Journal of Petrology 45(7) © Oxford University Press 2004; all rights reserved
Structural and P–T Evolution of a Major Cross Fold in the Central Zone of the Limpopo High-Grade Terrain, South Africa
1 DEPARTMENT OF GEOLOGY, RAND AFRIKAANS UNIVERSITY, P.O. BOX 524, AUCKLAND PARK, 2006, JOHANNESBURG, SOUTH AFRICA
2 DEPARTMENT OF PETROLOGY, GEOLOGICAL FACULTY, MOSCOW STATE UNIVERSITY, VOROBIEVY GORY, MOSCOW, 119899 RUSSIA
3 INSTITUTE OF EXPERIMENTAL MINERALOGY, RUSSIAN ACADEMY OF SCIENCES, CHERNOGOLOVKA, MOSCOW DISTRICT, 142432, RUSSIA
4 INSTITUT FÜR GEOLOGIE, MINERALOGIE UND GEOPHYSIK, FAKULTÄT FÜR GEOWISSENSCHAFTEN, RUHR-UNIVERSITÄT-BOCHUM, 44780 BOCHUM, GERMANY
RECEIVED NOVEMBER 15, 2002; ACCEPTED FEBRUARY 26, 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGEOLOGICAL FRAMEWORK OF THE...METAMORPHISMTHERMODYNAMICS OF METAMORPHISMCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The Central Zone (CZ) of the Limpopo Complex of southern Africa is characterized by a complex deformational pattern dominated by two types of fold geometries: large sheath folds and cross folds. The sheath folds are steeply SSW-plunging closed structures whereas the cross folds are north–south-oriented with near-horizontal fold axes. In the area south of Messina this complexly folded terrain grades continuously towards the south into a crustal-scale ENE–WSW-trending ductile shear zone with moderate dip towards the WSW. All sheath folds document consistent top-to-the-NE thrust movement of high-grade material. The timing of this shear deformational event (D2) and thus of the gneissic fabric (S2) is constrained (at
2·6 Ga) by the syntectonic emplacement throughout the CZ of precursors to quartzo-feldspathic gneisses (Singelele-type gneisses). Cross folds deform the S2 fabric and are characterized by a near-vertical axial planar cleavage (S3). Recent single-phase Pb–Pb dating of garnet from a metapelitic gneiss with an S3 fabric from one of the largest cross folds in the CZ constrains the timing of the deformational (D3) and metamorphic (M3) event at 2·0 Ga. Mineral chemistry for metapelites from this cross fold shows a single peak on an NMg histogram for garnet reflecting a single phase of mineral growth. Metapelites from this cross fold also preserve evidence for only one well-developed reaction texture, Grt + Sil + Qtz 
Crd. This reaction is accompanied by the simultaneously operating reaction Grt + Fsp + H2O = Bt + Sil + Qtz. Both these divariant reactions belong to the univariant KFMASH equilibrium Crd + Grt + Fsp + H2O Bt + Sil + Qtz. The progress of the two divariant retrograde reactions leads to the consumption of Grt and Fsp: K-feldspar (Or94–100) never occurs with both cordierite and garnet. Microprobe profiling coupled with calculated isopleths for Bt, Grt and Crd in divariant equilibria define a decompression-cooling P–T path that reflects a single (M3) high-grade metamorphic event during the evolution of the cross fold. This decompression-cooling P–T path traverses from 780°C, 5·7 kbar to 600°C, 3·3 kbar. KEY WORDS: deformation; granulites; Limpopo Complex; geothermobarometry; mineral equilibrium; mineral thermodynamics; P–T path
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL FRAMEWORK OF THE...METAMORPHISMTHERMODYNAMICS OF METAMORPHISMCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The 700 km x 250 km Limpopo Complex of Southern Africa is a unique field laboratory for studying fundamental problems of the formation and exhumation of granulite-facies terrains located between Precambrian cratons (e.g. Van Reenen et al., 1992
). Mason (1973) was the first to subdivide the complex into three tectonic units: the Late Archaean Southern (SMZ) and Northern (NMZ) Marginal Zones (e.g. Mkweli et al., 1995; Kreissig et al., 2001) and a Central Zone (CZ) located between the SMZ and NMZ. Most workers agree that the two marginal zones were exhumed and emplaced in the Late Archaean. The age of the formation and P–T history of the CZ is still being debated.
Barton et al. (1994), Kamber et al. (1995), Holzer et al. (1998), and Schaller et al. (1999) published evidence based on single-phase dating of metamorphic minerals for three distinct events in the CZ: (1) a Mid-Archaean (3·2 Ga) event mainly reflected by magmatic activity with relic traces of an early deformational and metamorphic event (here termed D1/M1); (2) a Late Archaean (2·57–2·65 Ga) low-pressure ‘anticlockwise’ granulite-facies event (here termed D2/M2) associated with granitic/charnockitic magmatism that resulted from vertical crustal growth of the Zimbabwe Craton; (3) a Palaeoproterozoic (2·05–1·95 Ga) HP–HT clockwise granulite-facies event (here termed D3/M3) caused by the collision of the Zimbabwe and Kaapvaal cratons. In this interpretation the complex deformational pattern of the CZ mainly reflects the Palaeoprotorozoic event (D3/M3) that obliterated most of the evidence for the two Archaean events.
Hofmann et al. (1998) and Kröner et al. (1998, 1999), on the basis of single-zircon dating of syntectonic granitic magmatism, also documented evidence for three distinct events, but with a different emphasis on the relative significance of the last two events. Those workers concluded that the late Archaean D2/M2 event represents the major fabric-forming episode in the CZ, and other deformation events are subordinate: the 2·0 Ga Palaeoprotorozoic event is thus considered to be mainly a high-grade static metamorphic event (e.g. Hofmann et al., 1998; McCourt & Armstrong, 1998), suggesting that the complex deformational pattern of the CZ mainly reflects the late Archaean D2/M2 event.
Published data can therefore be interpreted to suggest that the deformational pattern of the CZ, dominated by two major folded structures—cross folds and sheath folds (Figs 1 and 2)—reflects a major high-grade deformational event, either in the Late Archaean (Hofmann et al., 1998; Kröner et al., 1998, 1999) or in the Palaeoprotorozoic (Holzer et al., 1998; Schaller et al., 1999). New age data (Boshoff, 2004), however, are not in accordance with the formation of the CFZ as the result of a single geodynamic event. These new data constrain the age of (1) a major sheath fold (the Avoca sheath fold), based on zircon sensitive high-resolution ion microprobe (SHRIMP) data from syn- to late tectonic precursors to the Singelele-type gneisses, to 2·57 Ga, and (2) a major cross fold (the Baklykraal fold), based on single-phase dating of garnet from metapelitic gneisses, to 2·0 Ga.
|
|
The thermodynamic (P–T–fluid) history of the CZ is also still the subject of debate. Several studies (e.g. Horrocks, 1983
; Harris & Holland, 1984; Droop, 1989; Van Reenen et al., 1990; Hisada & Miyano, 1996) proposed essentially isothermal decompression (ITD) of the high-grade rocks, which implies an extremely high speed of exhumation of the CZ: the velocity of upward movement should be much higher than the rate of heat loss. In addition, none of these studies integrated detailed structural data with the proposed P–T trajectories. In contrast, structural, metamorphic and physicochemical studies of major structures in the SMZ (Smit et al., 2001), supported by limited data from the CZ by Perchuk et al. (1996, 2000), have demonstrated that the decompressional-cooling (DC) of the high-grade rocks of these two zones suggests a broad match between the rate of exhumation and associated heat loss (Gerya et al., 2000). These results are consistent with the concept (Perchuk, 1989) of the gravitational redistribution of the initial cratonic rocks that resulted in the formation of the entire Limpopo high-grade terrain at 2·65 Ga. The aim of this paper is twofold: (1) to demonstrate the coherent structural and metamorphic evolution of one of the largest cross folds in the CZ (Baklykraal fold, Figs 1 and 2b); (2) to correctly derive the P–T path for this giant structure.
We use the following mineral abbreviations and thermodynamic symbols.
Mineral abbreviations used. Act, actinolite; And, andalusite; Ann, annite; Bt, biotite; Cal, calcite; Crd, cordierite; Di, diopside; East, eastonite; Fsp, feldspar; Grt, garnet; Kfs, K-feldspar; Ky, kyanite; Oam, orthoamphibole; Opx, orthopyroxene; Phl, phlogopite; Pl, plagioclase; Prp, pyrope; Qtz, quartz; Rut, rutile; San, sanidine; Sil, sillimanite; Spl, spinel.
Thermodynamic symbols used. T, temperature (K or°C); P, pressure (bar or kbar); 
G°, the change of standard Gibbs free energies in a reaction; R = 1·987 cal (1 cal = 4·186 J) is the gas constant; Xi, mole fraction of component i in a given phase; XMg = Mg/(Mg + Fe); Ni = 100Xi; NMg, Mg-number; aCrd, activity of dry cordierite in the system Crddry – CrdH2O–CO2, 
; 
, chemical potential of pure water as function of P and T; 
, water fugacity; 
, H2O activity in the fluid phase where 
is H2O activity coefficient in a fluid phase; 
, partial molar mixing Gibbs free energy for component i of a solid solution; 
, partial molar excess Gibbs free energy for component i in a solid solution; Ge, integral molar excess Gibbs free energy of a solid solution; Wij, the Margules interaction parameter (cal/mol).
|
GEOLOGICAL FRAMEWORK OF THE CENTRAL ZONE |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL FRAMEWORK OF THE...METAMORPHISMTHERMODYNAMICS OF METAMORPHISMCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The Limpopo Complex comprises, as mentioned above, three crustal domains (Mason, 1973
): the CZ, SMZ and NMZ. The latter two domains are separated from the adjacent cratons by 2·6 Ga (Kreissig et al., 2001) inward-dipping crustal-scale ductile dip-slip shear zones along which the granulite-facies rocks have been thrust onto the adjacent cratons (Van Reenen et al., 1987, 1992; Roering et al., 1992; Smit & Van Reenen, 1997): the Hout River Shear Zone in the south, and the North Marginal Thrust Zone in the north (e.g. Mkweli et al., 1995). The high-grade rocks from the two marginal zones have lithological and geochemical similarities to lower-grade rocks from the adjacent granite–greenstone cratons (Van Reenen & Smit, 1996; Kreissig et al., 2000), suggesting their derivation through metamorphism of the latter rocks under lower-crustal conditions.
