni), Cp

J37

amphibolite

Alm

Py

J38

amphibolite

Alm

Py( Cu), Cp,

Contacts observed

minerals

Py(Cu), MP

Py Py

Py þ Bn

MP, MR, MN, Bn,

Bn, Dg(Fe)

Py( Cu), [Cv þ Bn þ Dg]

73112

amphibolite

Alm

Po( Ni), Cp, MR

E803

glaucophanite

Alm

Py

MP, MM, MR

J37a

metasedimentary schist

Alm

Py( Co, Ni), Po

J43

metasedimentary schist

Alm

72908

amphibolite

Py, Po(Ni þ Co),Cp

Py þ Po

72909

glaucophanite

Py( Co), Po, Cp,

Py þ Po þ Cp

72915

amphibolite

Po( Co, Ni), Py( Co,

Cp MP, MR

Py

MN

Cp

MP, Bn, Dg, Cv( Fe) Py þ Po þ Cp þ Ttd

Ni, Cu), Cp, Bn, Ttd

*No sulfide inclusions in amphibolite 72908, glaucophanite 72909. No matrix sulfide in amphibolites J38, 73112, 72915. MN, bulk composition close to iss with metal/sulfur ¼ 0·95–1·05 and Fe 5 Cu. Ttd, tetradymite.

(1) Sulfide inclusion assemblages and the overall distribution of inclusion compositions are significantly different from those of matrix sulfide. (2) Inclusion compositions tend to be richer in Cu than matrix sulfides. This may imply that Fe has been lost to a growing silicate phase (garnet) during prograde metamorphism. In evidence, we note that it is possible to write balanced reactions such as

which did not yield any sulfide inclusions, have rather diverse matrix sulfide assemblages including pyrite and pyrrhotite (both of which sometimes contain Ni and/or Co), bornite, covellite, digenite, chalcopyrite and, in 72915, tetradymite (Bi2Te2S), which is intergrown with Py, Po and Cp.

Metasedimentary rocks Samples J37a and J43 are glaucophane-rich metasedimentary schists collected along the boundary between the Diahot and Poue¤bo terranes (Fig. 2). The only garnet inclusions large enough for point analysis were found in sample J37a and are again of pyrite, sometimes with small amounts of Co or Ni (50·4 wt %). Area scans were also obtained of chalcopyrite or cubanite composition in J43, but departing in either direction from 1:1 cation:sulfur ratio in J37a, implying the presence of pyrrhotite þ chalcopyrite þ pyrite and/or a metal-rich sulfide such as bornite (Fig. 11d). Only pyrite forms grains large enough for single-phase point analysis. Matrix sulfide is pyrite in J37a, and chalcopyrite in J43 (Table 7).

DISCUSSION In general, two major features show that sulfide inclusions were captured by porphyroblasts as the slab subducted, as follows.

25CuFeS2 þ 8Ca2 Al3 Si3 O12 ðOHÞ þ 12SiO2 þ14O2 ! Cp Czo Qtz  5Cu5 FeS4 þ 12 Ca4=3 Fe5=3 Al2 Si3 O12 þ15S2 þ 4H2 O: Bn

Grt ð1Þ

Such sulfide^silicate reactions are analogues of those discussed for very different bulk compositions and P^Tconditions by Thompson (1972) and Tracy & Robinson (1988).

Lawsonite zone Tables 2, 5 and 7 show that whereas chalcopyrite frequently occurs as a matrix phase in all zones, the Cu-rich phases covellite, bornite and digenite do not occur as matrix phases. The latter minerals were observed in the matrix in only two hornblende-zone rocks from one locality, which also show other unusual mineralization. Many inclusions cluster around chalcopyrite-like compositions in metabasalt J15a and spessartine-bearing metasediments J16b and c. The scatter of analyses in Fig. 4 is

662

BROWN et al.

PROGRADE SULFIDE METAMORPHISM

Table 8: Representative analyses from area scans and sulfide inclusions for hornblende-zone sulfides Representative analyses extrapolated from area scans (normalized to 100%) for hornblende-zone sulfide Sample:

J22

J22

J38

J38

J38

73112

73112

73112

J37a

J37a

J37a

Location:

S2

S3

S21

S17

S8

S4

S1

S2

S1

S2

S3

J43 S1

Silicate host:

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

at. % Fe

32·85

25·97

31·73

12·64

14·63

28·90

31·89

31·29

34·70

32·78

37·14

14·86

at. % Co

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

at. % Ni

0·00

0·00

0·00

0·00

0·00

5·15

0·00

0·00

0·00

0·00

0·00

0·00

at. % Cu

1·58

22·99

2·71

47·85

34·40

15·23

12·09

20·48

17·97

4·18

10·03

35·50

at. % S

65·57

51·04

65·56

39·51

50·97

50·72

56·02

48·24

47·33

63·03

52·84

49·64

Fe/M

0·95

0·53

0·92

0·21

0·30

0·59

0·73

0·60

0·66

0·89

0·79

0·30

Co/M

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

Ni/M

0·00

0·00

0·00

0·00

0·00

0·10

0·00

0·00

0·00

0·00

0·00

0·00

Cu/M

0·05

0·47

0·08

0·79

0·70

0·31

0·27

0·40

0·34

0·11

0·21

0·70

M/S

0·53

0·96

0·53

1·53

0·96

0·97

0·79

1·07

1·11

0·59

0·89

1·01

Identity

Py(Cu)

MP

Py(Cu)

Bn

MM

MM

MP

MR

MR

MP

MP

MM

J37a

Representative point analyses for hornblende-zone sulfide inclusions Sample:

J22

J22

E803

J38

J38

J38

J38

J38

73112

73112

J37a

Location:

