| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology | Volume 43 | Number 9 | Pages 1673-1705 | 2002
© Oxford University Press 2002
Characterization and P–T Evolution of Melt-bearing Ultrahigh-temperature Granulites: an Example from the Anápolis–Itauçu Complex of the Brasília Fold Belt, Brazil
1LABORATORY FOR CRUSTAL PETROLOGY, DEPARTMENT OF GEOLOGY, UNIVERSITY OF MARYLAND, COLLEGE PARK, MD 20742, USA
2INSTITUTO DE GEOCIÊNCIAS, UNIVERSIDADE DE BRASÍLIA, BRASÍLIA, DF, 709110-900, BRAZIL
3COMPANHIA DE PESQUISA DE RECURSOS MINERAIS (CPRM), RUA S-02, 463 APTO 101, SETOR BELA VISTA GOIÂNIA, GO 74823-430, BRAZIL
4COMPANHIA DE PESQUISA DE RECURSOS MINERAIS (CPRM) SGAN 603 CONJ. ‘J’, PARTE ‘A’, 1°, ANDAR BRASÍLIA, DF 70830 030, BRAZIL
Received September 10, 2001; Revised typescript accepted March 8, 2002
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONREGIONAL GEOLOGICAL SETTING OF...PETROGRAPHYMINERAL CHEMISTRYINTERPRETATION OF...DISCUSSION OF THE ROLE...TECTONIC SETTINGCONCLUSIONSREFERENCES |
|---|
Sapphirine–quartz occurs in orthopyroxene–garnet granulites in two areas
22 km apart within ‘common’ granulites of the Anápolis–Itauçu Complex, recording extreme P–T conditions: a minimum of 1030–1050°C at 10 kbar. In one area, the post-peak evolution is constrained to begin at P > 10 kbar by down-temperature stability of garnet–orthopyroxene–sillimanite–quartz and the absence of cordierite. In a second area, the post-peak evolution is constrained by a succession of melt-present reactions that occur at P < 10 kbar, inferred from microstructural relations between garnet, orthopyroxene and sillimanite, and coronae or symplectites involving corundum, sapphirine, spinel, cordierite, plagioclase, sillimanite, orthopyroxene, biotite and ilmenite. The retrograde segment of the P–T path was further constrained by biotite-producing reactions. A composite P–T path was constructed from samples that exhibit different amounts of retrograde reaction, reflecting different amounts of melt retention. Using back-calculated compositions for garnet and orthopyroxene, thermobarometry yields 1012–960°C and 9·7–8·1 kbar; these P–T results underestimate ‘peak’ conditions, in part as a result of modification of garnet compositions in rocks in which some melt was retained. In samples from both areas, orthopyroxene porphyroblasts have high Al2O3 (12·9–9·7 wt %) in cores, which suggests maximum T > 1150°C in both areas. After decompression at the metamorphic peak, the P–T path for both areas followed a near-isobaric cooling stage to <900°C, but these two paths are separated by 3 kbar. This ultrahigh-temperature metamorphism occurred in a collisional tectonic setting, in which the extreme thermal perturbation probably was a result of asthenosphere replacement of mantle lithosphere. KEY WORDS: Brasília Fold Belt; granulite; melting; P–Tevolution; sapphirine–quartz; ultrahigh-temperature metamorphism
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGICAL SETTING OF...PETROGRAPHYMINERAL CHEMISTRYINTERPRETATION OF...DISCUSSION OF THE ROLE...TECTONIC SETTINGCONCLUSIONSREFERENCES |
|---|
Granulite-facies rocks record extreme thermal perturbation of Earth’s lithosphere; they are common in both Precambrian shields and Phanerozoic orogens. Characterizing granulite-facies rocks yields information critical to our understanding of the P–T–X environment of formation and subsequent modification of the continental crust (e.g. Harley, 1989, 1992). Fifteen years ago it was suggested that peak metamorphic conditions in most granulite terranes fit within a P–T window of 7·5 ± 1 kbar and 800 ± 50°C (Bohlen, 1987). However, even at that time several exceptions to this generalization had been documented, including Enderby Land, Antarctica (Dallwitz, 1968; Ellis, 1980; Harley, 1985), Scourie, Scotland (O’Hara & Yarwood, 1978; Barnicoat & O’Hara, 1979) and the Labwor Hills, Uganda (Sandiford et al., 1987), and other examples were reinterpreted by Hensen (1987), who proposed higher temperatures of metamorphism than the original researchers postulated.
Since 1987, the number of localities has greatly increased where temperatures higher than 900°C have been retrieved from the mineralogical and mineral chemical evidence preserved in appropriate samples. More recently documented examples include the Eastern Ghats (Dasgupta et al., 1995) and Palni Hill Ranges in India (Brown & Raith, 1996; Raith et al., 1997), the In Ouzzal Complex in Algeria (Mouri et al., 1996), the Barro Alto Complex in Brazil (Moraes & Fuck, 2000), and the Saxon and Moldanubian Granulite Complexes of the Bohemian Massif in Central Europe (Cooke & O’Brien, 2001; Rötzler & Romer, 2001). The term ultrahigh-temperature (UHT) metamorphism has been introduced for rocks that record evidence of such extreme thermal conditions in the crust (Harley, 1998a). These conditions are known to exceed 1100°C in the Napier Complex, Antarctica (Harley & Motoyoshi, 2000; Hokada, 2001; Motoyoshi & Hensen, 2001), and similar temperatures are recorded in rocks of this study from central Brazil.
During the 1980s, granulite terranes were grouped into two types according to the post-peak P–T evolution: those that record evidence of near-isobaric cooling P–T paths and those that exhibit near-isothermal decompression P–T paths (see review by Harley, 1989). This generated debate about clockwise vs counterclockwise P–T paths and crustal thickening vs magmatic underplating for the origin of the extreme thermal perturbation required by granulite-facies metamorphism (e.g. Bohlen, 1987; Ellis, 1987). Additionally, retrograde paths have been documented from both ‘common’ granulites and UHT granulites that are stepped, being composed of both near-isothermal decompression and near-isobaric cooling segments (e.g. Harley et al., 1990; Brown & Dallmeyer, 1996; Raith et al., 1997). Such a P–T evolution is believed to be the post-thermal peak segment of a clockwise P–T path, and commonly this evolution has been related to crustal thickening, possibly at a convergent plate margin (Brown, 2001).
The role of melt in the retrogression of UHT assemblages
Mineral associations characteristic of UHT granulite-facies metamorphic conditions include aluminous orthopyroxene–sillimanite–quartz, sapphirine–quartz and spinel–quartz (Harley, 1998a, 1998b). At the temperatures attained during UHT metamorphism, melt probably coexisted with the mineral assemblages (Carrington & Harley, 1995a, 1995b; Holland et al., 1996). A genetic connection between granulite-facies mineral assemblages and melting began to receive serious attention in the early 1970s, when depleted granulites were interpreted as residues after melt extraction (e.g. Fyfe, 1973). Granulite-facies assemblages in metasedimentary rock protoliths may be generated by biotite dehydration melting, with production of melt and peritectic phases, such as garnet, cordierite, aluminous orthopyroxene, sillimanite, sapphirine and spinel. Melanosome is not formed along the prograde evolution because biotite is almost completely consumed and the peritectic minerals tend to be intimately associated with leucosome (Kriegsman, 2001; Brown, 2002); some biotite may remain stable with rutile or ilmenite to temperatures higher than 1000°C, but it will be rich in Ti and F (Hensen & Osanai, 1994; Motoyoshi & Hensen, 2001).
In the field, the generally migmatitic aspect of metapelitic gneisses and metagreywackes that record granulite-facies conditions confirms that these rocks were once melt bearing. However, survival of peak mineral assemblages and the depleted bulk compositions of many granulite terranes suggest that a substantial proportion of melt generated commonly has been lost from the system (Waters, 1988; Stüwe & Powell, 1989; Harley & Hensen, 1990; Raith et al., 1997; Sawyer, 2001; Guernina & Sawyer, 2002). During the past 15 years, the role of melt loss in the preservation of granulite-facies mineral assemblages and of melt retention in the retrogression of granulite-facies rocks has become better appreciated (e.g. Powell & Downes, 1990; Kriegsman, 2001; White et al., 2001; Brown, 2002; White & Powell, 2002). It is now clear that a significant amount of retrogression may occur during cooling as melt-consuming reactions are crossed, unless most of the melt produced is extracted (White & Powell, 2002).
Where UHT mineral assemblages are preserved, they occur locally within much larger areas of ‘common’ granulite (Harley, 1989; Harley & Hensen, 1990; Moraes & Fuck, 2000; Cooke & O’Brien, 2001). This raises the possibility that melt-present retrogression under granulite-facies conditions may have returned many UHT mineral assemblages to ‘common’ granulite-facies mineral assemblages.
Questions addressed by this study
The restricted occurrence of UHT mineral assemblages raises two important questions about the metamorphic processes involved. First, did the extreme P–T conditions required by UHT metamorphism occur regionally? Second, if so, was some feature of the protolith or the retention of some melt responsible for the widespread retrograde reaction to ‘common’ granulites? Additionally, the clockwise P–T path inferred for many granulite-facies terranes raises two important questions about the geodynamic processes involved. First, what is an appropriate tectonic setting to generate the clockwise P–Tpaths characteristic of most UHT metamorphism? Second, how is the extreme thermal perturbation required by UHT mineral assemblages achieved? We address these questions using evidence from the Anápolis–Itauçu Complex, which occurs in Goiás, central Brazil, within the Neoproterozoic Brasília Fold Belt.
|
REGIONAL GEOLOGICAL SETTING OF THE BRASÍLIA FOLD BELT |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGICAL SETTING OF...PETROGRAPHYMINERAL CHEMISTRYINTERPRETATION OF...DISCUSSION OF THE ROLE...TECTONIC SETTINGCONCLUSIONSREFERENCES |
|---|
The Anápolis–Itauçu Complex (AIC) occurs within the internal zone of the Neoproterozoic Brasília Fold Belt (Fuck et al., 1994), which borders the western margin of the São Francisco Craton (Fig. 1a). The São Francisco Craton is an Archaean and Palaeoproterozoic granite–gneiss terrane that is covered by sedimentary rocks of Neoproterozoic and Phanerozoic age.
|
In the external zone of the Brasília Fold Belt, adjacent to the São Francisco Craton, continental margin, platform-type sediments underwent greenschist-facies metamorphism during the Neoproterozoic Brasiliano Orogeny. Further west, the internal portion of the Brasília Fold Belt consists of two units (Fig. 1b and c). The Araxá Group is composed of predominantly turbidite-type sediments, volcanic rocks and ophiolite mélange, which underwent Barrovian-type metamorphism from greenschist- to amphibolite-facies conditions during the Brasiliano Orogeny. The AIC forms a NW–SE elongate area (260 km x 70 km) of outcrop composed of mainly granulite-facies rocks, separated by a mylonite zone from the Araxá Group. To the NW is the Goiás Massif, a poorly understood crustal fragment made up of Archaean and Palaeoproterozoic rocks that was reworked by the Brasiliano Orogeny. Southwest of the Goiás Massif is the Goiás magmatic arc, formed between c. 900 and c. 600 Ma (Pimentel & Fuck, 1992; Pimentel et al., 1997), which consists of volcano-sedimentary sequences and gneisses with geochemical and isotopic signatures typical of modern intra-oceanic arcs.
The Anápolis–Itauçu Complex
The AIC comprises several lithological groups, elongated NW–SE, similar to the complex as a whole (Fig. 1b and c). These are: (1) an orthogranulite unit, derived from protoliths of tonalite and granodiorite composition; (2) a unit of predominantly supracrustal granulites, including aluminous granulite, leptinite and garnet gneiss, with minor marble, calc-silicate rock, quartzite and fine-grained mafic granulite; (3) mafic–ultramafic rocks; (4) granites; (5) volcano-sedimentary sequences comprising amphibolite, micaschist, felsic metavolcanic rock, metachert and iron formation.