The CZ is bounded by crustal-scale inward-dipping 2·0 Ga strike-slip mylonitic/cataclastic shear zones, the Palala Shear Zone in the south, and the Triangle Shear Zone in the north (e.g. McCourt & Vearncombe, 1992; Kamber et al., 1995; Schaller et al., 1999). The CZ consists of a wide range of lithologies, including high-grade metamorphic platform sediments (the Beit Bridge Complex), intruded at 3·3 Ga by granitoids, named the Sand River Gneiss (SRG) and by mafic–ultramafic rocks of the Messina Suite (MS), related to the Mid-Archaean (D1/M1) event (e.g. Kröner et al., 1999). Subsequently (at 2·6 Ga) granitic intrusions of various ages, including the precursors to the Alldays- and Singelele Gneisses and the Bulai pluton, were emplaced in the Beit Bridge Complex, SRG and MS (e.g. Van Reenen et al., 1990; Hofmann et al., 1998; Kröner et al., 1998, 1999). Syntectonic magmatism (related to the Alldays and Singelele Gneisses) occurred between 2·67 and 2·57 Ga and accurately dates the D2/M2 Late Archaean event (e.g. Kröner et al., 1999). The Bulai pluton, on the other hand, was emplaced into high-grade ductilely deformed rocks of the Beit Bridge Complex in the area west of Messina at 2·57 Ga (Barton et al., 1994) (Fig. 2a), and shows little evidence of the D2 deformation event (McCourt & Armstrong, 1998).
The D2/M2 event is thus related to an 100 Myr period of extensive granitic magmatism that occurs throughout the Limpopo Complex (Kröner et al., 1999). Clear evidence for a deformational event (D3) that postdates D2 in the area SE and east of Messina, according to Hoffman et al. (1998) and Kröner et al. (1998, 1999), is expressed by various sets of shear zones. Similar shear zones are also developed in the Bulai pluton (Watkeys, 1983), and various undeformed granitic patches and pegmatite veins are dated at 2·0 Ga (e.g. Holzer et al., 1998; Kröner et al., 1999).
The complex regional deformation pattern that characterizes the entire CZ (Fig. 1) comprises two major structural domains, namely a high-grade ENE–WSW-directed ‘Straightening Zone’ in the south, and the complex folded area north of the Straightening Zone referred to as the ‘Cross Folded Zone’ (CFZ) (Fig. 1) (Söhnge, 1946; Bahneman, 1972; Watkeys, 1983). The Tshipise Straightening Zone is characterized by a monotonous ENE–WSW-trending foliation pattern that dips steeply towards the SSE. Fold axes and lineations dip moderately towards the WSW (Fripp, 1983). Two major folded structures are recognized in the CFZ: (1) cross folds, i.e. a regional spread of large NNW–SSE-trending folds with near-horizontal fold axes perpendicular to the regional strike of the CZ (e.g. Söhnge, 1946; Bahneman, 1972; Pienaar, 1985; Pretorius, 1986; Feldtmann et al., 1995; Feldtmann, 1996); (2) sheath folds characterized by moderately WSW-plunging central fold axes (Roering et al., 1992). The best examples of sheath folds in the CZ in South Africa include the Avoca fold developed in quartzo-feldspathic gneisses (Singelele Gneiss) in the Baklykraal area west of Alldays (Fig. 2b, stereonet 1), and the Bellevue (Fig. 1) and Ha-Tshansi (Fig. 2a) folds developed in rocks of the Beit Bridge Complex near Messina. Two major cross folds are represented by the Baklykraal (Fig. 2b) and Campbell (Fig. 2a) folds.
All sheath folds in the CZ demonstrate identical fold geometries with central fold axes that show consistent top-to-the-NE movement (e.g. Fig. 2a and b). The Avoca sheath fold (Fig. 2b) is developed in Singelele-type gneisses about 20 km WNW of the Baklykraal cross fold and is an important time marker for the development of sheath folds in the CZ, and thus of the D2/M2 event. Boshoff (2004) have obtained sensitive high-resolution ion microprobe (SHRIMP) Pb–Pb zircon ages of 2651 ± 8 Ma from gneisses in the rim of the structure. These gneisses preserve a strong S2 shear fabric and a single population of mineral stretching lineations (L2, Fig. 2b, stereonet 1) that define a steeply SW-oriented central fold axis. These penetratively deformed gneisses grade continuously into an unfoliated, but strongly lineated variety of the same rock (the Avoca granite) in the core of the oval-shaped structure. Mineral stretching lineations in the Avoca granite (dated at 2570 ± 22 Ma, Boshoff, 2004) are also developed parallel to the central axis of the fold. The Avoca sheath fold therefore constrains the timing of the D2/M2 tectonometamorphic event within a relatively narrow time period in the Late Archaean.
Structural evolution of the Baklykraal cross fold
The 40 km x 80 km Baklykraal cross fold (Fig. 2b) offers an ideal opportunity to study the structural and metamorphic evolution of cross folds in the CZ as well as the timing of this important event based on single-phase dating of garnet in metapelitic gneisses (Boshoff, 2004). The Baklykraal cross fold is an ENE-trending synformal structure with a near-horizontal fold axis developed in gneissic rocks of the Beit Bridge Complex. These include marble, calc-silicates, mafic gneisses, quartzites, garnet–biotite gneisses and minor metapelitic gneisses. Two well-developed fabrics (S2, S3) are present within the rocks forming the Baklykraal fold. The S2 gneissic fabric that defines the form of this structure (Fig. 2b, stereonets 2–6) in outcrop is similar to the S2 gneissic fabric of the sheath folds. However, in outcrop S2 is often enhanced by trails of small garnet grains or by elongated garnet grains that are oriented parallel to the fold axis. This S2 gneissic fabric in the nose portion of the fold is cut by a near-vertical north–south-oriented axial planar S3 cleavage (Fig. 2b, stereonets 2–6). This relationship can be interpreted to suggest that the S2 fabric preserved in the Avoca and similar sheath folds in the CZ was reworked during a subsequent geological event that resulted in the formation of the cross folds. The reworked S2 fabric is thus referred to as S3.
Two populations of linear elements are also displayed. Feldtmann (1996) recorded a population of linear elements (boudinaged calc-silicates, cigar-shaped pegmatite boudins, quartz rods, minor fold axis, etc.) from the Baklykraal quarry (Fig. 2b, stereonet 2) developed in marble and calc-silicates near the centre of the fold. These elements demonstrate a consistent linear direction that plunges at 5–12° in a general north–south direction, defining the ß-fold axis of the synformal structure and are referred to as L3ß. A second population of moderately SW-plunging mineral stretching lineations (L3 s) are also developed in metaquartzite (Fig. 2b, stereonet projections 3 and 4), suggesting movement of material to the NE along planes that developed parallel to the original bedding planes, mainly within the limbs of the fold.
The presence of two fabric-forming events (D2, D3) in the Baklykraal cross fold is also supported by recent single-phase dating of garnet (Boshoff, 2004) from a metapelitic gneiss with an S3 fabric (Fig. 2b, sample T73). The results of this study suggest that the garnet in sample T73 grew during a high-grade metamorphic (M3) and deformational (D3) event at 2·0 Ga. The asymmetric Campbell cross fold (Figs 1 and 2a) near Messina displays a similar structural geometry and is, therefore, considered to have developed at the same time.
In the following sections we decipher the metamorphic evolution of metapelitic gneisses from the Baklykraal cross fold.
|
METAMORPHISM |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL FRAMEWORK OF THE...METAMORPHISMTHERMODYNAMICS OF METAMORPHISMCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Petrography
The main purpose of this section is (1) to characterize the mineral assemblages and their compositions in the rocks from the Baklykraal cross-fold, and (2) to show the relationships between the minerals in the various rock types. This allows us to create the basis for a detailed thermobarometric study.
With the exception of massive recrystallized almost pure quartzite all rock types from the Baklykraal structure are in outcrop characterized by a thinly banded gneissic layering often enhanced by strongly attenuated leucocratic zones. However, different rocks from this structure show different degrees of deformation, varying from unfoliated (e.g. J4, Table 1) to highly foliated (T73, T20, T53, Table 1). This observation is very common for high-grade gneisses over the entire Limpopo Complex (e.g. Van Reenen et al., 1990; Smit & Van Reenen, 1997; Smit et al., 2001). Recrystallized quartzite contains more than 90 modal % quartz with minor K-feldspar, albitic plagioclase, Fe-rich spinel, biotite, and accessory minerals such as zircon, apatite, sphene, and ilmenite. Garnet-bearing grey gneisses include three major varieties: (1) orthoamphibole–garnet–quartz–plagioclase gneisses; (2) biotite–garnet–plagioclase–quartz–sillimanite gneisses; (3) cordierite-bearing metapelitic gneisses (see Table 1) that are the major focus of this study. Up to several metres thick calc-silicate and marble units are intercalated with mafic and pelitic gneisses.
|
Mg-rich orthoamphibole-bearing rocks (Pienaar, 1986) do not show any reaction textures. All samples studied show either a weak foliation or no foliation at all (e.g. J4 in Table 1). In addition to orthoamphibole the rocks always contain orthopyroxene, plagioclase and quartz. In some varieties of this rock type high-aluminium orthoamphibole coexists with cordierite and orthopyroxene. Orthoamphibole-bearing rocks are locally characterized by the presence of micro shear zones composed of fine-grained mylonitic material that includes the assemblage biotite–sillimanite–quartz.
Biotite–garnet gneisses contain sillimanite, quartz, plagioclase and K-feldspar. Some gneisses (e.g. T53, Table 1) are poor in garnet but rich in sillimanite, whereas others show the opposite relation. Despite the fact that the gneisses are migmatitic there is no evidence for the reaction microtexture Bt + Sil + Qtz Crd + Grt + Kfs + melt that would indicate the fluid-absent melting of biotite. The gneisses also contain thin, <150–200 µm thick, micro shear zones composed of intergrowths of sillimanite and quartz. In comparison with the matrix of this rock, these micro shear zones are poor in potassium-bearing minerals such as K-feldspar and biotite, which are concentrated along the margins of the zones.
