The Proterozoic P–T–t Evolution of the Kemp Land Coast, East Antarctica; Constraints from Si-saturated and Si-undersaturated Metapelites
1School of Geosciences, University of Sydney, Sydney, Nsw 2006, Australia
2School of Earth Sciences, University of Melbourne, Melbourne, Vic 3010, Australia
3Present Address: Institute of geoscience, University of Mainz, D-55099 Mainz, germany
4Continental Evolution Research Group, School of Earth & Environmental Sciences, University of Adelaide, Adelaide, Sa 5005, Australia
RECEIVED JUNE 29, 2006; ACCEPTED MARCH 29, 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONREGIONAL GEOLOGYMETAMORPHIC GEOLOGYMINERAL CHEMISTRYP-T METAMORPHIC CONDITIONSMONAZITE (Th + U)-Pb...DISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Integrated metamorphic and geochronological data place new constraints on the metamorphic evolution of a Neoproterozoic orogen in east Antarctica. Granulite-facies rocks from a 150 km stretch of the Kemp Land coast reflect peak conditions involving T
870–990°C at P 7·4–10 kbar, with pressure increasing westward towards an Archaean craton. Electron microprobe-derived (Th + U)–Pb monazite ages from metapelitic assemblages indicate that the major mineral textures in these rocks developed during the c. 940 Ma Rayner Orogeny. Complex compositional zoning in monazite suggests high-T recrystallization over c. 25 Myr. Diversity in metapelitic reaction textures reflects silica and ferromagnesian content: Si-saturated Fe-rich metapelites contain garnet that is partially pseudomorphed by biotite and sillimanite, whereas Si-saturated Mg-rich metapelites and Si-undersaturated metapelitic pods have reaction microstructures involving cordierite enclosing orthopyroxene, garnet and/or sapphirine, cordierite + sapphirine symplectites around sillimanite and coarse-grained orthopyroxene + corundum separated by sapphirine coronae. Interpretations based on P–T pseudosections provide integrated bulk-rock constraints and indicate a clockwise P–T–t path characterized by a post-peak P–T trajectory with dP/dT 15–20 bar/ °C. This moderately sloped decompressive-cooling P–T path is in contrast to near-isothermal decompression P–T paths commonly cited for this region of the Rayner Complex, with implications for the post-collisional tectonic response of the mid- to lower crust within this orogenic belt. KEY WORDS: electron microprobe monazite dating; granulite facies; Rayner Complex; sapphirine; THERMOCALC
Minerals abbreviations: q, quartz; g, garnet; sill, sillimanite; ky, kyanite; opx, orthopyroxene; cd, cordierite; ksp, alkali feldspar; pl, plagioclase; bi, biotite; sp, spinel; ilm, ilmenite; mt, magnetite; ru, rutile; sa, sapphirine; cor, corundum; osm, osumilite; liq, silicate melt; mnz, monazite
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGYMETAMORPHIC GEOLOGYMINERAL CHEMISTRYP-T METAMORPHIC CONDITIONSMONAZITE (Th + U)-Pb...DISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Characterization of the metamorphic history for middle to lower crustal rocks is largely based on the interpretation of mineral reaction microstructures. However, different P–T paths may produce the same observed texture in a given rock, and vice versa a mineral reaction microstructure does not uniquely define a P–T path (e.g. Vernon, 1996
; Kelsey et al., 2003b). Furthermore, the inferred P–T trajectory is valid only if all the metamorphic reaction textures developed during the same event (e.g. Hand et al., 1992; Vernon, 1996; Goncalves et al., 2004). The calculation of P–T pseudosections as a technique for deciphering P–T paths is becoming increasingly common as sophisticated thermodynamic datasets become available for chemical systems that closely approximate actual rock compositions. Many workers have successfully used P–T pseudosections to constrain prograde and retrograde paths via the calculation of mineral mode contours (e.g. Stüwe & Powell, 1995; White et al., 2002; Boger & White, 2003; Johnston et al., 2003; Kelsey et al., 2003c, 2005; Ouzegane et al., 2003; Zeh et al., 2004). In this way mineral reaction microstructures involving the production of one or more minerals at the expense of an earlier mineral assemblage can be quantified, and a P–T trajectory would cross modal contours representative of such a reaction. Integrated with geochronological data and a structural context, this approach provides a powerful means for the estimation of P–T–t paths.
This paper investigates the Proterozoic P–T–t evolution of a reworked Archaean terrane exposed along the Kemp Land coastline in east Antarctica. Granulite-facies rocks in this area record evidence of the large-scale, c. 1000–900 Ma, Rayner Orogeny (locally known as the Rayner Structural Episode: Sandiford & Wilson, 1984) induced by the convergence of the Napier Complex and parts of the Indian Peninsula, with a continental block that now constitutes much of east Antarctica (e.g. Mezger & Cosca, 1999). The thermo-physical evolution of the orogenic belt in this region of east Antarctica is still poorly understood; access to the area is difficult, extrapolation between ice-covered regions is necessary, and previous work has mostly relied on conventional thermobarometry (e.g. Ellis, 1983; Black et al., 1987). Recent work (e.g. Kelly & Harley, 2004a) suggests a ‘clockwise’ P–T path and near-isothermal decompression (ITD) from high to ultra-high temperatures in rocks from western Kemp Land during the Neoproterozoic. In this study we estimate peak P–T conditions and the post-peak evolution of the Kemp Land rocks via modelling of the mineralogical evolution of a suite of metapelitic assemblages from the Stillwell Hills, Broka and Havstein Islands and the Oygarden Group, three locations that constitute the most accessible outcrop along the Kemp Land coast (Fig. 1). Si-saturated (quartz-present) gneisses provide ‘peak’ P–T estimations via pseudosections calculated in the system Na2O–CaO–K2O–FeO–MgO–Al2O3–SiO2–H2O–TiO2–Fe2O3 (NCKFMASHTO). The trajectory of the P–T path for rocks from the Stillwell Hills is constrained via the interpretation of several stages of sapphirine growth and associated reaction textures in Si-undersaturated (quartz-absent) metapelites, combined with the constraints from post-peak assemblages in Si-saturated assemblages. Finally, the timing of the inferred metamorphic evolution is constrained by in situ electron microprobe (EMP) dating of coexisting monazite grains. This ‘bulk rock integrated’ approach incorporates constraints from diverse metamorphic assemblages and allows for a well-constrained P–T–t path for Neoproterozoic orogenesis in the Stillwell Hills. This revised metamorphic evolution may be more widely applicable to the Kemp Land terrane with implications for the thermo-physical evolution of this region of the East Antarctic Shield.