P1

P2

P1

P17

P16

P9

P12

P14

P2

P1

P1

P2

Silicate host:

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

Grt

wt % Fe

45·55

30·53

45·89

46·18

46·45

13·17

13·68

3·84

60·38

31·11

46·72

46·42

wt % Co

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·33

0·00

wt % Ni

0·36

0·00

0·23

0·00

0·00

0·00

0·00

0·00

0·39

0·31

0·00

0·32

wt % Cu

0·26

34·13

0·21

0·00

0·68

60·93

60·30

74·52

0·00

34·29

0·00

0·35

wt % S

53·14

34·07

52·46

53·61

52·79

25·68

25·15

21·53

37·39

34·24

53·46

53·19

Total

99·31

98·73

98·79

99·79

99·92

99·77

99·13

99·89

98·16

99·95

100·51

100·28

at. % Fe

32·85

25·47

33·34

33·08

33·41

11·82

12·38

3·60

47·97

25·67

33·34

33·24

at. % Co

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·22

0·00

at. % Ni

0·25

0·00

0·16

0·00

0·00

0·00

0·00

0·00

0·29

0·24

0·00

0·22

at. % Cu

0·16

25·02

0·13

0·00

0·43

48·05

47·97

61·30

0·00

24·87

0·00

0·22

at. % S

66·74

49·51

66·37

66·92

66·15

40·13

39·65

35·10

51·74

49·21

66·44

66·33

Fe/M

0·99

0·50

0·99

1·00

0·99

0·20

0·21

0·06

0·99

0·51

0·99

0·99

Co/M

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·01

0·00

Ni/M

0·01

0·00

0·00

0·00

0·00

0·00

0·00

0·00

0·01

0·00

0·00

0·01

Cu/M

0·00

0·50

0·98

0·00

0·01

0·80

0·79

0·94

0·00

0·49

0·00

0·01

M/S

0·50

1·02

0·94

0·49

0·51

1·49

1·52

1·85

0·93

1·03

0·51

0·51

Py(Ni)

Cp

Cv(Fe)

Py

Py(Cu)

Bn

Bn

Dg(Fe)

Po(Ni)

Cp

Py(Co)

Py(Ni)

Identity

dispersed over a range of Cu:Fe and metal:sulfur ratios around the chalcopyrite composition and does not concentrate along trend lines to pyrrhotite or pyrite. We interpret this scatter to represent multiphase intergrowths derived by breakdown of original single-phase intermediate solid solution. In contrast, sulfide inclusions in almandine-

bearing metasediment J15 are dominated by pyrite, except for one area analysis that is interpreted as Fe-contaminated covellite (maybe covellite þ nukundamite) and one point analysis that is close to bornite. The metabasalt J15a contains covellite as the most Cu-rich inclusion sulfide. Inclusions in J15a show Py þ Cp and Py þ Cv contacts

663

JOURNAL OF PETROLOGY

VOLUME 55

NUMBER 3

MARCH 2014

Fig. 11. Compositions of inclusion sulfide from the hornblende zone. (a) Diahot terrane metamafic rocks J22, J37 and E803. Dashed lines enclose bulk composition field that could correspond to original pyrite þ iss. (b) Poue¤bo terrane metamafic rock J38. Inset photomicrograph shows coexisting pyrite and bornite. Inclusion at highest cation/sulfur ratio is probably an oxidized bornite pseudo-inclusion. (c) Pue¤bo terrane metamafic rock 73112. Inset photomicrograph shows pyrrhotite surrounding chalcopyrite with mottled texture in backscattered electron contrast, implying multiphase intergrowth at submicron scale owing to iss breakdown. (d) Diahot terrane metasedimentary samples J37a and J43. It should be noted that average compositions are confined below the Py^Dg line (dashed).

(Fig. 4). However, the stability of a Py þ Bn tie line and of nukundamite over the temperature range from ambient to the peak of metamorphism would imply that Cp þ Cv are not mutually stable at atmospheric pressure (Fig. 6), and that the current assemblage in the inclusion of Fig. 5b is a disequilibrium one, preserved by rapid cooling or uplift. The assemblage may have been stable at peak pressure, or result from the breakdown of a different peak assemblage. Mutual stability of Py þ Cp þ Cv would require the following vapour-absent reactions to be favoured with increasing P: destabilization of nukundamite, Nk ! Py þ Bn þ Cu338 Fe062 S4 ! 0  4133FeS2 þ 0  2067Cu5 FeS4 þ

:

Cv 2  3467CuS ð2Þ

and Py 2FeS2

þ þ

Bn Cu5 FeS4

! !

Cp 2CuFeS2

þ þ

Cv : ð3Þ 3CuS

The molar volume data in Table 9 show that V for reaction (2) is 5·20 cm3 mol1, a 6·2% decrease. Therefore, we propose that nukundamite is strongly destabilized relative to the dense Py þ Bn assemblage with increasing pressure, and is unlikely to be a stable phase under subduction-zone conditions. However, the V for reaction (3) is þ2·12 cm3 mol1, a 1·4% increase, so Cp þ Cv, already unstable relative to Py þ Bn at 1atm, will not become any more stable at depth. The shift of bornite to more Cu-rich compositions discussed by Seal et al. (2001) will act to enhance the stoichiometric ratio of bornite relative to Nk/Cv in the reactions, increasing the magnitude of these volume changes. Although high-P crystallographic data for covellite have been collected at 30 kbar by

664

BROWN et al.