Marbles and calc-silicate rocks are well exposed in the Baklykraal quarry in three dimensions. The following three petrographic groups (see also Feldtmann, 1996) of consistent mineral compositions occur among marbles (all minerals are given in modal %): (1) calcite marble with rare phlogopite and diopside; (2) phlogopite marble (Cal 70–80, Phl 5–15, Di 3–5, Qtz 3 and secondary actinolite after Di); (3) diopside–phlogopite marble (Cal 85, Di 3–5, Phl 3–1, and also scapolite, quartz, K-feldspar, dolomite and tourmaline). The calc-silicate rocks are severely boudinaged, a feature that is well exposed in both the vertical and horizontal plane in the quarry. The compositional banding is characterized by alternating layers of fine-grained Di–Phl- and Qtz–Kfs–Cal-rich zones on millimetre to centimetre scale. The most prominent feature of the rock is the strong orientation of phlogopite parallel to the banding. Both the marbles and calc-silicates contain small (centimetre size) boudins composed of quartzite, metapelite and pegmatite. The long axes of all kinematic elements are oriented parallel to the fold axis of the cross fold (Fig. 2b, stereonet projection 2).
Metapelites occur as lenses, boudins and thin layers (up to a few metres thick) in biotite–garnet and mafic gneisses, and marble, as well as enclaves in quartzite. Typical mineral assemblages and microtextures of metapelites are shown in Table 1 (samples T18, T20, T73) and Figs 3–6, respectively. In outcrop metapelites display leucosomes, migmatitic patches and narrow cross-cutting granitic veins. The well-developed S3 gneissic foliation is often enhanced by the parallel orientation of biotite, sillimanite and by the presence of very narrow trains of small D3 garnet grains or elongated D3 garnet porphyroblasts that are oriented parallel to the fold axis (L3ß) of the cross fold (see Fig. 4e and f). Small, up to 1·5 cm, Qtz-rich lenses also occur along the foliation. The lenses are coated by biotite and cordierite, and, in places, contain almost euhedral Grt porphyroblasts. Rare Bt–Sil–Qtz micro shear zones always follow the foliation (see Fig. 5d). The leucocratic material of metapelites does not preserve evidence for the fluid-absent melting reaction Bt + Sil + Qtz Crd + Grt + Kfs + melt. This suggests that at the peak of metamorphic conditions the metapelites did not experience partial melting. However, to avoid possible effects related to late migmatitic events, only leucosome-free pelitic samples were used in our studies.
|
|
|
|
Porphyroblastic grains of quartz and garnet occur in a granoblastic matrix composed of biotite, feldspars, quartz and sillimanite. Apart from the matrix sillimanite (Sil1) a second generation, Sil2, occurs with Qtz and Bt as a part of the reaction textures developed around the garnet porphyroblasts. In addition, sillimanite (Sil3) together with Qtz forms up to 150–200 µm thin shear zones showing ambiguous relationships with the reaction textures (see Figs 3a and 5a). Cordierite occurs in slightly altered well-developed corona textures replacing garnet (Fig. 4a), or as euhedral porphyroblasts included in a fine-grained quartz–sillimanite–plagioclase schistose granoblastic matrix (Figs 3 and 4). Fresh cordierite forms porphyroblasts up to a few millimetres in size with well-developed twins. In other places cordierite mimics the morphology of the replaced garnet porphyroblasts (Fig. 4a). The cordierite porphyroblasts are commonly filled with sillimanite (Sil1) and/or quartz inclusions (Fig. 4d). Garnet forms both inclusion-free asymmetric grains elongated along the S3 schistosity (Figs 3a and 5a–c) and atoll-shaped crystals (Fig. 3c). The formation of the asymmetric shape of garnet resulted from the replacement of elongated oval-shaped garnet porphyroblasts by cordierite from one side only, whereas the reaction-free side in contact with quartz and/or feldspar preserved the rounded primary shape. In sample T20 garnet is overgrown by well-developed cordierite–quartz symplectites (Fig. 4c and d). Backscattered images (BSI) in Fig. 4a–c demonstrate a systematic replacement of garnet by cordierite after the net-transfer divariant reaction
 | (1) |
; Van Reenen et al., 1990). In some thin sections garnet porphyroblasts contain inclusions of quartz, plagioclase, and biotite. Micron-size inclusions of sillimanite are present in quartz porphyroblasts and millimetre-size sillimanite (Sil1) inclusions occur in both garnet and cordierite porphyroblasts. In sample T20 a few grains of Sil2 occur in corona textures, whereas Sil3 composes up to 90 modal % of the above-mentioned micro shear zones (e.g. Fig. 4f). Rare K-feldspar, Or94–100 (Table 2), is present in all studied metapelites together with biotite, sillimanite (Sil2) and quartz. In a few cases we observed K-feldspar in paragenesis with garnet, but never with both cordierite and garnet, suggesting the displacement of the net-transfer reaction (1) and the reaction
 | (2) |
|
In summary, the metapelites from the Baklykraal cross fold show the following three distinct features that developed during D3/M3: (1) systematic replacement of garnet, quartz and Sil1 by cordierite after reaction (1); (2) the simultaneous development of the Bt + Qtz + Sil2 assemblage instead of the Grt + Kfs paragenesis suggesting reaction (2), a typical retrograde hydration reaction; (3) the formation of the Bt + Qtz + Sil3 assemblage that characterizes the micro shear zones, which probably relate to the latest stage of a single-stage retrograde metamorphism (M3).
Reaction textures and mineral chemistry
From the previous section we can conclude that the relationship between minerals in metapelites from the Baklykraal cross fold is characterized by the widespread development of at least two major types of reaction textures: (1) replacement of garnet (± Sil1 + Qtz) by cordierite (Figs 3–5); (2) replacement of the assemblage Grt + Kfs by the paragenesis Bt + Qtz + Sil2 (Fig. 3). Both these ‘classical’ reaction textures reflect the retrograde stage of metamorphism (e.g. Harley, 1989) and provide a key for estimating the 
conditions during exhumation of the granulite-facies rocks. In the Petrography section we showed that these shear zones also occur in other rock types of the Baklykraal cross fold. Similar micro shear zones also are characteristic of many granulite-facies terrains including the SMZ and the NMZ of the Limpopo Complex.
To calculate the thermodynamic parameters we used the method of systematic microprobe profiling of the minerals forming a reaction texture (e.g. Perchuk et al., 2000). Minerals were analysed using a CAMECA electron microprobe at the Rand Afrikaans University (Johannesburg, South Africa) and a CAMECA SX50 at both the Ruhr University (Bochum, Germany) and the Department of Petrology at Moscow State University (Russia).
Selected microprobe analyses for representative profiles, the majority of which relate to the reaction textures, or to coexisting Fe–Mg minerals that are in contact with each other, are provided in Tables A1 and A2, which can be downloaded from the Journal of Petrology website at http://www.oupjournals.org. Figures 3, 5d and 6 demonstrate chemical profiles of contacting cordierites and garnets. In addition, for the purpose of thermobarometry we have analysed garnet and cordierite porphyroblasts from the sillimanite (Sil1) quartz matrix to avoid possible Fe–Mg diffusional exchange in the course of the net-transfer reaction (1).
Zoning profiles (Fig. 3) reflect the redistribution of Mg and Fe between cordierite and garnet according to the exchange reaction
 | (3) |
, 1977). The shapes of the compositional profiles for garnet and cordierite in contact with each other reflect the crystallization of minerals in the four-phase field (Grt + Sil1 + Qtz + Crd) during the progress of reactions (1) and (3). For example, a large grain of garnet along profile IV (Fig. 3g) is isolated from cordierite by a narrow Qtz + Sil2 ± Bt symplectitic zone (Fig. 3c), which results in a decrease of the Mg-numbers of garnet and cordierite towards their rims (Perchuk, 1969, 1977). This profile also may reflect a shift to the right of both the net-transfer reaction (2) and the associated (simultaneously operating) exchange reaction
 | (4) |
The Mg-number of garnets from samples T73 (Table A1) and T20 (see Table 4) varies between 13 (rims) and 24 (cores), whereas in metapelite T18 marginal portions of Grt porphyroblasts have slightly higher NMg (20) whereas the same NMg (24) is preserved in cores (see Table 4). The Ca content of garnet is low and decreases from core to rim. This is a common feature for retrograde granulites because of the breakdown of the grossularite component of garnet to produce anorthitic plagioclase, quartz and sillimanite (Newton & Haselton, 1981). In contrast, the low Mn content of garnet shows no systematic change from cores to rims of Grt porphyroblasts. Similar to garnet, cordierite in sample T18 shows a higher NMg (66–70, for rim to core, respectively) compared with samples T73 and T20 (61–66, for rim to core, respectively). Thus, the rims of the T18 cordierite have the same NMg as the cordierite cores in samples T73 and T20, a very unusual feature of metapelites sampled from the same geological structure. This compositional difference reflects a significant difference in temperature of the Grt–Crd equilibrium (Perchuk, 1977) because the 
(24) in the cores of all samples reflects identical pressure conditions. The direct contacts of garnet with cordierite are characterized by a slight decrease of (1–2), and an increase of 
(62–66) (e.g. Fig. 3d–f, profiles I–III). A large, about 5 mm, atoll garnet was analysed from core to rim along Profile III in Fig. 3f. varies from 24 (core) to 16 (rim), whereas is fairly constant in the core (62–63) and increases to the rim (up to 66) in contact with garnet. Several contacting grains of cordierite and garnet were analysed along profile IV (Fig. 3g). The major features of this profile are the following: (1) a decrease of and where the garnet and cordierite are separated by thin (50 µm) intergrowths of Qtz and Sil2 (Fig. 3g), demonstrating a shift to the right of reaction (1); (2) absence of systematic chemical zoning in cordierite in contact with garnet (Fig. 3g).