The AIC was considered to be basement to the Araxá Group, based on its high metamorphic grade (Marini et al., 1984; Wolff, 1991; Lacerda Filho & Oliveira, 1995; Winge, 1995). However, recently determined TDM model ages for supracrustal granulites from the AIC (Fischel et al., 1998; Sato, 1998; Pimentel et al., 1999) and for metasedimentary rocks of the Araxá Group (Fischel et al., 1999a, 1999b; Pimentel et al., 1999) are between 1·3 and 1·6 Ga in both cases. This limits the maximum age for the protoliths of the supracrustal granulites, and suggests that at least part of the AIC might be the high-grade equivalent of the Araxá Group. Although no geochronological data are available for the UHT rocks, the age of metamorphism for surrounding ‘common’ granulite was determined as 630 Ma using the Sm–Nd method on whole rock and garnet, with monazite, biotite or amphibole (Fischel et al., 1998, 1999a). Similar ages were obtained from metamorphic zircon separated from granulites from the Interlândia quarry (Fig. 1b) using sensitive high-resolution ion microprobe (SHRIMP) (Tassinari et al., 1999).
|
PETROGRAPHY |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGICAL SETTING OF...PETROGRAPHYMINERAL CHEMISTRYINTERPRETATION OF...DISCUSSION OF THE ROLE...TECTONIC SETTINGCONCLUSIONSREFERENCES |
|---|
In the AIC, UHT mineral assemblages are recognized in two areas, at one locality immediately north of Goiânia (ML-67), and at two closely spaced localities
22 km NNW of Goiânia, one immediately WSW of Damolândia (PT-62) and the other in the Monjolo stream (ANA-287) 5 km west of Damolândia (Fig. 1c). In the Damolândia area, UHT mineral assemblages occur in rocks of two different compositions, quartz rich and quartz poor. For several samples from each of the three locations, phases were identified initially using optical microscopy. Preliminary identification of phases that occur as small inclusions or in complex coronae and symplectites was confirmed subsequently by electron probe microanalysis as part of a detailed programme of element mapping and precise spot chemical analysis.
Goiânia
In the first area, at locality ML-67 (Fig. 1c), loose centimetre-scale blocks of a banded impure quartzite are associated with blocks of metapyroxenite and banded hornblende-bearing dioritic granulite; the structural relationship between the rocks is unknown. Samples of the impure quartzite have macroscopic compositional layering that is defined by different proportions of quartz, garnet, orthopyroxene and sillimanite (Table 1). They exhibit a pervasive mylonitic foliation marked by very fine-grained quartz in ribbons, within which occur elongate to almond-shaped grains of garnet, orthopyroxene and sillimanite, resembling mica fish (see Hanmer, 2000); garnet and orthopyroxene porphyroblasts vary from 6 to 18 mm in length, whereas prismatic sillimanite is up to 4 mm in length. Fine-grained recrystallized orthopyroxene forms monomineralic layers (see Hanmer, 2000). Microscopically, coarser grains of quartz display undulose extinction and development of sub-grains, and have serrated to lobate boundaries with other quartz grains. This mylonitic microstructure reflects deformation under granulite-facies metamorphic conditions.
|
Sapphirine occurs as inclusions in garnet, orthopyroxene and sillimanite. The mineral pair sapphirine–quartz occurs in mutual contact in the matrix, and sapphirine grains may show crystallographic faces against quartz to suggest stable coexistence (Fig. 2a); the presence of sapphirine–quartz constrains minimum P–T conditions for the rocks at this locality. There are rare inclusions of idiomorphic garnet in orthopyroxene (Fig. 2b). More commonly, garnet (located against quartz) occurs with sillimanite (located against sapphirine) along sapphirine–quartz grain boundaries (Fig. 2c); this microstructure is particularly common where sapphirine and quartz occur as multigrain inclusions in xenomorphic garnet. Finally, sapphirine may be separated from quartz by sillimanite, which surrounds sapphirine, and orthopyroxene, which surrounds sillimanite and garnet and separates them from quartz (Fig. 2d).
|
Grt + Spr + Qtz and (c) is interpreted to record reaction (2) Spr + Qtz 
plate (long dimension is 2 mm). (y) CL image of plagioclase–quartz intergrowth interpreted as a magmatic microstructure and representing crystallization of a pocket of residual melt. PT-62-A (long dimension 2·0 mm). (z) Quartz–plagioclase inclusion in garnet where the low dihedral angle between them is interpreted to mimic a solid–liquid relationship as a result of crystallization of a pocket of melt (long dimension is 0·6 mm). (aa) Partial replacement of garnet–cordierite by sillimanite–biotite as a result of reaction with residual melt by reaction (17) Grt + Crd ( ± Kfs)
|
K-feldspar, plagioclase, cordierite and spinel do not occur in samples from this locality (Table 1). Up to 2 vol. % biotite occurs as platy grains close to or against orthopyroxene, sillimanite or garnet; it is parallel to the mylonitic foliation. We infer that this biotite was developed as a retrograde phase. Abundant rutile and sillimanite needles are present in orthopyroxene porphyroblasts, commonly disposed parallel to cleavage planes. Granular rutile is common, occurring in quartz, garnet and orthopyroxene; zircon, ilmenite and pyrite are present in accessory amounts.
Damolândia and the Monjolo stream
In the second area, at locality PT-62 (Fig. 1c), sapphirine occurs in rocks with two different bulk compositions, quartz poor (PT-62-F) and quartz rich (PT-62-A), 20 m apart, but it is not possible to establish any structural relationship between the two outcrops because of lack of continuous exposure between them. None the less, the fact that sapphirine occurs in contact with quartz as multigrain inclusions in garnet in the quartz-rich rock type unambiguously constrains minimum P–T conditions for these rocks. In the Monjolo stream, at locality ANA-287 (Fig. 1c), the rocks are strongly deformed. We distinguish orthopyroxene–garnet tonalite, granodiorite gneiss and migmatites that comprise metapelitic gneiss with centimetre-scale leucosomes of granite and orthopyroxene–garnet felsic granulite (probably a metagreywacke protolith) with patchy leucosomes of granite. In this paper, we report data from leucosome and host metapelitic gneiss, and felsic granulite.
Quartz-poor rock type
Samples of the quartz-poor rock have a massive microstructure with a grain size that varies from 5 to 40 mm (Fig. 2e). These samples are composed predominantly of garnet and orthopyroxene porphyroblasts with cordierite and minor amounts of sapphirine, spinel, plagioclase and biotite; also there are aggregates of plagioclase, cordierite and orthopyroxene (Table 1). Sillimanite, quartz, corundum, ilmenite, pyrite, zircon and apatite occur in accessory amounts. Orthopyroxene porphyroblasts are commonly recrystallized to smaller grains at their edges; rutile and sillimanite needles are common in orthopyroxene, disposed parallel to the cleavage planes. Quartz occurs as rounded inclusions inside garnet or rarely in the matrix in some domains with plagioclase and biotite; it is never in contact with sapphirine, spinel or corundum.
Sillimanite occurs in the cores of some large garnet porphyroblasts, where complex coronae and symplectites may separate it from the garnet (Fig. 2f). Commonly, sillimanite and garnet are separated by two layers. First, against sillimanite there is a complex symplectite composed of dominantly sapphirine on the inside giving way to fine-grained sapphirine and cordierite on the outside (Fig. 2g), with minor symplectites of spinel and cordierite (Fig. 2h); these symplectites may include minor amounts of plagioclase and ilmenite. Second, against garnet there is a moat of cordierite (Fig. 2f).
In garnet, there are rare examples of corona microstructure that have a core of corundum and spinel surrounded by massive sapphirine in contact with the host garnet (Fig. 2i). In sillimanite, there are complex corona microstructures composed of successive layers of spinel, massive sapphirine and a sapphirine–cordierite–plagioclase symplectite (Fig. 2j). In one example, the corona microstructure has corundum in the innermost portion (identified by electron probe microanalyser, and visible in Fig. 2f). In another example, the coronae microstructure is continuous with the main volume of symplectite described above (located in the lower centre of Fig. 2f), suggesting that the enclosing sillimanite has been replaced completely. Spinel–sapphirine–cordierite–plagioclase symplectites also occur inside garnet porphyroblasts (Fig. 2k); although the mineral proportions in these symplectites may be different from one porphyroblast to another, we interpret these features as pseudomorphs after inclusions of sillimanite.
In turn, garnet is separated from matrix orthopyroxene porphyroblasts by extensive development of cordierite, which may be intergrown with new orthopyroxene (Fig. 2f). At the rims, garnet is replaced by several different symplectites involving orthopyroxene (Fig. 2l)—in lamellar symplectites with sapphirine or spinel and minor cordierite (Fig. 2m), and in blebby symplectites with cordierite (Fig. 2n). Plagioclase occurs associated with these symplectites, mainly when cordierite also is present. Symplectites of orthopyroxene with sapphirine or spinel also may develop along rims of orthopyroxene, particularly along contacts with garnet, and at some garnet–orthopyroxene contacts a symplectite of orthopyroxene and cordierite has partially replaced both minerals (Fig. 2o and p).
In the matrix, orthopyroxene–sapphirine symplectites are overprinted by a granular aggregate of spinel, cordierite and biotite, largely dominated by biotite (Fig. 2q). Most biotite occurs in a network of weakly oriented grains around garnet porphyroblasts. Ilmenite forms partial corona with embayed contacts around rutile, and occurs in rare symplectites with orthopyroxene along rims of orthopyroxene porphyroblasts.
Quartz-rich rock type
Samples of the quartz-rich rock have a gneissic structure with millimetre-scale layers defined by different proportions of garnet and orthopyroxene; grain size varies from 2 to 15 mm. The mineral assemblage is quartz, plagioclase, biotite, orthopyroxene, cordierite, garnet, spinel, sillimanite and rutile (Table 1). Sapphirine, ilmenite, pyrite, apatite and zircon are present in accessory amounts. Orthoclase may occur, but only rarely. As in the other samples, sillimanite and rutile needles occur oriented in the orthopyroxene cleavage. Ilmenite forms coronae around rutile.
Sapphirine is rare, occurring in garnet as multigrain inclusions in direct contact with quartz (Fig. 2r), sometimes with accessory spinel and apatite. Cordierite with intergrown spinel and inclusions of sillimanite and rutile (± neighbouring plagioclase) has partially replaced garnet (Fig. 2s and t), and there are aggregates of cordierite with spinel, sillimanite and rutile that occur in irregular patches and layers throughout the rock (Fig. 2u). Spinel and sillimanite are predominantly xenomorphic, and although some square basal sections of sillimanite are observed, the edges of these grains tend to be embayed or rounded (Fig. 2t and u).
Some garnet porphyroblasts are replaced along the rims by coronae of orthopyroxene and cordierite (Fig. 2v), possibly with ilmenite and biotite, or symplectites with both minerals, similar to those in the quartz-poor rock (compare Fig. 2m). Biotite may replace orthopyroxene and cordierite, sometimes replacing the entire corona except for minute orthopyroxene relicts (Fig. 2w); although rare, small euhedral garnets may be associated with the biotite.
Plagioclase occurs as multigrain aggregates in the matrix, in general around or between garnet grains, and has straight grain boundaries. However, where plagioclase is in contact with quartz the grain boundaries may be cuspate (Fig. 2x), in which quartz shows low apparent dihedral angles that resemble those between feldspar and siliceous melt in quenched experiments (Laporte et al., 1997). Plagioclase may form complex intergrowths with quartz in the matrix (Fig. 2y) or it may be present with quartz as inclusions in garnet, where cuspate terminations with low dihedral angles are also observed (Fig. 2z).
Migmatite
In the migmatitic metapelitic gneiss, centimetre-scale leucosomes are dominantly composed of quartz, perthitic orthoclase and plagioclase; biotite, sillimanite (coarse grains and fibrolite), garnet, cordierite, rutile, tourmaline, zircon, ilmenite, and minor amounts of kyanite, muscovite and chlorite are present in accessory amounts (Table 1). Quartz is concentrated in layers 2–3 mm thick, in which coarser grains show sub-grains, deformation bands and lobate contacts. Where quartz is in contact with large orthoclase grains, it has recrystallized to finer grains that surround a coarser relict core. Orthoclase is perthitic, with exsolved albite forming either rounded drops or fine films; undulose extinction, sub-grains and micro-faults are common. Plagioclase shows similar microstructures to K-feldspar, and some grains are antiperthitic with exsolved albite. Kyanite occurs as sporadic individual grains, but whether it developed along the prograde or retrograde part of the P–T evolution is unclear from the microstructural relations. Garnet occurs as isolated rounded porphyroblasts that may be replaced along the rims by cordierite and biotite. Biotite and sillimanite (fibrolite or coarse grains) may replace garnet and/or cordierite (Fig. 2aa). Chlorite replaces garnet, biotite and cordierite, and late muscovite replaces biotite. Rutile and ilmenite occur within the main foliation. Commonly there is a melanosome 2–3 mm thick composed of biotite and rare garnet between leucosomes and host metapelitic gneiss (Fig. 2ab).