PROGRADE SULFIDE METAMORPHISM

Table 9: Unit cell and molar volume data for some Cu^Fe^S phases Phase

Formula unit

Hf (kJ mol1)

S (J mol1 K1)

Crystal

a (A˚)

b (A˚)

c (A˚)

Z

Molar volume

Reference

(cm3 mol1)

system

Bornite

Cu5FeS4

–371·6

398·5

Orth.

10·95

16

98·66

1

Chalcopyrite

CuFeS2

–194·9

116·2

Tetr.

5·281

10·401

4

43·67

2

Covellite

CuS





Hex.

3·7938

16·341

6

20·44

3

Nukundamite

Cu3·38Fe0·62S4





Trig.

3·782

11·187

1

83·45

4

171·5

52·9

Cubic

5·4179

4

23·94

5

97·5

60·7

Hex.

6·8673

17·062

24

17·48

6

102·6

60·3

Trig.

3·452

5·762

2

17·90

7

Pyrite

FeS2

Pyrrhotite

Fe0·875S

Troilite

FeS

10·95

21·862

Standard enthalpies of formation and entropies are from Robie & Hemingway (1995); unit cell and molar volume references are as indicated. References: 1, Koto & Morimoto (1975); 2, Hall & Stewart (1973); 3, Evans & Konnert (1976); 4, Sugaki et al. (1981); 5, Brostigen & Kjekshus (1969); 6, Fleet (1971); 7, Ska´la et al. (2006).

Take¤uchi et al. (1985) and Peiris et al. (1995), the phase may have persisted only because of slow decomposition kinetics at low temperature. It should be noted that that very (Cu,S)-rich overall inclusion compositions dominated by Cv  Py, such as analysis J15-a1 in Table 3 and illustrated in Fig. 4 for J15, cannot be achieved using any of the Cu-rich sulfide phases considered in Figs 4 and 6, other than covellite. Once a Cv þ Py inclusion becomes a closed system, its S-rich bulk composition must include either (1) covellite, (2) digenite (or bornite) þ S2, to which covellite decomposes at low P and high T (5078C at 1atm: Kullerud, 1965; Roseboom, 1966; Fleet, 2006), or (3) villamaninite, if P becomes high enough for this phase to be stable. If both covellite and villamaninite are unstable at the time of enclosure, such (Cu,S)-rich bulk compositions cannot be represented by all-solid assemblages and hence cannot be incorporated as inclusions in growing silicates. The pseudoinclusion identified optically as ‘digenite’ in sample J15a (Fig. 5a) has an unusually Cu-rich composition, close to djurleite/chalcocite; it is likely to have formed by additional Fe loss owing to interaction with metamorphic fluids. Matrix sulfide in the lawsonite zone does not include the Cu-rich phases seen in inclusions, but is dominated by pyrite with some chalcopyrite. The predominance of pyrite and absence of pyrrhotite in the matrix is distinct from the inclusion sulfide suite, implying a lower cation/S ratio and higher temperature-adjusted fS2 in the matrix fluids in comparison with any fluid present at the time of inclusion entrapment. Matrix pyrite in J16 metasediments also contains inclusions of galena, implying that the matrix experienced an input of Pb from which the inclusions were isolated. We infer that matrix and inclusion sulfide were essentially different systems during

metamorphism, and experienced different changes in their bulk chemistry.

Epidote zone Pyrite and chalcopyrite are ubiquitous as matrix sulfide minerals in the epidote-zone metasediments, as they are in the lawsonite zone, again implying the presence of late fluids with low Fe/S ratios. Inclusions in quartz are of the same species, consistent with the hypothesis that these inclusions were vulnerable to re-equilibration with the matrix. However, the observation of Po þ Cp in albite implies that an earlier sulfide suite with a lower cation/ sulfur ratio has been preserved here, as in the lawsonite zone. The matrix of glaucophanite 73004a also contains pyrrhotite and pyrite, recording a different fluid composition from that of the metasediments. Matrix sulfide minerals in samples 73004a and 23956 show detectable Ni ( Co), potentially derived from the breakdown of peridotite olivine under conditions where it was not incorporated into phyllosilicates such as serpentine minerals or talc.

Omphacite zone A major difference between the omphacite zone and the lower-grade rocks is that pyrrhotite rather than pyrite becomes the dominant Fe sulfide in the matrix, irrespective of host-rock lithology. In the matrix, pyrrhotite is associated with pyrite and chalcopyrite, but not with the more diverse range of sulfides observed in inclusions, discussed below. Garnet served to isolate inclusions from matrix fluids during retrogression, as in the lower-grade rocks, but the matrix sulfides themselves are different from those of the lawsonite and epidote zones, owing to either differing fluid compositions at the time of crystallization (lower fS2) or a greater degree of subsequent sulfur loss.