The studied mineral assemblages and their chemical features are thus well suited to correctly decipher the P–T history of the Baklykraal cross fold, if the potential pitfalls are taken into consideration. The samples represent both high- and low-strain coarse-grained rocks in which Grt and Crd grain sizes may exceed 1 mm, and contain reaction textures extensively developed after garnet (Figs 2 and 3). Histograms in Figure 7a–c show that in spite of the near-normal (Fig. 7a and b) or log-normal (Fig. 7c) distributions of NMg, the data demonstrate only one peak. This suggests that the garnet in the studied metapelites represents a single generation. In addition, zoning profiles of Grt and Crd have clear central plateaux with consistent compositions (Figs 4–6) that are considered to represent the highest P–T conditions. Whereas both the garnet and K-feldspar are being consumed in reactions (1) and (2), some other minerals are present in the reaction textures in two continuous generations (e.g. Bt1 Bt2, Sil1 Sil2). This is clearly exemplified by Fig. 7d, where the TiO2 content of biotite decreases systematically with 
, and, therefore, with fall in temperature (see Fig. 12c and d). This also follows from chemical zoning of biotites showing a decrease in TiO2 from the core through the margin to the rim; for example, 4·5 3·6 3·0 wt % (see Table A2). Thus, the composition of biotite changes continuously from the peak Bt1 to Bt2 in the course of the progress of reaction (4).
|
In the Petrography section we mentioned that no reaction textures related to reaction (2) were observed because of almost total consumption of K-feldspar presented by Or100–95 (Table A2). We measured several biotite–garnet pairs in contact with Qtz and Sil2 that are associated with K-feldspar (Tables A1 and A2). Biotite has a relatively narrow range in composition and shows a strong negative Ti–Mg correlation (Fig. 7d): TiO2 (wt %) = –15·951 XMg + 11·064 (within the limits of data from Table A2), whereas the Al content shows no correlation with
(Fig. 7e) at an almost constant composition of K-feldspar (Table A2). Profiles I–III (Fig. 3) reflect the progress of reaction (3), whereas profile IV demonstrates a very complex chemical zoning because the quartz–sillimanite intergrowth around cordierite may not be in direct contact with garnet, although some sillimanite (Sil1) and quartz occur as inclusions in garnet. The coexistence of minerals with both types of zoning profiles reflects the simultaneous operation of both reactions (1) and (3).
|
THERMODYNAMICS OF METAMORPHISM |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL FRAMEWORK OF THE...METAMORPHISMTHERMODYNAMICS OF METAMORPHISMCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
In this section we utilize the mineral chemistry from the studied rocks to understand the thermal and dynamic history of exhumation and emplacement of the Baklykraal cross fold. For this purpose we will use mineral relationships in the metapelites (Figs 3–6) and the observed changes in their compositions with pressure and temperature on the basis of the method described in detail by Perchuk et al. (2000
, and references therein) and Smit et al. (2001).
Mineral thermodynamics
We use an internally consistent dataset on mineral thermodynamics for the system FeO–MgO–SiO2–Al2O3–K2O–Na2O–H2O, which is derived from detailed experimental data aimed at the calibration of a set of geothermobarometers (Zyrianov et al., 1978; Perchuk & Lavrent'eva, 1983; Aranovich & Podlesskii, 1989). Standard state data for the end-member reactions (1)–(4) are given in Table 2. Mixing properties of the mineral solid solutions of the reactions are taken from Gerya & Perchuk (1994).
Mg–Fe cordierite
It should be mentioned that reaction (1) noticeably depends on the water content of cordierite (Newton, 1972). To account for H2O and CO2 solubility in cordierite Aranovich & Podlesskii (1989) suggested a garnet-like solid solution model:
 | (5) |
 | (6) |
 | (7) |
because the concentration of water in Crd is unknown. Substituting the term Q in formulae (5) and (6) using equation (7) we obtain
 | (8) |
 | (9) |
Pyrope–almandine solid solution
 | (10) |
 | (11) |
Biotite
Partial molar Gibbs mixing free energies of biotite solid solution
 | (12) |
 | (13) |
):
 | (14) |
and 
, respectively. Molar fractions XPhl, XAnn and XEas were calculated with the following formulae:
 | (15) |
 | (16) |
 | (17) |
.
K-feldspar
A sub-regular model for the high albite–sanidine solid solution (Zyrianov et al., 1978) can be used for calculations of 
via reaction (2) by the use of compositions of K-feldspar (Table 3) in direct contact with the Bt + Sil2 + Qtz assemblage developed around garnet (e.g. Fig. 3g).
|
Using these mixing properties and standard state thermodynamic data from Table 2 we have calculated the T–P–NMg phase diagram for divariant equilibrium (1). Figure 8 demonstrates an isothermal (700°C) P–NMg diagram. All the discussed properties and data of Table 2 allow calculation of the
parameters for local mineral equilibria in metapelites of the Baklykraal cross fold.
|
P–T paths
Because P–T paths are the basis for geodynamic speculations, the method of their derivation will always be important. For the past two decades, workers have used different approaches for constructing P–T paths with the result that published P–T trajectories for the same geological unit may have different shapes. Also, P–T paths from complexly deformed high-grade terrains cannot be accurately derived without integrating petrological and structural data. The CZ of the Limpopo Complex is one of the best examples of controversies concerning both the shapes of published P–T paths and their geodynamic interpretation.
Detailed paragenetic analysis of widely developed retrograde Crd-bearing mineral assemblages in aluminous rocks from different parts of the CZ in South Africa and Botswana led to the conclusion that isothermal-decompression (ITD) (Horrocks, 1983; Harris & Holland, 1984; Windley et al., 1984; Droop, 1989; Van Reenen et al., 1990), isobaric cooling (IBC) combined with ITD (Tsunogai et al., 1992; Hisada & Miyano, 1996), or decompressional-cooling (DC) (Perchuk et al., 1996, 2000) P–T paths reflect a single exhumation event. Holzer et al. (1998), on the other hand, proposed two contrasting P–T paths for the evolution of the Central Zone: an anticlockwise P–T path reflecting a Late Archaean (2·6–2·55 Ga) low-pressure granulite-facies metamorphism associated with voluminous granitic and charnockitic plutonism, and a Proterozoic (2·0 Ga) high-pressure ITD path. Evidence for the anticlockwise P–T path is based mainly on the results of unpublished data (Holzer, 1995) from metapelitic gneisses from the ‘Three Sisters’ area on the farm Boston, about 20 km WNW of Messina, in which rectangular sillimanite was interpreted to reflect pseudomorphs after andalusite.
In the following section we will derive a P–T path for metapelites from the Baklykraal cross fold based on reaction textures and mineral chemistry (Tables 2 and A2), and on the dataset (Tables 3 and 4) to demonstrate the coherence of the thermodynamic results with the discussed structural evolution of this cross fold.
|
Thermobarometric studies and H2O activity calculations
Thermobarometry of granulites is recognized as a complex problem because of the high-temperature conditions and strong recrystallization of rocks at peak metamorphic conditions that obliterate most compositional and textural evidence for the pre-peak history. Therefore, in most cases only retrograde P–T paths are documented for granulites based on reaction textures and chemical zoning of coexisting minerals (e.g. Perchuk, 1985
, 1989; Harley, 1989). In addition, many ion-exchange geothermometers have apparent closure temperatures below those of granulite-facies conditions, causing difficulties in obtaining P–T parameters of peak metamorphic conditions for granulite-facies terrains. Several workers (e.g. Frost & Chacko, 1989; Harley, 1989; Spear & Florence, 1992; Spear, 1993) have also noted that geothermometers based on Fe–Mg exchange reactions are not likely to quench at the same P–T conditions as geobarometers that are based on net-transfer reactions. Therefore, in several cases, retrograde P–T paths of granulites deduced from thermobarometry using mineral compositions affected by late Fe–Mg exchange may give misleading results, representing apparent P–T trajectories (e.g. Frost & Chacko, 1989; Spear & Florence, 1992; Spear, 1993). Frost & Chacko (1989) suggested several procedures to maximize the correct retrieval of peak metamorphic conditions and retrograde P–T paths for granulites. The approach used by our group (Perchuk et al., 1985, 1996, 2000; Smit et al., 2001) for derivation of a P–T path not only involves the majority of the above-mentioned procedures but also accounts for other methods such as numerical modelling of P–T paths for rocks of high-grade terrains (e.g. Gerya et al., 2000, 2002). Taking into account that different dT/dP slopes of P–T paths are commonly associated with specific mineral reactions observed in granulites (e.g. Harley, 1989), a crucial step is the use of reaction textures to test P–T trajectories of granulites (Perchuk et al., 1989, 1996, 2000; Smit et al., 2001).
For thermobarometric calculations we used an internally consistent thermodynamic dataset of Perchuk et al. (1985; see also Perchuk, 1990), calibrated on the basis of original experimental data on the mineral equilibria involving solid solutions of garnet, cordierite and biotite (Perchuk & Lavrent'eva, 1983; Aranovich & Podlesskii, 1989). The cordierite–garnet [reaction (3)] and the biotite–garnet [reaction (4)] exchange equilibria were used for the estimation of temperatures parameters, whereas reactions (1) and (3) were utilized for calculating the pressure and , respectively, for local mineral equilibria. At equilibrium conditions the Gibbs free energy is the following:
 | (18) |
can be calculated from the H2O activity in a fluid. Because all studied metapelites of the Baklykraal area contain the Bt + Sil + Qtz + Kfs + Grt reaction texture, indicating reaction (2), we can apply the equation
 |
 | (19) |
. Therefore to calculate external thermodynamic parameters (T, P and ) of local equilibria for the derivation of the P–T path, we have to use (1) an internally consistent thermodynamic database, and (2) the correct method of investigating the equilibrium compositions of coexisting minerals in metapelites from the Baklykraal cross fold.
The standard state thermodynamic data from Table 2, mixing properties from formulae (10)–(14) and the Ab–San solid solution model allow calculation of with equation (19) at a given T and P. There are two ways to do this. The first one is the joint solution of equations (18) and (19) at the temperature calculated with the Bt–Grt thermometer. The second way is based on the negligible contribution of the PV° term in G(3): the variation of P within 5 kbar results in a change of of not more than 0·005. Both methods show similar, or even identical results. Table 3 demonstrates the results for 29 Bt + Sil + Qtz + Kfs + Grt assemblages, which were described in the previous section (see Table A2).
However, because the H2O activity depends on temperature, it is more practical to find analytical correlations between T and using data from Table 3.
