Elsevier

Chemical Geology

Volume 546, 20 July 2020, 119647
Chemical Geology

Zircon formation in mafic and felsic rocks of the Bushveld Complex, South Africa: Constraints from composition, zoning, Th/U ratios, morphology, and modelling

https://doi.org/10.1016/j.chemgeo.2020.119647Get rights and content

Highlights

  • Zircon in felsic and mafic rocks of the Bushveld Complex grew mostly at 800–900 °C.

  • Zircon in rutile-free mafic and felsic rocks crystallized at aTiO2 = 0.3.

  • Zircon in felsic rocks was formed after 20% fractional crystallization (FC).

  • In mafic rocks after 75% FC from B1 and B2 parental magmas, and UUMZ (+25% felsic).

  • Zircon zoning is controlled by zircon/melt ratio, assemblage, and Bt-in reaction.

Abstract

Zircon is a potential petrogenetic indicator that can be used to derive physicochemical conditions during magma crystallization. In this study, such conditions are obtained from zircon of both felsic and mafic rocks of the Bushveld Complex (BC), which are characterized by a wide range in bulk-rock Zr contents (4–552 ppm). For that, information from bulk-rock compositions, petrography, zircon trace element data and morphologies are combined with results of thermodynamic modelling using the software packages rhyolite-MELTS and Perple_X. In felsic rocks (Lebowa Granite, Rashoop Granophyre), zircon is formed in rutile-free assemblages together with olivine-clinopyroxene-amphibole-biotite-ilmenite-titanite-apatite after ≥20% fractional crystallization at 761–935 °C (mean: 860 °C), based on Ti-in-zircon thermometry using aTiO2 = 0.3, in agreement with independent geothermometers and modelling results. The resulting zircon populations show {100}- and {101}-dominated morphologies as well as high ƩREE (mean: 651 ppm) and low Ti contents (mean: 9.5 ppm), and only minor zoning in Th/U (0.3–0.8) and Nb/Ta (1.2–4.4). Identical zircon characteristics in gabbros and diorites of the Upper Zone within the Rustenburg Layered Suite (RLS) suggest admixing of felsic melts during ingression of mafic parental magmas. In contrast, zircons in mafic rocks of the lower RLS (Basal Ultramafic Sequence to lower Main Zone) show significantly lower ƩREE (mean: 324 ppm), commonly higher Ti contents (8–60 ppm), as well as large variations in Ti, U, Th contents, Th/U (0.2–24) and Nb/Ta ratios (0.15–18) as well as in zircon morphology. These zircons mostly occur in rutile-bearing intercumulus domains associated with orthopyroxene-biotite-amphibole-plagioclase-quartz-rutile and are formed at 690–962 °C (mean: 835 °C; aTiO2 = 1.0) based on Ti-in-zircon thermometry. These temperatures are in good agreement with zircon morphologies, but mostly higher than those obtained by rhyolite-MELTS modelling, suggesting zircon growth at <810 °C after >75% fractional crystallization of high Mg andesitic (B1) and tholeiitic parental magma (B2). These lower temperatures perhaps result from oversimplified modelling parameters or may reflect variable mixing of parental magma with evolved resident magma. Zircon populations in rutile-bearing mafic rocks of the lower RLS reveal two distinct zoning trends: an early trend at high Ti (>20 ppm), characterized by increasing Th/U (0.5 to 18) at decreasing U (175 to 10 ppm) from core to rim and a reverse trend at lower Ti contents. Both trends require zircon formation in intercumulus melt pockets at high zircon/melt ratios. The high-Ti trend can be explained by Rayleigh-like fractionation due to zircon growth together with rutile, both having highly different partition coefficients for U ≫ Th. The low-Ti trend results from zircon growth after onset of the biotite-in reaction, causing breakdown of previously formed rutile, thereby releasing U but no Th. The absence of pronounced Th/U zoning of zircon in felsic rocks reflects zircon growth in less fractionated melts, resulting in Thsingle bondU fractionation compensated by coeval crystallization of abundant rock-forming minerals, all being highly incompatible for Th and U.