665

JOURNAL OF PETROLOGY

VOLUME 55

Pyrrhotite is also by far the most abundant inclusion mineral, observed as both an inclusion and a matrix phase in all lithologies. Nickel ( Co) was detected much more frequently than in the lower grades in both inclusion and matrix pyrrhotite, suggesting a greater input of Ni to the system, presumably from fluids derived from serpentinized mantle olivine. A broad range of metal-rich ‘MR’ (with M/S 1·05; probably containing digenite or bornite), medium-metal ‘MM’ (with 0·955M/S51·05 and Fe4Cu; probable breakdown products from intermediate solid solution) and metal-poor ‘MP’ (with M/S  0·95; probably containing pyrite) analyses imply frequent intergrowth with other phases in both inclusion and matrix sulfide in metamafic rocks; this variety was not observed in metasediment sulfide. Metamafic sample 96312k contains pyrrhotite, pyrite, chalcopyrite and the cation-rich Cu sulfide minerals digenite and bornite. The only sulfide minerals clearly identified in other metamafic samples are chalcopyrite and (in 73101 only) thiospinels of the polydymite^ violarite^greigite series (Ni3S4^NiFe2S4^Fe3S4). However, the wide spread of Cu and Ni contents and cation:sulfur ratios in the analyses may also imply intergrowth with cation-rich sulfide minerals such as pentlandite and bornite or digenite: the latter were observed in 96312k. Inclusions in sample 96312k show contacts between Py þ Dg and Py þ Cv (Fig. 8). The latter association implies that a (Cu,S)-rich solid phase, either covellite or villamaninite, was stable at the time of incorporation. It is noteworthy that this sample, whose peak temperature is near the 1atm thermal stability limit for covellite, is the highest-grade rock in which covellite was observed in inclusions. It should be noted that the overall envelope of analysis points in Fig. 10 lies below the line from Py to Bn and/or Dg. This suggests that (Cu,S)-rich phases were not stable, either because fluid fS2 was intrinsically too low for them, or because the pressure at the time of garnet growth was too high for covellite/nukundamite stability and too low for villamaninite, so pyrite þ bornite/digenite was the most S-rich sulfide assemblage possible. The fresh eclogites of Fig. 10a and more altered rocks of Fig. 10b show a similar pattern in which the vast majority of inclusions are close to pyrrhotite in composition, with subordinate pyrite and chalcopyrite, as discussed above. There is no cluster of analyses around chalcopyrite/iss compositions, as seen in the lawsonite zone. However, the possibly higher-T metamafic rocks of Fig. 10c do show a more even scatter of analyses across the Py^Po^Cp triangle, suggesting that these inclusions tend to be more Cu-rich. Conversely, the metasediment inclusions of Fig. 10d are dominated by Py þ Po and almost devoid of Cu. Temperature estimates for the omphacite zone are higher than the maximum temperature of stability of ordered tetragonal chalcopyrite at 1atm (Kullerud, 1965).

NUMBER 3

MARCH 2014

At T45478C, 1atm, the disordered isometric intermediate solid solution (iss) extends far enough to include the composition of chalcopyrite. Because this transformation is order^disorder driven and has a large S/V, the temperature is likely to be similar at high P, in which case any chalcopyrite observed in inclusions of this zone is a breakdown product of a variable-composition iss phase. The broad scatter of analyses from Py þ Po to Cp compositions in Fig. 10c, with no extension towards Bn, suggests that the most Cu-rich phase at peak conditions in these rocks was the intermediate solid solution.

Hornblende zone Rocks of the hornblende zone differ again in that the dominant inclusion sulfide is pyrite rather than pyrrhotite. Given the positive slope of the Py^Po equilibrium in fS2^ T space (Toulmin & Barton, 1964), this is most probably a consequence of much higher fS2 in the associated fluids relative to the omphacite zone, although differences in bulk-rock composition may also be a factor. Small amounts of Co and Ni were sometimes detected in both pyrite and pyrrhotite, but not as frequently as in the omphacite zone. Bulk inclusion compositions again indicate crystallization of prograde sulfide from different fluids in different suites of rocks (Fig. 11): the analyses of sulfide contained in garnet from the Diahot terrane metamafics and Poue¤bo terrane metamafic 73112 lie in the Py^Po^iss triangle, with a tendency to higher Cu content for the inclusions in 73112. The tendency to Cu-rich compositions is even more marked for the Diahot terrane metasediments and Poue¤bo terrane amphibolite J38, where compositions tend towards bornite or the pyrite^digenite line. There is no obvious difference between matrix and inclusion populations for the Diahot terrane samples (Fig. 11a and d), which is inconsistent with the hypothesis of copper enrichment of inclusions owing to garnet growth. The matrix sulfide minerals Cv, Bn and Dg (Table 7; Poue¤bo terrane samples 72909, 72915) presumably reflect the presence of a late, Cu-rich fluid during retrogression. Enrichment of the fluid in other chalcophile elemenys is suggested by the presence of tetradymite (Bi2Te2S) in the matrix. It should be noted that covellite probably requires P considerably below conditions of peak metamorphism for stability.

Cu sulfide phase equilibria at high pressure Although thermodynamically consistent data are available for pyrite, pyrrhotite^troilite and major S-bearing fluid phases (Evans et al., 2010), this is not the case for Cu-bearing sulfides, so phase equilibria cannot be calculated reliably. As discussed above, molar volume data (Table 9) suggest that covellite and nukundamite will destabilize by volatile-absent reactions with increasing P, making it impossible to produce sulfide inclusions that are rich in both Cu and S, in line with our observations. Such compositions

666

BROWN et al.

PROGRADE SULFIDE METAMORPHISM

will again be accessible if villamaninite becomes stable at higher P. The role of chalcopyrite/iss and its significance under subducting slab conditions remains in question. From the data in Table 9, it is possible to estimate approximate conditions for the volatile-absent breakdown of Cp: Cp 5CuFeS2

! Py ! 2FeS2

þ þ

Tr þ 2FeS þ

Bn Cu5 FeS4

ð4Þ

or, with a Po composition (Fe0·875S ¼ Fe7S8) that is richer in S: Cp ! Py þ Po þ Bn : 5CuFeS2 ! 1  667FeS2 þ 2  667FeS þ Cu5 FeS4 ð5Þ Equations (4) and (5) have, respectively, H ¼ þ54·7 and 56·98 kJ mol1, S ¼ þ43·9 and 67·57 kJ mol1 K1, and V ¼ 36·00 and 33·16 cm3 mol1 (data of Robie & Hemingway, 1995). Therefore, Cp would be expected to decompose by equation (4) at P ¼11·4 kbar at 258C, 6·9 kbar at 4008C, or by equation (5) at 11·0 kbar (258C) to 3·4 kbar (4008C). Cation-disordered, higher-entropy iss would persist to slightly higher P and T, but it is possible that Cp/iss would be replaced by Py þ Po þ Bn/Dg under blueschist- or eclogite-facies conditions. Although the clustering of analyses in Figs 4, 10c and 11c suggests that a chalcopyrite/iss phase was incorporated in the inclusions, the breakdown of such a phase may explain the absence of such clustering in Fig. 11b and d.

