The most reliable microstructural criterion for the former presence of felsic melt in regional migmatites is a three-mineral (quartz, K-feldspar, and sodic plagioclase) aggregate in veinlets. Several other criteria are potentially reliable, namely: (1) euhedral crystals of feldspar (precipitated from liquid) or peritectic minerals (e.g., garnet, cordierite, orthopyroxene, K-feldspar) lining felsic “protoleucosomes”; (2) inclusion-free euhedral overgrowths of feldspar (precipitated from liquid) or peritectic minerals (e.g., garnet, cordierite, orthopyroxene, K-feldspar) on residual grains of the same minerals with abundant inclusions in the mesosome; (3) aligned, euhedral feldspar crystals; (4) simple twinning in K-feldspar; (5) dihedral angles of ≤ 60° subtended where a grain of feldspar and/or quartz (inferred to have pseudomorphed former melt) meets two grains of other minerals; (6) cuspate volumes of quartz, K-feldspar or sodic plagioclase, especially where surrounded by grains inferred to have been residual during melting; (7) veinlets of inferred former melt (now mineral pseudomorphs consisting of one of quartz, K-feldspar or sodic plagioclase, preferably, though less commonly, involving two or three of these minerals) along grain boundaries or along inferred former intragranular fractures; (8) biotite pseudomorphed by feldspar; (9) veinlets of plagioclase that is more sodic than plagioclase grains in the adjacent rock; (10) plagioclase with oscillatory zoning; (11) microgranophyric intergrowths of quartz and alkali feldspar in patches or veinlets between primary grains; (12) symplectic replacement aggregates that can be explained by reactions between peritectic grains and cooling melt; and (13) melanosome patches and layers, from which leucosome has been extracted. However, all these criteria must be interpreted with care. Some other proposed criteria are questionable, for example: (1) random mineral distributions; (2) grain-size increase; (3) interstitial grains; (4) corroded relics of inferred reactant mineral grains surrounded by areas of quartz, K-feldspar, or sodic plagioclase; (5) projections into a mineral grain; (6) lobes of myrmekite; and (7) plagioclase rims with a constant sodic composition occurring on plagioclase cores that are more calcic and/or of variable composition.

The current role of rare-earth-element- (REE-) and actinide-bearing accessory minerals as both geochronological markers and recorders of geochemical processes is reviewed in this paper. The minerals covered include the most common REE- and actinide-bearing accessory minerals found in high-grade rocks—i.e., monazite, xeno-time, apatite, huttonite, thorite, zircon, allanite, and titanite. The goal of this review is to describe the most recent research developments regarding these minerals and their interaction with each other as well as their role as geochronometers. These results are then applied to two cross sections of lower crust from Rogaland–Vest Agder, SW Norway, and Tamil Nadu, S India (granulite and amphibolite facies) with regard to seeing how REE-accessory minerals relate to each other as a function of metamorphic grade. In either traverse the same relationships are seen between monazite, fluorapatite, allanite, and titanite. Namely, in the amphibolite-facies zone, the REE are hosted by titanite and allanite, whereas in the clinopyroxene-in transition zone between the amphibolite-facies and granulite-facies zones, monazite is the stable REE-bearing phase in the region of the orthopyroxene-in isograd. In the granulite-facies zone the REE are hosted by fluorapatite as opposed to monazite. These similarities suggest that mineral hosting of REE follows certain general trends, which are a function of metamorphic grade, whole rock chemistry, and intergranular fluid chemistry.