Muscovite is a critical mineral in the studied metapelites in terms of the lowest temperature limits of the retrograde stage of metamorphism: it occurs very rarely as a secondary phase after Fsp and Sil2. The upper temperature limit cannot be higher than 800°C (Spear et al., 1999) because partial melting is not characteristic of the metapelites studied, and orthopyroxene is absent. Therefore reaction (2) proceeds within a relatively narrow range. We used this temperature range to make corrections for calculations of both cordierite and garnet isopleths and the determination of for each particular local equilibrium at temperatures estimated by the Bt–Grt thermometer. Using the data from Table 3 (some selected analyses are given in Tables A1 and A2) and taking into account that pressure has a negligible effect on the biotite–garnet thermometer, and in particular on the H2O activity, we have approximated for sample T73 as a linear function of temperature (Fig. 9) within the interval 550–750°C:
 | (20) |
and ) on the P–T plane of the system. A similar procedure was used for calculating for sample T20: it is approximated by the following equation (see also Fig. 9):
 | (21) |
varies within the limits 0·294–0·260, being very close to the calculated for sample T20. This very small change allows (1) assigning an average value, i.e. 0·282, or (2) integrating all values for sample T18 from Table 3 with the 
data for sample T20. This integration gives a result that is almost identical to approximation (21):
 | (22) |
and isopleths and the P–T path derived from chemical zoning of coexisting minerals in the studied metapelites. Such a simple approach allows the calculation of Crd and Grt isopleths in the P–T field for reaction (1) at a given , as well as P–T parameters for the cordierite–garnet assemblages from local equilibria in the reaction textures (see Fig. 4). We also used the temperature interval of the evolution of the biotite–garnet equilibrium (see Table 3). Selected P–T parameters for some local mineral equilibria are given in Tables 3 and 4.
|
). The Grt (continuous lines) and Crd (dashed lines) isopleths show the corresponding Mg-numbers. Both the isopleths and data points for each sample in (a)–(c) are calculated on the basis of thermodynamic data from The method
For the accurate derivation of a P–T path for the studied metapelites from the Baklykraal cross fold we used the procedure described in several papers (Perchuk et al., 1985
, 1989, 1996, 2000; Perchuk, 1987, 1989; Gerya & Perchuk, 1997; Gerya, 1999; Smit et al., 2001) that involves the following major steps. (1) Microstructure studies. The sequence of events related to the simultaneous growth of minerals in the reaction textures [e.g. reactions (1) and (2)] should be established.
(2) Microprobe studies. The chemical evolution of rock-forming minerals, involved in different types of reactions (e.g. net-transfer and exchange reactions), must be based on detailed microprobe profiling. This, however, should only be done after a preliminary microprobe study of relatively large porphyroblasts of garnet and other coexisting minerals involved in the reaction textures. A flat profile through the central portions of such porphyroblasts may indicate a peak of metamorphic conditions (e.g. Spear & Florence, 1992). This allows an understanding of the contribution of exchange [e.g. reactions (3) and (4)] and net-transfer reactions [e.g. reactions (1) and (2)] during the coherent change of mineral compositions in the course of exhumation.
(3) Geothermobarometric calculations. P–T parameters for local equilibria should be calculated for the different stages of the formation of mineral assemblages using the ‘core–core rim–rim’ method. Mineral compositions of cores and rims can be selected using the results of the microprobe profiling, to avoid the influence of the latest exchange diffusion on barometry (e.g. Frost & Chacko, 1989; Spear & Florence, 1992). In the case of a cordierite-bearing equilibrium 
must be calculated via a simultaneously operating reaction.
(4) Derivation of a P–T path. A P–T trajectory should be derived on the basis of individually calculated P–T data for local mineral equilibria. The consistency of the P–T trajectory must be tested with both mineral zoning (i.e. the relationship of the P–T path with calculated compositional isopleths in observed divariant assemblages) and reaction textures.
(5) Fluid inclusions test. If fluid inclusion data for the studied samples are available, the combination of densities of the inclusions with mineral thermometry data can be used for the independent derivation of a P–T path. Its comparison with the P–T trajectory, derived from mineral thermobarometry only, may help to demonstrate the accuracy of the derivation (e.g. Perchuk et al., 1985, 1989; Perchuk, 1987, 1989).
(6) Mineral mode test. It is recommended that the derived P–T path be tested by calculating the Grt mode isopleths on the basis of bulk-rock compositions (e.g. Gerya, 1991, 1999; Perchuk et al., 1996, 2000; Smit et al., 2001), or using the polymineral modes (e.g. Carson et al., 1999; Wei et al., 2003).
(7) Numerical test. An additional test is that the derived P–T path should conform to the results of numerical modelling (e.g. Gerya et al., 2000, 2002).
(8) Application. The correct geological and geodynamic interpretation of the derived P–T path.
Step (5) is useful, if well-developed reaction textures are absent from the particular metamorphic rock. Steps (7) and (8) can be extremely useful, if detailed structural and petrological data from major geological structures are integrated, such as is the case with the Baklykraal cross fold in the CZ of the Limpopo Complex.
Derivation of a P–T path for the Baklykraal metapelites
Detailed microstructural studies are used to define the most appropriate reaction textures (see Figs 3–6) reflecting changing thermodynamic parameters. The systematic profiling of garnet porphyroblasts has been done for each studied sample with the aim of defining a maximal NMg along a flat plateau [see step (2) in the previous subsection]. Histograms (a)–(c) in Fig. 7 demonstrate the distribution of NMg for the central portions of relatively large (>0·5–1 mm) garnet porphyroblasts (e.g. Spear & Florence, 1992), and profiles in Figs 3, 5 and 7 show the NMg plateaux for both Grt and Crd. Thus, we are confident that the mineral compositions, equilibrated at near-peak metamorphic conditions, are correctly determined. Equilibrium compositions of rims were assumed for grains that are unaffected by very late exchange diffusion (Lasaga, 1983). Such rims are either isolated by the inert products of mineral reactions (e.g. Fig. 3g), or include rims that do not show typical diffusion profiles (Figs 3d–f, 6a, b and d).
A P–T trajectory can be derived on the basis of data calculated individually for a particular local equilibrium. Therefore, the consistency of a P–T path is controlled both by the mineral zoning (i.e. the relationship of the P–T path with calculated isopleths for Grt, Crd, and Bt in the observed divariant assemblages), and by the reaction textures. However, in the cases of reactions (1) and (2) the isopleths on the P–T plane should be projected from , i.e. the H2O activity must vary along each particular isopleth in accordance with equations (20)–(22).
Thus, using individual P–T estimates we deduced P–T paths for three samples, T18, T73, and T20. Figure 9 demonstrates the results. All data points in the diagrams represent individual measurements of P–T parameters and for the assemblage Crd + Grt + Sil + Qtz + Bt + Kfs, as given in Table 4. These P–T paths are compared with and isopleths (Fig. 9) calculated using thermodynamic data for reactions (1) and (3) (Table 2) and our calibrations of as a function of temperature (corresponding equations are shown in Fig. 9a–c).
The P–T path for sample T18 (Fig. 9a) slightly intersects both the garnet and the cordierite isopleths. The path for sample T73 (Fig. 9b) was derived on the basis of the P–T–data from Table 4. It follows the cordierite isopleth and cross-cuts the garnet isopleth, reflecting changes in along a profile from the core to the rim (e.g. Fig. 3e and f). The T73 P–T path is very similar to that of sample T18. As mentioned above, all three diagrams are a complex projection of the Crd–Grt–Sil–Qtz system onto the P–T plane from and NMg of the Fe–Mg minerals. The cordieritein sample T20 demonstrates two types of chemical zoning (Figs 5 and 6): (1) an increase of in direct contact with garnet in a reaction texture (Figs 5d and 6a); (2) a decrease of in porphyroblasts included in the quartz–sillimanite (Sil1) matrix (Fig. 6b). Thus, we have the opportunity to check our statement on the simultaneous operation of reactions (1) and (3) in the studied samples by the derivation of P–T paths on the basis of analytical data on: (1) profiling of the reaction texture (Fig. 4, contacting Grt and Crd with diffusion-like zoning); (2) non-contacting cordierite and garnet porphyroblasts (Fig. 6, with opposite NMg zoning) from sample T20. Both P–T paths for this sample were derived at an identical and overlap each other completely, resulting in a single P–T path (Fig. 9c). The P–T paths also cross-cut three garnet isopleths reflecting changes in along the profile from the core to rim of garnet (e.g. Figs 5d and 6a), whereas a decrease in the is negligible (Fig. 9c, Table 4). The P–T path is similar to that of samples T18 and T73 (Fig. 9a and b).
The P–T diagram of Fig. 10 integrates all three trajectories into a single P–T path for the metapelites from the Baklykraal cross fold. The diagram also locates the P–T path between major reactions that limit the field of stability at relatively low temperature by the Fsp + Sil paragenesis, and at relatively high temperatures by partial melting of the rocks. In the Petrography section we mentioned the presence of rare muscovite and the absence of partial melting reactions in the metapelites, thus supporting the correct position of the P–T path in the diagram (Fig. 10).
|
. Data points (
, samples T20 and T73; (
, sample T18) and corresponding P–T trajectories (fine arrowed lines) are taken from 
), calculated after Discussion of the derived P–T path
In this section we discuss two of the most important problems related to the calculated P–T paths: (1) the reason for the linearity of the P–T trajectory in Fig. 10; (2) the shape and direction of the P–T path, i.e. the question as to whether the trajectory resulted from DC or IBD.