CONC LUSIONS We have surveyed the sulfide minerals present in a suite of high-pressure rocks from New Caledonia that experienced metamorphic conditions across a wide range of P^T space. Sulfide minerals included in porphyroblasts such as garnet, lawsonite and clinozoisite differ from matrix sulfide, demonstrating that enclosure can protect the inclusions from reacting with fluids during retrogression. The composition ranges of sulfide inclusions vary with both metamorphic grade and bulk composition of the hostrock. We interpret the inclusion compositions to correspond to solid sulfide assemblages that were stable at the time of enclosure. At atmospheric pressure, the range of possible compositions extends to high contents of Cu and S, given the stability of pyrite þ nukundamite/covellite. However, those Cu sulfide minerals are low-density phases and are expected to destabilize at moderate pressures; overall inclusion compositions would then be poorer than the Py^Dg line. This pattern was observed in the rocks we studied, as no covellite was observed in sulfide inclusions for peak P conditions above the lower omphacite zone. We predict that (Cu,S)-rich inclusion compositions might return at pressures higher than those of this study, if the Cu-rich pyrite mineral villamaninite becomes

stable. Furthermore, there is some evidence that at high P and T inclusion compositions may cease to cluster near Fe:Cu ¼1:1 owing to the replacement of Cp/iss by the denser Py þ Po þ Bn assemblage.

AC K N O W L E D G E M E N T S C. Spandler, J. Fitzherbert, G. Clarke and P. Black provided samples that greatly helped to complete this study. We thank K. Evans, T. Kawakami, A. Tomkins and an anonymous reviewer for their incisive and thorough reviews. J.L.B. acknowledges the support of an Australian International Postgraduate Research Scholarship and an Australian Postgraduate Award from the Australian National University.

R EF ER ENC ES Aitchison, J. C., Clarke, G. L., Meffre, S. & Cluzel, D. (1995). Eocene arc^continent collision in New Caledonia and implications for regional southwest Pacific tectonic evolution. Geology 23, 161^164. Aitchison, J. C., Ireland, T. R., Clarke, G. L., Cluzel, D., Davis, A. M. & Meffre, S. (1998). Regional implications of U/Pb SHRIMP age constraints on the tectonic evolution of New Caledonia. Tectonophysics 299, 333^343. Barton, P. B. (1973). Solid-solutions in system Cu^Fe^S Part 1. Cu^S and CuFe^S joins. Economic Geology 68, 455^465. Barton, P. B. (1974). Sulfide petrology. In: Ribbe, P. H. (ed.) Sulfide Mineralogy. Mineralogical Society of America, Reviews in Mineralogy 1, B1^B11. Bayliss, P. (1989). Crystal chemistry and crystallography of some minerals within the pyrite group. American Mineralogist 74, 1168^1176. Black, P. M. (1977). Regional high-pressure metamorphism in New Caledoniaçphase equilibria in Ouegoa district. Tectonophysics 43, 89^107. Black, P. M. & Brothers, R. N. (1977). Blueschist ophiolites in melange zone, northern New Caledonia. Contributions to Mineralogy and Petrology 65, 69^78. Briggs, R. M., Kobe, H. W. & Black, P. M. (1977). High-pressure metamorphism of stratiform sulfide deposits from Diahot Region, New Caledonia. Mineralium Deposita 12, 263^279. Brostigen, G. & Kjekshus, A. (1969). Redetermined crystal structure of FeS2 (pyrite). Acta Chemica Scandinavica 23, 2186^2188. Brothers, R. N. (1974). High-pressure schists in northern New Caledonia. Contributions to Mineralogy and Petrology 46, 109^127. Brothers, R. N. & Yokoyama, K. (1982). Comparison of the high-pressure schist belts of New Caledonia and Sanbagawa, Japan. Contributions to Mineralogy and Petrology 79, 219^229. Brown, J. L. (2007). The deep sulfur cycle: Insights from prograde sulfide metamorphism in blueschist and eclogite, northeastern New Caledonia, PhD thesis, Australian National University, Canberra. Cabri, L. J. (1973). New data on phase relations in Cu^Fe^S system. Economic Geology 68, 443^454. Cabri, L. J. & Hall, S. R. (1972). Mooihoekite and haycockite two new copper^iron sulfides and their relationship to chalcopyrite and talnakhite. American Mineralogist 57, 689^708. Carson, C. J., Powell, R. & Clarke, G. L. (1999). Calculated mineral equilibria for eclogites in CaO^Na2O^FeO^MgO^Al2O3^SiO2^ H2O: application to the Poue¤bo Terrane, Pam Peninsula, New Caledonia. Journal of Metamorphic Geology 17, 9^24.

