| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology | Volume 44 | Number 3 | Pages 517-545 | 2003
© Oxford University Press 2003
Prograde and Retrograde Evolution in High-temperature Corundum Granulites (FMAS and KFMASH Systems) from In Ouzzal Terrane (NW Hoggar, Algeria)
1 INSTITUT DES SCIENCES DE LA TERRE, USTHB, BP 32, DAR EL BEIDA, ALGIERS, ALGERIA
2 LABORATOIRE DE MINÉRALOGIE, FR CNRS 32, MUSÉUM NATIONAL D'HISTOIRE NATURELLE, 61 RUE BUFFON, 75005 PARIS, FRANCE
3 LABORATOIRE DE GÉOSCIENCES MARINES, UMR 7097, INSTITUT DE PHYSIQUE DU GLOBE DE PARIS, TOUR 26-0, 4 PLACE JUSSIEU, 75252, PARIS, FRANCE
E-mail: mguiraud{at}mnhn.fr
RECEIVED FEBRUARY 24, 2001; ACCEPTED SEPTEMBER 16, 2002
 |
ABSTRACT |
|---|
The mineralogy and phase relationships of corundum-bearing granulites from In Ouzzal (Hoggar, Algeria) are analysed within the context of the present knowledge of In Ouzzal metamorphism, as independently deduced from quartz-bearing rocks: a clockwise evolution characterized by peak temperatures ranging from 850°C to more than 1100°C at about 10 kbar, followed by decompression and cooling to about 5 kbar and 750°C. The corundum-bearing rocks have been divided into three types (A–B, C and D) according to their mineralogy. Type A–B is characterized by the occurrence of sapphirine–sillimanite–orthopyroxene–phlogopite with or without garnet, Type C by the occurrence of spinel–sapphirine–garnet, and Type D by the occurrence of garnet–spinel–sillimanite. Thermodynamic data for sapphirine have been adjusted from the current THERMOCALC dataset to fit in with available experimental data and current theoretical phase diagrams. MAS, KMASH, FMAS and KFMASH petrogenetic grids that connect quartz-present and corundum-present grids are presented. Pseudosections derived from these grids account well for the observed textures. The three types of rock agree with the decompression path experienced by the quartz-bearing rocks. The occurrence of sapphirine corresponds to peak and decompression conditions, and trends of sapphirine and orthopyroxene compositions are consistent with this evolution.
KEY WORDS: corundum; In Ouzzal; sapphirine; UHT metamorphism
|
INTRODUCTION |
|---|
Granulites that preserve prograde sequences of mineral reactions are rare. Outstanding examples of such rocks are found in corundum-bearing silica-undersaturated granulites from the In Ouzzal granulitic unit (NW Hoggar) and have also been reported in Central Australia (Warren, 1983
; Goscombe, 1992), the Gruf complex, Italian Central Alps (Droop & Bucher-Nurminen, 1984), the Wilmington complex, USA (Strogi et al., 1993), the Limpopo belt (Ackermand et al., 1982; Horrocks, 1983; Windley et al., 1984; Droop, 1989), Namaqualand, South Africa (Waters, 1986) and Central Sri Lanka (Kriegsman & Schumacher, 1999). Kienast & Ouzegane (1987) and Bertrand et al. (1992) carried out preliminary work on quartz-absent rocks from In Ouzzal but their detailed mineralogy and a study of phase stability within model systems remains to be established. This paper aims to complement this previous work by providing detailed petrological and chemical data for a variety of corundum-bearing granulites collected at several localities from In Ouzzal (NW Hoggar, Algeria) and by working out the reaction history from the rock textures as completely as possible. |
GEOLOGICAL SETTING ANDPREVIOUS WORK |
|---|
 TOPABSTRACTINTRODUCTION GEOLOGICAL SETTING ANDPREVIOUS... FIELD RELATIONSHIPSWHOLE-ROCK CHEMISTRYMINERAL ASSEMBLAGES ANDREACTION...MINERAL CHEMISTRYP-T-X PSEUDOSECTIONS AND...CONCLUSIONREFERENCES |
|---|
Comprehensive data on the In Ouzzal granulitic terrane have been published elsewhere (Kienast et al., 1996
) and therefore only the main features are presented here. The In Ouzzal series belongs to the western Hoggar, which is situated between the western and eastern branches of the Pan-African belt (Fig. 1). Both the western and eastern shear zones represent sutures as evidenced by occurrences of eclogites in the Tin Zebane region, and of granulitic metagabbros with relict glaucophane SE of Ahnet (Caby, 1996). The In Ouzzal granulite terrane has undergone Pan-African greenschist-facies retrogression on its boundary, close to each shear zone; however, in the inner part of the terrane, granulite-facies rocks are preserved and unusual remnants exhibiting a great diversity of mineral reactions have been observed (Kienast & Ouzegane, 1987; Bertrand et al., 1992; Ouzegane & Kienast, 1996).
|
The In Ouzzal granulite terrane is composed of a charnockitic orthogneiss basement overlain by metasedimentary rocks, including marbles, magnetite-bearing quartzites (banded iron formations) and Al–Mg and Al–Fe granulites intercalated with mafic, ultramafic and anorthositic lenses; this kind of intercalation has been described elsewhere in Archaean supracrustal series [e.g. in South Africa (Ackermand et al., 1982
; Horrocks, 1983; Windley et al., 1984; Waters, 1986; Droop, 1989) and Sri Lanka (Kriegsman & Schumacher, 1999)]. The continental crust of the In Ouzzal granulitic terrane was formed by successive periods of orogenic activity between 3·5 and 2·0 Ga (Haddoum et al., 1994; Ouzegane & Kienast, 1996; Peucat et al., 1996). After sediment deposition, estimated at between 2·7 and 2·65 Ga (Bernard-Griffiths et al., 1996), alkali granite intrusion occurred at 2·65 Ga, probably in an extensional tectonic setting. Calc-alkaline magmatism, characterized by granodioritic and monzogranitic suites, followed at about 2·5 Ga, probably associated with crustal thickening processes (Peucat et al., 1996). Ultrahigh-temperature metamorphism occurred at 2·0 Ga and was associated with two main deformation phases (Bertrand et al., 1992; Haddoum et al., 1994; Ouzegane & Boumaza, 1996). The first was responsible for the development of a regional foliation, which was refolded during the second phase (Haddoum et al., 1994). The most pronounced structures in the field are ENE-trending folds shifting progressively to the NNE in the north of the In Ouzzal terrane. Generally, these large asymmetric folds are steeply plunging and affect a pre-existing foliation (S1) and lineation (L1) (Haddoum et al., 1994). Kinematic structures in the Al–Mg granulites include snowball garnets and pressure shadows made of orthopyroxene–sillimanite, which demonstrate the synkinematic character of mineral growth. Thus development of the main foliation (S1) during the D1 deformation stage is contemporaneous with the highest P–T conditions. D1 corresponds to progressive deformation resulting in crustal shortening and over-thickening by fold repetition during an Eburnean event (Haddoum et al., 1994; Peucat et al., 1996). The post-foliation folds (P2) are present on all scales with steeply dipping axes. Field observations suggest that the late, unoriented crystallization of granulite-facies mineral assemblages (symplectites) is contemporaneous with the P2 folding, as well as with the emplacement of monzogranites (Haddoum et al., 1994) and of anorthosites. Anorthosites are partly deformed and mainly located at the P2 hinge of the P2 fold. Reaction textures and the progressive development of mineral assemblages suggest a clockwise P–T trajectory (Kienast & Ouzegane, 1987; Bertrand et al., 1992; Mouri et al., 1996; Ouzegane & Boumaza, 1996) reaching very high temperatures (1050°C) followed by a decompression path. The very high temperatures are probably related to an increase in heat flow in the lower crust as suggested by the presence of plurimetric bodies of anorthosites (Figs 1 and 2) dated at 
2·0 Ga (Peucat et al., 1996).
|
Despite the very high temperature conditions, the In Ouzzal terrane displays only rare cordierite-bearing monzogranitic gneisses and veinlets of mesoperthite–quartz leucosomes (Ouzegane & Kienast, 1996
; Peucat et al., 1996). This very dry metamorphic environment, with little evidence of melting, could have originated as a consequence of polymetamorphic evolution, with dehydration during the earlier high-grade event at 2·5 Ga (Peucat et al., 1996) resulting in a dry restite that subsequently underwent very high temperature metamorphism without further melting at 2·0 Ga. |
FIELD RELATIONSHIPS |
|---|
Most of the samples described in this paper came from three localities, In Roccan and Tin Tchik Tchik in the northern part of the In Ouzzal and Tekhamalt in the central part (Fig. 1). The Tekhamalt and Tin Tchik Tchik areas mostly consist of charnockitic gneisses, whereas the In Roccan area is mostly supracrustal rocks and is one of the main localities for corundum-bearing Al–Mg granulites (Figs 1 and 2).