Since the discovery of CO 2 fluid inclusions in granulites, the role of fluids in the formation of these rocks has been widely studied. Owing to the complexity of the tectono-metamorphic history of granulite terrains, fluid inclusion data alone are not sufficient. They need to be integrated with geochemical and mineralogical studies done on the same rock samples. A clear understanding of the tectono-metamorphic history of granulite terranes is also indispensable. The widespread occurrence of CO 2 and the later discovered high-salinity aqueous fluid inclusions support the idea that the lower crust underwent fluid flow and that both carbonic and brine fluids played a role in its formation. Both low-H 2 O-activity fluids play a similar role in destabilizing hydrous mineral phases. Furthermore, experimental studies have shown that brine fluids have a much larger geochemical effect on granulites than initially expected. These fluids are far more mobile in the lower crust compared with CO 2 and also have the capability for dissolving numerous minerals. As in the example of the Limpopo Complex, fluid inclusions and many metasomatic features observed in granulite terranes can thus be explained only by large-scale movement of high-salinity aqueous fluids and, to a lesser extent, CO 2 , implying that lower-crustal granulites are not as dry as previously assumed. Similar brines and CO 2 -rich fluids are also found in mantle material, most likely derived from deeply subducted supracrustal protoliths.

An extensive database of published cordierite volatile contents from twenty-eight different high-grade terranes is used to investigate whether CO 2 streaming or fluid-absent melting processes prevail during the formation and evolution of granulites. In most case studies the cordierite volatile contents, calculated activities, and melt-H 2 O contents are entirely consistent with a mode of origin dominated by fluid-absent melting processes. In accordance with published experimental and theoretical evidence these data suggest that CO 2 is not a prerequisite for granulite formation. Even in cases in which cordierite preserves high CO 2 contents this does not necessarily imply that fluid-saturated conditions prevailed. A cordierite may be CO 2 -rich but can still be fluid undersaturated and preserve H 2 O contents that equilibrated with melts formed from fluid-absent melting. In fluid-saturated case studies the mechanism and relative timing of saturation should be evaluated. It is evident from the data that most fluid-saturated granulites formed initially under fluid-absent conditions but subsequently became fluid saturated along a retrograde path.

This paper is a review of fundamental principles and rules for reconstructing pressure-temperature (P-T) paths followed by crystalline rocks. The fundamental principles are (1) the Prigogine-Korzhinskii principle of local equilibrium, and (2) the phase correspondence principle for coexisting minerals (Perchuk, 1971, 1977). These principles have consequences (rules) that are particularly useful for the complex assemblages that occur in Precambrian high-grade terranes. The resulting methodology is the only rigorous means of unraveling the history of polymetamorphic rocks. To assure the accuracy of any thermobarometric methodology it is essential to consider geological structures, microstructural analysis, calculations of mineral modes in terms of temperature and pressure, fluid inclusion data, and numerical modeling of P-T paths. The sequence in which these different approaches are considered has important consequences for the accurate reconstruction of P-T paths for polymetamorphic high-grade rocks. Several examples of P-T path reconstructions for mono- and polymetamorphic granulite facies complexes are considered in detail.

The results of 50 years of geochronological work in the Limpopo Complex are reviewed. The data define three main age clusters. The oldest, at ca. 3.3 Ga, exists in the Central and Southern Marginal Zones and is defined by magmatic zircon dates. The second, with a genuine spread between 2.7 and 2.55 Ga, occurs in all three zones. It was a period of high-grade regional metamorphism with intense deformation and widespread anatexis, dated also mainly (but not exclusively) by zircon U-Pb. The third cluster is well constrained at 2.02 ± 0.02 Ga in the Central Zone by zircon overgrowths, sparse magmatic zircons, monazite, apatite, Sm-Nd and Lu-Hf garnet dating, Pb/Pb discrete phase and stepwise leaching dating of garnet and titanite, and hornblende Ar/Ar dating. The Paleoproterozoic dates from metamorphic minerals are particularly associated with zones of intense transcurrent shearing at high-grade metamorphism. In the Northern Marginal Zone this event is more protracted, from 2.08 to 1.94 Ga, and defined in medium- to low-grade shear zones. In the Southern Marginal Zone it is absent. The evidence for both Neoarchean and Paleoproterozoic mineral ages, both defining high-grade tectono-metamorphic events, is in part paradoxical and has led to controversies as to the age of a proposed collisional orogeny. Studying the mineral dates in their tectonic context leads to the conclusion that fluid access in deformation, rather than mere reheating, mainly caused their partial resetting in the Paleoproterozoic event. This allows the controversy to be resolved.