Introduction

Zircon is commonly used as a geochronometer, but also as a petrogenetic indicator to constrain physicochemical conditions during magma crystallization. In recent studies, trace element (TE) contents in zircon were used to determine zircon crystallization temperatures by Ti-in-zircon thermometry (Ferry and Watson, 2007; Fu et al., 2008), and the redox state of the host magma using Ce- and Eu-anomalies (Trail et al., 2012). Furthermore, trace elements were applied to monitor magmatic differentiation processes (Miles et al., 2013) using experimentally calibrated zircon/melt partition coefficients (Rubatto and Hermann, 2007), to assess the amount of fractional vs. equilibrium crystallization during zircon growth (Kirkland et al., 2015), and to fingerprint tectono-magmatic provenances (Belousova et al., 2002; Carley et al., 2014; Grimes et al., 2015). The results of these studies indicate that incorporation of TE into zircon is complex and controlled by several factors, comprising P-T conditions, oxygen fugacities and bulk rock compositions. These parameters in combination determine in which mineral assemblage zircon is likely to be formed, and consequently the bulk distribution coefficient during magma crystallization.

In a thermodynamic context, zircon starts crystallizing when an evolving melt becomes zircon-saturated, whereby the saturation temperature is dependent on bulk-rock composition and can be calculated by applying calibrated zircon saturation geothermometers (Watson and Harrison, 1983; Boehnke et al., 2013). During subsequent magma cooling, zircon growth commonly continues within changing mineral assemblages, melt fractions, and melt compositions, resulting in zircon grains with different morphologies, compositions and zoning patterns. The successive change in mineral assemblage during magma crystallization can be traced by petrographic observations and thermodynamic modelling, such as by using the software packages rhyolite-MELTS (Gualda et al., 2012) and Perple_X (Connolly, 2009). Zircon formation temperatures can be obtained either by applying Ti-in-zircon thermometry (Ferry and Watson, 2007; Fu et al., 2008) or by using the classification scheme of Pupin (1980), based on zircon {100}:{110} prism ratios.

However, each of the aforementioned thermometric methods has several disadvantages, which make them more or less useful in obtaining reliable zircon formation temperatures. For example, temperatures estimated by using Pupin's classification scheme are imprecise (50 °C steps) and might be biased due to enhanced and/or suppressed growth of certain crystal faces dependent on U, Th and REE contents and/or the degree of zircon supersaturation of the melt, as discussed by Benisek and Finger (1993) and Vavra (1990). Titanium-in-zircon thermometry only yields precise and accurate zircon crystallization temperatures if zircon was formed in equilibrium with quartz and rutile (indicating aSiO2 = aTiO2 = 1), a prerequisite that is rarely fulfilled for magmatic rocks (Fu et al., 2008). Most magmatic rocks commonly contain Ti-bearing phases like titanite, ilmenite, titanomagnetite and biotite, requiring the use of aTiO2 < 1 for Ti-in-zircon thermometry, which otherwise produces minimum temperatures (Ferry and Watson, 2007; Fu et al., 2008; Siégel et al., 2018; Schiller and Finger, 2019). Zircon saturation geothermometry has the advantage that it is independent on mineral assemblages. However, it only yields reliable temperatures if the measured bulk rock composition is close to the melt composition at the time of zircon growth. This is not the case for mafic cumulate rocks, where zircon crystallization occurs late during the fractionation history, if at all (Siégel et al., 2018). This illustrates that the different zircon-based thermometers cannot be used without limitation for all rock types.

In this study, the consistency between different zircon-based thermometers and their reliability will be tested by applying them to a wide range of geochemically and petrologically well-characterized rocks of the BC. This complex is an ideal testing ground as it contains zircon in a wide range of rock types comprising both mafic cumulates and felsic rocks, all being essentially coeval but having formed from different parental magmas. Furthermore, we will place new constraints on the formation or absence of spectacular zoning patterns, in particular of Th/U ratios of zircon in felsic and mafic rocks. Finally, implications of the observed zircon systematics for the BC magma chamber evolution will be discussed.