In the central zone of the In Roccan area (Fig. 2), remnants of ultramafic rocks occur as folded discontinuous lenses within the metasediments ranging in size from 1–2 m to a few tens of metres. The metasediments consist mainly of marbles and quartzites, which are often folded together with Al–Mg granulites. Some quartzites grade into the magnetite-rich banded iron formation facies. The occurrence of corundum–magnetite-bearing quartzite from Ihouhaouene, in the northern part of the In Ouzzal area, suggests that very high temperatures (up to 1200°C) could have been reached locally at the contact with the anorthosites and the carbonatites (Guiraud et al., 1996).
Chemically, Al–Mg granulites range from silica under-saturated (corundum- or spinel-bearing granulites) to varieties saturated and oversaturated in silica. The corundum-bearing granulites represent a very small proportion of the exposed rocks and are found as lenses immediately in contact with quartz-bearing Al–Mg granulites (Fig. 2). The reaction textures and P–T evolution of the silica-saturated Al–Mg granulites have been discussed previously by Kienast & Ouzegane (1987), Bertrand et al. (1992), Mouri et al. (1996) and Ouzegane & Boumaza (1996). Prograde metamorphism occurred at pressures of 10 kbar, with maximum temperatures ranging from 850°C to more than 1100°C according to the location. This was followed by decompression and cooling to 4–5 kbar and 750°C.
The corundum-bearing granulites studied here are very inhomogeneous with respect to their mineral and chemical compositions. Generally, corundum is a very minor phase in the rocks, but it occurs consistently and therefore we use the term corundum-bearing to describe these samples. Four rock types have been distinguished in the field according to the main mineral assemblage and the presence or absence of sapphirine:
- Type A: garnet–absent, corundum–orthopyroxene–sillimanite–phlogopite–sapphirine–bearing Al–Mg granulites;
- Type B: garnet–corundum–orthopyroxene–sillimanite– phlogopite–sapphirine–bearing Al–Mg granu-lites;
- Type C: garnet–spinel–corundum–sapphirine Al–Mg granulites that lack both orthopyroxene and sillimanite;
- Type D: garnet–spinel–sillimanite–corundum Al–Fe granulites that lack both orthopyroxene and sapphirine.
- Type B: garnet–corundum–orthopyroxene–sillimanite– phlogopite–sapphirine–bearing Al–Mg granu-lites;
Type A granulites are predominantly composed of coarse-grained porphyroblasts of orthopyroxene (25–60 vol. %), phlogopite (8–30%), sillimanite (2–10%) and sapphirine (30–50%). Corundum (<1%) is not evenly distributed in the rock; it is concentrated as small colourless grains within sapphirine porphyroblasts. Foliated (S1) corundum–orthopyroxene–sillimanite–phlogopite–sapphirine granulites differ from massive samples by the presence of higher amounts of phlogopite and a prominent lineation L1 made of aggregates of sillimanite and sapphirine.
Type B granulites lack a well-developed foliation and are very diverse in terms of mineral assemblages and mineral proportions. Some samples contain more than 10 distinct mineral phases. Modal proportions are typically 1–25% garnet, 20–35% orthopyroxene, 8–28% biotite, 1–20% sillimanite, 20–40% sapphirine, 4–15% cordierite, 4–15% K-feldspar; corundum, plagioclase, spinel, rutile and ilmenite represent altogether only 1% of the mode. Sapphirine is dark blue, contrasting with pale blue sapphirine in garnet-absent samples. Garnet occurs as coarse porphyroblasts.
Types A and B occur only in the In Roccan area as lens-shaped bodies of no more than a few metres length within quartz-bearing Al–Mg granulites (Fig. 2). The distinction between garnet-bearing and garnet-absent lithologies corresponds to differences in bulk-rock composition, the former being more Fe rich [Mg/(Mg + Fe) <0·80; Ouzegane & Kienast, 1996]. However, the observed modal mineralogy is highly variable and appears to be related to grain size. Thus, although the distinction between garnet-absent and garnet-bearing rocks is convenient in the field, it corresponds to two end-members of the same rock type, which will be referred to subsequently as Type A–B unless otherwise specified.
Type C granulites are found in the Tin Tchik Tchik area (Fig. 1) as nodules of no more than few centimetres long within garnet–spinel–orthopyroxene–biotite granulites. Typically these samples consist of sapphirine containing blebs of spinel, corundum, garnet and rare wagnerite.
Type D granulites are a rare rock type among the In Ouzzal corundum-bearing granulites. They occur as dark lenses (up to 1 m long) closely associated with leucocratic, highly foliated quartz–K feldspar–biotite–garnet granulites in the Tekhamalt area (Tan Afella and the northern part of Khanfuss, Fig. 1). There is a sheared border zone 5 mm thick along the sharp and concordant contact with the quartz-bearing granulite. A typical mode is 1–2% corundum, 16–17% spinel, 4–7% sillimanite, 40–42% garnet, 5–7% biotite, 21–26% K-feldspar, 2–3% plagioclase, 1–2% ilmenite and accessory rutile and zircon. In addition to their lack of orthopyroxene and sapphirine, one of the features of these Al–Fe granulites is that corundum can occur as a primary phase included in biotite and spinel, as well as coronae around spinel.
|
WHOLE-ROCK CHEMISTRY |
|---|
Representative samples of Types A, B and D were analysed by X-ray fluorescence (XRF) at CRPG (Nancy, France) for major and trace elements (Table 1). SiO2, Al2O3, FeO and MgO represent altogether 95 wt % of the total. All these rocks have K2O > Na2O and are characterized by relatively low CaO contents (<1·3 wt %). The striking chemical feature is their wide XMg range [molecular MgO/(MgO + total FeO)], from 0·31 to 0·96. With respect to trace elements, strong, although variable, enrichment in Cr (55–2471 ppm) and Ni (53–522 ppm) matches observations made in equivalent peraluminous metamorphic rocks from other areas (Schreyer et al., 1981
; Kerrich et al., 1987; Bernards-Griffiths et al., 1996). The corundum-bearing granulites have variable Zr contents (18–467 ppm), reflecting the presence of significant amounts of zircon in some samples. Sr concentration is unusually low (12–129 ppm) and the Rb/Sr ratio (0·5–13) shows a large scatter. Fluorine content is relatively high (0·1–0·6 wt %) and chlorine reaches its maximum value at 455 ppm in Type D, the least magnesian rock.
|
The term Al–Mg granulite stems from its unusual chemistry—poor in silica and alkali—with respect to normal pelites. This peculiar chemistry has led to several hypotheses about the origin of Al–Mg granulites. Schreyer et al. (1981)
reviewed some of the mechanisms by which corundum-bearing rocks may be generated. Numerous explanations have been proposed, such as metasomatism of mafic to ultramafic rocks or other precursors, pre-metamorphic exhalative alteration of ultramafic komatiitic lavas, or partial melting of sedimentary materials, mainly shales, the removal of the granitic component from which leaves a residue enriched in MgO and Al2O3 and depleted in alkali elements (Grant, 1968; Lal et al., 1978). Origins related to pre-metamorphic history have also been considered (Mouri et al., 1994; Bernards-Griffiths et al., 1996; Ouzegane & Kienast, 1996). Experimental studies (Wiggering et al., 1992) have emphasized the links between hydrothermal fluids and the genesis of an aluminous palaeosol. Al, Mg, Cr, Co and Ni are generally accepted as being relatively immobile during weathering and are largely incorporated in clay minerals or in the chlorite component (Bernards-Griffiths et al., 1996). Also, most weathering reactions are characterized by removal of elements such as Na, Ca, Si, Fe and Mn, and therefore the relatively immobile elements Al, Ti, Mg, Zr, Hf, Cr, Ni, V and Co are concentrated in the weathered residue from the original ultramafic or mafic rocks (Kerrich et al., 1987). For example, the high concentrations of Ni and Cr are consistent with a genetic relationship between the corundum-bearing granulites and mafic or ultramafic rocks. Lower concentrations of Sr and variable Rb/Sr ratios also support an origin dominated by hydrothermal alteration of mafic or ultramafic protoliths. Trace element patterns and oxygen isotope compositions in Al–Mg granulites are best explained by variable mixtures of hydrothermally altered mafic or ultramafic and felsic protoliths (Bernard-Griffiths et al., 1996; Fourcade et al., 1996). |
MINERAL ASSEMBLAGES ANDREACTION TEXTURES |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTING ANDPREVIOUS...FIELD RELATIONSHIPSWHOLE-ROCK CHEMISTRYMINERAL ASSEMBLAGES ANDREACTION...MINERAL CHEMISTRYP-T-X PSEUDOSECTIONS AND...CONCLUSIONREFERENCES |
|---|
The interpretation of the relative timing of the various mineral parageneses is based on the textural development of coronae or symplectite growth and resorption textures. In the quartz-bearing Al–Mg granulites, coarse-grained crystals, even where preserved within coronae, are early features compared with fine-grained textures; the former are associated with the prograde stage and the peak metamorphism, and the latter with late-stage decompression assemblages (Bertrand et al., 1992
; Ouzegane & Boumaza, 1996). The same textural interpretation has been adopted as a starting point for the corundum-bearing rocks.