|
|
REGIONAL GEOLOGY |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGYMETAMORPHIC GEOLOGYMINERAL CHEMISTRYP-T METAMORPHIC CONDITIONSMONAZITE (Th + U)-Pb...DISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Granulite-facies rocks in Kemp Land, together with those exposed in parts of Enderby Land (to the west) and MacRobertson Land (to the east), constitute the greater Rayner Complex in east Antarctica, originally defined as distinct from the Napier Complex in Enderby Land by Kamenev (1972
; Fig. 1). Archaean components of Kemp Land are preserved in the Stillwell Hills, Broka and Havstein Islands and the Oygarden Group (Fig. 1; Clarke, 1988; Kelly et al., 2004; Halpin et al., 2005) and Hf model ages indicate a c. 3900–3600 Ma source, similar to model ages reported for parts of the Napier Complex (e.g. Rippon Point; Halpin et al., 2005). At least part of Kemp Land is therefore Archaean cratonic material reworked during the Rayner Orogeny (e.g. Sheraton et al., 1980, 1987; Grew et al., 1988).
Effects of the Rayner Orogeny at c. 940–930 Ma in rocks along the Kemp Land coast (Grew et al., 1988; Kelly et al., 2002; Halpin et al., 2005) have been interpreted to result from ductile thrusting of Archaean gneiss westward onto a craton, now represented by the Napier Complex (Clarke, 1988). A macroscopic strain gradient developed during the Rayner Orogeny increases to the east, recorded by metamorphosed basic dykes that cut Archaean foliations in extensive layered orthogneiss in Kemp Land (Stillwell Gneiss; Trail, 1970). Extensive c. 990–960 Ma charnockitic and felsic magmatism associated with the Rayner Orogeny, which post-date peak P–T conditions along the Mawson Coast in MacRobertson Land (Young & Black, 1991; Dunkley et al., 2002), are not obvious in Kemp Land. Kelly et al. (2002) suggested this may be indicative of a two-stage structural evolution for the Rayner Complex during the Rayner Orogeny.
P–T data for rocks in western Kemp Land are limited. Estimates made by Black et al. (1987) in the Dismal Mountains (Fig. 1) yielded peak conditions of T = 800–850°C at P = 5–7 kbar. Ellis (1983) obtained estimates of T = 700–800°C at P = 8·4–11 kbar from rocks at Turbulence Bluffs (Fig. 1a). Mafic granulites in western Kemp Land commonly contain plagioclase + orthopyroxene symplectites that enclose garnet; P–T estimates for these textures fall in the range P = 3–5·9 kbar at T = 660°C and have been interpreted to suggest partial equilibration at lower pressures (Ellis, 1983). P–T estimates based on mafic and metapelitic granulites from the Oygarden Group (Fig. 1) indicate peak or near-peak conditions of P > 9–10 kbar at T = 800–950°C (Kelly et al., 2000; Kelly & Harley, 2004a), followed by near-ITD to P 5 kbar (Kelly et al., 2000; Kelly & Harley, 2004a). Schröter (2006) modelled the mineralogical evolution of coronae in a mafic granulites from Broka and Havstein Islands (Fig. 1) and inferred peak P–T conditions of P = 8·4 kbar at T = 770°C followed by near-ITD to P = 6 ± 2 kbar at T = 750 ± 50°C. Average P–T estimates for mafic assemblages made by Schröter (2006) suggest conditions of P = 8·4 ± 1·3 kbar at T = 760 ± 90°C for the Stillwell Hills. These data suggest that a P–T path dominated by near-ITD characterizes the post-peak metamorphic evolution of Kemp Land, similar to P–T paths for the Rayner Complex in Enderby Land (reviewed by Harley & Hensen, 1990).
|
METAMORPHIC GEOLOGY |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGYMETAMORPHIC GEOLOGYMINERAL CHEMISTRYP-T METAMORPHIC CONDITIONSMONAZITE (Th + U)-Pb...DISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Outcrop setting
The layered Stillwell Gneiss dominates rock outcrop in Kemp Land west of the Scoble Glacier (Fig. 1), and forms extensive exposures in the Stillwell Hills, Broka and Havstein Islands and the Oygarden Group. The majority of this gneiss comprises alternating pyroxene–plagioclase-rich and quartz–feldspar-rich layers that record several high-grade deformation phases (Table 1: D1–D4 of Kelly et al., 2000
, 2002). Low-angle cross-cutting relationships between early felsic-dominated S1 layers and mafic layers defining S2 are consistent with parts of the mafic layering originating from deformed post-S1 mafic dykes (Clarke, 1988; James et al., 1991). D1–D2 are interpreted to have occurred prior to the Rayner Orogeny (James et al., 1991; Kelly et al., 2000). In the Oygarden Group of islands, effects of the Rayner Orogeny are characterized by macroscopic recumbent F3 folds and east-trending thrusts commonly localized on mafic dyke–gneiss contacts (Kelly et al., 2000). Steeply dipping D4 shear zones locally transposed earlier fabrics (Kelly et al., 2000). At Broka and Havstein Islands and the Stillwell Hills, pervasive sub-horizontal, simple shear-dominated D3 deformation produced inclined to recumbent structures and invariably transposed earlier fabrics (Table 1). Open to tight, upright to inclined folding characterizes later Rayner Orogeny deformation in these areas.