The metapelitic gneiss is formed of quartz, plagioclase, orthopyroxene, garnet, cordierite, sillimanite, spinel, rutile, zircon and ilmenite (Table 1). Despite lacking sapphirine, observed microstructural relations include granular aggregates of spinel, cordierite and sillimanite and orthopyroxene–cordierite intergrowths replacing garnet rims, similar to microstructural relations in PT-62-A. Small idiomorphic garnet grains occur in the matrix, sometimes associated with biotite (Fig. 2ac). Rare garnet–quartz symplectites are associated with biotite–quartz intergrowths around garnet (Fig. 2ad). Biotite–quartz symplectites are common at the rims of large biotite grains (Fig. 2ae). Where rutile occurs close to spinel, a corona of ilmenite is developed between them.
The felsic granulite layers in the same outcrop are composed of quartz, plagioclase (antiperthite), orthoclase (perthite), orthopyroxene, garnet, ilmenite, minor biotite and muscovite. Quartzo-feldspathic domains in felsic granulite have a granoblastic fabric, with xenomorphic grains of orthopyroxene and garnet, commonly surrounded by biotite or biotite–quartz symplectites. Late randomly distributed muscovite is widespread, and sometimes occurs in symplectites with quartz (Fig. 2af).
|
MINERAL CHEMISTRY |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGICAL SETTING OF...PETROGRAPHYMINERAL CHEMISTRYINTERPRETATION OF...DISCUSSION OF THE ROLE...TECTONIC SETTINGCONCLUSIONSREFERENCES |
|---|
Chemical compositions of minerals were obtained using an electron probe microanalyser (JEOL JXA-8900 SuperProbe) in the Center for Microanalysis and Microscopy at the University of Maryland. For elemental mapping, operating conditions were 15 kV and 40–50 nA, and for precise spot chemical analysis of mineral composition, operating conditions were 15 kV and 20 nA, with a 5–10 µm beam size. Natural materials were used as standards for all elements.
Bearing in mind the coarse grain size of porphyroblasts in these rocks, in the following description of mineral chemistry, ‘core’ refers to the centre of cuts through the largest grains in any single thin section. All samples have microstructural features that suggest disequilibrium and preserve some evidence of retrograde reaction. For this reason, the technique used to select compositions for thermobarometric calculations was as follows. First, elemental maps of porphyroblasts of garnet (Fe, Mg, Ca and Mn) and orthopyroxene (Fe, Mg, Mn, Ca and Al) were produced to evaluate compositional zoning (e.g. Fig. 2ad, ag and ah). Second, based on this information, high-precision spot chemical analyses were obtained from regions with the most refractory compositions.
A summary of mineral compositions is given in Table 1 with the mineral assemblages, and representative analyses of major phases are given in Tables 2–8345678.
|
|
|
|
|
|
|
Garnet
Garnet is a pyrope–almandine solid solution, with minor grossular and spessartine components (Table 2). In ML-67, small idiomorphic garnets have compositions that are less rich in Mg (core prp 51·0–49·8 mol % and Alm 46·1–44·1 mol %) in comparison with larger garnets that are partly replaced by orthopyroxene. These larger garnets have homogeneous cores that are richer in Mg than rims, which are richer in Fe, Mn and Ca. The maximum pyrope content is 57·5 mol %, with 38·0 mol % almandine, whereas the rims have pyrope and almandine contents between 54·8 and 47·3 mol % and 47·9 and 40·7 mol %, respectively; 
varies from 0·60 to 0·55 from core to rim. Grossular and spessartine vary from 4·0 to 0·7 mol % and 2·7 to 1·6 mol %, respectively, from core to rim.
In PT-62-A, compositional mapping shows complex zoning patterns of Ca depletion around plagioclase–quartz inclusions in garnet (Fig. 2ag), which is compensated by higher concentrations of Fe and Mg (Fig. 3a and b). In general, close to rims, Fe is enriched in relation to cores and the reverse is observed for Mg (core: prp 48·4–42·2 mol % and alm 52·4–44·4 mol %; rim: prp 47·4–41·5 mol % and alm 55·9–47·6 mol %). Grossular can have values up to 12·6 mol % close to the core, zoning to a minimum of 0·5 mol % at the rim.
In PT-62-F, garnet shows homogeneous cores enriched in Mg in relation to rims (core: prp 51·0–45·2 mol % and alm 49·5–42·8 mol %; rim: prp 49·5–38·5 mol % and alm 55·7–45·2 mol %). Although some variation in Ca and Mn contents is observed from grain to grain, concentrations are uniform inside the same grain, varying from 5·2 to 1·2 and 1·2 to 0·7 mol %, respectively, with slightly higher concentrations close to rims.
In ANA-287, two garnet generations are recognized: porphyroblasts and smaller idiomorphic grains that may be associated with biotite. Porphyroblasts have cores richer in Mg and Ca than rims, which are richer in Fe and Mn (core: prp 48·0 mol %, alm 47·2 mol %, grs 6·2–3·6 mol % and sps 0·81 mol %, 
; rim: prp 41·4 mol %, alm 53·3 mol %, grs 4·1–1·4 mol % and sps 1·2 mol %, 
). The composition of the small garnets is similar to that of porphyroblast rims, or more Fe rich. varies from 0·44 to 0·38 from grain to grain and differences inside a grain are negligible, although a slight enrichment in Ca and Mg at rims is accompanied by decrease of Fe and Mn. Grossular content is 4·1–3·4 mol %, and spessartine content is 1·4–1·1 mol %.
Orthopyroxene
Orthopyroxene occurs as porphyroblasts and in symplectites with other minerals (mainly sapphirine, spinel, cordierite and plagioclase). The composition of porphyroblasts and symplectite lamellae is different (Table 3). In porphyroblasts, Al2O3 varies from 12·9 to 6·1 wt %, decreasing rimward in all the analysed samples, and this feature is accompanied by increase of MgO (Table 3). The high Al2O3 in porphyroblast cores is a typical feature of UHT granulites (Harley, 1998a, 1998b). Trends in orthopyroxene composition are shown in Fig. 4.
|
Orthopyroxene porphyroblasts in ML-67 exhibit a narrow rim of lower Al in comparison with homogeneous cores of higher Al. These rims are continuous around the orthopyroxene regardless of whether the orthopyroxene is in contact with garnet or quartz. In contrast, Mg enrichment in the orthopyroxene rim occurs only at grain boundaries with garnet (Fig. 2ah and ai). These features indicate that the zoning to lower Al in the orthopyroxene rim reflects growth of this orthopyroxene during cooling immediately after the metamorphic peak, whereas high Al in the homogeneous core reflects growth of that orthopyroxene during decompression at the metamorphic peak. In contrast, high Mg in the orthopyroxene rims reflects diffusive exchange with garnet during cooling. Of all samples, the highest Al2O3, 12·9–11·5 wt %, occurs in the core of orthopyroxene porphyroblasts in ML-67 (Tables 1 and 3; Fig. 4a,) whereas the Al2O3 varies from 10·4 to 7·2 wt % across the narrow rims (Tables 1 and 3; Fig. 4a).
In PT-62-A, orthopyroxene occurs as porphyroblasts and as lamellar intergrowths with cordierite in symplectites replacing garnet rims. In the cores, Al2O3 varies from 11·9 to 7·5 wt % decreasing rimward to 9·9–4·5 wt %, which is higher than or similar to the composition of orthopyroxene in the symplectites, which have 7·6–5·0 wt % Al2O3 (Table 3). Two compositional groups of orthopyroxene are observed, with higher or lower Mg at constant Al (Fig. 4b), which probably reflect differences in bulk composition from layer to layer.
In PT-62-F, porphyroblasts and several generations of symplectite have different compositions (Tables 1 and 3; Fig. 4c). Cores of porphyroblasts have the highest Al2O3, 11·0–9·5 wt %, with Al2O3 content decreasing rimward (Table 3). Orthopyroxene in symplectites has small differences from one association to another, and in a simple way the Al decreases and the Mg increases from symplectites with spinel and sapphirine to those with cordierite (Fig. 4c).
Orthopyroxene in sample ANA-287 is compositionally similar to orthopyroxene in PT-62-A, with core composition of 9·7–7·8 wt % Al2O3 and rim composition of 9·5–6·1 wt % Al2O3 (Table 3). There is some compositional overlap between grains in symplectites of orthopyroxene with cordierite and porphyroblast rims, but compositions of the orthopyroxene in symplectites extend to lower Al and higher Mg (Fig. 4a).
Cordierite
Although more than one generation of cordierite is recognized on the basis of microstructural relations, chemical variations in cordierite from one microstructural relation to another are subtle (Tables 1 and 4). This reflects FeMg-1 exchange with other ferromagnesian phases, which is the principal factor affecting compositional change in cordierite. In PT-62-A and ANA-287, there is a shift in composition of cordierite involved in multiple replacement of garnet (symplectites and coronae with orthopyroxene or granular intergrowths with sillimanite, spinel and biotite), and it is possible to distinguish some general trends. Matrix porphyroblasts of cordierite are richer in Fe (
is 0·82 in PT-62-A and 0·83 in ANA-287) than the cordierite in symplectites and coronae ( is 0·87–0·86 in PT-62-A and 0·84 in ANA-287). In PT-62-F, matrix porphyroblasts of cordierite are richer in Fe ( is 0·87) than cordierite involved in symplectites and coronae, for which compositions overlap and no clear compositional trends are evident.
Spinel
In PT-62-F, spinel is present in four different microstructural relations (see Fig. 2f, h–m, p and q), and has a different composition in each of these microstructures (Tables 1 and 5; Fig. 5). Spinel inside sillimanite (Fig. 2f) and in symplectites formed between garnet and sillimanite (Fig. 2f and h) is similar in composition, although spinel in the former microstructure has exsolved ilmenite lamellae and probably Fe3+ and Ti were higher. 
varies between 0·50 and 0·47, Zn content is low (0·001–0·002 a.p.f.u.) and Ti content is negligible. Spinel in symplectite with orthopyroxene (Fig. 2l, m and p) has the highest (0·53), and higher Zn (0·003–0·004 a.p.f.u.), whereas spinel coexisting with cordierite and biotite (Fig. 2q) has the lowest (0·45), but similar Zn (0·003–0·005 a.p.f.u.).
|
In PT-62-A and ANA-287, spinel is associated with sillimanite and cordierite (Fig. 2p–r). Differences in spinel composition between PT-62-A (
) and ANA-287 (
) are small (Tables 1 and 5; Fig. 5).
Sapphirine
Sapphirine varies in composition between samples (Table 6), although all compositions have Al in excess of the 2:2:1 end member, as a result of Tschermak exchange. Sapphirine in ML-67 has a wide range in composition, although no clear variation between core and rim was detected. 
varies between 0·86 and 0·74 and the formula becomes close to the 7:9:3 end member in some rims (Fig. 6a). Sapphirine in PT-62-A has the highest Al content and a formula close to the ideal 7:9:3 end member (Fig. 6a), assuming Fe and Mg exchange on sites. Calculated Fe3+ is low, between 0·05 and 0·005 a.p.f.u., and varies between 0·78 and 0·76.
|
In PT-62-F, sapphirine in symplectites with cordierite and spinel that are inferred to replace sillimanite and garnet (Fig. 2f, h and i) overlap in values (0·75–0·72; Fig. 6b). The grains in the inner part of the symplectite are less aluminous in comparison with those in the outer part of the symplectite (Fig. 6b). Lamellar sapphirine in symplectites with orthopyroxene (Fig. 2l, m and o) has higher (0·78–0·75) and tends to be more Al rich than lamellae in other symplectites (Fig. 6b). In all symplectites, 
(Table 1).