The Bushveld Complex (BC) in South Africa is the Earth's largest mafic layered intrusion and hosts the world's largest resources for platinum group elements, Cr, V and Ni (Cawthorn and Walraven, 1998; Cawthorn, 2015). It comprises three large igneous provinces which were formed successively: (1) Rooiberg Group volcanics (mostly felsic), (2) mafic cumulate rocks of the RLS, and (3) granitic rocks of the Lebowa Granite and the Rashoop Granophyre Suites (Fig. 1) (Molyneux, 1970; Von Gruenewaldt, 1972; Molyneux, 1974; Twist and French, 1983; Walraven, 1987; Schweitzer et al., 1997; Mathez et al., 2013; VanTongeren and Mathez, 2015). The RLS and related granites crystallized between 1.5 and 3.0 ± 0.5 kbar (Kaneko and Miyano, 1990; Pitra and De Waal, 2001; Waters and Lovegrove, 2002), and the parental magmas of the RLS intruded at 1150–1300 °C (Cawthorn and Walraven, 1998). The felsic rocks of the Rooiberg Group and Lebowa Granite Suite had liquidus temperatures >1100 °C (Günther et al., 2018).

The Rooiberg Group consists of an up to 5 km thick pile of mostly very fine-grained to glassy lava flows (felsites) and minor pyroclastic rocks (Twist and French, 1983; Twist and Harmer, 1987; Buchanan et al., 1999; Buchanan et al., 2002, Buchanan et al., 2004; Lenhardt and Eriksson, 2012). It comprises from bottom to the top the Dullstroom, Damwal, Kwaggasnek and Schrikkloof Formations.

The RLS is made up of a 7500–9000 m thick stack of mafic cumulate rocks. Based on stratigraphy, gravimetric data, and xenoliths in kimberlites, it is concluded that the RLS developed in a continuous magma chamber of at least 400 km extent in an E-W direction (Cawthorn and Walraven, 1998). Presently, it is exposed in five distinct limbs: Northern, Eastern, Southern (mostly under cover), Western and Far Western (Fig. 1), and subdivided from bottom to top into 6 stratigraphic units: Basal Ultramafic Sequence (BUS), Marginal Zone, Lower Zone, Critical Zone, Main Zone and Upper Zone (for details see Hall, 1932 and Wilson, 2012, Wilson, 2015). The BUS represents relics of a sill complex, which developed in three distinct compartments prior to formation of a coherent magma chamber, hosting the other 5 units (Wilson, 2015). The RLS stratigraphy is further subdivided by several prominent marker horizons comprising prominent chromitite and magnetite layers, the Merensky Reef, and the Pyroxenite Marker (see Fig. 1; and summary in Cawthorn et al., 2006; Wilson et al., 2017).

Geochemical and petrological data suggest that the cumulate rocks of the RLS result from crystallization of different parental magmas (UM, B1, B2, B3; e.g. Barnes et al., 2010, Wilson, 2012), which intruded the RLS magma chamber periodically, and led to the formation of cyclic units, which are most spectacular in the Upper Critical Zone (Eales and Costin, 2012; Latypov et al., 2017). It is suggested that UM (ultramafic) and B1 (high-Mg andesite) magmas played a major role during formation of the BUS, Marginal Zone, and Lower Zone, mixed B1 and B2 (tholeiite) magmas for the Critical and lower Main zones, and the B3 (aluminous basalt) magma for the Main Zone below the Pyroxenite Marker. For the rocks above the Pyroxenite Marker, comprising the uppermost Main Zone and Upper Zone (UUMZ), two parental magmas are suggested. The first magma is represented by the average bulk rock composition of the UUMZ rocks (Tegner et al., 2006), and the second magma by a mixture of UUMZ with a felsic component, which after unmixing led to the formation of the overlying Rashoop granophyres (VanTongeren et al., 2010). Results of high-precision Usingle bondPb dating suggest that accretion of the RLS (from Marginal Zone to Upper Zone) and subsequent cooling below 700 °C lasted less than 1 million years, between 2055.91 ± 0.26 Ma and 2054.89 ± 0.37 Ma (Zeh et al., 2015) and is in accord with the modelled emplacement duration of ca. 70,000 years (Cawthorn and Walraven, 1998).