667

JOURNAL OF PETROLOGY

VOLUME 55

Carson, C. J., Clarke, G. L. & Powell, R. (2000). Hydration of eclogite, Pam Peninsula, New Caledonia. Journal of Metamorphic Geology 18, 79^90. Caye, R., Cervelle, B., Cesbron, F., Oudin, E., Picot, P. & Pillard, F. (1988). Isocubanite, a new definition of the cubic polymorph of cubanite CuFe2S3. Mineralogical Magazine 52, 509^514. Clarke, G. L., Aitchison, J. C. & Cluzel, D. (1997). Eclogites and blueschists of the Pam Peninsula, NE New Caledonia: a reappraisal. Journal of Petrology 38, 843^876. Cluzel, D., Aitchison, J. C. & Picard, C. (2001). Tectonic accretion and underplating of mafic terranes in the Late Eocene intraoceanic fore-arc of New Caledonia (Southwest Pacific): geodynamic implications. Tectonophysics 340, 23^59. Cluzel, D., Aitchison, J., Clarke, G., Meffre, S. & Picard, C. (1994). The tectonic and geodynamic evolution of New-Caledonia - a view. Comptes Rendus de l’academie des sciences serie II 319, 683^690. Constantinou, G. (1975). Idaite from the Skouriotissa massive sulfide orebody, Cyprus: its composition and conditions of formation. American Mineralogist 60, 1013^1018. Craig, J. R. & Scott, S. D. (1974). Sulfide Phase Equilibria. Mineralogical Society of America, Reviews in Mineralogy 1. Evans, H. T., Jr & Konnert, J. A. (1976). Crystal structure refinement of covellite. American Mineralogist 61, 996^1000. Evans, K. A., Powell, R. & Holland, T. J. B. (2010). Internally consistent data for sulfur-bearing phases and application to the construction of pseudosections for mafic greenschist facies rocks in Na2O^CaO^K2O^FeO^MgO^Al2O3^SiO2^CO2^O^S^H2O. Journal of Metamorphic Geology 28, 667^687. Fitzherbert, J. A., Clarke, G. L. & Powell, R. (2003). Lawsonite^ omphacite-bearing metabasites of the Pam Peninsula, NE New Caledonia: Evidence for disrupted blueschist- to eclogite-facies conditions. Journal of Petrology 44, 1805^1831. Fitzherbert, J. A., Clarke, G. L. & Powell, R. (2005). Preferential retrogression of high-P metasediments and the preservation of blueschist to eclogite facies metabasite during exhumation, Diahot terrane, NE New Caledonia. Lithos 83, 67^96. Fleet, M. E. (1971). The crystal structure of a pyrrhotite (Fe7S8). Acta Crystallographica B27, 1864^1867. Fleet, M. E. (2006). Phase equilibria at high temperatures. In: Vaughan, D. J. (ed.) Sulfide Mineralogy and Geochemistry. Mineralogical Society of America and Geochemical Society, Reviews in Mineralogy and Geochemistry 61, 365^419. Frenzel, G. (1959). Ein neues Mineral: Idait, natu«rliches Cu5FeS6. NeuesJahrbuch fu«r Mineralogie, Abhandlungen 93, 87^114. Frost, B. R. (1991). Stability of oxide minerals in metamorphic rocks. In: Lindsley, D. H. (ed.) Oxide Minerals. Mineralogical Society of America. Reviews in Mineralogy 25, 469^487. Grguric, B. A., Harrison, R. J. & Putnis, A. (2000). A revised phase diagram for the bornite^digenite join from in situ neutron diffraction and DSC experiments. Mineralogical Magazine 64, 213^231. Hall, S. R. & Stewart, J. M. (1973). The crystal structure refinement of chalcopyrite, CuFeS2. Acta Crystallographica 29, 579^585. Inan, E. E. & Einaudi, M. T. (2002). Nukundamite (Cu3·38Fe0·62S4)bearing copper ore in the Bingham porphyry deposit, Utah: Result of upflow through quartzite. Economic Geology 97, 499^515. Itaya, T., Brothers, R. N. & Black, P. M. (1985). Sulfides, oxides and sphene in high-pressure schists from New Caledonia. Contributions to Mineralogy and Petrology 91, 151^162. Kawakami, T., Ellis, D. J. & Christy, A. G. (2006). Sulfide evolution in high-temperature to ultrahigh-temperature metamorphic rocks from Lutzow^Holm Complex, East Antarctica. Lithos 92, 431^446.