We report new petrological data for granulites from the Central Zone of the Limpopo Complex, southern Africa, and construct a prograde P-T path that traverses from high-pressure granulite-facies metamorphism to peak ultrahigh-temperature (UHT) metamorphism by rapid decompression, which was followed by further decompression and cooling. Mg-rich ( X Mg ~0.58) staurolite enclosed within poikiloblastic garnet in an Mg-Al-rich rock from the Beit Bridge area is rarely mantled by a sapphirine + quartz corona, suggesting the progress of the prograde dehydration reaction: staurolite + garnet → sapphirine + quartz + H 2 O. The symplectic sapphirine + quartz developed around staurolite probably implies decompression from P >14 kbar toward the stability of sapphirine + quartz at T ~1000 °C along a clockwise P-T path. The orthopyroxene + sillimanite + quartz assemblage mantled by cordierite aggregates in a pelitic granulite from the same area also suggests extreme metamorphism and subsequent further decompression. Various corona textures such as kyanite + sapphirine, sapphirine + cordierite, and orthopyroxene + cordierite were probably formed as a result of decompression cooling events. The prograde high-pressure metamorphism and the following UHT event relate to the collisional tectonics of the Zimbabwe and Kaapvaal Cratons, which are associated with the amalgamation of microcontinents during the Neoarchean.

Petrological and fluid-inclusion data of high-grade metapelitic gneisses that occur as enclaves and in the immediate surroundings of the 2.612 Ga old Bulai granitoid intrusive are presented in this chapter. The Bulai intrusive is an important time marker in the tectono-metamorphic evolution of the Central Zone of the Limpopo Complex. The host-rock gneisses show one generation of garnet, cordierite, and sillimanite, whereas the enclave gneisses show two different generations of garnet (Grt 1,2 ), cordierite (Crd 1,2 ), and sillimanite (Sil 1,2 ). The first generation defines a gneissic texture, whereas the second generation shows a random mineral orientation. Grt 1 and Crd 1 show a higher Mg content compared with Grt 2 and Crd 2 . Host rock garnet and Grt 1 show K-feldspar micro-veins at the contact with quartz as a result of high-temperature metasomatism. Host rock garnet, Grt 1 , and Grt 2 are zoned and participate in two simultaneously operating reactions: sillimanite + garnet + quartz = cordierite and garnet + K-feldspar + H 2 O = biotite + sillimanite + quartz. The combination of petrographic, geothermobarometric, and fluid-inclusion results shows evidence of two different pressure-temperature (P-T) paths in the enclave and a single P-T path in the host rocks. The decompressional cooling P-T path in the host rock is typical of the country rocks throughout the Central Zone. The high-pressure part of the host-rock P-T path overlaps with the Grt 1 -Crd 1 -Sil 1 P-T path found in the enclave rocks. The second P-T path is calculated from the Grt 2 -Crd 2 -Sil 2 assemblage and is found only in the enclave rocks. The two P-T paths in the enclave rocks can be connected by a sub-isobaric heating event of ~50 °C at 5.5 kbar. This increase in temperature is followed by decompressional cooling but with a lower P-T gradient compared with that of the country rocks caused by the emplacement of the Bulai Pluton. Fluids present during granulite metamorphism include CO 2 and brines. Retrograde infiltration of water in graphite-bearing country rocks under relatively reduced conditions resulted in the formation of a methane-rich fluid.