The Lebowa Granite Suite is the largest exposure (ca. 65.000 km2) of anorogenic granite on Earth with a ferroan (A-type) composition (Hill et al., 1996). It is intrusive into mafic cumulate rocks of the RLS, volcanic rocks of the Rooiberg Group, and Rashoop granophyres (Von Gruenewaldt, 1972; Walraven, 1985, Walraven, 1987; VanTongeren and Mathez, 2015). Its major and trace element compositions overlap with those of the Rashoop granophyres and rhyolites of the Rooiberg Group (Fig. 2), as noted by several authors (Twist and French, 1983; Twist and Harmer, 1987; Hatton and Schweitzer, 1995; Hill et al., 1996; Schweitzer et al., 1997; Buchanan et al., 2002, Buchanan et al., 2004; Mathez et al., 2013).

In previous studies, zircon of the BC was used for high-precision Usingle bondPb dating (Scoates and Friedman, 2008; Zeh et al., 2015; Mungall et al., 2016) and Hf isotope analyses to obtain information about magma sources (Zirakparvar et al., 2014; VanTongeren et al., 2016; Zeh et al., 2020). The first systematic TE study on zircon was carried out by Yudovskaya et al. (2013) on intercumulus rocks of the Critical Zone, revealing significant variation in Ti-in-zircon temperatures (760–930 °C), and zircon trace element contents (Hf, U, REE). An observed decrease in Th/U ratio (from 4 to 0.5) and Ti content at increasing U content from core to rim was interpreted in that study to result from zircon growth during cooling accompanied by intercumulus percolation of U-rich melts. Zeh et al. (2015) additionally reported zircon grains with increasing Th/U from core to rim, which they interpreted to result from early Rayleigh fractionation during zircon growth. Zircon populations with even more extreme variations in Th/U (from 0.5 to 77) in Critical Zone rocks were recently reported by Ver Hoeve et al. (2018) but detailed core-rim systematics of the investigated zircons was not provided. These authors speculate that high Th/U ratios result from the selective loss of U from the intercumulus melt, caused by the migration of oxidized, chlorine-rich fluids. Ver Hoeve et al. (2018) presented the first TE systematics of zircon in both mafic and felsic rocks of the BC. Furthermore, based on Ti-in-zircon geothermometry they suggested that zircon in Rt-bearing mafic rocks of the Lower and Critical Zone of the RLS were formed on average at 60 °C higher temperatures than in Rt-free mafic rocks of the Upper Zone and overlying granites thereby having important petrogenetic implications for the emplacement of the BC and associated rock units.

Section snippets

Samples and analytical methods

In this study, zircon populations were investigated from 41 samples with different bulk rock compositions. Most samples were taken from the Eastern Limb of the BC, and a few from the Northern, Western and Southern limbs (for sample locations and coordinates see Fig. 1 and Table 1). The samples comprise cumulate rocks from all stratigraphic units of the RLS (except for the Lower Zone), as well as from the overlying Lebowa Granite Suite, Rashoop Granophyre Suite, and Rooiberg Group. Furthermore,

Bulk rock composition

The investigated samples comprise magmatic rocks of very different composition, as is reflected by highly variable bulk-rock SiO2 and Zr contents ranging from 51 to 76 wt%, and from 4 to 552 ppm, respectively (Table 2, Fig. 2a). The compositions of the investigated magmatic rocks overlap with those analyzed in previous studies of the RLS, Rooiberg Group, Lebowa Granite and Rashoop Granophyres (Fig. 2a–d). We note that mafic rocks of the RLS commonly show much lower Zr contents (4–204 ppm;

Zircon crystallization temperatures

The results of this study indicate that zircon crystallization temperatures for different rock types throughout the BC stratigraphy are very similar and average at ca. 850 ± 30 °C, irrespective of applying different thermometric methods. These methods are either based on bulk rock compositions (i.e. zircon saturation temperatures and thermodynamic modelling) or on the composition and shapes of individual zircon grains (i.e. Ti-in-zircon and zircon morphology geothermometry) - (Table 4).

For

Conclusions

  • 1.