NUMBER 3

MARCH 2014

Koto, K. & Morimoto, N. (1975). Superstructure investigation of bornite, Cu5FeS4, by the modified partial Patterson function. Acta Crystallographica B 31, 2268^2273. Kullerud, G. (1965). Covellite stability relations in the Cu^S system. Freiberger Forschungshefte B 186C, 145^160. Kullerud, G. (1967). Sulfide studies. In: Abelson, P. H. (ed.) Researches in Geochemistry 2. John Wiley, pp. 268^321. Kullerud, G. & Yoder, H. S. (1959). Pyrite stability relations in the Fe^S system. Economic Geology Monograph 54, 533^572. Lennie, A. R. & Vaughan, D. J. (1996). Spectroscopic studies of iron sulfide formation and phase relations at low temperatures. In: Dyar, M. D., McCammon, C. & Schaeffer, M. W. (eds) Mineral Spectroscopy: A Tribute to Roger G. Burns. Special Publication of the Geochemical Society 5, 117^131. Lillie, A. R. (1975). Structures in lawsonite^glaucophane schists of New Caledonia. Geological Magazine 112, 225^240. Liou, J. G. (1971). P^T stabilities of laumontite, wairakite, lawsonite, and related minerals in system CaAl2Si2O8^SiO2^H2O. Journal of Petrology 12, 379^411. Marmo, B. A., Clarke, G. L. & Powell, R. (2002). Fractionation of bulk rock composition due to porphyroblast growth: effects on eclogite facies mineral equilibria, Pam Peninsula, New Caledonia. Journal of Metamorphic Geology 20, 151^165. Maurizot, P., Eberle, J. M., Habault, C. & Tessarollo, C. (1989). Carte ge¤ologique a' l’e¤chelle du 1/50 000, feuillet Pam^Ouegoa, Map sheet and explanatory notes. Paris: Bureau des Recherches Ge¤ologiques et Minie'res. Misra, K. C. & Fleet, M. E. (1973). The chemical compositions of synthetic and natural pentlandite assemblages. Economic Geology 68, 518^539. Munson, R. A. (1966). The synthesis of copper disulfide. Inorganic Chemistry 5, 1296^1297. Paris, J. F. (1981). Ge¤ ologie de la Nouvelle Cale¤ donie. Bureau des Recherches Ge¤ ologiques et Minie' res, Me¤ moire 113, 279 p. Peiris, S. M., Sweeney, J. S., Campbell, A. J. & Heinz, D. L. (1995). Pressure-induced amorphization of covellite, CuS. Journal of Chemical Physics 104, 11^16. Potter, R. & Evans, H. (1976). Definitive X-ray powder data for covellite, anilite, djurleite and chalcocite. Journal of Research of the US Geological Survey 4, 205^212. Rawling, T. J. & Lister, G. S. (1999). Oscillating models of orogeny in the Southwest Pacific and the tectonic evolution of New Caledonia. In: Ring, U., Brandon, M. T., Lister, G. S. & Willett, S. D. (eds) Exhumation Processes: Normal Faulting, Ductile Flow and Erosion. Geological Society, London, Special Publications 154, 109^127. Rawling, T. J. & Lister, G. S. (2002). Large-scale structure of the eclogite^blueschist belt of New Caledonia. Journal of Structural Geology 24, 1239^1258. Rice, C. M., Atkin, D., Bowles, J. F. W. & Criddle, A. J. (1979). Nukundamite, a new mineral and idaite. Mineralogical Magazine 43, 193^200. Robie, R. A. & Hemingway, B. S. (1995). Thermodynamic properties of minerals and related substances at 298·15 K and 1bar pressure and at higher temperature. US Geological Survey Bulletin 2131, 461 p. Roseboom, E. H. (1966). An investigation of system Cu^S and some natural copper sulfides between 258 and 7008C. Economic Geology 61, 641^672. Schmidt, M. W. & Poli, S. (1998). Experimentally based water budgets for dehydrating slabs and consequences for arc magma generation. Earth and Planetary Science Letters 163, 361^379. Sdrolias, M., Muller, R. D., Mauffret, A. & Bernadel, G. (2004). Enigmatic formation of the Norfolk Basin, SW Pacific: a plume

668

BROWN et al.

PROGRADE SULFIDE METAMORPHISM

influence on back-arc extension. Geochemistry, Geophysics, Geosystems 5, article Q06005. Seal, R. R., Inan, E. E. & Hemingway, B. S. (2001). The Gibbs free energy of nukundamite (Cu3·38Fe0·62S4): a correction and implications for phase equilibria. Canadian Mineralogist 39, 1635^1640. Ska¤la, R., C|¤ sar› ova¤, I. & Dra¤bek, M. (2006). Inversion twinning in troilite. American Mineralogist 91, 917^921. Spandler, C. J., Rubatto, D. & Hermann, J. (2005). Late Cretaceous^ Tertiary tectonics of the southwest Pacific; insights from U^Pb SHRIMP dating of eclogite-facies rocks from New Caledonia. Tectonics 24, TC3003. Sugaki, A., Shima, H., Kitakaze, A. & Harada, H. (1975). Isothermal phase relations in the system Cu^Fe^S under hydrothermal conditions at 3508C and 3008C. Economic Geology 70, 806^823. Sugaki, A., Shima, H., Kitakaze, A. & Mizota, T. (1981). Hydrothermal synthesis of nukundamite and its crystal structure. American Mineralogist 66, 398^402. Take¤uchi, Y., Kudoh, Y. & Sato, G. (1985). The crystal structure of covellite under high pressure up to 33 kbar. Zeitschrift fu«r Kristallographie 173, 119^128. Thompson, J. B., Jr (1972). Oxides and sulfides in regional metamorphism of pelitic schists. Proceedings of the 24th International Geological Conference, Section 10, 27^35. Toulmin, P. & Barton, P. B. (1964). A thermodynamic study of pyrite and pyrrhotite. Geochimica et Cosmochimica Acta 28, 641^671. Tracy, R. J. & Robinson, P. (1988). Silicate^sulfide^oxide fluid reactions in granulite-grade pelitic rocks, central Massachussetts. AmericanJournal of Science 288-A, 45^74. Vernon, R. H., White, R. W. & Clarke, G. L. (2008). False metamorphic events inferred from misinterpretation of microstructural evidence and P^Tdata. Journal of Metamorphic Geology 26, 437^449. Wang, L.-W. (2012). High chalcocite Cu2S: a solid^liquid hybrid phase. Physical Review Letters 108, article 085703. Wang, N. (1984). A contribution to the Cu^Fe^S system: the sulfidization of bornite at low temperatures. Neues Jahrbuch fu«r Mineralogie, Monatshefte 346^352. Yardley, B. W. D. (1977). An empirical study of diffusion in garnet. American Mineralogist 62, 793^800.