The major periods of metamorphism in the Central Zone (CZ) of the Limpopo Complex occurred at 2.0 Ga and in the time range between ca. 2.7 and ca. 2.55 Ga. We investigate intracrustal radioactivity as a possible heat source for the earlier of these episodes. Available airborne radiometric surveys that cover the South African part of the CZ, combined with rock analyses, yield 2.15 μg/g U, 12.3 μg/g Th, and 12,650 μg/g K as a weighted regional average. The corresponding heat production rate at 2.65 Ga is 2.6 μW × m −3 . A steady-state geotherm, calculated assuming uniform [U], [Th], and [K] throughout the crustal column and its thickening to 45 km during the ca. 2.65 Ga event (both arguable on the basis of peak metamorphic pressure-temperature [P-T] data), surpasses temperatures of the peak metamorphism at middle and lower crustal levels, which cluster around the fluid-absent biotite dehydration solidus. Intracrustal radioactivity thus provided a sufficient heat source to account for the metamorphism at ca. 2.65 Ga, and partial melting acted as a lower crustal thermostat. After crustal thickening, up to more than 100 m.y. (dependent on U, Th, and K concentrations) would be needed to approach a new steady state. Predicted regional variations thus account for the long duration of the ca. 2.65 Ga metamorphism. Lower crustal partial melting could have led to diapirism, yielding the steep structures in the CZ, which are not aligned to a regional fabric. Metamorphism ceased after crustal thinning to a normal 30 km. The metamorphic event at 2.0 Ga cannot be explained by this type of process.

Published whole-rock Sm-Nd and zircon Lu-Hf data from the Limpopo Complex and adjoining areas of the Zimbabwe and Kaapvaal Cratons provide insight into the regional crustal evolution and tectonic processes that shaped the complex. The Northern Marginal Zone of the complex, and the Francistown area of the Zimbabwe craton, represent an accretionary margin (active at 2.6–2.7 Ga) at the southern edge of that craton, at deep and shallow crustal levels, respectively. The Southern Marginal Zone represents a deep crustal level of the northern Kaapvaal Craton and was not an accretionary margin at the time of high-grade metamorphism (2.72–2.65 Ga). The syntectonic Matok granite was produced by crustal anatexis. In the Central Zone, the presence of ca. 3.5–3.3 Ga crust is indicated throughout its E-W extent by T Nd,DM model ages of metapelites and by zircon xenocrysts and their T Hf,DM model ages. The ca. 2.65 Ga granitoids in the Central Zone (the Singelele-type quartzofeldspathic gneisses in the Musina area, granitoids in the Phikwe Complex, Botswana, the so-called gray gneisses, and the Bulai charnockite) were formed by anatexis of such old crust, whereas 2.6 Ga juvenile (arc-related?) magmatism produced the Bulai enderbite, and may be a component in the Zanzibar gneiss. The Mahalapye granitoid complex in Botswana was formed by crustal anatexis at 2.0 Ga, but mafic and hybrid rocks of this age have a mantle-derived component. The data do not prohibit a collisional model for the Neoarchean high-grade metamorphic event in the Central Zone and Southern Marginal Zone of the Limpopo Complex.

Integrated structural, metamorphic, and geochronological data indicate that the evolution of the Southern Marginal Zone (SMZ) of the Limpopo Complex of southern Africa was controlled by a single Neoarchean high-grade tectono-metamorphic event. The exhumation history reflected by the high-grade rocks is determined by their location relative to the contact with the low-grade rocks of the Kaapvaal Craton. Exhumation of granulites far north from this contact is recorded by a decompression-cooling (DC) pressure-temperature (P-T) path linked to steep southward-verging thrusts related to the Hout River Shear Zone. This P-T path traverses from P ~8 kbar, T ~825 °C to P ~5 kbar, T ~550 °C and reflects exhumation of the SMZ in the interval ca. 2.68–2.64 Ga. P-T paths for granulites close to this contact are characterized by a distinct inflection at P ~6 kbar, T ~700 °C that exhibits near-isobaric cooling (IC) to T ~580 °C. The IC stage is linked to low-angle, out-of-sequence, southward-verging thrusts that developed in the interval 2.63–2.6 Ga. The thrust-controlled exhumation of the SMZ furthermore is demonstrated by the convergence at P ~6 kbar, T ~700 °C of DC P-T paths in the hanging wall with prograde P-T loops in the footwall of the steeply southward-verging Hout River Shear Zone, and by the establishment of a retrograde isograd and zone of rehydrated granulites in the hanging wall derived from the dehydration of the low-grade rocks in the footwall. A composite deformation-pressure-temperature-time (D-P-T-t) diagram provides evidence in support of a tectonic model for the evolution of the Limpopo Complex that involves early crustal thickening and peak metamorphic conditions followed by doming and diapirism related to gravitational redistribution mechanisms.