Feldspar
Plagioclase composition does not vary extensively in quartz-rich rocks or domains (Table 7), and is andesine–labradorite with 1 mol % Or. Compositional zoning is weak, with cores enriched in Na, except in ANA-287 where Na increases rimward. In quartz-rich samples, there is no difference in composition between plagioclase inclusions in garnet and grains in the matrix. Small drops of exsolved orthoclase may occur, but always in small proportions.
In PT-62-F, plagioclase occurs with two different compositions. The plagioclase associated with spinel–cordierite–sapphirine and orthopyroxene–sapphirine symplectites is Ca rich (An88·4–81·4Ab17·0–11·4Or0·5–0·0), whereas plagioclase associated with quartz and biotite is slightly more sodic (An86·4–80·6Ab19·3–13·6Or0·0). In both cases, Ca increases toward the rim. Orthoclase occurs only in leucosome layers in ANA-287; its composition is Or97·4–73·7Ab24·9–2·6An3·1–0·0, with exsolved albite occurring in some grains.
Biotite
Biotite has high 
, Ti and F, and low AlVI (Tables 1 and 8), which are typical features of the granulite facies (Guidotti, 1984). Si, Al and have a negative correlation with Ti (Fig. 7). Matrix biotite in PT-62-F has of 0·81–0·78, and 0·65–0·40 F a.p.f.u. and 0·50–0·32 Ti a.p.f.u., whereas inclusions in porphyroblast phases are richer in Mg ( of 0·87–0·85), with similar F but lower Ti (0·35–0·26 Ti a.p.f.u.). In both cases, AlVI is low (0·29–0·08 a.p.f.u.). In PT-62-A, there are three types of biotite. In general, is 0·81–0·75, with 0·67–0·52 F a.p.f.u. and 0·69–0·42 Ti a.p.f.u., whereas in Mg-richer layers, biotite has of 0·84–0·81, with 0·95–0·68 F a.p.f.u. and 0·51–0·42 Ti a.p.f.u., and almost no AlVI (0·06 a.p.f.u.). Composition of biotite inclusions in garnet shows considerable variation; some grains have the highest of 0·84–0·82, with F up to 1·01 a.p.f.u. (1·01–0·52) and Ti up to 0·75 a.p.f.u. (0·75–0·50) but no AlVI. Biotite in ANA-287 has of 0·80–0·75, 0·73–0·33 F a.p.f.u., 0·59–0·25 Ti a.p.f.u. and AlVI 0·18 a.p.f.u.; no compositional difference is observed between grains in the matrix and in the melanosome.
|
|
INTERPRETATION OF MICROSTRUCTURAL RELATIONS AND P–T EVOLUTION |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGICAL SETTING OF...PETROGRAPHYMINERAL CHEMISTRYINTERPRETATION OF...DISCUSSION OF THE ROLE...TECTONIC SETTINGCONCLUSIONSREFERENCES |
|---|
Sapphirine-bearing UHT granulites commonly preserve microstructures such as coronae and symplectites that record evidence of overstepped mineral reactions crossed during the P–T evolution (e.g. Harley et al., 1990; Brown & Raith, 1996; Mouri et al., 1996; Raith et al., 1997; Harley, 1998c). Development and preservation of these microstructures depend on reaction kinetics, diffusion rates and the presence of an appropriate grain boundary medium to enhance diffusion (supercritical fluid or melt; Brown, 2002). Microstructures such as these may preserve information about several points along or portions of the retrograde segment of the P–T path. However, evidence of the prograde evolution commonly is overprinted by the metamorphic peak mineral assemblage, making the distinctions between late prograde and metamorphic peak mineral growth difficult to distinguish in many cases.
In the rocks of this study, there are examples where the close-to-peak UHT mineral assemblages are preserved, but little evidence of the prograde P–T path remains. In other examples, the close-to-peak UHT mineral assemblages are largely overprinted by reaction during the retrograde evolution. We relate this difference in degree of retrograde reaction to the variable retention of residual melt in these rocks after the peak of metamorphism. In both cases, important information can be retrieved and used to trace the metamorphic peak and retrograde segments of the P–T path.
For residual granulite-facies assemblages derived from metapelitic gneiss and felsic granulite (metagreywacke) compositions and other Mg–Al-rich protoliths, assuming a common set of excess phases, interpretation of microstructural relations and variations in mineral chemistry inferred to record reactions may be made using petrogenetic grids constructed for the simplified model systems FMAS and KFMASH (e.g. Hensen & Green, 1973; Harley, 1989; Harley et al., 1990; Hensen & Harley, 1990; Spear et al., 1999; McDade & Harley, 2001). This is possible because, in general, the mineral assemblages have low variance.
It is unlikely that additional components will change the interpretation significantly, because these are incorporated as essential structural components in phases present in accessory amounts or they require the addition of a phase component (e.g. Ca in garnet) or they occur in a major rock-forming mineral that is present only in low abundance (e.g. all samples contain a small amount of plagioclase in the mode, requiring consideration of Ca and Na in the natural system).
The petrogenetic grids used in this paper (e.g. Fig. 8) have been constructed from the grid of Harley (1998c), which is based on the experimental data in the FMAS system of Hensen & Green (1973) and Bertrand et al. (1991), in which melt (± K-feldspar) is a phase in excess, and melt-producing reactions involving biotite in the KFMASH system as calculated by Spear et al. (1999). Abbreviations follow Kretz (1983), plus Spr for sapphirine and L for melt.
|
Goiânia
The banded impure quartzite (sample ML-67) has a simple mineral assemblage composed of quartz, sapphirine, garnet, orthopyroxene and sillimanite. Some orthopyroxene porphyroblasts may represent the residual peritectic product of a prograde biotite dehydration melting reaction, possibly with some sillimanite. The presence of the mineral pair sapphirine–quartz (Fig. 2a) is sufficient to establish minimum conditions of 1030–1050°C at P of 10 kbar for the thermal peak in the FMAS model system (Fig. 8a). In natural systems, additional components, particularly Fe3+, may expand the stability field of the mineral pair sapphirine–quartz to lower temperatures, although the effect is not expected to be large. However, in ML-67, neither haematite nor any other phase rich in Fe3+ is present, and the amount of Fe3+ calculated by charge balance for the Fe–Mg phases is low. These two features suggest that the ambient 
was not significantly above that of the experiments on which the FMAS model system phase relations are based. We will return to the issue of peak temperature in a subsequent section.
Idiomorphic garnet occurs as inclusions in orthopyroxene (Fig. 2b), sapphirine–quartz are separated by garnet and sillimanite developed along mutual grain boundaries (Fig. 2c), and coronae of sillimanite and orthopyroxene separate sapphirine, garnet and quartz (Fig. 2d). These microstructural relations may be explained by the FMAS univariant reaction
 |
 |
Garnet, orthopyroxene, sillimanite and quartz appear to have been a stable assemblage along the high-temperature retrograde segment of the P–T path, consistent with lower Al2O3 in orthopyroxene rims, and there is no microstructural evidence of further reaction until biotite was produced at T of <900°C. The amount of biotite produced was small, implying that if melt was retained down to this temperature, the volume was small. The microstructural relations of biotite do not allow us to specify the specific biotite-producing reaction.
These features support a retrograde evolution involving near-isobaric cooling from >1050°C to <900°C at P above the [Spl] invariant, the Opx + Sil + Qtz 
Grt + Crd univariant reaction, and the [Spr, Spl] invariant, respectively (Fig. 8a), as constrained by the absence of cordierite from samples of the banded impure quartzite at this locality.
Damolândia and the Monjolo stream
Quartz-poor rock type
In the quartz-poor rock (sample PT-62-F), there is no microstructural evidence preserved to allow identification of reactions that occurred along the prograde segment of the P–T path. However, we postulate that the UHT mineral assemblage probably was formed by biotite dehydration melting, producing peritectic orthopyroxene and sillimanite with residual F–Ti-rich biotite (F 0·65–0·50 a.p.f.u. and Ti 0·35 a.p.f.u.; Tables 1 and 8; Fig. 7) and possibly garnet. Rutile inclusions in garnet, and rutile and sillimanite needles in orthopyroxene could be formed at this biotite dehydration melting stage, with the crystallographic orientation of the acicular rutile and sillimanite being controlled by the cleavage in peritectic orthopyroxene.
Quartz occurs only in minor amounts as inclusions in garnet or in small pockets in the matrix associated with plagioclase (An54–51) and biotite, and inferred to record the former presence of melt. We infer that quartz was almost completely consumed before the metamorphic peak was achieved. This is consistent with melting reactions consuming quartz along the prograde evolution, and suggests loss of melt.
Coronae microstructure with a core of corundum and spinel surrounded by massive sapphirine in garnet (Fig. 2i) could be interpreted to represent early-formed corundum–spinel multigrain inclusions in garnet that have reacted to form sapphirine by the discontinuous FMAS reaction
 |
Large inclusions of sillimanite and small rounded grains of quartz are present in garnet porphyroblasts and, together with orthopyroxene porphyroblasts in the matrix, these features suggest that garnet could have grown at the expense of orthopyroxene and sillimanite through the FMAS univariant reaction (1). Sapphirine is an expected product of this reaction, and may be represented by sapphirine inclusions in sillimanite that have subsequently reacted with sillimanite to produce corundum, spinel and cordierite through the discontinuous FMAS reaction
 |
Cordierite moats against garnet and symplectites of dominantly sapphirine, sapphirine–cordierite–plagioclase and spinel–cordierite–plagioclase against sillimanite (Fig. 2f) were formed by reaction between garnet and sillimanite in the absence of quartz. These relationships are inferred to record the following FMAS divariant reactions (Fig. 8b):
 |
 |
Garnet is replaced along the rims by symplectites composed of lamellar orthopyroxene and sapphirine (Fig. 2l and m), sometimes with cordierite and plagioclase, lamellar orthopyroxene and spinel (Fig. 2l and m), sometimes with cordierite and plagioclase, and blebby orthopyroxene and cordierite (Fig. 2l and n), sometimes with plagioclase. We infer the following FMAS reactions based on these microstructures (Fig. 8b):
 |
 |
 |
with reaction progress (Fig. 9). Plagioclase has a composition similar to the earlier grains (An85-80).
|
Porphyroblastic orthopyroxene at contacts with garnet is replaced by lamellar symplectites of orthopyroxene and sapphirine or spinel (Fig. 2o and p). The extent of garnet involvement in the reaction is unclear, as the symplectite replaces the orthopyroxene, and compatibility relations suggest it may not be a necessary reactant (Fig. 9). Orthopyroxene in symplectite is always less aluminous than the porphyroblastic orthopyroxene (Table 1). We infer that the following FMAS reactions are recorded by these microstructures (Fig. 8b):
 |
 |
 |
with reaction progress (Fig. 9). In effect, there are domains in which predominantly orthopyroxene has reacted, but only where in contact with garnet, and domains in which garnet and orthopyroxene together have reacted, in all cases to produce symplectic intergrowths. These microstructures probably reflect concurrent operation of several FMAS reactions that vary from domain to domain.
Orthopyroxene–sapphirine reacted with melt to produce granular spinel–cordierite–biotite (Fig. 2q; Table 1), via the (dis)continuous KFMASH reaction
 |
Reaction (13) emanates from the [Grt, Qtz] invariant in the KFMASH model system, which invariant is shown as stable in the petrogenetic grid proposed by Hensen & Harley (1990). Recently, McDade & Harley (2001) have proposed an alternative topology for the KFMASH model system in which the [Grt, Qtz] invariant is considered to be metastable. Microstructural relations recorded by sample PT-62-F in this study are inconsistent with the alternative topology of McDade & Harley (2001), but appear to be consistent with the topology proposed by Hensen & Harley (1990). Until new experimental data become available, both topologies are potentially acceptable, and the topology of Hensen & Harley (1990) is followed in this work.
In general, rutile is the Ti phase included in garnet and orthopyroxene, whereas in all symplectites ilmenite is present instead, including some symplectites of ilmenite and orthopyroxene at orthopyroxene porphyroblast rims, which suggests that at lower P and T ilmenite is the stable Ti-rich phase. This rock does not appear to have recorded any further information about the lower-temperature part of the P–T evolution.