(1) The shape of the P–T path derived in Fig. 10. As mentioned before, near-linear P–T trajectories have been the subject of much discussion over the last 15 years (e.g. Frost & Chacko, 1989; Harley, 1989; Spear & Florence, 1992; Spear, 1993; Smit et al., 2001). Frost & Chacko (1989) attempted to prove that the linearity of a P–T trajectory is an artefact as such paths simply follow a Crd isopleth for any equilibria involving cordierite. Those workers suggested several approaches to avoid this problem. The majority of these approaches were discussed in the previous section of this paper and used in our study. Our previous studies (Perchuk et al., 1985, 1989, 1996, 2000; Smit et al., 2001) have shown not only that the linear trajectories often cross a Crd isopleth, but also that in many cases a linear P–T path can reflect an inflection caused by a change in the exhumation process (e.g. Smit et al., 2001). In the majority of cases this inflection is recorded in the formation of a new reaction texture (e.g. the intergrowth of Grt + Sil + Qtz after Crd, or the formation of Grt + Qtz symplectites after Crd + Opx) reflecting isobaric or sub-isobaric cooling of the metapelite. In addition, in our numerical modelling of P–T paths for high-grade terrains (Gerya et al., 2000, 2002) we reproduced the linear or near-linear shapes of such paths reflecting synchronous changes in the rheological properties (viscosity and density) of a metamorphic rock during its decompression and cooling (i.e. exhumation). However, a theoretical argument for the linearity of a P–T path has never been presented before.
As mentioned above, reactions (1)–(4) suggest a coherent change in compositions of all coexisting minerals, particularly Fe–Mg minerals, with variation of external thermodynamic parameters, i.e. P, T, and . If this coherency exists, the shift of these parameters of reactions (1) and (3) is recorded by the mineral equilibria. The compositional behaviour of minerals from all simultaneously operating reactions (1) –(4) with changing P–T conditions can also be predicted from the theory of phase correspondence (Perchuk, 1969, 1977). For example, a fall in temperature leads to the redistribution of Mg from H2O-free minerals to H2O-bearing coexisting minerals, whereas Fe shows the reverse redistribution (Perchuk & Ryabchikov, 1976), indicating that the exchange reaction (3) should be displaced to the right side with falling temperature. With reference to the metapelites from the Baklykraal cross fold this effect is exemplified by the profiles in Figs 3d–f, 5d and 6a. At isothermal conditions the net-transfer reaction (1), i.e. the divariant equilibrium, is displaced towards Fe-rich compositions (Fig. 8) with a decreasing pressure, therefore smoothing the effect of T on the exchange equilibrium (3). However, at a constant P the temperature effect on the coherent reaction (3) is shown in displacement of the phase boundaries on the P–NMg plane to the Mg-side (see Fig. 8). Thus, pressure and temperature affect the exchange reaction (3) in opposite ways. As a result, the synchronous fall of T and decrease of P must ultimately produce a negligible change in the composition of the cordierite (Table 4, Fig. 10). Thus, the linearity of a P–T trajectory is the direct result of the opposing influence respectively of P and T on the simultaneously operating reactions (1) and (3), respectively (Fig. 8).
(2) Direction of the P–T path as a reflection of exhumation. The decompression-cooling P–T trajectory derived in Fig. 10 reflects the exhumation of the studied rocks from the middle portion of the crust to the level of about 9–10 km, where the Baklykraal cross fold was emplaced. This DC path is in agreement with our previous results on both the CZ and the SMZ of the Limpopo Complex (Perchuk et al., 1996, 2000; Smit et al., 2001) but differs from published ITD paths (Horrocks, 1983; Harris & Holland, 1984; Droop, 1989; Van Reenen et al., 1990; Hisada & Miyano, 1996) for other parts of the CZ.
The accuracy of the trajectory of the DC path in Fig. 10 can also be proved based on the independent biotite–garnet equilibria (2) and (4). Both divariant equilibria are independent, as neither has been directly involved in the calculation of P–T paths in Fig. 9: we used reaction (2) to calculate the H2O activity so as to correctly position Crd–Grt isopleths and for individual P–T estimates. However, being part of the univariant equilibrium Grt + Crd + Fsp + H2O = Sil + Qtz + Bi, and thus operating simultaneously with reactions (1) and (3), reactions (2) and (4) can serve as good indicators of the accuracy of the P–T path derived in Fig. 10. We can prove this statement by using calculated Bt isopleths for reaction (2) as a control of the direction of the P–T path of Fig. 10.
The compositional relationships of minerals for the Bt-bearing system, i.e. reactions (2) and (4), should be similar to those of the Crd–Grt equilibrium, but with one very important difference: temperature affects reactions (2) and (4) more strongly than reaction (1). Therefore, a fall in T should result in a spectrum of iron-rich coexisting biotites and garnets. At a constant pressure, this conclusion is correct and supported by a large statistical dataset (e.g. Perchuk, 1977; Perchuk & Lavrent'eva, 1983) on compositions of coexisting biotite and garnet (more than 200 pairs) from metapelites of different metamorphic complexes around the world (see Fig. 11). Figure 11a shows a comprehensive empirical phase diagram for the KCFMASH system projected onto the T–XMg plane from P and H2O. According to the diagram both 
and 
decrease while temperature drops. On the other hand, the arrowed lines in Fig. 12c and d tend to intersect phase boundaries copied from Fig. 11a, demonstrating an increase of with fall in temperature. There it appears that the two diagrams of Fig. 12c and d and Fig. 11a contradict each other. However, this is only an apparent contradiction, because Fig. 11a reflects phase relations at unknown pressure and H2O activity. In contrast, the paths in Fig. 12c and d resulted from the simultaneous operation of reactions (1) and (3) as is controlled by changes in both the pressure and temperature as controlled by the exhumation history, i.e. the P–T trajectory in Fig. 10. The Korzhinskii AFM diagrams (Fig. 13) solve this contradiction.
|
) for
|
|
diagram (d) constructed by using the Marakushev approach (Marakushev & Perchuk, 1966Accounting for the discussed competition between exchange and net-transfer reactions, the AFM diagrams in Fig. 13a–c demonstrate the strong effect of falling temperature on exchange reactions (3) and (4), and, in addition, on the Crd–Bt exchange equilibrium (5) in Table 2. This leads to an increase of
in the univariant paragenesis Grt + Crd + Fsp + Sil + Qtz + Bt, interfering with the displacement of reaction (2) to the Fe-rich side (see Fig. 11a). The resultant effect is seen in Fig. 13c, which demonstrates an extremely small, or no change in the composition of cordierite, and an appreciable increase of . This is in contrast to the relatively large decrease of the (compare with Fig. 12b–d and Fig. 9). In the IBD case the P–T path would intersect more and more Fe-rich biotite isopleths.
The Korzhinskii method, in general, appears to be useful for the solution of such complex problems. Marakushev & Perchuk (1966) have strengthened this method by the semi-quantitative calculations of uni- and divariant lines on the P–T plane from a 
point within the 
space, accounting for the opposite effects of the H2O chemical potential and temperature on a hydration–dehydration reaction. Therefore, any line in the pseudoprojection of Fig. 13 suggests a systematic and coherent change of the water fugacity with falling T at a given . This pseudoprojection thus demonstrates, on the one hand, an increase of along the P–T path, and on the other hand the bounding of divariant reactions (1) and (2) by the univariant equilibrium Grt + Kfs + Crd + H2O = Bt + Qtz + Sil.
Using the standard thermodynamic data from Table 2, thermodynamics of solid solutions described by equations (5)–(19), and equations (20)–(22) for each metapelitic sample we can also calculate the quantitative Bt isopleths for the assemblage Grt + Bt + Kfs + Qtz + Sil on the P–T plane. We calculated isopleths for both net-transfer reactions (1) and (2) by using average compositions (formulae) of the Fe–Mg minerals coexisting in the samples (Table 5) and bounded by the integrated P–T path (Fig. 10). The calculations were carried out by varying both the XMg (n, m and k) and aH2O = –0·0002247T + 0·5017 at constant 
(see Table A2). The limits of XMg correspond to the Mg-numbers of Crd, Grt and Bt from the samples studied. The equation for the Gibbs potential, derived from the first principles (Gerya & Perchuk, 1997), has been used for calculating the water fugacity 
. The results are shown in the P–T diagram of Fig. 14, which is a pseudoprojection from the H2O chemical potential () onto the P–T plane within the 
space of the KFMASH system. The P–T path, taken from Fig. 10, slightly intersects the Crd63 isopleth and cross-cuts both the Grt and Bt isopleths, leaving no room for any alternative interpretation (e.g. IBD) of the decompression-cooling exhumation of rocks from the Baklykraal cross fold.
|
to 
. This corresponds to data from
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL FRAMEWORK OF THE...METAMORPHISMTHERMODYNAMICS OF METAMORPHISMCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The Central Zone of the Limpopo Complex is characterized by the presence of two major fold types reflecting distinctly different fold geometries and intensities of deformation: (1) large steeply WSW-plunging sheath folds; (2) large NNW-SSE-trending cross folds with near-horizontal fold axis. The Avoca sheath fold and Baklykraal cross fold are respectively developed in quartzo-feldspathic gneisses (Singelele Gneiss) and rocks of the Beit Bridge Complex west of Alldays. Gneisses in the rim of the Avoca sheath fold display a strong gneissic shear fabric (S2) that appears to be reworked during the subsequent formation of the gneissic fabric (S3) that defines the form of the Baklykraal cross fold.
Structural, petrographic and mineral chemistry data for metapelitic gneisses from the Baklykraal fold reflect a distinct stage of deformation (D3) and mineral growth during the post-peak metamorphic evolution (M3) of this structure. A systematic and monotonous change of mineral compositions in retrograde reaction textures in which garnet is being replaced by the assemblage Bt + Qtz + Sil and by Crd is the only characteristic of metapelites. In addition, no polymetamorphic signatures have been observed in other rocks forming this folded structure. Using thermobarometric estimations and calculated isopleths for the corresponding divariant reactions Bt + Qtz + Sil Grt + Kfs + H2O and Qtz + Sil + Grt Crd, belonging to the univariant equilibrium Grt + Kfs + Crd + H2O = Bt + Qtz + Sil, both the linearity and type (decompression-cooling) of a single P–T path are proved. The possibility of sub-isothermal decompression is completely ruled out. These data allow only one conclusion, namely for a single decompression-cooling metamorphic (M3) event accompanied by a distinct deformational (D3) event during the emplacement of the rocks forming the Baklykraal cross fold. This (D3/M3) event obliterated all petrological evidence for the early D2/M2 event in the studied metapelitic gneisses from this major fold.