Integrated geological studies in the Central Zone of the Limpopo Complex formed the basis for the construction of a composite deformation (D)–pressure (P)– temperature (T)–time (t) (D-P-T-t) diagram that shows the following: First, in the Neoarchean the Central Zone probably underwent high-pressure (HP) (P >14 kbar, T ~950 °C) conditions followed by near isothermal decompression to ultrahigh- temperature conditions (UHT) (T ~1000 °C, P ~10 kbar), before ca. 2.68 Ga. Second, the post-peak exhumation history linked to two distinct decompression cooling stages commenced at ca. 2.68 Ga and ended before the emplacement of the Bulai Pluton at ca. 2.61 Ga. Stage 1 started at P ~9 kbar, T = 900 °C, and culminated with the emplacement of leucocratic anatectic granitoids at ca. 2.65 Ga. Stage 2, linked to the development of major SW-plunging sheath folds and related shear zones, started at P ~6 kbar, T ~700 °C and ended at P ~5 kbar, T ~550 °C, before ca. 2.61 Ga. The rocks resided at the mid-crustal level for more than 600 m.y. before they were again reworked at ca. 2.02 Ga by a Paleoproterozoic event. This event commenced with isobaric (P ~5 kbar) reheating (T ~150 °C) of the rocks related to the emplacement at ca. 2.05 Ga of magma linked to the Bushveld Igneous Complex. This was followed by final exhumation of the Central Zone. The Neoarchean high-grade event that affected the Limpopo Complex is linked to a Himalayan-type collision of the Kaapvaal and Zimbabwe Cratons that resulted in over-thickened unstable crust and the establishment of HP and UHT conditions. This unstable crust initially responded to the compressional event by thrust-driven uplift and spreading of the marginal zones onto the two adjacent granite-greenstone cratons. The post-peak exhumation history was probably driven by a doming-diapiric mechanism (gravitational redistribution).

The Northern Marginal Zone (NMZ) of the Limpopo Belt, southern Africa, is a high-grade gneiss belt dominated by magmatic granulites of the charnoenderbite suite, which intruded minor mafic-ultramafic and metasedimentary rocks between 2.74 and 2.57 Ga. The intrusive rocks have crustal and mantle components, and occur as elliptical bodies interpreted as diapirs. Peak metamorphism (P ≤800 MPa, T = 800–850 °C) occurred at ca. 2.59 Ga. The highly radiogenic nature of the rocks in the NMZ, supplemented by heat from mantle melts, led to heating and diapirism, culminating in the intrusion of distinctive porphyritic charnockites and granites. Horizontal shortening and steep extrusion of the NMZ, during which crustal thickening was limited by high geothermal gradients, contrast with overthickening and gravitational collapse observed particularly in more recent orogens. The granulites were exhumed by the end of the Archean. The pervasive late Archean shortening over the whole of the NMZ contrasts with limited deformation on the Zimbabwe Craton, possibly owing to the strengthening effect of early crust in the craton. In the southeast of the NMZ, strike-slip kinematic indicators occur within the Transition Zone and the Triangle Shear Zone, where dextral shearing reworked the Archean crust at ca. 1.97 Ga.