Corundum–orthopyroxene–sillimanite–phlogopite–sapphirine–garnet-bearing and garnet-free granulites (Types A and B)
Complex primary reactions produce coarse-grained sapphirine associated with sillimanite and/or orthopyroxene and/or K-feldspar, leaving relicts of corundum associated with garnet, phlogopite and early orthopyroxene. Such primary textures are followed by development of successive coronae and symplectites through reactions with the matrix, characterized by cordierite associated with secondary sapphirine, orthopyroxene, K-feldspar and rare spinel.
Inclusions within prograde phases
Coarse-grained sapphirine porphyroblasts contain numerous inclusions either of individual grains of orthopyroxene, garnet, biotite, sillimanite, corundum or rutile, or of assemblages such as orthopyroxene–phlogopite, corundum–sillimanite or garnet–phlogopite–corundum. Corundum is the only inclusion mineral not found in the matrix assemblage. Corundum is found as colourless tablets that commonly show mechanical twinning. It is almost always armoured by sapphirine (Fig. 3a–c) breaking any former equilibrium contact, for example with orthopyroxene. Garnet occurs either as grains enclosed in sillimanite or sapphirine or as individual porphyroblasts containing inclusions of phlogopite, corundum (as microinclusions), K-feldspar, plagioclase and rutile.
|
Coronae and symplectites:corundum-bearing domains
Corundum (0·2–0·5 mm) is commonly surrounded by two distinct elongate coronae (Fig. 3b and c). The inner one is composed of blue, strained, fractured grains of sapphirine (0·5–3 cm). This corona is composed of recrystallized subhedral sapphirine and sillimanite grains, suggesting that sillimanite and sapphirine were initially in equilibrium during deformation. The outer corona close to orthopyroxene (B) ± biotite is commonly 0·2–1 cm wide and is composed of sillimanite. Sillimanite forms fine-grained polygonal aggregates of recrystallized prisms. Sapphirine coronae appear to be in textural equilibrium with coarse-grained matrix sillimanite. These textures support the reaction orthopyroxene + corundum
sapphirine1 + sillimanite (3b, Table 2) or biotite + orthopyroxene + corundum sapphirine1 + sillimanite + K-feldspar in biotite-bearing microdomains. Moreover, some sillimanite–sapphirine and sillimanite–orthopyroxene contacts display fine-grained (0·05–0·01 mm) cordierite–sapphirine symplectites, suggesting the reactions sillimanite + sapphirine1 cordierite + sapphirine2 and orthopyroxene + sillimanite sapphirine + cordierite (Fig. 3b and c, Table 2). Another type of complex corona is also recognized between corundum and phlogopite ± garnet with the inner margin, close to corundum, made of sapphirine enclosing corundum and sillimanite ± garnet and the outer margin, close to phlogopite, made of K-feldspar (Fig. 3a). These textures suggest the reactions biotite + corundum + sillimanite sapphirine + K-feldspar or garnet + biotite + corundum + sillimanite sapphirine + K-feldspar in garnet-bearing domains (Table 2).
|
|
|
|
Coronae and symplectite assemblages: corundum-free, phlogopite- and garnet-bearing domains
Corroded primary garnets are replaced by coarse symplectitic intergrowths of orthopyroxene–sillimanite–sapphirine according to the reaction garnet
orthopyroxene + sillimanite + sapphirine (Fig. 4a). This reaction occurs either as a consequence of heating or on decompression (Harley et al., 1990; Harley, 1998b). Additionally, coarse-grained orthopyroxene–sapphirine–K-feldspar symplectites are interpreted as the prograde breakdown of phlogopite ± garnet according to the reaction biotite orthopyroxene1 + sapphirine1 + K-feldspar (Fig. 5a) and garnet + biotite orthopyroxene1 + sapphirine1 + K-feldspar (Table 2). All these reaction products break down further to secondary cordierite–sapphirine–orthopyroxene symplectites (Table 2). For example, orthopyroxene1 and sapphirine1, which develop either inside phlogopite or at its margin, give way to fine-grained composite intergrowths of sapphirine2, orthopyroxene2 and cordierite. Such coronitic textures can be very complex (e.g. Fig. 5b): within sapphirine, the K-feldspar core is surrounded by a cordierite corona, which outlines the original K-feldspar shape. Sapphirine is itself surrounded by a corona of cordierite, which separates it from orthopyroxene and K-feldspar. Fine symplectites and coronae are characteristic of the post-peak metamorphic evolution of these rocks. Fine intergrowths of orthopyroxene–sapphirine–cordierite–K-feldspar occur in microcracks within biotite (Fig. 5c) or at its margin (Fig. 5d) according to the reaction biotite orthopyroxene + cordierite + sapphirine + K-feldspar. Another type of corona corresponds to well-developed symplectites of sapphirine and cordierite around sillimanite in locally Al-rich bulk compositions at the contact with phlogopite (Fig. 6a), orthopyroxene (Fig. 3b) or coarse-grained sapphirine (Fig. 5c and d). These textures support the reaction biotite + sillimanite cordierite + sapphirine + K-feldspar, orthopyroxene + sillimanite cordierite + sapphirine and sapphirine1 + sillimanite sapphirine2 + cordierite. In some areas, fine-grained symplectites of orthopyroxene–cordierite–sapphirine occur at the margin of or within garnet, indicating that garnet is the main reactive phase (Fig. 4b). Cordierite–spinel symplectites separating orthopyroxene from sapphirine are observed rarely, in late textures.
Garnet–spinel–corundum granulite(Type C)
Most of the phases garnet, corundum, spinel and sapphirine reach 1·5 cm in size and are therefore macroscopically visible. The most remarkable feature in this rock type is that garnet, spinel and sapphirine occur respectively as elongate aggregates and large coronae of euhedral grains, which give the rock a granoblastic polygonal texture. Garnet is always separated from spinel and corundum by large (3 mm) sapphirine coronae, whereas spinel and corundum always occur together (Fig. 6b, Table 2). These textures support the reaction garnet + corundum + spinel sapphirine. This exceptional reaction texture has not been preserved from retrogression in any other sample. A similar mineral assemblage consisting of corundum, garnet, spinel, sapphirine, cordierite, biotite and orthopyroxene was described by Horrocks (1983) from the Limpopo Mobile Belt and by Kriegsman & Schumacher (1999) from Sri Lanka. The most prevalent reaction in the Limpopo Belt area is garnet + corundum cordierite + sapphirine + spinel, which differs from the sample described here only by the presence of cordierite.
Garnet (2–5 mm length) appears as elongate aggregates of limpid, pseudohexagonal grains (Fig. 6b). This shape is attributed to recrystallization of a previously deformed grain during D1 in combination with grain boundary movement, which allows the initial curved interfaces to straighten and the interface junctions to adjust to minimum energy positions (Spry, 1969). Sapphirine inclusions preferentially occur at triple points and joins where low total boundary energy is achieved.
Spinel occurs as large (1·5 cm diameter) dark green globular grains. The cores and rims of grains are structurally different: the cores show areas separated by deformation bands whereas the rims exhibit a distinct polygonization attributed to a larger strain gradient. Irregular garnet grain boundaries are considered to be related to reaction processes. Spinels contain many exsolved lamellae of magnetite and titanomagnetite, and some inclusions of ilmenite, corundum and the magnesian fluorophosphate wagnerite. The last phase occurs as elliptical to irregular colourless grains that show a slight cleavage. Wagnerite has rarely been described in rocks (Heggemann & Steinmetz, 1927; Fin'ko, 1972; Sheridan et al., 1976; Irouschek & Armbruster, 1984; Novak & Povondra, 1984; Simmat & Rickers, 2000). It has been synthesized by Raade (1990) and has a wide stability field. Sapphirine has formed large coronae (2–4 mm) of polygonal blue grains around the other phases. It is worth noting that sapphirine coronae grow systematically between garnet, spinel and corundum (Fig. 6b). The sapphirine grains contain many cracks and inclusions, mostly represented by wagnerite and rarely ilmenite.
Complete recrystallization of garnet and sapphirine indicates dynamic recrystallization at high temperature (Poirier, 1985), subsequent to sapphirine growth at the expense of garnet, spinel and corundum. Spinel shows evidence of recrystallization, although dynamic recrystallization affects only its margin because of its larger size.