Quartz-rich rock type
The quartz-rich rock (sample PT-62-A) is from the same outcrop as PT-62-F, and although the outcrop is not continuous, we infer that these samples are likely to have followed the same P–T path. As in the case of the quartz-poor sample, no reliable microstructural relations that might be related to the prograde stage are preserved. However, biotite dehydration melting is inferred to have occurred, leaving residual F–Ti-rich biotite (F 1·0–0·58 a.p.f.u. and Ti 0·75–0·57 a.p.f.u., Table 1) and possibly residual garnet, and peritectic orthopyroxene and sillimanite. Some crystallization of garnet occurred concomitantly with or later than sapphirine–quartz, as sapphirine occurs in contact with quartz as inclusions in garnet (Fig. 2r), but only rarely as part of the matrix assemblage. We infer that sillimanite was involved as a reactant, because it occurs as anhedral inclusions in garnet. We interpret these features to suggest that the FMAS univariant reaction (1) was crossed in a clockwise sense (Fig. 8c).
Mineral aggregates dominated by cordierite in association with sillimanite, spinel and rutile (Fig. 2s, t and u) are inferred to record the FMAS divariant reaction (6) or, possibly, the FMAS univariant reaction with sapphirine if any sapphirine remained in the assemblage. Sillimanite required by reaction (6) was present as inclusions in garnet and most probably in the matrix, but garnet also includes sapphirine with quartz so that some sillimanite could have been produced with cordierite from sapphirine and quartz in the matrix by the FMAS divariant reaction
 |
Subsequently, garnet rims were replaced by an orthopyroxene–cordierite (± ilmenite) corona (Fig. 2v), inferred to record the FMAS divariant reaction
 |
 |
Migmatite
In the metapelitic gneiss (ANA-287), no sapphirine has been recognized. Cordierite with spinel and sillimanite occurs in a microstructure where host cordierite has partially replaced porphyroblasts of garnet, which may imply the sequential operation of reactions (14) and (6). Thus, sapphirine and quartz may have occurred in ANA-287, and we expect that rocks at this locality followed a similar P–T evolution to the samples from locality PT-62. Alternatively, this microstructure may record the FMAS divariant reaction between garnet and sillimanite to produce cordierite and spinel. Samples from this locality also allow the lower-T stage of the evolution to be evaluated, using reactions inferred to have occurred between Fe–Mg phases and melt, the remnants of which are now represented by leucosome layers.
A thin melanosome composed of biotite (± garnet) (Fig. 2ab) and small grains of garnet associated with biotite in the matrix (Fig. 2ac) are inferred to have been generated by reaction of orthopyroxene and cordierite with melt, according to the KFMASH univariant reaction (16). Inside leucosome, biotite and sillimanite overgrew garnet and cordierite (Fig. 2aa), as a result of reaction with melt according to the univariant KFMASH reaction (Fig. 8c)
 |
Thermobarometry
Thermodynamic data for sapphirine, particularly the entropy, are not well constrained (Harley & Motoyoshi, 2000), which limits the use of internally consistent thermodynamic datasets in thermobarometry of sapphirine-bearing assemblages. Additionally, diffusive ion exchange during cooling modifies peak mineral compositions and limits maximum retrievable temperatures. None the less, methods have been developed to maximize the temperature retrieved by back-calculating peak mineral compositions (Fitzsimons & Harley, 1994; Pattison & Bégin, 1994) and utilizing thermobarometric calibrations for garnet–orthopyroxene. We have used the Al-orthopyroxene–garnet barometers (Harley & Green, 1982; Harley, 1984b) and the Fe–Mg garnet–orthopyroxene thermometer (Harley, 1984a), because they are based on the same set of experiments. The regression method of Pattison & Bégin (1994) was used to retrieve close-to-peak mineral compositions. Calculated P–T conditions are listed in Table 9, and plotted in Fig. 10.
|
|
Core compositions of two garnet–orthopyroxene pairs were chosen from ML-67 for thermobarometric calculations (Table 9). Calculated P is 9·1 and 9·4 kbar with T of 1000°C and 1012°C, respectively (±2 kbar and ±40–60°C; assuming ±0·1 is the maximum error in the calculated KD; Harley, 1984a, 1984b). The mineral pair sapphirine–quartz occurs in mutual contact in this rock, which requires a minimum T of 1030–1050°C in the FMAS model system. The retrograde evolution requires peak pressure to have been greater than the [Spl] invariant or 10 kbar, based on the stability of the mineral assemblage garnet–orthopyroxene–sillimanite–quartz in the FMAS model system. These results are consistent within the uncertainties of the analytical data, the thermobarometers and the grids.
In the quartz-rich rock (PT-62-A), the chemical composition of garnet porphyroblasts is disturbed around multigrain quartz–plagioclase inclusions, which are inferred to have crystallized from inclusions of melt resulting in Ca depletion and modified Fe and Mg (Fig. 3). Calculated P of 8·1 kbar for this rock leads to calculated T of 960°C. In the quartz-poor rock (PT-62-F), no Ca perturbation was observed in garnet and P–T retrieved are 8·5 kbar and 975°C. The mineral pair sapphirine–quartz occurs in mutual contact in the quartz-rich rock, which requires a minimum T of 1030–1050°C at 10 kbar based on the FMAS model system. We conclude that thermobarometry based on samples from locality PT-62 fails to retrieve metamorphic peak conditions consistent with those estimated from the mineral assemblages.
In ANA-287, one garnet porphyroblast was used to calculate P and T against two porphyroblasts of orthopyroxene. Calculated values are 9·6 kbar and 960°C and 9·7 kbar and 975°C, respectively. In this sample, the presence of the mineral pair sapphirine–quartz was inferred from the breakdown products of reaction between them, and subsequent reaction to produce cordierite and spinel. If this is a correct inference, the presence of sapphirine–quartz would require a minimum temperature of 1030–1050°C at 10 kbar based on the FMAS model system.
Aluminium content in orthopyroxene in the FMAS divariant assemblages Opx + Sil + Grt + Qtz, Opx + Grt + Spr + Qtz and Opx + Crd + Grt + Qtz can be used to estimate temperatures in the samples with quartz, using Al2O3 isopleths in P–T space as calculated using internally consistent datasets and from combinations of experimental data and interpolation (Aranovich & Berman, 1996; Harley, 1998c; Harley & Motoyoshi, 2000). Estimated temperatures based on isopleths for the Al2O3 content in orthopyroxene in the FMAS assemblage Opx + Grt +Spr + Qtz (from Harley & Motoyoshi, 2000) are listed in Table 9, and shown in Fig. 10.
Maximum Al2O3 in orthopyroxene in ML-67 and PT-62-A is 12·9 and 11·9 wt %, respectively, which suggests temperatures apparently higher than 1150°C, although these will be maxima because of presence of minor Fe3+ in orthopyroxene. In ANA-287, the maximum Al2O3 content in orthopyroxene is 9·7 wt %, which implies a maximum temperature around 1070°C. PT-62-F is quartz poor, so we have not calculated a temperature based on Al2O3 in orthopyroxene, but the maximum Al2O3 content in orthopyroxene cores in this rock is 10·9 wt %, which implies higher temperatures than those calculated using Fe–Mg exchange between orthopyroxene and garnet. Although these calculated temperatures are maxima, they are significantly above the minimum of 1030–1050°C implied by sapphirine–quartz stability and the position of the [Spl] invariant in the FMAS model system.
|
DISCUSSION OF THE ROLE OF MELT IN THE EVOLUTION OF THE AIC UHT GRANULITES |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGICAL SETTING OF...PETROGRAPHYMINERAL CHEMISTRYINTERPRETATION OF...DISCUSSION OF THE ROLE...TECTONIC SETTINGCONCLUSIONSREFERENCES |
|---|
Samples from the two areas of study appear to show different amounts of retrogression to ‘common’ granulite- and amphibolite-facies mineral assemblages. We suggest that this might relate to different amounts of residual melt retained by the samples, and we evaluate this postulate below.
The impure quartzite (ML-67) has a simple mineral assemblage (sapphirine, quartz, garnet, orthopyroxene, sillimanite) that is consistent with biotite dehydration melting having occurred during prograde-to-peak metamorphism, but no feldspar or evidence of leucosome is present. It is not clear whether the bulk-rock composition was inappropriate to produce melt or if melt was generated and essentially extracted completely. The peak UHT assemblage is preserved and the rock records very little information about the retrograde P–T evolution, as garnet, orthopyroxene, sillimanite and quartz appear to be stable along the high-temperature part of the retrograde segment of the P–T path. The small amount of biotite in this rock suggests that if melt was produced on the prograde path, it was mostly lost from this sample before significant retrogression could occur, and only a minimum amount of biotite was formed during cooling.
In outcrop PT-62, UHT assemblages in the quartz-rich rock (PT-62-A) are controlled by local bulk composition, as the rock has a gneissic structure defined by different proportions of garnet and orthopyroxene from layer to layer. It is not possible to distinguish the presence of macroscopic leucosome and K-feldspar is only rarely present, but plagioclase appears to mimic melt–solid contacts (Fig. 2u–w). However, although plagioclase is widespread, there is no apparent connection between the grains, and no implication that an interconnected melt network is preserved. This implies that a significant proportion of melt was extracted, possibly leaving behind pockets of residual melt (compare Sawyer, 2001).
The bulk composition of the quartz-poor rock (PT-62-F) is unusual in comparison with typical metapelite bulk compositions; it produces an assemblage composed essentially of garnet and orthopyroxene porphyroblasts with cordierite and minor amounts of plagioclase and other refractory phases (sapphirine, corundum, spinel, and rutile), but no K-feldspar. If melt was generated from this rock, a significant amount is inferred to have been extracted, leaving behind a refractory composition.
Retrograde biotite (Table 1) was produced in both the quartz-rich and quartz-poor rocks by consumption of UHT minerals until residual melt was exhausted. Under these conditions of partial retrogression, the high-temperature part of the retrograde segment of the P–T path can be established with confidence using microstructural relations that record incomplete reaction progress. However, in the case of these samples, the fact that the high-temperature retrograde evolution occurred at P below the [Spl] invariant also places this crust in a reactive part of P–T space.
Outcrop at the ANA-287 locality is migmatitic, and differences in bulk composition from one layer to another mean that UHT mineral assemblages occur only in some layers. UHT conditions are inferred from the replacement of garnet by cordierite with spinel and sillimanite. Metapelitic gneiss layers with UHT mineral assemblages do not contain K-feldspar and may be melt depleted. Reaction between melt (inferred from the presence of leucosomes) and metapelitic gneiss layers is marked by thin selvedges of biotite, sometimes with minor amounts of garnet. These represent late melanosomes formed during cooling by a melt-consuming reaction (Kriegsman, 2001; Brown, 2002). Additional biotite and sillimanite were formed in the same way. Preservation of garnet and cordierite porphyroblasts within leucosome without significant retrograde reaction indicates that some melt loss occurred, as proposed for similar features in other metamorphic belts (Powell & Downes, 1990; Brown, 1998). Late muscovite and symplectites of garnet, biotite or muscovite with quartz are widely distributed within leucosome and metapelitic gneiss layers, and reflect continued reaction with melt during cooling (Kriegsman, 2001; Brown, 2002).
On the basis of a progression in samples from the AIC, in which the rocks with evidence for a higher proportion of retained melt are more extensively retrogressed, we suggest that many of the enclosing ‘common’ granulites may also have experienced the UHT metamorphic conditions and that any record of such conditions has been erased by extensive retrogression. This may reflect diffusion fast enough at temperatures above 900°C that equilibrium domains are not isolated but ‘overlap’; that is, most of the system under consideration is within a common equilibration domain.