A P P E N D I X : A N A LY T I C A L M ET HODS Major element analyses of all minerals were obtained using a JEOL 6400 scanning electron microscope equipped with an energy-dispersive spectrometer and Oxford Link-ISIS quantification software, in the Electron Microscopy Unit (now the Centre for Advanced Microscopy) of the Australian National University. An accelerating voltage of 15 kV, a beam current of 1nA, and a counting time of 180 live seconds were used for each analysis. Elements analysed that were above detection limit in sulfide or silicate minerals were Mg, Al, Si, S, K, Ca, Ti, Mn, Fe, Co, Ni and Cu. Quantification was conducted in all-elements mode, which also returned a weight percentage for oxygen. However, this was not regarded as accurate, given sensitivity to coating quality and overlap from transition metal L peaks. Data were collected for sulfide inclusions, along with data for pseudo-inclusions and matrix sulfide. Where

grains were large enough (45 mm), point analyses were obtained. Analyses of smaller grains were necessarily contaminated by contributions from the host silicate. This was also true for area scans used to obtain an estimate of the average compositions of polyphase inclusions. In these cases, the sulfide composition was derived as follows. A set of area scans were obtained, beginning with a relatively large area containing the inclusion and host silicate, and then successively spanning smaller areas. The overall compositions of such areas are weighted averages of the sulfide and silicate contributions. In the simplest case, where sulfide is in contact with only one silicate phase (usually almandine garnet), linear regression of all weight percentages plotted against a component that is present in silicate but not sulfide (such as Si or Ca) allowed calculation of the true sulfide composition, as percentages in the sulfide were given by the y-intercepts of the regression lines. Where more than one silicate phase was present, it was necessary to extract the contributions owing to minor silicate phases first, using data for components that were significant in one silicate but not another. For instance, the titanite contribution to a titanite þ garnet mixture could be estimated from the overall content of Ti, which is an essential major component in titanite but negligible in the garnets and sulfides of this study. In any given analysis, we collected data from only a planar section of a 3D inclusion. In the case of a sulfide inclusion that is a polyphase intergrowth, the proportions of phases that are visible at the polished surface may not be exactly representative of the true proportions in three dimensions, and, furthermore, phases that appear isolated from one another in one cross-section may actually be in contact in the third dimension. To obtain the truest overview of inclusion makeup, we analysed numerous inclusions where possible. Repolishing inclusions to reveal progressively deeper surfaces was not successful owing to the small grain sizes. The Cu^Fe^S system at low temperature is known to contain a very large number of CuxS phases with x ¼ 0·50^2·00, and also CuxFeyS phases with x þ y 1 (see Fleet, 2006). In the absence of diffraction data, we were reliant for this study on optical appearance and compositions alone to identify the major phase or phases present, and have erred on the conservative side in our assignment of likely mineral species identifications. Atomic ratios of Fe:Co:Ni:Cu and M/S (where M ¼ Fe þ Co þ Ni þ Cu) were examined, and where appropriate, were attributed to phases that are known to persist at T42008C. The corresponding phases were generally pyrrhotite, pyrite, chalcopyrite, bornite, digenite and covellite. Pyrrhotite and pyrite occasionally showed minor Co or Ni above detection limit, interpreted to be solid solution components, and Cu, which may have been in solid solution or may have been contamination from an

669

JOURNAL OF PETROLOGY

VOLUME 55

adjacent sulfide phase. Similarly, minor Fe was frequently observed in the very Cu-rich phases covellite and digenite. These were readily recognized by their blue colours in reflected light microscopy (and strong anisotropy in the case of covellite), but the (Cu þ Fe):S ratio was frequently far from ideal, suggesting unresolved intergrowth with phases of different metal:sulfur ratio. Rare instances of analyses with Ni as the dominant cation and M/S close to 0·75 were identified as thiospinels in the polydymite^linnaeite^violarite series (Ni3S4^CoNi2S4^FeNi2S4). Many analyses lay in the Cu^Fe^S subsystem, but at Cu:Fe:S ratios removed from the known or likely solid solution ranges of the phases listed above. These analyses are interpreted as unresolved fine intergrowths, which may result from breakdown of high-P/T solid solutions. Such compositions were classified as ‘MP’ ¼ ‘metal-poor’ (M/S  0·95; probably containing pyrite), ‘MR’ ¼ ‘metalrich’ (M/S 1·05; probably containing digenite or bornite), ‘MM’ ¼ ‘medium-metal’ (0·955M/S51·05 and Fe4Cu; probable breakdown products from intermediate solid solution, dominated by pyrrhotite þ chalcopyrite or maybe cubanite) and ‘MN’ ¼ ‘nukundamite-like’ (0·955M/ S51·05 and Cu4Fe; probable breakdown products of a solid solution, dominated by chalcopyrite þ covellite or maybe nukundamite). Some ‘MM’ analyses corresponded closely to the CuFe2S3 composition of cubanite, but these

NUMBER 3

MARCH 2014

were rare enough for this to appear coincidental. The ‘MN’ analyses clustered at Fe/(Cu þ Fe) ¼ 0·2^0·3, which is an Fe content appreciably higher than that expected for nukundamite [Fe/(Cu þ Fe) ¼ 0·17]. We endeavour to maintain a distinction between the phases and compositions that are detected in the specimens under ambient conditions and those that were likely to have been present at peak conditions. For instance, an area scan that is within analytical error of the ideal chalcopyrite composition is referred to as ‘Cp’ in the results, as it is extremely likely to correspond to an area that is singlephase Cp or close to it, Cp being a phase that is stable at ambient conditions. Conversely, there is much evidence that Cp was not a distinct phase under peak conditions, where the same composition was merely part of the iss field, so in the discussion, the same analysis may be inferred to imply the former presence of iss. The interpretation of matrix and pseudo-inclusion sulfide compositions was in some cases further complicated by the effect of weathering. Oxidation causes loss of sulfur, and an increase in overall M/S ratio of the resulting sulfide^oxide mixture. However, obvious oxide contamination was rare, and the majority of sulfide inclusions do not have in situ weathering rims. Samples with highly weathered sulfide minerals were rejected from further consideration.

670

Suggest Documents