Corundum–garnet–spinel–sillimanite Al–Fe granulites (Type D)
The corundum-bearing Al–Fe granulites consist of coarse-grained (1–10 mm) garnet, spinel, K-feldspar, sillimanite and biotite, with variable amounts of corundum, plagioclase, rutile and ilmenite. All these minerals have been affected by ductile deformation at high temperature (D1), characterized by textures with high strain zones and preserved zones that contain abundant coronae and symplectites. Relict plagioclase grains with undulose extinction and elongate subgrains grade laterally into recrystallized grains with grain boundaries that are generally gently curved and with 120° triple junctions. K-feldspar is sheared and partly recrystallized into small lenses with sutured boundaries. Reddish brown flakes of biotite are commonly surrounded by fine-grained secondary biotite with plagioclase locally developed in the cleavage planes. Rounded biotite crystals also occur enclosed in large garnet porphyroblasts. Garnet also has inclusions of spinel and plagioclase and is fractured and elongated in the high-strain zones. Locally, small subhedral grains of secondary garnet grow at the expense of pre-existing coarse-grained garnets. The most common reaction texture in sillimanite-absent low-strain zones is a corona of corundum around spinel, which is in turn enclosed by a corona of biotite associated with garnet adjacent to K-feldspar; rutile is included in ilmenite in this microdomain (Fig. 6c). This suggests the reaction spinel + K-feldspar + rutile garnet + corundum + biotite + ilmenite (Table 2). Three reaction assemblages can be recognized in sillimanite-bearing zones. The first assemblage contains garnet and corundum separated by symplectites of sillimanite and spinel, suggesting the reaction garnet + corundum spinel + sillimanite (Table 2). The second assemblage contains garnet–corundum, K-feldspar and rutile separated by coarse-grained complex intergrowths of sillimanite, biotite and ilmenite. This texture supports the reaction garnet + corundum + K-feldspar + rutile sillimanite + biotite + ilmenite (Table 2). The third primary assemblage is garnet–corundum–K-feldspar overprinted by coarse-grained biotite–spinel–sillimanite symplectites (Fig. 6d), suggesting the reaction garnet + corundum + K-feldspar biotite + spinel + sillimanite.
|
MINERAL CHEMISTRY |
|---|
Representative analyses of the main mineral phases are presented in Tables 3
–10. Analyses were performed on an SX50 CAMECA electron microprobe at the University of Paris 6. The operating conditions were 15 kV accelerating voltage and 10 nA sample current. Natural silicates and synthetic oxides were used as standards for all elements, except for fluorine and zinc, which were calibrated on fluorite and sphalerite, respectively.
|
|
|
|
|
|
|
|
Biotite
The main variables are the XMg [Mg/(Mg + Fe)] ratio and Al, Ti, F and Cl content (Table 3). There is no compositional zoning within grains but there are large variations between grains in a single thin section, especially with respect to Ti, Al and XMg. Octahedral Al content in all biotites is <0·40 per six octahedral sites, which is consistent with the data from other granulite-facies rocks (Guidotti, 1984
). Differences in Al depend upon the type of mineral adjacent to biotite: biotite grains that are enclosed in corundum or sillimanite always contain more Al2O3 (15–16%) than biotite grains in the matrix (13–14%). Al decreases as XMg increases in primary biotites. Differences among biotites from Types A, B and D are shown in Fig. 7. The XMg ratio (0·64–0·97) is related to the whole-rock composition and increases as follows: (1) orthopyroxene-free Type D (0·64–0·75); (2) garnet-bearing Type B (0·78–0·85); (3) garnet-free Type A (0·86–0·97). The highest TiO2 content is found in Types B and D (1·41–6%) and the lowest content in Type A (0·2–2·8%); TiO2 decreases as XMg increases. Biotite is brown in Types B and D and becomes paler in Type A as the TiO2 content decreases. Despite the presence of rutile and ilmenite, phlogopite in the most magnesian samples (Inh 324, Inh 1002) contains very little Ti. The Cl content is low, except in Type D, where it reaches 1·25 wt %, and always lower than the F content (1·56–4·4 wt %) (Table 3). The maximum F value corresponds to the maximum Mg value. In the two Type A–B samples (Inh 131, Inh 317), biotite has higher F content in the garnet-free sites (Fig. 8).
|
|
Orthopyroxenes
Orthopyroxenes range in Mg/(Mg + Fe) value from 0·74 to 0·95 (Table 4). Large compositional variations in Al, Mg, Fe and Si are observed within single thin sections. Zoning in individual grains is due to (Mg,Fe)Si = AlVIAlIV substitution, with decreasing Al2O3 from core to rim; as Al decreases in coarse primary orthopyroxenes the XMg ratio remains almost constant (Fig. 9). XMg is related to bulk composition in Type A and B granulites (Table 4 and Fig. 9). Primary orthopyroxene in garnet-bearing rocks has a higher Al2O3 content (maximum 10·6 wt %) and lower XMg (<0·80) (Fig. 9) than in garnet-free granulites (Al2O3 up to 9 wt %). Significant differences have also been observed between fine-grained orthopyroxene embayments in biotite rims (Al2O3 values as low as 4 wt %; Fig. 5c; Table 4) and those in garnet rims (6–8 wt % Al2O3). Small subgrains in symplectites between coarse- grained orthopyroxene and sapphirine have Al2O3 contents ranging from 4 to 6 wt %, lower than the associated coarse-grained orthopyroxene (9–10·6 wt %).
|
Sapphirine
Sapphirine compositions lie on the ideal line joining sapphirine 2(Mg,Fe)O:2(Al,Fe3+,Cr)2O3:1SiO2 and sapphirine 7(Mg,Fe)O:9(Al,Fe3+,Cr)2O3:3SiO2 according to the Tschermakitic substitution Si(Mg, Fe)=Al(Al,Cr,Fe3+). The Fe3+/(Fe3+ + Fe2+) value does not generally exceed 0·3 using the Fe3+ recalculation method of Higgins et al. (1979)
. Chromium generally occurs only in trace amounts but a chromium-rich variety with values ranging from 1 to 4·4 wt % is found in sample Inh 131 (Table 5). Such an unusual Cr-rich sapphirine (4·2–7·52 wt % Cr2O3) was reported by Friend (1982) in kornerupine–corundum–spinel granulites from Greenland. In Types A and B, coarse-grained sapphirines (Spr1 in Table 5 and Fig. 10a) are always higher in SiO2 and lower in Al2O3 than fine-grained secondary sapphirines (Spr2); this relationship is opposite to the one observed for orthopyroxene compositions. The cores of primary sapphirines commonly have higher SiO2 contents in garnet-bearing granulites (Spr1: maximum 14·3 wt %) than in garnet-free granulites (maximum SiO2 13·9 wt %, Fig. 10a). As for biotite and orthopyroxene, sapphirine compositions are also strongly related to the XMg bulk-rock composition: there is a significant difference between garnet-bearing (0·77–0·85, Types B and C) and garnet-free rocks (0·86–0·97, Type A) (Fig. 10a). Like many sapphirines from other granulites (Droop, 1989; Bertrand et al., 1992; Ouzegane & Boumaza, 1996) individual crystals show zoning from core to rim, and composition varies at the contact with mineral inclusions. Sapphirines immediately adjacent to corundum or sillimanite have the highest Al contents whereas sapphirines surrounding garnet have the lowest Al. In corundum–garnet–spinel granulites (Type C), sapphirine displays opposite trends in Tschermakitic substitution depending on whether it is in contact with garnet or with spinel (Figs 6b and 10b). Sapphirine in contact with spinel evolves towards the 7:9:3 composition whereas sapphirine in contact with garnet evolves towards 2:2:1 (Fig. 10b).
|
Garnets
Garnets from the three corundum-bearing granulites types (B, C and D) are almandine–pyrope with minor grossular (1–6 mol %) and spessartine (0·2–1·5 mol %) contents (Table 6). Their XMg ratio decreases in the following order: 0·47–0·68 (Type B); 0·54–0·59 (Type C); 0·25–0·33 (Type D). Compositional zoning is a common feature, with Fe enrichment and Mg depletion in the outer 50 µm of individual grains (Types B and D). Rim compositions of primary garnets are close to the compositions of secondary garnets in Type D rocks. Primary garnets in Type C are slightly zoned, probably as a result of homogenization during dynamic recrystallization. In Type B granulites the garnet armoured in sillimanite or sapphirine is richer in pyrope (60–66 mol %) than the cores of garnets found in the matrix (pyrope 53–55 mol %, Fig. 11). Garnets from the Type B rocks have higher pyrope contents (66 mol %) than values reported for garnets from corundum granulites in the Limpopo Belt (Horrocks, 1983
), Gruf complex (Droop & Bucher-Nurminen, 1984) and from quartz–sapphirine-bearing granulites in In Ouzzal (Bertrand et al., 1992; Mouri et al., 1996; Ouzegane & Boumaza, 1996) but lower than garnet (XMg = 0·71) analysed by Harley & Fitzsimons (1991) in the garnet–orthopyroxene–sillimanite–sapphirine assemblages from Long Point, Antarctica (Harley, 1998b).
|
Spinels
Spinels from Types B, C and D have XMg in the range of 0·37–0·56, 0·57–0·61 and 0·29–0·42, respectively (Table 7). Spinel contains small amounts of ZnO (1–2 wt %) and Fe2O3 (up to 3 wt %). In Type D rocks spinel inclusions in garnet are high in Mg and low in Cr (XMg = 0·40–0·42; Cr2O3 = 0·2–0·4 wt %) compared with spinel inclusions in corundum (XMg = 0·29–0·30; Cr2O3 = 1·7–1·9 wt %). Matrix spinels in Type D rocks are zoned, with a decrease in XMg from core (0·39) to rim (0·33), and are low in chromium compared with spinel in Type B rocks, which reach 11·9–22·4 wt % Cr2O3 (Table 7). In the latter sample (Inh 131) spinel occurs in symplectites with cordierite between orthopyroxene and chromium-rich sapphirine. To our knowledge, such a chromium-rich spinel has not been reported so far in other corundum–sapphirine-bearing rocks.