Using pseudosections in the NCKFMASH model system, White et al. (2001) and White & Powell (2002) have demonstrated that in a closed system (i.e. one from which no melt was lost) granulite-facies mineral assemblages coexisting with melt should be reacted out by the solidus. To avoid complete retrogression, some melt loss or separation from residual mineral assemblages must occur (e.g. Powell & Downes, 1990; Kriegsman, 2001; Brown, 2002; White & Powell, 2002). In the case of melt segregation rather than loss, retrogression depends on the size of the equilibration volume (Stüwe, 1997), which decreases with falling temperature. Segregation followed by melt loss is common in granulites, consistent with depleted bulk compositions (Fyfe, 1973) and residual mineral assemblages and microstructures that mimic melt–solid relations (Sawyer, 2001). On the basis of the pseudosections of White & Powell (2002), loss of 80 vol. % of the melt generated from an aluminous metapelite is necessary to preserve a granulite-facies mineral assemblage, although a granulite-facies mineral assemblage can be preserved with only 40% melt loss in a sub-aluminous metapelite.
Alternatively, preservation of UHT mineral assemblages may demand a more refractory protolith, such as greywacke or hydrothermally altered basalt, as suggested by the difference between the aluminous and sub-aluminous metapelite compositions modelled by White & Powell (2002). Of course, it is probable that some horizons within the granulites have inappropriate bulk-rock compositions to record mineralogical evidence of UHT conditions, and it remains possible that there is an unrecognized component of tectonic interleaving between slivers of UHT rocks and ‘common’ granulites.
As a simple test of the hypotheses discussed above, we suggest that detailed studies of granulites on a regional scale throughout the Brasília Fold Belt should produce many more examples of relict UHT mineral assemblages. Furthermore, detailed mapping in critical areas should allow the relationship between UHT granulites and ‘common’ granulites to be established; that is, to ensure continuity or identify tectonic interleaving.
P–T paths
For the sample from Goiânia (ML-67), the P–T path at the metamorphic peak is likely to have been steep, involving decompression along a path approximately parallel to an isopleth of constant Al2O3 in orthopyroxene. The post-metamorphic peak P–T path is constrained by the stability of the mineral assemblage garnet–orthopyroxene–sillimanite–quartz, and the absence of assemblages with cordierite. These features require the high-temperature retrograde evolution to follow a path at P > 10–9 kbar for the temperature interval down to <900°C (Fig. 11, dark grey path).
|
For samples expected to have a common P–T evolution, a composite P–T path may be established using reactions inferred from microstructural relations between mineral assemblages in rocks of different bulk composition, or those involving variable melt loss and retrograde reaction. The procedure has been applied to establish P–T paths with great success in several UHT terranes (e.g. Harley et al., 1990; Brown & Raith, 1996; Raith et al., 1997; Moraes & Fuck, 2000). The composite post-thermal peak P–T path constructed for the rocks from Damolândia (PT-62-A and PT-62-F) and the Monjolo stream (ANA-287), based on the observed microstructural relations and reaction relations in the FMAS model system, is shown in Fig. 11 (light grey path). It is characterized by a decompression segment from >12 to <8 kbar at 1100–1000°C, followed by a close-to-isobaric cooling segment from 1000°C to <600°C at <8 to <6 kbar.
|
TECTONIC SETTING |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGICAL SETTING OF...PETROGRAPHYMINERAL CHEMISTRYINTERPRETATION OF...DISCUSSION OF THE ROLE...TECTONIC SETTINGCONCLUSIONSREFERENCES |
|---|
One problem still to be fully understood in granulite terranes, especially those that include UHT metamorphic rocks, is the cause of the thermal perturbation responsible for the metamorphism. Magmatic underplating or emplacement of a large volume of basaltic magma has been proposed as a cause of additional heat for granulite-facies metamorphism and melting, mainly in circumstances where near-isobaric cooling paths are recorded by granulites (England & Richardson, 1977; Ellis, 1980; Wells, 1980; Bohlen, 1987, 1991; Bergantz, 1989). An example where granulite-facies metamorphism has been attributed to magmatic underplating is the Ivrea Zone, Italy. However, recent work has shown that the Mafic Complex, the putative heat source, immediately below the granulites, cuts across the regionally developed metamorphic isograds and imposes a contact metamorphic aureole on amphibolite-facies metasedimentary rocks in the roof, indicating that final emplacement occurred after the regional metamorphic peak (Barboza et al., 1999; Barboza & Bergantz, 2000). A significant problem with the underplating model is the volume of magma that needs to be accreted to continental crust to achieve the necessary thermal profile, which must be similar to the thickness of the crust that is to be metamorphosed (Oxburgh, 1990).
The origin of some granulite terranes has been related to collisional tectonics (e.g. Ellis, 1987; Raith et al., 1997). In general, numerical models of metamorphism as a result of thickening and thermal relaxation do not generate granulite-facies conditions at appropriate depths (e.g. Thompson & England, 1984; England & Thompson, 1986). This indicates that an additional heat source is necessary (Harley, 1989), and the problem resolves to a consideration of what additional heat sources may be involved during collisional tectonics. It has been argued that sufficient additional heat could be generated to drive granulite-facies metamorphism and crustal melting by accretion of radioactive material and concurrent erosion in orogens developed by basal traction at convergent plate margins (Harley, 1989; Jamieson et al., 1998, 2002; Huerta et al., 1999). However, the extreme thermal perturbation required by UHT metamorphism appears to be unachievable without additional heat from the asthenosphere.
The lithosphere is a thermal boundary layer part of which may be removed convectively (e.g. England, 1987, 1993; Anovitz & Chase, 1990; Platt & England, 1994). Such a process may be facilitated by collisional thickening of continental crust, which may be decoupled from lithospheric mantle, to generate a thick, dense and unstable orogenic root (Houseman et al., 1981; Sonder et al., 1987). Modelling studies suggest that after 10–50 Myr such an orogenic root may founder, detach and sink through the asthenosphere (Molnar et al., 1993; Houseman & Molnar, 1997). Detachment might occur by ductile necking, or by delamination (Bird, 1979; Schott & Schmeling, 1998). All scenarios predict syn-orogenic or post-orogenic lithospheric thinning, uplift, near-surface extension and substantial mantle strain. Thus, the continental lithosphere may be thinned and part of the lithospheric mantle may be replaced by the asthenosphere while the continental crust is still being thickened (Sandiford & Powell, 1990; Liu, 2001). Subsequent extensional collapse of the mountain belt is likely to be accompanied by granulite-facies metamorphism, decompression melting and generation of granites (Thompson & England, 1984; Sandiford, 1989; Sandiford & Powell, 1990; Ledru et al., 2001). The thermal peak of metamorphism will be attained during or after a decompression phase and, after isostatic equilibrium is re-established, rocks in the lower continental crust will cool isobarically (Harley, 1989; Sandiford, 1989).
An alternative way to bring the asthenosphere closer to the base of continental crust in convergent orogens is by slab break-off (Davies & von Blanckenburg, 1995; von Blanckenburg & Davies, 1995; Wortel & Spakman, 2000). Such a model predicts that after collision the slab may become detached and sink into the mantle, allowing hot asthenosphere to be emplaced beneath the thickened orogenic crust. There is a tectonic, metamorphic and magmatic response similar to the case of delamination. Thus, the thermal peak occurs during or after a decompression phase, and rocks in the lower continental crust follow a near-isobaric cooling after isostatic equilibrium is achieved.
The Brasília Fold Belt is recognized as a Neoproterozoic collisional orogen (Pimentel et al., 2000). Evolution of the Brasília Fold Belt is thought to involve closure of an ocean basin, loss of oceanic lithosphere by subduction at the Goiás magmatic arc and collision between the Amazon and São Francisco cratons. In the AIC, meta-igneous rocks are abundant; they have chemical characteristics that fit those required by models in which asthenospheric mantle is involved. The meta-igneous rocks in the AIC were affected by the high-grade metamorphism, which indicates that magmatism pre-dated or was contemporaneous with UHT metamorphic conditions. Precise ages for the crystallization of igneous protoliths or for the UHT metamorphism are not yet available. The P–T paths for the two areas of UHT granulite in the AIC both have a near-isothermal decompression stage, when the thermal peak was attained, followed by a near-isobaric cooling stage. P–T paths of this kind are similar to those generated in the lower plate of collision zones (Ellis, 1987). In this setting, slab break-off or detachment of an orogenic root could have been responsible for allowing asthenospheric mantle closer to the base of the thickened continental crust. In such a tectonic scenario, the asthenosphere is the source of the thermal perturbation recorded by the UHT metamorphism.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGICAL SETTING OF...PETROGRAPHYMINERAL CHEMISTRYINTERPRETATION OF...DISCUSSION OF THE ROLE...TECTONIC SETTINGCONCLUSIONSREFERENCES |
|---|
Coexisting sapphirine–quartz in garnet–orthopyroxene–sillimanite granulites from two areas within ‘common’ granulites of the Anápolis–Itauçu Complex (AIC) demonstrates that UHT metamorphic conditions were achieved in the Brasília Fold Belt during the Neoproterozoic Brasiliano Orogeny. Minimum conditions at the metamorphic peak were 1030–1050°C at
10 kbar, but Al2O3 content of orthopyroxene suggests temperatures at the metamorphic peak exceeded 1100°C, and possibly exceeded 1150°C. Additionally, the AIC appears to have followed a clockwise P–T path, so that maximum pressure was likely to have been substantially greater than 10 kbar and most probably pre-dated the attainment of maximum temperature.
The retrograde reaction history inferred from microstructural relations indicates decompression followed by near-isobaric cooling after the metamorphic peak, but samples from two areas 22 km apart followed retrograde paths separated by 3 kbar. Down-temperature reaction with residual melt may partially to completely overprint evidence of the UHT metamorphic history, and ‘common’ granulites may represent rocks where evidence of the UHT history has been obliterated along the retrograde segment of the P–T path.
It is likely that the Brasília Fold Belt was formed during closure of an ocean basin and consequent collision between the Amazon Craton, the Goiás magmatic arc and the São Francisco Craton. The extreme thermal perturbation required to generate UHT conditions of the AIC may have been due to replacement of mantle lithosphere by asthenosphere following detachment of an orogenic root or slab break-off during collision.
|
ACKNOWLEDGEMENTS |
|---|
We thank Philip M. Piccoli for his help in the Laboratory for Microscopy and Microanalysis; the EPMA used in this study was purchased with grants from Department of Defense-Army/ARO (DAAG 559710383) and NSF (EAR-9810244), and UMD funding. The help of Jeanne Martin during final preparation of this manuscript is appreciated. We acknowledge stimulating discussions with Nathalie Marchildon, Thomas Zack, Roger Powell, Paddy O’Brien, Ed Grew and Bas Hensen. Constructive journal reviews by Simon Harley, Hassina Mouri and Yasu Osanai were appreciated. We acknowledge differences of opinion in the interpretation of the more ambiguous microstructures and we hope that we have expressed our interpretation with greater clarity in the revised paper. Cathodoluminescence images were obtained at George Washington University by kind permission of John Hanchar. Danielle Piuzana is thanked for discussion and help during fieldwork. R.M. acknowledges a post-doctoral fellowship from CNPq, Brazil (20.0602/00-4) held at the Laboratory for Crustal Petrology at University of Maryland; R.M. and M.B. acknowledge support from the University of Maryland. Fieldwork was supported by CNPq, Brazil (grants 42.0081/99-2 and 46.0408/00-2 to R.A.F.).
|
FOOTNOTES |
|---|
*Corresponding author. Telephone: (301) 405-4080. Fax: (301) 314-7970. E-mail: mbrown{at}geol.umd.edu
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL GEOLOGICAL SETTING OF...PETROGRAPHYMINERAL CHEMISTRYINTERPRETATION OF...DISCUSSION OF THE ROLE...TECTONIC SETTINGCONCLUSIONSREFERENCES |
|---|
Ackermand, D., Seifert, F. & Schreyer, W. (1975). Instability of sapphirine at high pressures. Contributions to Mineralogy and Petrology 50, 79–92.[Web of Science]
Anovitz, L. M. & Chase, C. G. (1990). Implications of post-thrusting extension and underplating for P–T paths in granulite terrains: a Greenville example. Geology 18, 466–469.
Aranovich, L. Ya. & Berman, R. G. (1996). Optimized standard state and solution properties of minerals. II. Comparisons, predictions, and applications. Contributions to Mineralogy and Petrology 126, 25–37.[Web of Science]
Barboza, S. A. & Bergantz, G. W. (2000). Metamorphism and anatexis in the mafic complex contact aureole, Ivrea Zone, northern Italy. Journal of Petrology 41, 1307–1327.