|
Subordinate metapelitic gneiss occurs as discontinuous lenses or pods in c. 3650 Ma Stillwell Gneiss in the Oygarden Group (zircon U–Pb minimum emplacement age; Kelly et al., 2004
). Similar relationships occur at Broka and Havstein Islands, where metapelitic gneiss is interlayered with c. 3540 Ma orthogneiss (zircon U–Pb minimum emplacement age; Halpin et al., 2005). Field relationships differ in the Stillwell Hills, where a distinctive iron-stained metapelitic package, hereafter referred to as the rusty gneiss, is infolded with, but not cut by, c. 3490 Ma Stillwell Hills Orthogneiss (zircon U–Pb minimum emplacement age; Halpin et al., 2005). Though the rusty gneiss is tentatively correlated with metapelitic gneiss exposed throughout MacRobertson Land (McLeod et al., 1966), age relationships are unclear because of the pervasive overprint during the Rayner Orogeny.
Metapelitic mineral assemblages
The investigated rocks can be divided into two compositional groups based on their distinct bulk compositions: Si-saturated and Si-undersaturated metapelites. The Si-saturated rocks are further subdivided into Fe-rich (Assemblage I) rocks, which contain garnet and sillimanite, and Mg-rich (Assemblage II) rocks, which additionally contain cordierite and orthopyroxene. The descriptions below are based on observations from a large number of thin sections from the Stillwell Hills (SW- sample numbers), Broka and Havstein Islands (BH- sample numbers) and the Oygarden Group (OG- sample numbers). Particular reference is made to samples used in the phase diagram calculations and summary information is presented in Tables 2 and 3.
|
|
Petrography of Si-saturated metapelites
Assemblage I (samples SW113, SW118, BH108, OG554)
Coarse-grained assemblages. Garnet (1–8 mm across) is enveloped by prismatic sillimanite, with or without biotite, rutile and opaque minerals (Fig. 2a). Larger garnet porphyroblasts commonly contain inclusions of quartz, sillimanite, ilmenite, plagioclase and lesser biotite and rutile, whereas garnet grains in more Fe-rich samples may contain inclusions of spinel and magnetite (e.g. BH108). Elongate garnet grains have sillimanite inclusions defining an internal fabric that is commonly orientated parallel to the dominant external fabric (Fig. 2a). Small, idioblastic (
0·3 mm) and inclusion-poor garnet may also occur. Perthite (1·5 mm across) is elongate and may have fine-grained inclusions of sillimanite, biotite or rutile. Plagioclase is common only in leucocratic segregations, where it occurs with alkali feldspar and quartz. Sample OG554 is an exception to this generalization, where both alkali feldspar and plagioclase (1 mm) occur in textural equilibrium with other coarse-grained minerals in the matrix including quartz, biotite, sillimanite and garnet.
|
Fine-grained assemblages. Fine-grained sillimanite occurs at garnet–garnet grain boundaries (Fig. 2b). Minor fine-grained sillimanite is also present on the margins of prismatic sillimanite (Fig. 2b). Fine-grained biotite rims garnet in strain shadows, or cuts S3. Minor ilmenite is exsolved from biotite. Myrmekite is present at some feldspar grain boundaries.
Assemblage II (samples BH134, OG589)
Coarse-grained assemblages. Garnet (<5 mm) and orthopyroxene (< 2 mm) with or without cordierite and perthite, is enveloped by S3 biotite with or without sillimanite and rutile (Fig. 2c and d). Garnet may contain inclusions of biotite, quartz, sillimanite and rutile and also occurs as small granoblastic grains (1 mm). Quartz, minor sillimanite and biotite occur as inclusions in orthopyroxene, whereas smaller (1 mm) grains are commonly inclusion-free. Perthite and cordierite (1·5 mm) are part of the coarse-grained assemblage in sample BH134. The presence of cordierite and plagioclase at peak conditions for sample OG589 is uncertain.
Fine-grained assemblages. Cordierite or plagioclase may rim garnet and/or orthopyroxene (Fig. 2d). Orthopyroxene + cordierite symplectites may enclose garnet in sample OG589. Coarse-grained sillimanite is partially mantled by fine-grained sillimanite. Fine-grained biotite may occur at the margin of porphyroblastic garnet in sample BH134 (Fig. 2c).
Petrography of Si-undersaturated metapelites
Coarse-grained assemblages. These samples have two domains. The first is dominated by coarse-grained sapphirine (<18 mm; Fig. 2e) and orthopyroxene (<8 mm; Fig. 2f) with or without sillimanite (<4 mm, e.g. SW235) and corundum (<8 mm, e.g. SW158). The second is biotite-rich, where biotite occurs as large poorly orientated laths (<2 mm in length) intergrown with orthopyroxene and sapphirine. Cordierite (<2·5 mm) occurs throughout the matrix, predominantly in biotite-poor domains, and embays the porphyroblastic mineral grains. Biotite, with or without sillimanite, defines an S3 foliation that envelops pods of the two domains. Sapphirine is commonly tabular or intimately intergrown with orthopyroxene. Orthopyroxene contains sapphirine as bleb-like inclusions (Fig. 2f) or as coronae that enclose corundum inclusions (Fig. 2g). Orthopyroxene also contains inclusions of cordierite, biotite, rutile and plagioclase. Corundum occurs as colourless tabular to rounded grains with rutile and sapphirine inclusions; it is always enclosed by coronae of sapphirine (Fig. 2g). Locally these sapphirine coronae (0·2–2·4 mm wide) pseudomorph corundum (Fig. 2h).