Cordierite
XMg is higher than for coexisting orthopyroxene, sapphirine, phlogopite, spinel or garnet. XMg is 0·88–0·91 in Type B (Table 8) and 0·92–0·98 in Type A rocks. Cordierite shows little chemical variation either within large crystals or between grains in contact with different mineral assemblages.
Other minerals
Sillimanite contains 0·4–1 wt % Fe2O3. Associated corundum shows a maximum of 0·7 wt % of Cr2O3 and Fe2O3. K-feldspar and plagioclase occur in Type A, B and D granulites. The K-feldspar is perthitic with a composition that varies between Or84Ab16 An1 and Or88Ab11An1. The anorthite content in plagioclase ranges from 28 to 43% (Table 9). The main elements in wagnerite are MgO (45·83–47·32 wt %), P2O5 (43 wt %) and F (6·9–7·7 wt %); it also contains 3 wt % FeO and between 1·5 and 1·7 wt % TiO2 (Table 10). This composition is close to that of wagnerite from Czechoslovakia (Novak & Povondra 1984) and India (Simmat & Rickers, 2000).
|
P–T–X PSEUDOSECTIONS AND INTERPRETATION OF REACTION TEXTURES |
|---|
Textures indicate that the main reaction in corundum-bearing Al–Mg granulites is the production of sapphirine from corundum + orthopyroxene mineral assemblages. The thin sections also show that, except in one sample, phlogopite is stable during the prograde metamorphic stage and part of the decompression stage. During the decompression stage it reacts out to form cordierite and K-feldspar. For the K2O-free assemblages, reaction textures can be interpreted in the simple FMAS system involving sapphirine, garnet, spinel, corundum, orthopyroxene and sillimanite. However, to interpret biotite-bearing textures, an attempt to study these rocks within the KFMASH model system using THERMOCALC (Holland & Powell 1998
) was made and is discussed below.
Appropriate thermodynamic data for sapphirine are required as it is the main mineral observed in the textures. Therefore we compared the calculated phase equilibria with MAS system experimental runs involving sapphirine. The thermodynamic data in THERMOCALC (Holland & Powell, 1998) for sapphirine generate the two stable invariant points (Spl, Crd) and (Sil, Crd) in MAS found by Hensen & Green (1973). However, they are located at different P–T conditions than the experimental data. The purpose of this paper is to derive pseudosections suitable for the studied rocks and the estimated P–T conditions; therefore, we empirically studied the effect of changing enthalpies on the bulk composition of sapphirine. We found that Mg end-members had to be made more negative to improve the fit of the experimental data, a feature already found by Viellard (1992) and Harley & Motoyoshi (2000). However, this change had to be small as we kept the high entropy proposed by Holland & Powell (1998). We also found that data for Fe Spr had to be adjusted to fit expected chemical compositions for sapphirine. Slightly changing the enthalpy for sapphirine end-members (-5 kJ for MgSpr221, -10 kJ for MgSpr793 and +10 kJ for FeSpr; Table 11) using a simple ideal activity model has the effect of stabilizing the two MAS invariant points at much lower temperature, thus making the location of the Grt + Spl + Crn = Spr univariant curve in agreement with the experimental data of Ackermand et al. (1975). The location of the Spl-absent invariant point that involves the phases Grt, Opx, Spr, Qtz, Sil and CrdA in the FMAS system is at 7 kbar, whereas, in the FMASH system, with CrdH (hydrous cordierite) instead of CrdA (anhydrous cordierite), it is at 13·5 kbar, in fairly good agreement with experimental runs of Hensen & Green (1973) (Fig. 12b). Hensen & Harley (1990), Bertrand et al. (1991, 1992) and Harley & Motoyoshi (2000) placed the FMASH invariant point (Spl) at P–T conditions of 10 ± 1 kbar and 1020°C, 9·5 ± 1 kbar and 1030°C, 10·8 ± 1 kbar and 1040°C, and 10·5 ± 1 kbar and 1050°C, respectively. However, our calculated FMASH invariant point is metastable with respect to melting as suggested by the experiments of Bertrand et al. (1992). Therefore the melt-present [Spl] invariant point should be located between 7 and 13·5 kbar: this point is calculated at 1050°C and 10·5 kbar assuming a(H2O) = 0·5 using the thermodynamic data given in Table 11, in agreement with these estimations. Moreover, the calculated compositions for sapphirine are within the range of those analysed in our rocks. The set of calculated grids is presented in Fig. 12a–c. The grids involve two cordierites, hydrous (CrdH) and anhydrous (CrdA). Figure 12a displays the MAS and MASH grids with Crn-bearing and Qtz-bearing systems. The Crn-bearing system is located at lower temperature and pressure and connects to the Qtz-bearing system through (Qtz, Crn)-absent reactions as previously predicted by Hensen & Harley (1990). The main effect is to expand the Crn–Spr stability field to lower P and T. Figure 12b displays the grid for the FMAS and FMASH systems, showing the stability field of sapphirine + quartz with respect to cordierite and sillimanite + orthopyroxene. FMAS invariant points move along Crd-absent reactions, which are the only stable univariant, non-degenerate, reactions in this part of the FMASH system. The stability of hydrous cordierite is limited up-pressure by H2O appearing as a phase in the mineral assemblages (Fig. 12b). Figure 12c shows the KFMASH system and gives the relationships involving biotite, K-feldspar and hydrous cordierite. The three paths in Fig. 12c summarize the different P–T evolutions recorded in the different types of quartz-bearing rocks (Kienast & Ouzegane, 1987; Bertrand et al., 1992; Mouri et al., 1996; Ouzegane & Boumaza, 1996). H2O and melt are not included as phases in this grid and therefore all the usual ‘dehydration’ reactions are replaced by reactions of the type Bi Crd + Kfs as given by Mouri et al. (1996). This is consistent with the textures observed in thin sections. In the pure KFMASH system, the absence of melt implies that the ratio of modal proportions of Crd and Ksp is fixed by the ratio K/H2O in biotite, the other elements (Si, Al, Fe and Mg) being balanced using the other minerals. In the natural system, biotite persists and often become enriched in F, which therefore affects the observed modal proportions of these three minerals. However, even in this H2O-undersaturated system, the grid should connect with melting reactions. We compared our grid with the experimentally derived grid by Carrington & Harley (1995), which involves melt and osumilite. From their work, Bi + Qtz + Ksp either with Grt or Sil is not stable above 900°C: the two reactions connect at their [osumilite] invariant point from which the liquid-absent reaction Bi + Opx + Sil + Qtz = Grt + Crd + Ksp emanates. This reaction is calculated in our grid, although with a steeper slope (Fig. 12c). Therefore the upper part of the reaction is metastable and our [L] invariant point is also metastable. As a consequence, the upper part of the (Qtz, Crn) absent reactions are also metastable, to follow Schreinemakers' rules. However, we did not calculate the full grid with melt and osumilite: in the studied rocks Qtz or Ksp is absent (for example, Qtz disappears instead of Bi at the melt-absent reaction) and therefore Bi remains stable. As a consequence, the bulk compositions are such that textures are controlled by melt-absent reactions and the pseudosections are not affected by the stability of melt at these temperatures.
|
|
Mineral chemical studies have shown that Fe3+ and Ti are present in significant amounts in some phases and therefore that KFMASHTO would be the most appropriate system. Fe–Ti oxides are required to balance reactions if biotite or spinel is involved. However, the current thermodynamic data are inadequate to allow the calculation of reliable petrogenetic grids for this complex system at high temperature. Therefore, we have focused on the KFMASH system, which is close enough to the natural system to allow interpretation of the reaction textures. Additionally, the thermodynamics of sapphirine introduces enough uncertainty in the calculations that accounting for the FeO–Fe2O3–TiO2 relationships would make the whole approach unreliable.