Barboza, S. A., Bergantz, G. W. & Brown, M. (1999). Regional granulite facies metamorphism in the Ivrea zone: is the Mafic Complex the smoking gun or a red herring? Geology 27, 447–450.
Barnicoat, A. C. & O’Hara, M. J. (1979). High temperature pyroxenes from an ironstone at Scourie, Sutherland. Mineralogical Magazine 43, 371–375.[Web of Science]
Bergantz, G. W. (1989). Underplating and partial melting; implications for melt generation and extraction. Science 245, 1093–1095.
Bertrand, P., Ellis, D. J. & Green, D. H. (1991). The stability of sapphirine–quartz and hypersthene–sillimanite–quartz assemblages: an experimental investigation in the system FeO–MgO–Al2O3–SiO2 under H2O and CO2 conditions. Contributions to Mineralogy and Petrology 108, 55–71.[Web of Science]
Bird, P. (1979). Continental delamination and the Colorado Plateau. Journal of Geophysical Research 84, 7561–7571.
Bohlen, S. R. (1987). Pressure–temperature time paths and a tectonic model for the evolution of granulites. Journal of Geology 95, 617–632.[Web of Science]
Bohlen, S. R. (1991). On the formation of granulites. Journal of Metamorphic Geology 9, 223–229.[Web of Science]
Brown, M. (1998). Unpairing metamorphic belts: P–T paths and a tectonic model for the Ryoke Belt, southwest Japan. Journal of Metamorphic Geology 16, 3–22.[Web of Science]
Brown, M. (2001). From microscope to mountain belts: 150 years of petrology and its contribution to understanding geodynamics, particularly the tectonics of orogens. Journal of Geodynamics 32, 115–164.[Web of Science]
Brown, M. (2002). Retrograde processes in migmatites and granulites revisited. Journal of Metamorphic Geology 20, 25–40.[Web of Science]
Brown, M. & Dallmeyer, R. D. (1996). Rapid Variscan exhumation and role of magma in core complex formation: Southern Brittany metamorphic belt, France. Journal of Metamorphic Geology 14, 361–379.[Web of Science]
Brown, M. & Raith, M. (1996). First evidence of ultrahigh-temperature decompression from the granulite province of southern India. Journal of the Geological Society, London 153, 819–822.
Carrington, D. P. & Harley, S. L. (1995a). Partial melting and phase relations in high-grade metapelites: an experimental petrogenetic grid in the KFMASH system. Contributions to Mineralogy and Petrology 120, 270–291.[Web of Science]
Carrington, D. P. & Harley, S. L. (1995b). The stability of osumilite in metapelitic granulites. Journal of Metamorphic Geology 13, 613–625.[Web of Science]
Cooke, R. A. & O’Brien, P. J. (2001). Resolving the relationship between high P–T rocks and gneisses in collisional terranes: an example from the Gföhl gneiss–granulite association in the Moldanubian Zone, Austria. Lithos 58, 33–54.[Web of Science]
Dallwitz, W. B. (1968). Co-existing sapphirine and quartz in granulite from Enderby Land, Antarctica. Nature 219, 476–477.
Dasgupta, S., Sengupta, P., Ehl, J., Raith, M. & Bardhan, S. (1995). Reaction textures in a suite of spinel granulites from the Eastern Ghats Belt, India: evidence for polymetamorphism, a partial petrogenetic grid in the system KFMASH and the roles of ZnO and Fe2O3. Journal of Petrology 36, 435–461.
Davies, J. H. & von Blanckenburg, F. (1995). Slab breakoff: a model of lithosphere detachment and its test in the magmatism and deformation of collisional orogens. Earth and Planetary Science Letters 129, 85–102.[Web of Science]
Ellis, D. J. (1980). Osumilite–sapphirine–quartz granulites from Enderby Land, Antarctica: P–T conditions of metamorphism, implications for garnet–cordierite equilibria and the evolution of deep crust. Contributions to Mineralogy and Petrology 74, 201–210.[Web of Science]
Ellis, D. J. (1987). Origin and evolution of granulites in normal and thickened crusts. Geology 15, 167–170.
England, P. C. (1987). Diffusive continental deformation: length scales, rates and metamorphic evolution. Philosophical Transactions of the Royal Society of London, Series A 321, 3–22.
England, P. C. (1993). Convective removal of thermal boundary layer of thickened continental lithosphere: a brief summary of causes and consequences with special reference to the Cenozoic tectonics of the Tibetan Plateau and surrounding regions. Tectonophysics 223, 67–73.[Web of Science]
England, P. C. & Richardson, S. W. (1977). The influence of erosion upon the mineral facies of rocks from different metamorphic environments. Journal of the Geological Society, London 134, 201–213.
England, P. C. & Thompson, A. B. (1986). Some thermal and tectonic models for crustal melting in continental collision zones. In: Coward, M. P. & Ries, A. C. (eds) Collision Tectonics. Geological Society, London, Special Publications 19, 83–94.
Fischel, D. P., Pimentel, M. M. & Fuck, R. A. (1998). Idade do metamorfismo de alto grau no Complexo Anápolis–Itauçu, Goiás, determinada pelo método Sm–Nd. Revista Brasileira de Geociências 28, 543–544.
Fischel, D. P., Fuck, R. A. & Pimentel, M. M. (1999a). Sm–Nd mineral isochrons and model ages of the Anápolis–Itauçu Complex and associated granites, Goiás: new insights into the geological evolution of the Brasília Belt. Boletim de Resumos Expandidos do 7° Simpósio de Geologia do Centro Oeste e 10° Simpósio de Geologia de Minas Gerais, Brasilia, SBG, p. 88.
Fischel, D. P., Pimentel, M. M. & Fuck, R. A. (1999b). Preliminary Sm–Nd isotopic study of the Araxá Group, central Goiás. In: Simpósio de Geologia do Centro Oeste, 7, Simpósio de Geologia de Minas Gerais, 10, Brasília, 1999. Boletim de Resumos. SBG, p. 87.
Fitzsimons, I. C. W. & Harley, S. L. (1994). The influence of retrograde cation exchange on granulite P–T estimates and a convergence technique for the recovery of peak metamorphic conditions. Journal of Petrology 35, 543–576.
Fuck, R. A., Pimentel, M. M. & Silva, L. J. H. D. (1994). Compartimentação tectônica da porção oriental da Província Tocantins. 38° Congresso Brasileiro de Geologia, Balneário Camboriú, SBG 1, 215–216.
Fyfe, W. S. (1973). The granulite facies, partial melting and the Archaean crus. Philosophical Transactions of the Royal Society of London, Series A 273, 457–461.
Guernina, S. & Sawyer, E. W. (2002). Large-scale melt-depletion in granulite terranes: an example from the Archaean Ashuanipi Subprovince of Quebec. Journal of Metamorphic Geology 20, 000–000.
Guidotti, C. V. (1984). Micas in metamorphic rocks. In Bailey S. W. (ed.) Mineralogical Society of America, Reviews in Mineralogy 13, 357–468. America)
Hanmer, S. (2000). Matrix mosaics, brittle deformation, and elongate porphyroclasts: granulite facies microstructures in the Striding–Athabasca mylonite zone, western Canada. Journal of Structural Geology 22, 947–967.[Web of Science]
Harley, S. L (1984a). An experimental study of the partitioning of the Fe and Mg between garnet and orthopyroxene. Contributions to Mineralogy and Petrology 86, 359–373.[Web of Science]
Harley, S. L. (1984b). The solubility of alumina in orthopyroxene coexisting with garnet in FeO–MgO–Al2O3–SiO2 and CaO–FeO–MgO–Al2O3–SiO2. Journal of Petrology 25(3), 665–696.
Harley, S. L. (1985). Garnet–orthopyroxene-bearing granulites from Enderby Land, Antarctica: metamorphic pressure–temperature–time evolution of the Archaean Napier Complex. Journal of Petrology 26, 819–856.
Harley, S. L. (1989). The origins of granulites: a metamorphic perspective. Geological Magazine 126, 215–331.[Abstract]
Harley, S. L. (1992). Proterozoic granulite terranes. Precambrian Geology 10, 301–360.
Harley, S. L. (1998a). On the occurrence and characterization of ultrahigh-temperature crustal metamorphism. In: Treloar, P. J. & O’Brien, P. J. (eds) What Drives Metamorphism and Metamorphic Reactions? Geological Society, London, Special Publications 138, 81–107.
Harley, S. L. (1998b). An appraisal of peak temperatures and thermal histories in ultrahigh-temperature (UHT) crustal metamorphism: the significance of aluminous orthopyroxene. Memoirs of National Institute of Polar Research, Special issue 53, 49–73.
Harley, S. L. (1998c). Ultrahigh temperature granulite metamorphism (1050°C, 12 kbar) and decompression in garnet (Mg70)–orthopyroxene–sillimanite gneisses from Rauer group, East Antarctica. Journal of Metamorphic Geology 16, 541–562.[Web of Science]
Harley, S. L. & Green, D. H. (1982). Garnet–orthopyroxene barometry for granulites and peridotites. Nature 300, 697–701.
Harley, S. L. & Hensen, B. J. (1990). Archaean and Proterozoic high-grade terranes of East Antarctica (40–80°E): a case study of diversity in granulite facies metamorphism. In: Ashworth, J. R. & Brown, M. (eds) High Temperature Metamorphism and Crustal Anatexis. London: Unwin Hyman, pp. 320–370.
Harley, S. L. & Motoyoshi, Y. (2000). Al zoning in orthopyroxene in sapphirine quartzite: evidence for >1120°C UHT metamorphism in the Napier Complex, Antarctica, and implications for the entropy of sapphirine. Contributions to Mineralogy and Petrology 138, 293–307.[Web of Science]
Harley, S. L., Hensen, B. J. & Sheraton, J. W. (1990). Two-stage decompression in orthopyroxene–sillimanite granulites from Forefinger Point, Enderby Land, Antarctica: implications for the evolution of the Archaean Napier Complex. Journal of Metamorphic Geology 8, 591–613.[Web of Science]
Hensen, B. J. (1987). P–T grids for silica-undersaturated granulites in the systems MAS (N + 4) and FMAS (N + 3)—tools for the derivation of P–T paths of metamorphism. Journal of Metamorphic Geology 5, 255–271.[Web of Science]
Hensen, B. J. & Green, D. H. (1973). Experimental study of the stability of cordierite and garnet in pelitic compositions at high pressures and temperatures. III. Synthesis of experimental data and geological applications. Contributions to Mineralogy and Petrology 38, 151–166.[Web of Science]
Hensen, B. J. & Harley, S. L. (1990). Graphical analysis of P–T–X relations in granulite facies metapelites. In: Ashworth, J. R. & Brown, M. (eds) High Temperature Metamorphism and Crustal Anatexis. London: Unwin Hyman, 427–457.
Hensen, B. J. & Osanai, Y. (1994). Experimental study of dehydration melting of F-bearing biotite in model pelitic compositions. Mineralogical Magazine 58A, 410–411.
Hokada, T. (2001). Feldspar thermometry in ultrahigh-temperature metamorphic rocks: evidence of crustal metamorphism attaining 1100°C in the Archean Napier Complex, East Antarctica. American Mineralogist 86, 932–938.
Holland, T. J. B., Babu, E. V. S. S. K. & Waters, D. J. (1996). Phase relations of osumilite and dehydration melting in pelitic rocks: a simple thermodynamic model for KFMASH system. Contributions to Mineralogy and Petrology 124, 383–394.[Web of Science]
Houseman, G. A. & Molnar, P. (1997). Gravitational (Rayleigh–Taylor) instability of a layer with non-linear viscosity and convective thinning of continental lithosphere. Geophysical Journal International 128, 125–150.[Web of Science]
Houseman, G. A., McKenzie, D. P. & Molnar, P. (1981). Convective instability of a thickened boundary layer and its relevance for the thermal evolution of continental convergent belts. Journal of Geophysical Research 86, 6115–6132.