Fine-grained assemblages. Symplectic intergrowths of sapphirine (sa2) and cordierite (cd2 < 0·3 mm in width) surround coarse-grained sillimanite (sill1) (Figs 2i and 3). Small blebs of alkali feldspar (ksp2) may be present within the symplectites of sapphirine and cordierite (Figs 2j and 3). Local fine-grained symplectic intergrowths (0·01–0·1 mm) of sillimanite (sill2) with or without orthopyroxene (opx2) and biotite (bi2) develop adjacent to the larger sillimanite grains (Fig. 3). Relict orthopyroxene (opx1) may be observed in the vicinity of these sillimanite breakdown textures and is embayed by coronal cordierite (<1·5 mm; Fig. 3). Fine-grained sillimanite (<0·05 mm) may be developed on the rims of tabular sapphirine, which may additionally be rimmed by cordierite (<0·15 mm; Fig. 2e) or plagioclase.
|
Inferred parageneses
We interpret ‘peak’ metapelitic assemblages to consist of coarse-grained minerals that appear to have texturally equilibrated at the thin-section scale. ‘Post-peak’ assemblages refer to the comparatively fine-grained minerals, coronal and/or symplectic textures associated with the retrograde history of the rock (Tables 2 and 3). Peak assemblages in Assemblage I metapelites are garnet, sillimanite, alkali feldspar, ilmenite, quartz with or without plagioclase, rutile, spinel and magnetite. Post-peak textures are limited but include fine-grained biotite, which rims porphyroblastic garnet or grows oblique to S3, fine sillimanite at garnet–garnet grain boundaries and as a breakdown product of coarse sillimanite, ilmenite exsolution from biotite, and magnetite exsolution in spinel. Metapelites of Assemblage II are interpreted to locally contain peak orthopyroxene, cordierite and sillimanite, together with garnet, rutile and quartz, with or without alkali feldspar and plagioclase. Common post-peak mineral development includes narrow rims of cordierite or plagioclase around ferromagnesian minerals and fine-grained biotite and sillimanite, which replaces garnet and sillimanite.
Sapphirine is common to all Si-undersaturated metapelites from the Stillwell Hills. Distinctions in textural context and the composition (below) of sapphirine (see Tables 6 and 8; Fig. 3) allow several stages of sapphirine growth to be recognized: association I—coarse-grained coronal sapphirine enclosing corundum that is inferred to preserve part of the prograde path; association II—coarse, tabular sapphirine grains that are inferred to reflect peak conditions; association III—sapphirine + cordierite symplectites that are inferred to form during the post-peak evolution. These textures are consistent with a ‘clockwise’ P–T evolution (i.e. the thermal maximum is reached after peak pressure), as discussed below. However, Si-undersaturated, high Mg–Al metapelites are commonly chemically and mineralogically domainal at outcrop to micrometre scales; complex mineral reaction microstructures may develop primarily as a function of chemical potential rather than changing P–T (e.g. Hensen, 1988; Dunkley et al., 1999). Therefore it can be difficult to confidently interpret the P–T trajectory of such rocks. If mineralogical changes are driven by changes in P–T, the domain-scale compositions may experience different reactions at different P–T conditions. The interpretation of complicated chemical and microstructural domains requires a composition-dependent approach and the consideration of chemical potential gradients.
|
MINERAL CHEMISTRY |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGYMETAMORPHIC GEOLOGYMINERAL CHEMISTRYP-T METAMORPHIC CONDITIONSMONAZITE (Th + U)-Pb...DISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
The majority of chemical analyses were collected using the Cameca SX-50 Camebax electron microprobe (EMP) housed at the University of New South Wales, operating with an accelerating voltage of 15 kV and a beam current of 20 nA. Additional samples were analysed using the Macquarie University Cameca SX-100 EMP under identical operating conditions. Both sets of EMP data were reduced using the PAP data reduction software supplied by the manufacturer. Fe3+ was estimated on the basis of ideal stoichiometric constraints (after Droop, 1987
) for garnet, orthopyroxene, spinel, sapphirine and the opaque minerals. Representative EMP analyses are presented in Tables 4–6. A summary of mineral chemistry for the samples used in pseudosection calculations is given below and summarized in Tables 7 and 8.
|
|
|
|
|
Si-saturated metapelites
Assemblage I garnet is an almandine-rich, almandine–pyrope solid solution with XAlm = Fetotal/(Fetotal + Mn + Mg + Ca) = 0·54–0·64 and XPyp = Mg/(Fetotal + Mn + Mg + Ca) = 0·31–0·43 with minor XGross = Ca/(Fetotal + Mn + Mg + Ca)
0·04 and XSpess = Mn/(Fetotal + Mn + Mg + Ca) 0·03 for most samples (Tables 4, 5 and 7). Garnet in Assemblage II is a pyrope–almandine solid solution with XPyp = 0·40–0·63 and XAlm = 0·35–0·57. Garnet may have a subtle rimward enrichment in Fe, and a decrease in Mg and Ca; the variation is limited and there are no systematic differences in the composition of garnet orientated in the foliation compared with porphyroblastic garnet. Assemblage I biotite inclusions in garnet [XMg = Mg/(Fe + Mg) = 0·73–0·74] are more magnesian than inclusions in alkali feldspar (XMg = 0·65–0·67). Matrix biotite in Assemblage I includes biotite in the S3 foliation (XMg = 0·69–0·77), and subtly more magnesian post-S3 biotite (XMg = 0·75–0·78). Assemblage I biotite contains 0·54–0·67 Ti cations per formula unit (p.f.u) based on 22 oxygen. Biotite in Assemblage II has XMg = 0·69–0·86 and Ti content of 0·44–0·73 p.f.u. The range in compositions within each sample is restricted to variation in XMg < 0·08 (Table 7). Biotite inclusions in garnet (e.g. BH134; XMg = 0·74–0·77) are commonly more magnesian than matrix biotite (XMg = 0·69–0·74).