Compatibility diagrams were drawn to interpret the textures and to work out the theoretical reactions in the KFMASH system. Representative electron microprobe analyses of coexisting phases have been projected from sillimanite and K-feldspar onto the Herc–Spl–Qtz plane and the parageneses have been represented in the rectangular diagram Spl/(Spl + Hc) vs Qtz/(Spl + Hc) (Fig. 13). These compatibility diagrams show the different stable assemblages derived from textural observations and mineral chemistry as well as the reaction sequences in Types A and B. They show that equilibrium prevailed during the metamorphic evolution and they also illustrate the influence of the bulk composition of microdomains on the mineral assemblages and mineral compositions. Garnet-bearing equilibria occur in Fe-rich bulk compositions, whereas orthopyroxene-bearing equilibria occur in the less aluminous bulk compositions. The sequence of mineral reactions has been worked out from Fig. 13: the breakdown of biotite to cordierite + sapphirine (Fig. 13d and Fig. 6a) should occur after the breakdown of garnet to orthopyroxene + sapphirine (Fig. 13c and Fig. 4a) and the breakdown of biotite + corundum to coarse-grained sapphirine (Fig. 13b). This metamorphic evolution can be better illustrated by P–T–X pseudosections, which show the variations of mineral assemblages and mineral compositions on pressure and temperature for a fixed bulk composition. The THERMOCALC software was used to calculate the pseudosections relevant to the different rock types from the KFMASH petrogenetic grid (Fig. 12c).
|
In the following section, P–T–X pseudosections are used to model metamorphic textures according to the P–T path recorded from Qtz-bearing Al–Mg granulites and compared with the thin-section observations. However, bulk compositions are generally not known in high-grade metamorphic facies as textural microdomains in which reactions take place are smaller than the volume used for bulk-rock analysis. Therefore bulk compositions used in the pseudosections were adapted from analyses in Table 1 to match mineral assemblages and mineral compositions. Figure 14 shows the pseudosection for an Mg-rich, Type A–B rock, Fig. 15 for the Fe-rich, Type C rock, and Fig. 16 for the Type D rock. Because the Type C rock does not involve a potassic or hydrous phase in the textures, the pseudosection has been calculated in the FMAS system (Fig. 15). Adding water to the system simply leads to expansion of cordierite-bearing fields, at least up to the stability of cordierite in the FMASH system (Fig. 12b).
|
|
|
Because the bulk-rock compositions have been chosen to match the mineral compositions and because a–X relationships are not well known, composition isopleths and mode isopleths cannot be used as absolute values. However, they can be used qualitatively as indicators of changes in mineral composition and modal proportions with P–T evolution. Comparison of these three pseudosections clearly illustrates that differences in the bulk composition can lead to considerable differences in the development of mineral assemblages in rocks as already stated by most workers (e.g. Hensen & Harley, 1990
; Harley, 1998a). The dominant feature in the KFMASH system is the prevalence of divariant, trivariant and quadrivariant fields and the scarcity of univariant reactions affecting the chosen bulk composition. Assuming a prograde path at 10 kbar, the large trivariant field involving corundum, orthopyroxene, sillimanite and biotite (with or without garnet) represents the P–T conditions at which the earlier assemblages of the Type A–B granulites developed (Fig. 14). The diagram also shows that corundum reacts out to sapphirine along the P–T path. This is consistent with the observation of sapphirine coronae shielding pre-peak corundum from further reaction (Fig. 3a–c). Garnet occurrence in this rock depends on the chosen XMg for the bulk rock. The more Fe rich the bulk composition, the lower the pressure of garnet appearance. Therefore, the prograde reaction Opx + Crn Spr + Sil (Fig. 3b), which has a positive slope, occurs in Mg-rich compositions before it occurs in Fe-rich ones, as noticed by Bertrand et al. (1992). In Fe-rich Type A–B bulk compositions, the prograde stage is similar except that Spr occurrence is the product of a divariant reaction characterized by the breakdown of corundum before sapphirine appearance. Corundum included in garnet, as observed in some thin sections, with no relation to sapphirine can be explained by such a divariant reaction occurring in Fe-rich bulk compositions. Garnet reacting out could correspond either to the prograde stage or the decompression stage. Both hypotheses match the observed garnet breakdown textures involving sillimanite in symplectitic intergrowths with orthopyroxene and sapphirine (Fig. 4a). However, increases in the size of these coronae, which can be attributed to annealing at high temperature (Fig. 6b), suggest the prograde stage rather than the decompression stage. The pseudosection shows that quartz can occur at high temperature, together with sapphirine. The absence of quartz in the studied rocks indicates that temperatures did not reach 1050°C in this rock. The decompression stage is characterized by cordierite forming at the expense of biotite (Figs 5c and d, 6a and 14) and by a decrease in the abundance of sillimanite. Both features match very well the thin-section observations. The pseudosection shows that secondary corundum should crystallize at pressures below 5 kbar (Fig. 14). Calculated compositional isopleths show only significant variation in XTs in both orthopyroxene and sapphirine (Fig. 14). This feature matches the observed compositions. Moreover, the variations in the Al content of sapphirine and orthopyroxene result mainly from changes in P–T conditions rather than variations in bulk composition. XTs isopleths for orthopyroxene and sapphirine have opposite slopes and therefore increasing XTs in sapphirine in conjunction with decreasing XTs in orthopyroxene corresponds to isothermal decompression in the Sil + Opx + Spr + Bt field and to decompression and cooling in the Opx + Sil + Spr + Crd field. This is consistent with the path derived from the quartz-bearing rocks. Decompression is also characterized by the appearance of cordierite. The pseudosection (Fig. 14) matches the observation of biotite disappearing to form symplectitic Opx + Spr + Crd + Kfs (Fig. 5c). The last stage of decompression is characterized by crystallization of spinel + cordierite. Decompression is also marked by a decrease in modal sillimanite, to the point where sillimanite disappears to form secondary corundum. Although the decrease in sillimanite proportion is clearly observed in thin section, secondary corundum has not been found. This means that either pressure did not go below 5 kbar or reaction kinetics were too slow to crystallize corundum.
In Type C rocks (Fig. 15), sapphirine should clearly occur during the prograde stage. This is consistent with the earliest textures displaying the growth of prismatic coronitic sapphirine in Tin Tchik Tchik. Again, increases in grain size of the sapphirine support the prograde evolution hypothesis. A decrease in the proportion of spinel once corundum has disappeared as reflected by garnet–sapphirine–spinel textures could also belong to the prograde stage. Decompression occurs almost completely in the trivariant field Spl + Grt + Spr, which matches the high-temperature recrystallization observed in thin section. Sillimanite and orthopyroxene should crystallize at the end of the P–T evolution but are not observed. Composition isopleths also fit in with the defined P–T path: XFe in garnet and in sapphirine and XTs in sapphirine increase with rise in temperature. That the variation in XFe in garnet composition is much larger than in sapphirine also matches the observed compositions.
In the Type D granulite (Fig. 16), sapphirine, orthopyroxene and cordierite are absent. The pseudosection using a very Fe-rich composition accounts for the same P–T evolution, with sillimanite and spinel crystallizing from corundum-bearing assemblages at approximately the peak metamorphic conditions. This is certainly consistent with the relatively large size of sillimanite (Fig. 6d). The occurrence of two wide divariant fields along the interpreted P–T evolution is consistent with the fact that this rock does not show much reaction texture.
|
CONCLUSION |
|---|
Corundum-bearing rocks are good indicators of metamorphic evolution in high-temperature granulite-facies metamorphism. They cover a wide range of compositions with many different mineral assemblages. The three pseudosections presented here indicate that quartz crystallizes in these rocks only at very high temperature, whereas corundum is present at the earliest (prograde) and latest (post-decompressional) stage of the evolution. Therefore, all bulk compositions mostly record the earliest stage of the P–T evolution and do not display the highest temperatures that are recorded in the quartz-bearing rocks from In Ouzzal. The metamorphic evolution is marked by the appearance of sapphirine at the expense of corundum-bearing assemblages. The compositional isopleths and mode isopleths calculated in the pseudosections match the observed textures, which are interpreted in terms of a clockwise P–T path. This is in agreement with the path defined by the quartz-bearing rocks. Therefore these rocks have bulk compositions suitable for tracking metamorphic evolution in high-temperature terranes and constitute good records of the thermal and geodynamic history of the lower continental crust.