Huerta, A. D., Royden, L. H. & Hodges, K. V. (1999). The effects of accretion, erosion and radiogenic heat on the metamorphic evolution of collisional orogens. Journal of Metamorphic Geology 17, 349–366.[Web of Science]
Jamieson, R., Beaumont, C., Pulsack, P. & Lee, B. (1998). Barrovian regional metamorphism: where’s the heat? In: Treloar, P. J & O’Brien, P. (eds) What Drives Metamorphism and Metamorphic Reactions? Geological Society, London, Special Publications 138, 23–51.
Jamieson, R., Beaumont, C., Nguyen, M. H. & Lee, B. (2002). Interaction of metamorphism, deformation and exhumation in large convergent orogens. Journal of Metamorphic Geology 20, 9–24.[Web of Science]
Kretz, R. (1983). Symbols for rock-forming minerals. American Mineralogist 68, 277–279.[Abstract]
Kriegsman, L. M. (2001). Partial melting, partial melting extraction and partial back reaction in anatectic migmatites. Lithos 56, 75–96.[Web of Science]
Lacerda Filho, J. V. & Oliveira, C. C. (1995). Geologia da região centro-sul de Goiás. Boletim de Geociências do Centro-Oeste 18(1/2), 3–19.
Lacerda Filho, J. V., Marques, V., Szislewiski, G., Jorge, L., Justo, E. C. & Oliveira, C. C. (1991). Projeto Geologia da Região Centro-Oeste, Folha Caraíba. Goiânia: CPRM, 145 pp.
Laporte, D., Rapaille, C. & Provost, A. (1997). Wetting angles, equilibrium melt geometry, and the permeability threshold of partially-molten crustal protoliths. In: Bouchez, J. L., Hutton, D. H. W. & Stephens, W. E. (eds) Granites: from Segregation of Melt to Emplacement Fabric. Dordrecht: Kluwer Academic, pp. 31–54.
Ledru, P., Courrioux, G., Dallain, C., Lardeaux, J. M., Montel, J. M., Vanderhaeghe, O. & Vitel, G. (2001). The Velay dome (French Massif Central): melt generation and granite emplacement during orogenic evolution. Tectonophysics 342, 207–237.[Web of Science]
Liu, M. (2001). Cenozoic extension and magmatism in the North American Cordillera: the role of gravitational collapse. Tectonophysics 342, 407–433.[Web of Science]
Marini, O. J., Fuck, R. A., Dardenne, M. A. & Danni, J. C. M. (1984). Proví ncia Tocantins: Setores Central e Sudeste. In: Almeida, F. F. M. & Hasui, Y. O. (eds) Pré-cambriano do Brasil. Sã o Paulo: Edgard Blücher, pp. 205–264.
McDade, P. & Harley, S. L. (2001). A petrogenetic grid for aluminous granulite facies metapelites in the KFMASH system. Journal of Metamorphic Geology 19, 45–59.[Web of Science]
Molnar, P., England, P. & Martinod, J. (1993). Mantle dynamics, uplift of the Tibetan Plateau, and the Indian monsoon. Reviews of Geophysics 31, 357–396.[Web of Science]
Moraes, R. & Fuck, R. A. (2000). Ultrahigh-temperature metamorphism in Central Brazil: the Barro Alto Complex. Journal of Metamorphic Geology 18, 345–358.[Web of Science]
Motoyoshi, Y. & Hensen, B. J. (2001). F-rich phlogopite stability in ultra-high-temperature metapelites from the Napier Complex, East Antarctica. American Mineralogist 86, 1404–1413.
Mouri, H., Guiraud, M. & Hensen, B. J. (1996). Petrology of phlogopite–sapphirine-bearing Al–Mg granulites from Ihouhaouene, In Ouzzal, Hoggar, Algeria: an example of phlogopite stability at high temperature. Journal of Metamorphic Geology 14, 755–761.[Web of Science]
O’Hara, M. J. & Yarwood, G. (1978). High pressure–temperature point on an Archaean geotherm, implied magma genesis by crustal anatexis, and consequences for garnet–pyroxene thermometry and barometry. Philosophical Transactions of the Royal Society of London, Series A 288, 441–456.
Oxburgh, E. R. (1990). Some thermal aspects of granulite history. In: Vielzeuf, D. & Vidal, Ph. (eds) Granulites and Crustal Evolution. Dordrecht: Kluwer Academic, pp. 569–580.
Pattison, D. R. M. & Bégin, N. J. (1994). Zoning patterns in orthopyroxene and garnet in granulites: implications for geothermometry. Journal of Metamorphic Geology 12, 387–410.[Web of Science]
Pimentel, M. M. & Fuck, R. A. (1992). Neoproterozoic crustal accretion in central Brazil. Geology 20, 375–379.
Pimentel, M. M., Whitehouse, M. J., Viana, M. G., Fuck, R. A. & Machado, N. (1997). The Mara Rosa arc in the Tocantins Province: further evidence for Neoproterozoic crustal accretion in central Brazil. Precambrian Research 81, 299–310.[Web of Science]
Pimentel, M. M., Fuck, R. A. & Fischel, D. P. (1999). Estudo isotópico Sm–Nd regional da porção central da Faixa Brasília: implicações para idade e origem dos granulitos do Complexo Anápolis–Itauçu e rochas metassedimentares do Grupo Araxá. Revista Brasileira de Geociências 29, 271–276.
Pimentel, M. M., Fuck, R. A., Jost, H., Ferreira Filho, C. F. & Araújo, S. M. (2000). The basement of the Brasília Fold Belt and the Goiás Magmatic Arc. In: Cordani, U. G., Milani, E. J., Thomaz Filho, A. & Campos, D. A. (eds) Tectonic Evolution of South America. Rio de Janeiro, 31st International Geological Congress, pp. 195–229.
Platt, J. P. & England, P. C. (1994). Convective removal of lithosphere beneath mountain belts: thermal and mechanical consequences. American Journal of Science 294, 307–336.
Powell, R. & Downes, J. (1990). Garnet porphyroblast-bearing leucosomes in metapelites: mechanisms, phase diagrams, and example from Broken Hill, Australia. In: Ashworth, J. R. & Brown, M. (eds) High Temperature Metamorphism and Crustal Anatexis. London: Unwin Hyman, pp. 105–123.
Raith, M., Karmakar, S. & Brown, M. (1997). Ultrahigh-temperature metamorphism and multi-stage decompressional evolution of sapphirine granulites from the Palni Hill Ranges, Southern India. Journal of Metamorphic Geology 15, 379–399.[Web of Science]
Rötzler, J. & Romer, R. L. (2001). P–T–t evolution of ultrahigh-temperature granulites from the Saxon Granulite Massif, Germany. Part I: Petrology. Journal of Petrology 42, 1995–2013.
Sandiford, M. (1989). Horizontal structures in granulite terrains: a record of mountain building or mountain collapse? Geology 17, 449–452.
Sandiford, M. & Powell, R. (1990). Some isostatic and thermal consequences of the vertical strain geometry in convergent orogens. Earth and Planetary Science Letters 98, 154–165.[Web of Science]
Sandiford, M., Neall, F. B. & Powell, R. (1987). Metamorphic evolution of aluminous granulites from Labwor Hills, Uganda. Contributions to Mineralogy and Petrology 95, 217–225.[Web of Science]
Sato, K. (1998). Evolução crustal da Plataforma Sul Americana, com base na geoquímica isotópica Sm–Nd. Ph.D. thesis, University of São Paulo, São Paulo, 290 pp.
Sawyer, E. W. (2001). Melt segregation in the continental crust: distribution and movement of melt in anatectic rocks. Journal of Metamorphic Geology 19, 291–309.[Web of Science]
Schott, B. & Schmeling, H. (1998). Delamination and detachment of a lithospheric root. Tectonophysics 296, 225–247.[Web of Science]
Sonder, L. J., England, P. C., Wernicke, B. P. & Chistiansen, R. L. (1987). A physical model for Cenozoic extension of western North America. In: Coward, M. P., Dewey, J. F. & Hancock, P. L. (eds) Continental Extensional Tectonics. Geological Society, London, Special Publications 28, 187–201.
Spear, F. S., Kohn, M. J. & Cheney, J. T. (1999). P–T paths from anatectic pelites. Contributions to Mineralogy and Petrology 134, 17–32.[Web of Science]
Stüwe, K. (1997). Effective bulk composition changes due to cooling: a model predicting complexities in retrograde reaction textures. Contributions to Mineralogy and Petrology 129, 43–52.[Web of Science]
Stüwe, K. & Powell, R. (1989). Metamorphic segregations associated with garnet and orthopyroxene porphyroblasts growth: two examples from Larsemann Hills, East Antarctica. Contributions to Mineralogy and Petrology 103, 523–530.[Web of Science]
Tassinari, C. C. G., Cordani, U. G., Correia, C. T., Nutman, A., Kinny, P. & Dias Neto, C. M. (1999). Dating of granulites by SHRIMP U–Pb systematics in Brazil: constraints for the age of the metamorphism of Proterozoic orogenies. Proceedings of 2nd South American Symposium on Isotope Geology, Cordoba, SEGEMAR, pp. 371–373.
Thompson, A. B. & England, P. C. (1984). Pressure–temperature–time paths of regional metamorphism II. Their inference and interpretation using mineral assemblages in metamorphic rocks. Journal of Petrology 25, 929–955.
von Blanckenburg, F. & Davies, J. H. (1995). Slab breakoff: a model for syncollisional magmatism and tectonics in the Alps. Tectonics 14, 120–131.[Web of Science]
Waters, D. J. (1988). Partial melting and the formation of granulite facies assemblages in Namaqualand, South Africa. Journal of Metamorphic Geology 6, 387–404.[Web of Science]
Wells, P. R. A. (1980). Thermal models for magmatic accretion and subsequent metamorphism of continental crust. Earth and Planetary Science Letters 46, 253–265.[Web of Science]
White, R. W. & Powell, R. (2002). Melt loss and the preservation of granulite facies mineral assemblages. Journal of Metamorphic Geology 20, in press
White, R. W., Powell, R. & Holland, T. J. B. (2001). Calculation of partial melting equilibria in the system Na2O–CaO–K2O–FeO–MgO–Al2O3–SiO2–H2O (NCKFMASH). Journal of Metamorphic Geology 19, 139–153.[Web of Science]
Winge, M. (1995). Evolução dos terrenos granulíticos da Província Estrutural Tocantins, Brasil central. Ph.D. thesis, University of Brasília, Brasília, 350 pp.
Wolff, F. (1991). Caracterização petrográfica, litogeoquímica, química mineral e geotermometria de rochas de alto grau metamórfico do Complexo Anápolis–Itauçu, Goiás. Masters dissertation, Federal University of Rio de Janeiro, Rio de Janeiro, 148 pp.
Wortel, M. J. R. & Spakman, W. (2000). Subduction and slab detachment in the Mediterranean–Carpathian Region. Science 290, 1910–1917.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  C. M. Valeriano, M. M. Pimentel, M. Heilbron, J. C. H. Almeida, and R. A. J. Trouw Tectonic evolution of the Brasilia Belt, Central Brazil, and early assembly of Gondwana Geological Society, London, Special Publications, January 1, 2008; 294(1): 197 - 210. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.A. Baldwin, M. Brown, and M.D. Schmitz First application of titanium-in-zircon thermometry to ultrahigh-temperature metamorphism Geology, April 1, 2007; 35(4): 295 - 298. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. K. BHOWMIK, A. B. SARBADHIKARI, B. SPIERING, and M. M. RAITH Mesoproterozoic Reworking of Palaeoproterozoic Ultrahigh-temperature Granulites in the Central Indian Tectonic Zone and its Implications J. Petrology, June 1, 2005; 46(6): 1085 - 1119. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. J. Henry, C. V. Guidotti, and J. A. Thomson The Ti-saturation surface for low-to-medium pressure metapelitic biotites: Implications for geothermometry and Ti-substitution mechanisms American Mineralogist, February 1, 2005; 90(2-3): 316 - 328. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. SAJEEV and Y. OSANAI Ultrahigh-temperature Metamorphism (1150{degrees}C, 12 kbar) and Multistage Evolution of Mg-, Al-rich Granulites from the Central Highland Complex, Sri Lanka J. Petrology, September 1, 2004; 45(9): 1821 - 1844. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Brown Hot orogens, tectonic switching, and creation of continental crust: Comment and Reply Geology, June 1, 2003; 31(6): e9 - e9. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||