The conclusion that the S2 gneissic fabric preserved in the Avoca sheath fold was reworked during the formation of the S3 gneissic fabric of the Baklykraal cross fold is supported by recent age dating: (1) zircon SHRIMP data (Boshoff, 2004) of syn- to late tectonic precursors to the Singelele gneisses that constrain the age of formation of the Avoca sheath fold and therefore the gneissic S2 fabric to 2·57 Ga; (2) single-phase garnet dating from a Baklykraal metapelitic gneiss (sample T73) that clearly shows that the gneissic (S3) fabric of the Baklykraal cross fold developed at 2·0 Ga (Boshoff, 2004).
If true for the entire CZ, the results of this study provide clear evidence that closely associated sheath and cross folds record a complex poly-deformational/metamorphic history for the CZ of the Limpopo Complex.
|
SUPPLEMENTARY DATA |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL FRAMEWORK OF THE...METAMORPHISMTHERMODYNAMICS OF METAMORPHISMCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Supplementary data for this paper are available on Journal of Petrology online.
|
ACKNOWLEDGEMENTS |
|---|
This work was carried out as part of the RF-RSA scientific collaboration and supported by grants from the National Research Foundation, South Africa, and the Rand Afrikaans University (to D.D.V.R.), RFBR grants 02-05-64025 (to L.L.P.) and 03-05-64633 (to T.V.G.), and the Russian State Leading Scientific Schools Program (project 1645.2003.5 to L.L.P.). Giles Droop and an anonymous reviewer are acknowledged for their detailed and very helpful reviews of the earlier version of this paper. Giles Droop also kindly reviewed the next version of the paper. We would like to thank Simon Harley for his helpful comments and for the final editing of the manuscript.
|
FOOTNOTES |
|---|
* Corresponding author. Fax: +27-11-4892309. E-mail: ddvr{at}na.rau.ac.za
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL FRAMEWORK OF THE...METAMORPHISMTHERMODYNAMICS OF METAMORPHISMCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Aranovich, L. Y. & Podlesskii, K. K. (1989). Geothermobarometry of high-grade metapelites: simultaneously operating reactions. In: Daly, J. S., Yardley, B. W. D. & Cliff, B. R. (eds) Evolution of Metamorphic Belts. Geological Society, London, Special Publications 42, 41–65.
Bahneman, K. P. (1972). A review of the structure, stratigraphy and the metamorphism of the basement rocks in the Messina district, Northern Transvaal. Ph.D. thesis, University of Pretoria, 156 pp.
Barton, J. M., Holzer, L., Kamber, B., Doig, R., Kramers, J. D. & Nyfeler, D. (1994). Discrete metamorphic events in the Limpopo belt, southern Africa: implications for the application of P T paths in complex metamorphic terranes. Geology 22, 1035–1038.
Boshoff, R. (2004). Formation of major fold types during distinct geological events in the Central Zone of the Limpopo Belt, South Africa: new structural, metamorphic and geochronological data. Ph.D. thesis, Rand Afrikaans University, Johannesburg, 121 pp.
Carson, C. J., Powell, R. & Clark, G. L. (1999). Calculated mineral equilibria for eclogites in the CaO–Na2O–FeO–MgO–Al2O3–SiO2–H2O system: application to the Pouébo Terrane, Pam Peninsula, New Caledonia. Journal of Metamorphic Geology 17, 9–24.[CrossRef][Web of Science]
Droop, G. T. R. (1989). Reaction history of garnet–sapphirine granulites and conditions of Archean high-pressure granulite-facies metamorphism in the Central Limpopo Mobile Belt, Zimbabwe. Journal of Metamorphic Geology 7, 383–403.[Web of Science]
Feldtmann, F. (1996). The structural–metamorphic evolution of the marble and calc-silicate rocks of the Baklykraal quarry near Alldays, Central Zone of the Limpopo Belt, South Africa. M.Sc. thesis, Rand Afrikaans University, Johannesburg, 132 pp.
Feldtmann, F., Smit, C. A. & Van Reenen, D. D. (1995). The significance of two major structural trends in the Central Zone of the Limpopo Belt. Centennial Geocongress, Johannesburg, Extended Abstracts 2, 596–599.
Fripp, R. E. P. (1983). The Precambrian geology of the area around the Sand River near Messina, Central Zone, Limpopo Mobile Belt. Geological Society of South Africa, Special Publication 8, 89–102.
Frost, B. R. & Chacko, T. (1989). The granulite uncertainty principle: limitations on thermobarometry in granulites. Journal of Geology 97, 435–450.[Web of Science]
Gerya, T. V. (1991). The method of physicochemical modeling of metamorphic reactions using mass-balance equations. Contributions to Physicochemical Petrology 16, 112–127 (in Russian).
Gerya, T. V. (1999). P–T paths and evolution granulite facies terrains. D.Sc. dissertation, Moscow State University, 453 pp. (in Russian).
Gerya, T. V. & Perchuk, L. L. (1994). A new thermodynamic database for thermobarometry. International Mineralogical Association, 16th General Meeting, Abstracts, Pisa, p. 142.
Gerya, T. V. & Perchuk, L. L. (1997). Equations of state of compressed gases for thermodynamic databases used in petrology. Petrology 5, 366–380.[Web of Science]
Gerya, T. V., Perchuk, L. L., Van Reenen, D. D. & Smit, C. A. (2000). Two-dimensional numerical modeling of pressure–temperature–time paths for the exhumation of some granulite facies terrains in the Precambrian. Journal of Geodynamics 30, 17–35.[CrossRef][Web of Science]
Gerya, T. V., Perchuk, L. L., Maresch, W. V., Willner, A. P., Van Reenen, D. D. & Smit, C. A. (2002). Thermal regime and gravitational instability of multi-layered continental crust: implications for the buoyant exhumation of high-grade metamorphic rocks. European Journal of Mineralogy 14, 687–699.
Harley, S. L. (1989). The origins of granulites: a metamorphic perspective. Geological Magazine 126, 215–247.[Abstract]
Harris, N. B. W. & Holland, T. J. B. (1984). The significance of cordierite–hypersthene assemblages from the Beitbridge region of the Limpopo Belt: evidence for rapid decompression in the Archaean. American Mineralogist 59, 1036–1049.
Hisada, K. & Miyano, T. (1996). Petrology and microthermometry of aluminous rocks in the Botswanan Limpopo: evidence for isothermal decompression and isobaric cooling. Journal of Metamorphic Geology 14, 183–197.[CrossRef][Web of Science]
Hofmann, A., Kröner, A. & Brandl, G. (1998). Field relationships of mid- to late Archaean high-grade gneisses of igneous and sedimentary parentage in the Sand River, Central Zone of the Limpopo Belt, South Africa. South African Journal of Geology 101, 185–200.[Abstract]
Holzer, L. (1995). The magmatic petrology of the Bulai Pluton and the tectono-metamorphic overprint at 2·0 Ga in the central zone of the Limpopo belt (Messina–Beitbridge area, southern Africa). Diploma thesis, University of Bern, 157 pp.
Holzer, L., Frey, R., Barton, J. M., Jr & Kramers, J. D. (1998). Unravelling the record of successive high-grade events in the Central Zone of the Limpopo belt using Pb single phase dating of metamorphic minerals. Precambrian Research 87, 87–115.[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]
Kamber, B. S., Blenkinsop, T. G., Villa, I. M. & Dahl, P. S. (1995). Proterozoic transpressive deformation in the Northern Marginal Zone, Limpopo Belt, Zimbabwe. Journal of Geology 103, 493–508.[Web of Science]
Kreissig, K., Nägler, T. F., Kramers, J. D., Van Reenen, D. D. & Smit, C. A. (2000). An isotopic and geochemical study of the Kaapvaal craton and the Southern Marginal Zone of the Limpopo Belt: are they juxtaposed terranes? Lithos 50, 1–25.[CrossRef][Web of Science]
Kreissig, K., Holzer, L., Frei, R., Villa, I. M., Kramers, J. D., Kröner, A., Smit, C. A. & Van Reenen, D. D. (2001). Geochronology of the Hout River Shear Zone and the metamorphism in the Southern Marginal Zone of the Limpopo Belt, Southern Africa. Precambrian Research 109, 145–173.[CrossRef][Web of Science]
Kröner, A., Jaeckel, P., Hofmann, A., Nemchin, A. A. & Brandl, G. (1998). Field relationships and age of supracrustal Beit Bridge Complex and associated granitoid gneisses in the Central Zone of the Limpopo Belt, South Africa. South African Journal of Geology 101, 201–213.[Abstract]
Kröner, A., Jaeckel, P., Brandl, G., Nemchin, A. A. & Pidgeon, R. T. (1999). Single zircon ages for granitoid gneisses in the Central Zone of the Limpopo Belt, southern Africa, and geodynamic significance. Precambrian Research 93, 299–337.[CrossRef][Web of Science]
Lasaga, A. C. (1983). Geospeedometry: an extension of geothermometry. Advances in Physical Geochemistry 3, 81–114.
Marakushev, A. A. & Perchuk, L. L. (1966). Physicochemical analyses of paragenesis of minerals: a review. Geological Journal 5, 67–94.
Mason, R. (1973). The Limpopo mobile belt—southern Africa. Philosophical Transactions of the Royal Society of London, Series A 273, 463–485.[CrossRef]
McCourt, S. & Armstrong, R. A. (1998). SHRIMP U–Pb zircon geochronology of granites from the Central Zone, Limpopo Belt, southern Africa: implications for the age of the Limpopo Belt. South African Journal of Geology 101, 329–338.[Abstract]
McCourt, S. & Vearncombe, J. R. (1992). Shear zones of the Limpopo Belt and adjacent granitoid–greenstone terrranes: implications for late Archaean collision tectonics in southern Africa. Precambrian Research 55, 553–570.[CrossRef][Web of Science]
Mkweli, S., Kamber, B. & Berger, M. (1995). Westward continuation of the craton–Limpopo Belt tectonic break in Zimbabwe and new age constraints on the timing of the thrusting. Journal of the Geological Society, London 152, 77–83.
Newton, R. C. (1972). An experimental determination of the high-pressure stability limits of magnesian cordierite under wet and dry conditions. Journal of Geology 80, 398–420.[Web of Science]
Newton, R. C. & Haselton, H. T. (1981). Thermodynamics of the garnet–plagioclase–Al2SiO5–quartz geobarometer. In: Newton, R. C., Navrotsky, A. & Wood, B. J. (eds) Thermodynamics of Minerals and Melts. Advances in Physical Geochemistry. New York: Springer, pp. 131–147.