It is generally agreed that granulites are formed in regions where the geothermal gradient exceeds normal continental values (Harley, 1989, 1998a; Sandiford & Powell, 1991). The origin of the thermal anomaly can be found below the crust, such as heat from mantle plumes or intraplated magmas. Mantle-derived magmas, such as anorthosites, are regarded as playing a major role in advecting heat into the crust to give ultrahigh-temperature metamorphism [for a review, see Ashwal (1993)]. The occurrence of some 2 Ga anorthositic rocks and carbonatitic bodies in the vicinity of the ultrahigh-temperature granulites could provide a link between magmas, hot fluids and metamorphism. However, the In Ouzzal metamorphism has a complex structure; field relationships show heterogeneously distributed isotherms corresponding to small-scale, steep thermal gradients. Therefore, an appropriate model in this case may be contact or hydrothermal metamorphism. At the top of a magmatic body, isotherms are organized with respect to the shape of the fluid and the magma conduits. For the In Ouzzal region, the scale of the heterogeneities with respect to the isotherms is 10–100 m for a difference of 200°C (Fig. 12c). This is similar to what is found in the Pyrenees where HT/LP metamorphism is associated with localized pull-apart basins and fluid channelling in a more general extensional environment (Golberg & Leyreloup, 1990). Therefore, a similar process can be envisaged at depth if heat sources are sufficiently important to provide similar temperature differences. Anorthosites and carbonatites crop out as dykes of 2–100 m width; this range is of the same order of magnitude as the heterogeneities observed in the metamorphic country rocks. Moreover, field observations provide good constraints on the relative timing of anorthosite and carbonatite emplacement and granulite-facies metamorphism. The anorthositic bodies have been dated at 2002 Ma, the same age as the granulitic metamorphism (Peucat et al., 1996), and are partly deformed by the main foliation but are preferentially concentrated in the hinge of D2 folds. The prograde event with peak temperatures from 800 to 1100°C in the quartz Al–Mg granulites occurred during this D2 event. The rise in temperature at In Ouzzal followed by near-isothermal decompression path is therefore interpreted as regional metamorphism in a crust thickened by collision followed by extension with extra and localized heat sources brought by anorthosites, carbonatites and related hot fluids.
|
ACKNOWLEDGEMENTS |
|---|
We would like to thank J. Hollis and an anonymous reviewer for their comments and thorough reviews, which helped us to improve the manuscript, and Simon Harley for his comments and handling of the manuscript. The work was supported by the CNRS and the French–Algerian co-operation programme MDU 476.
|
REFERENCES |
|---|
Ackermand, D., Seifert, F. & Schreyer, W. (1975). Instability of sapphirine at high pressures. Contributions to Mineralogy and Petrology 50, 79–92.[CrossRef][Web of Science]
Ackermand, D., Herd, R. K. & Windley, B. F. (1982). Chemographic relationships in sapphirine-bearing rocks of the Limpopo belt, Southern Africa. Revista Brasiliera de Geociencas, São Paulo 12, 292–300.
Ashwal, L. D. (1993). Anorthosites. Berlin: Springer, 422 pp.
Bernard-Griffiths, J., Fourcade, S., Kienast, J. R., Peucat, J. J., Martineau, F. & Rahmani, A. (1996). Geochemistry and isotope (Sr, Nd, O) study of Al–Mg granulites from the In Ouzzal Archaean block (Hoggar, Algeria). Journal of Metamorphic Geology 14, 709–724.[CrossRef][Web of Science]
Bertrand, P., Ellis, D. J. & Green, D. H. (1991). The stability of sapphirine–quartz and hypersthene–quartz assemblages: an experimental investigation of the system FeO–MgO–Al2O3–SiO2 under H2O and CO2 conditions. Contributions to Mineralogy and Petrology 108, 55–71.[CrossRef][Web of Science]
Bertrand, P., Ouzegane, K. & Kienast, J. R. (1992). P–T–X relationships in the Precambrian Al–Mg rich granulites from In Ouzzal, Hoggar, Algeria. Journal of Metamorphic Geology 10, 17–31.[Web of Science]
Caby, R. (1996). A review of the In Ouzzal granulitic terrane (Tuareg shield, Algeria): its significance within the Pan-African Trans-Saharan Belt. Journal of Metamorphic Geology 14, 659–666.[CrossRef][Web of Science]
Carrington, D. P. & Harley, S. L. (1995) 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]
Droop, G. T. R. (1989). Reaction history of garnet–sapphirine granulite-facies metamorphism in the Central Limpopo Mobile Belt, Zimbabwe. Journal of Metamorphic Geology 7, 383–403.[Web of Science]
Droop, G. T. R. & Bucher-Nurminen, K. (1984). Reaction textures and metamorphic evolution of sapphirine-bearing granulites from the Gruf Complex, Italian Central Alps. Journal of Petrology 25, 766–803.
Fin'ko, V. I. (1972). Pervaya nakhodka wagnerita, SSSR. Doklady Akademii Nauk SSSR 143, 1424–1427.
Fourcade, S., Kienast, J. R. & Ouzegane, K. (1996). Metasomatic effects related to channeled fluid streaming through deep crust: fenites and associated carbonatites (In Ouzzal Proterozoic granulites, Hoggar, Algeria). Journal of Metamorphic Geology 14, 763–781.[CrossRef][Web of Science]
Friend C. R. L. (1982). Al–Cr substitution in peraluminous sapphirines from the Bjonersund area, Fiskenaesset region, Southern West Greenland. Mineralogical Magazine 46, 323–328.[Web of Science]
Golberg, J. M. & Leyreloup, A. F. (1990). High temperature–low pressure Cretaceous metamorphism related to crustal thinning (Eastern North Pyrenean Zone, France). Contributions to Mineralogy and Petrology 104, 194–207.[CrossRef][Web of Science]
Goscombe, B. (1992). Silica-undersaturated sapphirine, spinel and kornerupine granulite facies rocks, NE Strangways Range, Central Australia. Journal of Metamorphic Geology 10, 181–201.[Web of Science]
Grant, J. A. (1968). Partial melting of common rocks as possible source of cordierite–anthophyllite bearing assemblages. American Journal of Science 266, 908–931.
Guidotti, C. V. (1984). Micas in metamorphic rocks. In: Baily, S. W. (ed.) Micas. Mineralogical Society of America, Reviews in Mineralogy 13, 357–456.
Guiraud, M., Kienast, J. R. & Ouzegane, K. (1996). Corundum–quartz bearing assemblage in the Ihouhaouene area (In Ouzzal, Algeria). Journal of Metamorphic Geology 14, 755–761.[CrossRef][Web of Science]
Haddoum, H., Choukroune, P. & Peucat, J. J. (1994). Structural evolution of the Precambrian In Ouzzal massif (Central Sahara, Algeria). Precambrian Research 65, 155–166.[CrossRef][Web of Science]
Harley, S. L. (1989). The origins of granulites: a metamorphic perspective. Geological Magazine 126, 215–247.[Abstract]
Harley, S. L. (1998a). On the occurrence and characterization of ultrahigh-temperature crustal metamorphism. In: Treolar, 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). Ultrahigh temperature metamorphism (1050°C, 12 kbar) and decompression in garnet (Mg70)–orthopyroxene–sillimanite gneisses from the Rauer group, East Antarctica. Journal of Metamorphic Geology 16, 541–562.[CrossRef][Web of Science]
Harley, S. L. & Fitzsimons, I. C. W. (1991). Pressure–temperature evolution of metapelitic granulites in a polymetamorphic terrane: the Rauer Group, East Antarctica. Journal of Metamorphic Geology 9, 231–243.[Web of Science]
Harley, S. L. & Motoyoshi, Y. (2000). Al zoning in orthopyroxene in a sapphirine quartzite: evidence for >1200°C UHT metamorphism in the Napier Complex, Antarctica, and implications for the entropy of sapphirine. Contributions to Mineralogy and Petrology 138, 293–307.[CrossRef][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]
Heggemann, F. & Steinmetz, H. (1927). Die Mineralgange von Werfen im Salzkammergut. Centralblatt für Mineralogie, Geologie und Paläontologie, Abteilung A 45–56.
Hensen, B. J. & Green, D. H. (1973). Experimental study of the stability of cordierite and garnet in pelitic compositions at high pressures and temperaratures. III. Synthesis of experimental data and geological applications. Contributions to Mineralogy and Petrology 38, 151–166.[CrossRef][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, pp. 19–56.
Higgins, J. B., Ribbe, P. H. & Herd, R. K. (1979). Sapphirine I: crystal chemical contributions. Contributions to Mineralogy and Petrology 68, 349–356.[CrossRef][Web of Science]
Holland, T. J. B. & Powell, R. (1998). An internally-consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology 16, 309–343.[CrossRef][Web of Science]
Horrocks, P. C. (1983). A corundum and sapphirine paragenesis from the Limpopo Mobile Belt, southern Africa. Journal of Metamorphic Geology 1, 13–23.[Web of Science]
Irouschek, A. & Armbruster, Th. (1984). Hydroxylhaltiger Wagnerit aus den Val Ambra (Tessin, Schweizland). Fortschritte für Mineralogie 62, 109–110.
Kerrich, R., Fyfe, W. S., Barnett, R. L., Blair, B. B. & Wilmore, L. M. (1987). Corundum, Cr-muscovite rocks at O'Briens, Zimbabwe: the conjunction of hydrothermal desilicification and LIL-element enrichment—geochemical and isotopic evidence. Contributions to Mineralogy and Petrology 95, 481–498.[CrossRef][Web of Science]
Kienast, J. R. & Ouzegane, K. (1987). Polymetamorphic Al–Mg rich parageneses in Archean rocks from Hoggar, Algeria. Geological Journal 22, 57–79.[Web of Science]
Kienast, J. R., Fourcade, S., Guiraud, M., Hensen, B. J. & Ouzegane, K. (eds) (1996). Special Issue on the In Ouzzal granulite unit, Hoggar, Algeria. Journal of Metamorphic Geology 14, 659–808.