Spinel (Assemblage I) has end-member compositions of XHerc = (Fe2+ – FeMt)/(Fe2+ + Mg + Zn) = 0·36–0·38, where FeMt is the amount of Fe2+ devoted to magnetite, XSpl = Mg/(Fe2+ + Mg + Zn) = 0·42, XGah = Zn/(Fe2+ + Mg + Zn) = 0·16–0·18 and XMt = FeMt/(Fe2+ + Mg + Zn) = 0·04 (Table 7). Ilmenite ranges between a pure end-member composition and an ilmenite–hematite solid solution with XHem = 0·02–0·14.
Cordierite (Assemblage II) has XMg = Mg/(Fetotal + Mg) = 0·86–0·93 (Table 6). Individual cordierite grains are not zoned. Orthopyroxene (Assemblage II) varies in the range XMg = Mg/(Fetotal + Mg) = 0·80–0·82 (Table 7). Al content ranges from 0·23 to 0·35 p.f.u. with no apparent zoning. Alkali feldspar is commonly microperthitic with XOr = K/(Ca + Na + K) = 0·77–0·90 (Table 7). Plagioclase varies in the range XAn = Ca/(Ca + Na) = 0·18–0·39 (Table 7). Sillimanite contains less than 0·5 wt % Fe2O3 (analysed as FeO) in most samples, but up to 1·3 wt % Fe2O3 in sample BH108. Rutile contains up to 0·2 wt % Fe2O3 (analysed as FeO).
Si-undersaturated metapelites
Sapphirine composition varies considerably with textural context with XMg = 0·82–0·88 and the Al content, or y(sa) 
XM1A1, ranges from 0·19 to 0·48 (Table 8). Figure 4 shows Al vs Si for sapphirine analyses, all of which lie below the ideal Tschermak's substitution line between (Mg:Al:Si) compositions 2:2:1 and 7:9:3 (Fig. 4a and b) implying considerable incorporation of Fe3+ in some cases (Tables 6 and 8). Compositions cluster closer to the 7:9:3 stoichiometric ratio (where Al = 4·5, Si = 0·75 p.f.u on the basis of 10 oxygen). Sapphirine blebs in large orthopyroxene grains show compositions close to symplectic sapphirine; however, they are more enriched in Fe3+ (Fig. 4b). A large compositional range in coronal sapphirine reflects the position within the corona and/or proximity to neighbouring minerals; inner coronal sapphirine against corundum is richest in Al and is Si-poor. The majority of coarse-grained porphyroblastic sapphirines are Si-rich and Al-poor compared with finer-grained symplectic sapphirine (Fig. 4a and b). The cores of porphyroblastic sapphirine have higher Si and lower Al than rim compositions, resulting in marked y(sa) profiles, with comparatively little variation in XMg across the grain (Fig. 4c).
|
Orthopyroxene compositions (Fig. 5) vary in the range XMg = 0·84–0·85 for corundum-bearing sample SW158 and XMg = 0·74–0·78 for samples SW235 and SW238 (Table 8). Zoning in Al content is apparent with y(opx)
XM1A1 = 0·12–0·20 (Table 8). Porphyroblast rim compositions have higher Si and lower Al than cores [y(opx)rim < y(opx)core] and trend towards Si-rich/Al-poor symplectic orthopyroxene (Fig. 5b and c). This relationship is the opposite of that observed for sapphirine compositions. Variation in XMg (opx) across orthopyroxene porphyroblasts is minimal (Fig. 5c).
|
Biotite is highly magnesian, reflecting the bulk-rock compositions (Table 8). Biotite inclusions in cordierite, sapphirine and orthopyroxene have slightly lower XMg (bi) (
0·02) than biotite in the matrix. Ti content varies between samples, presumably as a result of rock composition; sample SW158 has higher Ti (0·30–0·37 p.f.u) than samples SW235 and SW238 (0·01–0·29 p.f.u). Fine-grained biotite involved in symplectite textures in SW235 has the lowest Ti values (0·01–0·02 p.f.u), similar to the range of late biotite on the edge of orthopyroxene ± cordierite in SW238 (0·02–0·09 p.f.u).
Cordierite shows a restricted range of compositions between samples (XMg = 0·89–0·95), which is further restricted (0·02) within individual samples (Table 8). Alkali feldspar in the matrix in SW238 has XOr = 0·67–0·88. Small blebs of alkali feldspar intergrown with symplectic sapphirine in SW235 have XOr = 0·88–0·89 (Table 8). Inclusions of plagioclase in orthopyroxene have higher XAn values (0·50–0·51) than matrix plagioclase (0·44–0·45) in SW158. Rare plagioclase in SW238 rims sapphirine porphyroblasts and is close to pure anorthite (Table 8). Sillimanite contains up to 0·4 wt % Fe2O3, rutile contains up to 1· 4 wt % Fe2O3 and corundum up to 0·7 wt % Fe2O3 (all analysed as FeO) and 0·2 wt % Cr2O3.
|
P–T METAMORPHIC CONDITIONS |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGYMETAMORPHIC GEOLOGYMINERAL CHEMISTRYP-T METAMORPHIC CONDITIONSMONAZITE (Th + U)-Pb...DISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Methodology of P–T pseudosection calculations
To constrain the peak and retrograde history of the metapelites from Kemp Land, a series of P–T pseudosections relevant to the mineral assemblages preserved in these rocks are presented in the following section. Pseudosections are more useful for interpreting mineral textures than petrogenetic P–T grids as they portray and quantify multivariant equilibria. Mineral assemblage fields are contoured for mineral proportions, providing additional constraints on the interpretation of metamorphic microstructures.