Perchuk, L. L. (1969). The effect of temperature and pressure on the equilibria of natural Fe–Mg minerals. International Geological Review 11, 875–901.
Perchuk, L. L. (1977). Thermodynamic control of metamorphic processes. In: Saxena, S. K. & Bhattacharj, S. (eds) Energetics of Geological Processes. New York: Springer, pp. 285–352.
Perchuk, L. L. (1985). Metamorphic evolution of shields and fold-belts. Geologicky Zbornik—Geologica Carpathica 36, 179–189.
Perchuk, L. L. (1987). The course of metamorphism. International Geology Review 28, 1377–1400.
Perchuk L. L. (1989). P–T–fluid regimes of metamorphism and related magmatism with specific reference to the Baikal Lake granulites. In: Daly, J. S., Yardley, B. W. D. & Cliff, B. (eds) Evolution of Metamorphic Belts. Geological Society, London, Special Publications 42, 275–291.
Perchuk, L. L. (1990). Derivation of thermodynamically consistent system of geothermometers and geobarometers for metamorphic and magmatic rocks. In: Perchuk, L. L. (ed.) Progress in Metamorphic and Magmatic Petrology. Cambridge: Cambridge University Press, pp. 93–112.
Perchuk, L. L. & Lavrent'eva, I. V. (1983). Experimental investigation of exchange equilibria in the system cordierite–garnet–biotite. Advances in Physical Geochemistry 3, 199–239.
Perchuk, L. L. & Ryabchikov, I. D. (1976). Phase Correspondence in Mineral Systems. Moscow: Nedra, 312 pp. (in Russian).
Perchuk, L. L., Aranovich, L. Ya., Podlesskii, K. K., Lavrent'eva, I. V., Gerasimov, V. Yu., Fed'kin, V. V., Kitsul, V. N. & Karsakov, L. P. (1985). Precambrian granulites of the Aldan shield, eastern Siberia, USSR. Journal of Metamorphic Geology 3, 265–310.[Web of Science]
Perchuk, L. L., Gerya, T. V. & Nozhkin, A. D. (1989). Petrology and retrogression in granulites of the Kanskiy Formation, Yenisey Range, Eastern Siberia. Journal of Metamorphic Geology 7, 599–617.[Web of Science]
Perchuk, L. L., Gerya, T. V., Van Reenen, D. D., Safonov, O. G. & Smit, C. A. (1996). The Limpopo metamorphic complex, South Africa: 2. Decompression/cooling regimes of granulites and adjusted rocks of the Kaapvaal craton. Petrology 4, 571–599.[Web of Science]
Perchuk, L. L., Gerya, T. V., Van Reenen, D. D., Krotov, A. V., Safonov, O. G., Smit, C. A. & Shur, M. Yu. (2000). Comparative petrology and metamorphic evolution of the Limpopo (South Africa) and Lapland (Fennoscandia) high-grade terrains. Mineralogy and Petrology 69, 69–107.[CrossRef][Web of Science]
Pienaar, J. C. (1985). Die Geologie van die Alldays omgewing in noord Transvaal. M.Sc. thesis, Rand Afrikaans University, Johannesburg, 158 pp.
Pretorius, S. J. (1986). Die geologie in die omgewing van Swartwater en die Westelike deel van die Distrik Messina in die sentrale sone van die Limpopo gordel. M.Sc. thesis, Rand Afrikaans University, Johannesburg, 184 pp.
Roering, C., Van Reenen, D. D., Smit, C. A., Barton, J. M., Jr, De Beer, J. H., De Wit, M. J., Stettler, E. H., Van Schalkwyk, J. F., Stevens, G. & Pretorius, S. (1992). Tectonic model for the evolution of the Limpopo Belt. Precambrian Research 55, 539–552.[CrossRef][Web of Science]
Schaller, M., Steiner, O., Studer, I., Hozer, L., Herweg, M. & Kramers, J. D. (1999). Exhumation of Limpopo Central Zone granulites and dextral continent-scale transcurrent movement at 2·0 Ga along the Palala Shear Zone, Northern Province, South Africa. Precambrian Research 96, 263–298.[CrossRef][Web of Science]
Smit, C. A. & Van Reenen, D. D. (1997). Deep crustal shear zones, high-grade tectonites and associated alteration in the Limpopo belt, South Africa: implications for deep crustal processes. Journal of Geology 105, 37–57.[Web of Science]
Smit, C. A., Van Reenen, D. D., Gerya, T. V. & Perchuk, L. L. (2001). P–T conditions of decompression of the Limpopo high-grade terrain: record from shear zones. Journal of Metamorphic Geology 19, 249–268.[CrossRef][Web of Science]
Söhnge, P. G. (1946). The geology of the Messina copper mines and surrounding country. South African Geological Survey Memoirs 40, 280 pp.
Spear, F. S. (1993). Metamorphic Phase Equilibria and Pressure–Temperature–Time Paths. Mineralogical Society of America Monograph, 799 pp.
Spear, F. S. & Florence, F. P. (1992). Thermobarometry in granulites: pitfalls and new approaches. Precambrian Research 55, 209–241.[CrossRef][Web of Science]
Spear, F. S., Kohn, M. J. & Cheney, J. T. (1999). P–T paths from anatectic pelites. Contributions to Mineralogy and Petrology 134, 17–32.[CrossRef][Web of Science]
Tsunogai, T., Miyano, T. & Ridley, J. (1992). Metamorphic P–T profiles from Zimbabwe Craton to the Limpopo belt. Precambrian Research 55, 259–278.[CrossRef][Web of Science]
Van Reenen, D. D. & Smit, C. A. (1996). The Limpopo metamorphic complex, South Africa: 1. Geological setting and relationships between the granulite complex and the Kaapvaal and Zimbabwe cratons. Petrology 4, 562–570.[Web of Science]
Van Reenen, D. D., Barton, J. M., Jr, Roering, C., Smit, C. A. & Van Schalkwyk, J. F. (1987). Deep crustal response to continental collision: the Limpopo Belt of Southern Africa. Geology 15, 11–14.
Van Reenen, D. D., Roering, C., Brandl, G., Smit, C. A. & Barton, J. M. (1990). The granulite facies rocks of the Limpopo Belt, southern Africa. In: Vielzeuf, D. & Vidal, P. (eds) Granulites and Crustal Evolution. Dordrecht: Kluwer Academic, pp. 257–289.
Van Reenen, D. D., Roering, D., Ashwal, L. D. & De Wit, M. J. (eds) (1992). The Archaean Limpopo granulite belt: tectonics and deep crustal processes. Precambrian Research 55, 1–587.
Watkeys, M. K. (1983). A retrospective view of Central Zone of Limpopo Belt, Zimbabwe. Geological Society of South Africa, Special Publication 8, 65–80.
Wei, C. J. J., Powell, R. & Zhang, L. F. (2003). Eclogites from the South Tianshan, NW China: petrological characteristics and calculated mineral equilibria in the Na2O–CaO–FeO–MgO–Al2O3–SiO2–H2O system. Journal of Metamorphic Geology 21, 163–181.[CrossRef][Web of Science]
Windley, B. F., Ackermand, D. & Herd, R. K. (1984). Sapphirine/kornerupine-bearing rocks and crustal uplift history of the Limpopo belt, South Africa. Contributions to Mineralogy and Petrology 86, 342–358.[CrossRef][Web of Science]
Zyrianov, V. N., Perchuk, L. L. & Podlesskii, K. K. (1978). Nepheline–alkali feldspar equilibria: 1. Experimental data and thermodynamic calculations. Journal of Petrology 19, 1–44.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  H. Mouri, M.J. Whitehouse, G. Brandl, and H.M. Rajesh A magmatic age and four successive metamorphic events recorded in zircons from a single meta-anorthosite sample in the Central Zone of the Limpopo Belt, South Africa Journal of the Geological Society, September 1, 2009; 166(5): 827 - 830. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Zeh, A. Gerdes, and J. M. Barton Jr Archean Accretion and Crustal Evolution of the Kalahari Craton--the Zircon Age and Hf Isotope Record of Granitic Rocks from Barberton/Swaziland to the Francistown Arc J. Petrology, May 11, 2009; (2009) egp027v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T.C. Chudy, A. Zeh, A. Gerdes, R. Klemd, and J.M. Barton Jr. Palaeoarchaean (3.3 Ga) mafic magmatism and Palaeoproterozoic (2.02 Ga) amphibolite-facies metamorphism in the Central Zone of the Limpopo Belt: New geochronological, petrological and geochemical constraints from metabasic and metapelitic rocks from the Venetia area South African Journal of Geology, December 1, 2008; 111(4): 387 - 408. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Rigby, H. Mouri, and G. Brandl P-T conditions and the origin of quartzo-feldspathic veins in metasyenites from the Central Zone of the Limpopo Belt, South Africa South African Journal of Geology, September 1, 2008; 111(2-3): 313 - 332. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Zeh, A. Gerdes, R. Klemd, and J. M. Barton Jr Archaean to Proterozoic Crustal Evolution in the Central Zone of the Limpopo Belt (South Africa-Botswana): Constraints from Combined U-Pb and Lu-Hf Isotope Analyses of Zircon J. Petrology, August 1, 2007; 48(8): 1605 - 1639. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Alexandre, M. A.G. Andreoli, A. Jamison, and R. L. Gibson 40Ar/39Ar age constraints on low-grade metamorphism and cleavage development in the Transvaal Supergroup (central Kaapvaal craton, South Africa): implications for the tectonic setting of the Bushveld Igneous Complex South African Journal of Geology, September 1, 2006; 109(3): 393 - 410. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.M. Barton Jr., R. Klemd, and A. Zeh The Limpopo Belt: A result of Archean to Proterozoic, Turkic-type orogenesis? Geological Society of America Special Papers, January 1, 2006; 405(0): 315 - 332. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. ZEH, R. KLEMD, and J. M. BARTON JR Petrological evolution in the roof of the high-grade metamorphic Central Zone of the Limpopo Belt, South Africa Geological Magazine, May 1, 2005; 142(3): 229 - 240. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||