Kretz, R. (1973). Symbols for rock-forming minerals. American Mineralogist 68, 277–279.
Kriegsman, L. M. & Schumacher, J. C. (1999). Petrology of sapphirine-bearing and associated granulites from Central Sri Lanka. Journal of Petrology 40, 1211–1239.[CrossRef][Web of Science]
Lal, R. K., Ackermand, D., Seifert, F. & Haldar, S. K. (1978). Chemographic relations in sapphirine-bearing rocks from Sonapahar, Assam, India. Contributions to Mineralogy and Petrology 67, 169–187.[CrossRef][Web of Science]
Mouri, H., Guiraud, M. & Kienast, J. R. (1994). The origin of Al–Mg granulites of Ihouhaouene, Hoggar, Algeria: an example of phase relationships in the KFMASH system and melt absent equilibria. Comptes Rendus de l'Académie des Sciences 318, 941–948.
Mouri, H., Guiraud, M. & Hensen, B. J. (1996). Petrology of phlogopite–sapphirine–bearing Al–Mg granulites from Ihouhaouene (Mole In Ouzzal–Hoggar Algeria): an example of phlogopite stable at high temperature. Journal of Metamorphic Geology 14, 725–738.[CrossRef][Web of Science]
Novak, M. & Povondra, B. P. (1984). Wagnerite from Skrinarov, Central Czechoslovakia. Neues Jahrbuch für Mineralogie, Monatshefte 12, 536–542.
Ouzegane, K. & Boumaza, S. (1996). An example of very high temperature metamorphism: orthopyroxene–sillimanite–garnet, sapphirine–quartz, and spinel–quartz. Journal of Metamorphic Geology 14, 693–708.[CrossRef][Web of Science]
Ouzegane, K. & Kienast, J. R. (1996). Nature et évolution des séries métamorphiques de très haute température de l'Unité Granulitique de l'In Ouzzal (Ouest Hoggar). Bulletin du Service Géologique de l'Algérie 7 (fas. 2), 133–157.
Peucat, J. J., Capdevila, R., Drareni, A., Choukroune, P., Fanning, M., Bernard-Griffiths, J. & Fourcade, S. (1996). Major and trace element geochemistry and isotope (Sr, Nd, Pb, O) systematics of an Archaean basement involved in a 2·0 Ga VHT (1000°C) metamorphic event: In Ouzzal massif, Hoggar, Algeria. Journal of Metamorphic Geology 14, 667–692.[CrossRef][Web of Science]
Poirier, J. P. (1985). Creep in Crystals, High Temperature Deformation Processes in Metals, Ceramics and Minerals. Cambridge Earth Sciences Series. Cambridge: Cambridge University Press, 260 pp.
Raade, G. (1990). Hydrothermal syntheses of Mg2 PO4OH polymorphs. Neues Jahrbuch für Mineralogie, Monatshefte 7, 289–300.
Sandiford, M. & Powell, R. (1991). Some remarks on high temperature–low pressure metamorphism in convergent orogens. Journal of Metamorphic Geology 9, 333–340.[Web of Science]
Schreyer, W., Werding, C. & Abraham, K. (1981). Corundum–fuchsite rocks in greenstone belts of southern Africa: petrology geochemistry and possible origin. Journal of Petrology 22, 191–231.
Sheridan, D. M., Sheraman, P. M., Mrose, M. E. & Taylor, R. B. (1976). Mineralogy and geology of the wagnerite occurrence in Santa Fe montain, Front Range. US Geological Survey, Professional Papers 955, 23.
Simmat, R. & Rickers, K. (2000). Wagnerite in high Mg–Al granulites of Anakapalle, Eastern Ghats Belts, India. European Journal of Mineralogy 3, 661–666.
Spry, A. (1969). Metamorphic Textures. Commonwealth International Library of Science, Technology, Engineering and Liberal Studies, Oxford: Pergamon Press, 342 pp.
Strogi, L., Warner, M. E. & Lutz, T. M. (1993). Dehydration partial melting and disequilibrium in the granulite-facies Wilmington Complex, Pennsylvania Delaware Piedmont. American Journal of Science 293, 405–462.
Viellard, Ph. (1994) Prediction of enthalpy of formation based on refined crystal structures of multisite compounds. Part 2: application to minerals belonging to the system Li2O–Na2O–K2O–BeO–MgO–CaO–MnO–FeO–Fe2O3–Al2O3–TiO2–SiO2–H2O. Results and discussion. Geochimica et Cosmochimica Acta 58, 4064–4108.
Warren, R. G. (1983). Prograde and retrograde sapphirine in metamorphic rocks of central Australia. B.M.R. Journal of Australian Geology and Geophysics 8, 139–145.
Waters, D. J. (1986). Metamorphic history of sapphirine-bearing and related magnesian gneisses from Namaqualand, South Africa. Journal of Petrology 27, 541–565.
Wiggering, H., Neumann-Mahlkau, P. & Selbach, H. J. (1992). Experimental procedures to simulate weathering under atmospheres which may have characterized the early Archaean. In: Schidlowski, M. et al. (eds) Early Organic Evolution: Implications for Mineral and Energy Resources. Proceedings of the 9th Alfred Wegener Conference being the Final Meeting of IGCP Project 157, Maria Laach, FRG, 19–23 September 1988. Berlin: Springer, pp. 31–40.
Williams, K. L., Rock, N. M. S., & Carroll, G. W. (1990). Spinel and Spineltab: Macintosh program to plot spinel analyses in the three dimensional oxidized (magnetite) and reduced (ulvöspinel) prisms. American Mineralogist 75, 1428–1430.[Abstract]
Windley, B. F., Ackermand, D. & Herd, R. K. (1984). Sapphirine/kornerupine-bearing rocks and crustal uplift history of the Limpopo belt, Southern Africa. Contributions to Mineralogy and Petrology 86, 342–358.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. K. Podlesskii, L. Y. Aranovich, T. V. Gerya, and N. A. Kosyakova Sapphirine-bearing assemblages in the system MgO-Al2O3-SiO2: a continuing ambiguity European Journal of Mineralogy, October 1, 2008; 20(5): 721 - 734. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Pitra, P. Boulvais, V. Antonoff, and H. Diot Wagnerite in a cordierite-gedrite gneiss: Witness of long-term fluid-rock interaction in the continental crust (Ile d'Yeu, Armorican Massif, France) American Mineralogist, February 1, 2008; 93(2-3): 315 - 326. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Adjerid, K. Ouzegane, G. Godard, and J. R. Kienast First report of ultrahigh-temperature sapphirine + spinel + quartz and orthopyroxene + spinel + quartz parageneses discovered in Al-Mg granulites from the Khanfous area (In Ouzzal metacraton, Hoggar, Algeria) Geological Society, London, Special Publications, January 1, 2008; 297(1): 147 - 167. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Sato, T. Miyamoto, and T. Kawasaki Fe2+-Mg partitioning experiments between orthopyroxene and spinel using ultrahigh-temperature granulite from the Napier Complex, East Antarctica Geological Society, London, Special Publications, January 1, 2008; 308(1): 431 - 447. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. A. Halpin, R. W. White, G. L. Clarke, and D. E. Kelsey The Proterozoic P-T-t Evolution of the Kemp Land Coast, East Antarctica; Constraints from Si-saturated and Si-undersaturated Metapelites J. Petrology, July 1, 2007; 48(7): 1321 - 1349. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. ULIANOV and A. KALT Mg-Al Sapphirine- and Ca-Al Hibonite-bearing Granulite Xenoliths from the Chyulu Hills Volcanic Field, Kenya J. Petrology, May 1, 2006; 47(5): 901 - 927. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. DAS, A. BHATTACHARYA, M. M. RAITH, S. BHADRA, and M. BANERJEE Aluminous sapphirine granulites from the Eastern Ghats Belt (India): Phase relations and relevance to counterclockwise P-T history European Journal of Mineralogy, February 1, 2006; 18(1): 35 - 48. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. FEENSTRA, S. SAMANN, and B. WUNDER An Experimental Study of Fe-Al Solubility in the System Corundum-Hematite up to 40 kbar and 1300{degrees}C J. Petrology, September 1, 2005; 46(9): 1881 - 1892. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. GONCALVES, C. NICOLLET, and J.-M. MONTEL Petrology and in situ U-Th-Pb Monazite Geochronology of Ultrahigh-Temperature Metamorphism from the Andriamena Mafic Unit, North-Central Madagascar. Significance of a Petrographical P-T Path in a Polymetamorphic Context J. Petrology, October 1, 2004; 45(10): 1923 - 1957. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. M. KELLY and S. L. HARLEY Orthopyroxene-Corundum in Mg-Al-rich Granulites from the Oygarden Islands, East Antarctica J. Petrology, July 1, 2004; 45(7): 1481 - 1512. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||