The pseudosections presented in this study for the Si-saturated metapelites were calculated in the NCKFMASHTO chemical system using the software THERMOCALC v3.25 (Powell & Holland, 1988) and the internally consistent dataset of Holland & Powell (1998, updated 22 November 2003). The omission of MnO from the calculations is likely to overestimate the temperature at which garnet becomes stable by a small amount. New mixing models for garnet, biotite and liquid (silicate melt), have been utilized in this NCKFMASHTO system (White et al., 2007). The a–x relationships for orthopyroxene, spinel and magnetite are taken from White et al. (2002), ilmenite from White et al. (2000), osumilite from Holland et al. (1996) and the remaining minerals are those used by White et al. (2001). This model system incorporates TiO2 and Fe2O3 (White et al., 2002), which have been shown to affect the stability of biotite and spinel at high temperatures (e.g. Indares & Martignole, 1985; Sack & Ghiorso, 1991; Kleeman & Reinhardt, 1996). These components also allow the consideration of rutile and ilmenite, both minerals occurring within the peak assemblages in the Si-saturated metapelites in Kemp Land. P–T pseudosections for the Si-undersaturated metapelites are calculated in the NCKFMASH chemical system, as there is currently no a–x model for Fe3+-bearing sapphirine. The a–x relationships for sapphirine are those of Kelsey et al. (2004, 2005). NCKFMASH calculations here use the mineral and melt a–x relationships and non-ideal parameters of White et al. (2001).
A critical consideration in the calculation of these pseudosections is the estimation of an appropriate bulk chemical composition, which must take into account the equilibration volume of the mineral assemblage observed in thin section. In rocks with considerable mineral zoning it may be necessary to exclude certain areas for the effective bulk composition, such as the cores of porphyroblastic minerals (e.g. Marmo et al., 2002). For medium- to high-grade rocks, a representative bulk composition is commonly estimated using molecular proportions recalculated from X-ray fluorescence (XRF) analyses (e.g. Kelsey et al., 2005). This method may be applied if the rocks are relatively homogeneous and minerals largely unzoned. In this study, representative bulk compositions of the Assemblage I Si-saturated samples were estimated by XRF from cuts of the hand specimen that was used for thin-section preparation. Metapelites of Assemblage II are domainal with respect to the distribution of ferromagnesian minerals (and sillimanite in OG589). Although XRF analyses were used for peak P–T calculations for these Assemblage II rocks, a number of fields in the pseudosections appropriate to these domains are included as representative of the peak mineral assemblage. This will be discussed in further detail below. P–T pseudosections for Si-undersaturated metapelites were calculated with a bulk composition derived from XRF analyses where the samples are relatively coarse-grained (e.g. SW158, SW238). To model finer-grained, post-peak microstructures in sample SW235 an element map of a selected area of the thin section (Fig. 3) allowed for a bulk composition calculation after a matrix correction (after Clarke et al., 2001). The amount of H2O used for P–T pseudosection calculations was approximated by the ‘loss on ignition’ (LOI) value in the bulk analysis. An estimation of Fe2O3 (which is recast as ‘O’ in the NCKFMASHTO system, Table 9) was based on the abundance of the Fe3+-bearing minerals and modal constraints in the context of recalculated EMP analyses. In most cases the estimated amount of Fe3+ is low because of the paucity of minerals in the assemblage known to contain considerable Fe3+ (e.g. oxide minerals). Garnet and/or orthopyroxene in the analysed rocks have little Fe3+ (Tables 4, 5 and 7). In Fe-rich Assemblage I rocks, Fe3+-bearing minerals are somewhat more common, particularly in the case of the sample selected from Broka and Havstein Islands (BH108). The effect of altering the estimated Fe2O3 is discussed below.
|
P–T constraints from Si-saturated metapelites
Two Si-saturated metapelites from each of the three locations were chosen for the calculation of P–T pseudosections. Where possible, bulk-rock compositions at each location (Table 9) were chosen to provide a marked contrast in XMg (bulk; Table 7).
Estimation of peak conditions
The (inferred) peak metamorphic assemblages for the Si-saturated metapelites from the Kemp Land coast (Table 2) correspond to assemblage fields (in bold outline) in the pseudosections in Fig. 6a–f. As samples chosen for this study come from a small area at each location, the groupings are interpreted to have shared metamorphic histories and to have experienced similar peak P–T conditions (i.e. local field gradients, if present, were insignificant). The two bulk compositions from each location can therefore be used to constrain the peak P–T conditions for the area. This technique involves ‘overlapping’ the peak assemblage fields for the two samples (stippled areas in Fig. 6); the region of P–T space shared by these compositions may be considered as the peak P–T conditions for the area (Kelsey et al., 2003c). In the case of Assemblage II metapelites, several fields may be interpreted as peak assemblages, appropriate to different local textural domains. In all samples modelled, the inferred stability of biotite is pivotal to the interpretation of a lower limit to peak conditions. Biotite is excluded from the interpreted peak assemblage fields below because the petrographic observations are ambiguous and the NCKFMASHTO model system lacks F and Cl, and thus probably underestimates biotite stability (Peterson et al., 1991; Nair & Chacko, 2002).