Heterogeneous strain commonly serves as an important natural instrument for unraveling complex tectonic histories in polyphase metamorphic terranes. We present key examples of multi-scale heterogeneous deformation from two classic deep-crustal granulite terranes, the Athabasca Granulite Terrane in western Canada and the Limpopo Complex in southern Africa. These examples are chosen to illustrate how localized strain and attendant metamorphism played a key role in the development and preservation of important records of deep-crustal processes. In addition, several common characteristics of these terranes are identified through this analysis and include heterogeneous deep-crustal flow, regional-scale tectonic heterogeneity, and multistage exhumation with high-resolution records developed in locally hydrated shear zones. Better recognition of the fundamental spatial and temporal heterogeneity in these and other similar polymetamorphic terranes may help to reconcile apparently conflicting interpretations and tectonic models.

This paper proposes a revision of the gravitational redistribution model suggested by Leonid Perchuk for the formation, evolution, and exhumation of Precambrian high-grade terranes (HGTs) located between granite-greenstone cratons. Such HGTs are separated from greenstone belts by crustal-scale shear zones up to 10 km wide and several hundred kilometers long. Pelite samples far (>~50 km) from the bounding shear zones show coronitic and symplectitic textures that reflect a decompression-cooling (DC) pressure-temperature (P-T) path. On the other hand, samples from within ~50 km of the bounding shear zones are characterized by textures that reflect an isobaric or near-isobaric cooling (IC) path. Local mineral equilibria in the schists from the shear zones record hairpin-shaped clockwise P-T loops. The results of a numerical test of the gravitational redistribution model show the following plausible scenario: The diapiric rise of low-density, hot granulite upward in the crust causes the relatively high-density, predominantly mafic upper crust in the adjacent greenstone belt, consisting of metabasalt and komatiite, to move downward (subducted), cooling the base of the granulites along the intervening syntectonic shear zone. This causes (1) the formation of local convection cells that control the movement of some of the ascending granulite blocks near the contact with the cratonic rocks, and (2) near-isobaric cooling (IC) of the granulite blocks in the vicinity of the boundary with the colder wall rocks. Cooling of granulite blocks farther away from the contact is not arrested, and they ascend to the Earth's surface, recording DC P-T paths. In general, the results of numerical modeling provide support for the buoyant exhumation mechanism of granulites owing to gravitational redistribution within the metastable relatively hot and soft early Precambrian crust that was subjected to high-temperature (HT) and ultrahigh-temperature (UHT) metamorphism.

Published models for the Limpopo Complex as a whole include Neoarchean (ca. 2.65 Ga) continent-continent collision, Turkic-type terrane accretion, and plume-related gravitational redistribution within the crust. Hypotheses proposed for parts of the complex are Paleoproterozoic (ca. 2.0 Ga) dextral transpression for the Central Zone, westward emplacement of the Central Zone as a giant nappe, and gravitational redistribution scenarios. In this chapter these models and hypotheses are reviewed and tested against new data from geophysics (chiefly seismics and gravity), isotope geochemistry (mainly Sm-Nd and Lu-Hf data), geochronology, and petrology. Among the whole-complex models, the plume-related gravitational redistribution model and the Turkic-type terrane accretion model do not satisfy the constraints. The Neoarchean collision model remains as a viable working hypothesis, whereby (in contrast to published versions) the Zimbabwe Craton appears to be the overriding plate, with the Northern Marginal and Central Zones of the Limpopo Complex as its (possibly Andean-type) active margin and shelf, respectively. Of the partial models, gravitational redistribution in the context of crustal thickening is compatible with Neoarchean collision and can explain features at the Complex–Kaapvaal Craton boundary. Paleoproterozoic dextral transpression in the Central Zone can be superimposed on Neoarchean collision, provided that it does not itself entail a continent collision. The Paleoproterozoic metamorphism is characterized by near-isobaric prograde paths, which (along with combined teleseismic and gravity data) suggest magmatic underplating. This could be related to the Bushveld Complex, and may have weakened the crust, leading to the focusing of regional strain into transcurrent movement in the Central Zone.