|
The hatched areas on P–T pseudosections in Fig. 6 show the overlapping P–T space for peak metapelitic assemblages from the Stillwell Hills, Broka and Havstein Islands and the Oygarden Group. Overlapping peak assemblages for the Stillwell Hills (Fig. 6a and b) are the garnet–sillimanite–alkali feldspar–rutile–ilmenite–silicate melt–quartz (+ plagioclase for SW118) fields that border the rutile-out (to lower-P) and ilmenite-out (to higher-P) boundaries; the location of these fields varies with the ‘O’ content of the bulk composition. In both cases these boundaries shift up pressure with increasing oxygen values. The value chosen for the pseudosection calculations (O = 0·04) is relatively low, based on the low Fe3+ content of minerals in this sample (Tables 4 and 7). Peak metamorphic conditions for the Stillwell Hills as defined by the overlapping P–T space are P = 7·6–9·4 kbar at T
870°C. An upper-T constraint is particularly difficult to obtain as the interpreted peak assemblage fields extend to very high temperatures. The absence of spinel-bearing assemblages in Si-undersaturated sample SW238 (see below) constrains peak T to <1000°C. The contrast in the peak assemblages inferred for the Fe-rich (Fig. 6c) and Mg-rich (Fig. 6d) metapelites from Broka and Havstein Islands restricts the pressure range of high-grade conditions in that area. The peak assemblage for BH108 lies in the garnet–sillimanite–alkali feldspar–ilmenite–silicate melt–quartz ± spinel ± magnetite fields. An estimation of the amount of Fe3+ affects the stability of peak minerals spinel and magnetite. For this reason the larger spinel- and magnetite-absent field is included as part of the peak assemblage. For BH134, the narrow trivariant field garnet–cordierite–orthopyroxene–sillimanite–alkali feldspar–plagioclase–rutile–silicate melt–quartz best represents the peak assemblage. This sample locally lacks orthopyroxene and/or cordierite, and thus the orthopyroxene-absent (low-P) and cordierite-absent (high-P) fields are included here. Peak metamorphic conditions for Broka and Havstein Islands are estimated at P = 7·4–8·5 kbar at T = 900–1000°C.
The Mg-rich sample OG589 (Fig. 6f) contains orthopyroxene and garnet-rich domains with or without cordierite and sillimanite. Hence, cordierite and/or sillimanite may or may not have been part of the peak parageneses. The peak assemblage for OG589 is inferred to contain garnet–orthopyroxene–plagioclase–rutile–silicate melt–quartz ± cordierite ± sillimanite. The three (local) peak fields overlap with the garnet–sillimanite–alkali feldspar–plagioclase–rutile–silicate melt–quartz quinivariant field for the Fe-rich sample OG554 (Fig. 6e). The overlapping P–T space defines peak metamorphic conditions for the Oygarden Group of P = 7·7–10 kbar at T = 920–990°C.
Post-peak P–T evolution
The retrograde history of the rocks must be considered in the context of the estimated peak conditions. In the case of the Si-saturated metapelites, reaction microstructures are limited. Assemblage I metapelites were particularly mineralogically insensitive to changes in P or T from peak conditions, as corroborated by the large P–T extent of peak assemblage fields across which there are very small (<5 mol %) changes in mineral modal abundances (Fig. 7). Assemblage II metapelites show some reaction microstructures but as they are influenced by local domainal chemistry they cannot be confidently represented by the calculated pseudosections. Although the post-peak trajectory is difficult to quantitatively constrain using these rocks, several limits on the retrograde P–T paths for the Kemp Land coastline can be interpreted. Figure 7a–c shows the calculated Assemblage I P–T pseudosections for the three areas (SW118, BH108 and OG554) with mineral mode contours for the assemblage fields of relevance. Superimposed on these diagrams are the peak conditions inferred in the previous section, as well as the upper-P limits to a retrograde path inferred from the absence of cordierite from post-peak assemblages. A P–T path for each of the three areas must cross the mineral proportion contours in such a way as to explain the growth of biotite and minor sillimanite at the expense of garnet. It is not possible to form biotite along an isothermal or near-isothermal decompressive path from the peak conditions in these pseudosections. To more tightly constrain the magnitude of cooling vs decompression within metapelites from Kemp Land we have integrated pseudosections calculated for Si-undersaturated rocks with petrographic observations (below).
|
P–T constraints from Si-undersaturated metapelites in the Stillwell Hills
P–T pseudosections for three Si-undersaturated metapelites from the Stillwell Hills were calculated (Fig. 8) and satisfactorily account for the three stages of sapphirine growth previously described, as discussed below.
|
Prograde and peak P–T evolution
To investigate the development of coronal sapphirine (association I) on corundum, sample SW158 was modelled (Fig. 8a). These microstructures are very similar to those described from the Oygarden Group by Kelly & Harley (2004a
). The peak conditions deduced from the Si-saturated metapelites encompass the assemblage sapphirine–orthopyroxene–biotite–alkali feldspar–plagioclase ± silicate melt ± cordierite. Sapphirine coronae on corundum are interpreted to have formed from the reaction between orthopyroxene and corundum to form sapphirine and sillimanite, consistent with the relative changes in mineral modes (Fig. 8a). The (melt-absent) assemblage orthopyroxene + corundum is predicted to be stable over a wide range of P–T conditions on this pseudosection (Fig. 8a), with corundum reacting out across a steep positive-sloped (effective) univariant to produce sillimanite. However, this diagram can only be crudely used to interpret the development of the sapphirine coronae. The extensive orthopyroxene + corundum fields can be expected to be restricted if melt were present prior to peak conditions. For the estimated bulk-rock composition, orthopyroxene + corundum + melt-bearing fields are restricted to P > 9·6 kbar. Whereas it may be difficult to model prograde textures in such high-grade rocks, this pseudosection allows a qualitative assessment of the prograde P–T path in the Stillwell Hills, immediately prior to peak conditions. Coarse-grained sapphirine coronae that in some cases completely pseudomorph the corundum are inferred to have developed along a prograde heating path between 800°C and 900°C, to peak conditions. The absence of cordierite from the reaction microstructure probably indicates near-peak P > 7·6 kbar. Figure 8b is a P–T pseudosection appropriate to sample SW238, which contains sapphirine laths minimally consumed by overprinting assemblages involving fine-grained sillimanite and/or cordierite coronae (sapphirine of association II). The stability of garnet in this pseudosection at P > 8·4 kbar places a robust limit to upper-P conditions. The absence of spinel in these Mg-rich rocks places a further upper-T limit (<1000°C) on the conditions in the Stillwell Hills. More Fe-rich compositions would encounter garnet-bearing fields at lower P and spinel-bearing fields at lower T. The development of minor retrograde sillimanite and narrow cordierite coronae are consistent with a decompressive-cooling path. However, as the bulk composition is derived from XRF analysis, the interpretation of this diagram should be limited to peak conditions that coincide with peak P–T conditions constrained from Si-saturated metapelites (hatched area: Fig. 8).
Post-peak P–T evolution
The post-peak P–T trajectory for metapelites from the Stillwell Hills can be illustrated using the sillimanite reaction textures in SW235. Sapphirine + cordierite symplectites that enclose coarse-grained sillimanite are the most notable example of sapphirine in association III. An element map of the reaction texture between sillimanite and orthopyroxene (Fig. 3) was used for the calculation of an appropriate bulk composition. The pseudosection for this rock is shown in Fig. 8c and d. The observed sequence of mineral reactions is consistent with a decompressive-cooling path passing through the sapphirine–orthopyroxene–cordierite–sillimanite–plagioclase–alkali feldspar–silicate melt trivariant field to sub-solidus conditions (arrow in Fig. 8d). Increases in the abundance of sapphirine and cordierite across this field occur at the expense of sillimanite and orthopyroxene. This is consistent with observed textures involving the breakdown of peak sillimanite (sill1) and orthopyroxene (opx1) to produce sapphirine (sa2) and appreciable cordierite (cd2) (Fig. 3). The ratio of cordierite:sapphirine increases along this path to about 40:25, consistent with the observed cordierite moats around the cordierite + sapphirine symplectites.
Fine-grained orthopyroxene (opx2), sillimanite (sill2) and biotite (bi2; Fig. 3) indicate subsequent partial replacement of the symplectic minerals cordierite (cd2) + sapphirine (sa2) ± alkali feldspar (ksp2). We infer that this reaction was driven by the breakdown of hydrous cordierite and is predicted to occur across the trivariant field sapphirine–orthopyroxene–cordierite–sillimanite–biotite–plagioclase–alkali feldspar along the P–T path to T 840°C at P 7·5 kbar (Fig. 8d). Modal contours indicate the production of biotite, orthopyroxene and sillimanite at the expense of cordierite, sapphirine and alkali feldspar consistent with the observed texture.
An integrated P–T path
Si-saturated and Si-undersaturated samples from the Stillwell Hills provide consistent evidence for a prograde heating trajectory to peak conditions of T = 870–1000°C at P = 7·6–9·4 kbar, followed by decompression-cooling to T 840°C at P 7·5 kbar. As a range of peak conditions is representative of assemblages preserved in the Stillwell Hills, P–T paths with differing (including curved) slopes between peak metamorphic conditions and T 840°C, P 7·5 kbar are consistent with the changes in mineral proportions and growth of retrograde minerals. Shallow (cooling-dominated) retrograde paths do not allow for the increase in cordierite and sapphirine modes observed in the Si-undersaturated metapelites (path A in Fig. 8d). Alternatively, steep (decompression-dominated) retrograde paths do not intersect biotite-bearing fields and instead predict the consumption of orthopyroxene (path B in Fig. 8d). A decompressive-cooling path with dP/dT 15–20 bar/°C is consistent with the post-peak mineral parageneses in the Stillwell Hills (Fig. 9a and b). This P–T path also matches the trend of decreasing y(opx) for porphyroblastic orthopyroxene (Fig. 9b). The post-peak P–T trajectory for the Stillwell Hills is also consistent with the development of cordierite rims around garnet and/or orthopyroxene, and the growth of retrograde biotite in metapelites from Broka and Havstein Islands and the Oygarden Group, and thus may be more widely applicable for the Kemp Land terrane. A moderately sloped P–T path is in contrast to near-ITD paths inferred for Turbulence Bluffs (Ellis, 1983), the Oygarden Group (Kelly & Harley, 2004a) and Broka and Havstein Islands (Schröter, 2006).
|
|
MONAZITE (Th + U)–Pb GEOCHRONOLOGY |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGYMETAMORPHIC GEOLOGYMINERAL CHEMISTRYP-T METAMORPHIC CONDITIONSMONAZITE (Th + U)-Pb...DISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Chemical dating of monazite is now a well-established method (e.g. Montel et al., 1996
; Cocherie et al., 1998; Crowley & Ghent, 1999; Williams et al., 1999; Williams & Jercinovic, 2002; Pyle et al., 2005). The chemical composition of monazite grains (Fig. 10) was analysed in situ via mapping and spot analysis using a Cameca SX-51 EMP equipped with four wavelength-dispersive spectrometers at the University of Adelaide. Elemental mapping of Y, Th, U and Pb in monazite (Williams & Jercinovic, 2002; Yang & Pattison, 2006) was performed on larger monazite grains to detect chemical zonation (Figs 11 and 12). The position of spot analyses utilizes the chemical zonation patterns from elemental maps. Operating conditions and the methodology of monazite age calculations are detailed in the Appendix. Age calculations use the measured Th, U and Pb contents and involve a multi-stage procedure outlined by Kelsey et al. (2003a) and summarized in the Appendix.
|