    Zircon occurs abundantly in felsic and mafic rocks of all stratigraphic units of the BC, including mafic rocks of the Main and Upper Zones which, according to thermodynamic modelling based on B3 and UUMZ parental magma compositions, should be zircon-free.

  • 2.

    The occurrence of zircon in mafic rocks of the Main and Upper Zones most likely results from admixing of felsic melts during periodic parental magma ingression into a stratified magma chamber, as suggested by zircon morphologies and TE contents

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

DG and AZ thank the Deutsche Forschungsgemeinschaft (DFG grant ZE 424/12-1), and Linda Marko (University Frankfurt am Main, Germany) for support with sample preparation and LA-ICP-MS analyses. All authors are also indebted to late Joe Aphanane (University of the Witwatersrand) for zircon separation and preparation, David Schiller (Salzburg University, Austria) for support during thermodynamic modelling as well as to Samancor Chrome and Bokoni Platinum Mine Proprietary Ltd. for support during

References (86)

  • D. Rubatto et al.

    Experimental zircon/melt and zircon/garnet trace element partitioning and implications for the geochronology of crustal rocks

    Chem. Geol.

    (2007)
  • J.K. Schweitzer et al.

    Link between the granitic and volcanic rocks of the Bushveld Complex, South Africa

    J. Afri. Earth Sci.

    (1997)
  • C. Siégel et al.

    Use and abuse of zircon-based thermometers: a critical review and a recommended approach to identify antecrystic zircons

    Earth Sci. Rev.

    (2018)
  • D. Szymanowski et al.

    Isotope-dilution anchoring of zircon reference materials for accurate Ti-in-zircon thermometry

    Chem. Geol.

    (2018)
  • D. Trail et al.

    Ce and Eu anomalies in zircon as proxies for the oxidation state of magmas

    Geochim. Cosmochim. Acta

    (2012)
  • D. Twist et al.

    Geochemistry of contrasting siliceous magmatic suites in the Bushveld Complex: genetic aspects and implications for tectonic discrimination diagrams

    J. Volcanol. Geotherm. Res.

    (1987)
  • J.A. VanTongeren et al.

    Hf isotopic evidence for a cogenetic magma source for the Bushveld Complex and associated felsic magmas

    Lithos

    (2016)
  • E.B. Watson et al.

    Zircon saturation revisited: temperature and composition effects in a variety of crustal magma types

    Earth Planet. Sci. Lett.

    (1983)
  • A. Zeh et al.

    The Bushveld Complex was emplaced and cooled in less than one million years – results of zirconology, and geotectonic implications

    Earth Planet. Sci. Lett.

    (2015)
  • S.-J. Barnes et al.

    Composition of the marginal rocks and sills of the Rustenburg Layered Suite, Bushveld Complex, South Africa: implications for the formation of the platinum-group element deposits

    Econ. Geol.

    (2010)
  • E. Belousova et al.

    Igneous zircon: trace element composition as an indicator of source rock type

    Contrib. Mineral. Petrol.

    (2002)
  • A. Benisek et al.

    Factors controlling the development of prism faces in granite zircons: a microprobe study

    Contrib. Mineral. Petrol.

    (1993)
  • H. Brätz et al.

    Analysis of rare earth elements in geological samples by Laser Ablation – Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS)

    Agilent Technologies-Online Journal

    (2002)
  • P.C. Buchanan et al.

    Petrogenesis of the Dullstroom Formation, Bushveld Magmatic Province, South Africa

    Contrib. Mineral. Petrol.

    (1999)
  • P.C. Buchanan et al.

    Geochemistry of intermediate to siliceous volcanic rocks of the Rooiberg Group, Bushveld Magmatic Province, South Africa

    Contrib. Mineral. Petrol.

    (2002)
  • R.G. Cawthorn

    The Bushveld Complex, South Africa

  • R.G. Cawthorn et al.

    Emplacement and crystallization time for the Bushveld Complex

    J. Petrol.

    (1998)
  • R.G. Cawthorn et al.

    The Bushveld complex

  • J.A.D. Connolly

    The geodynamic equation of state: what and how

    Geochem. Geophys. Geosys.

    (2009)
  • J.F.A. Diener et al.

    A new thermodynamic model for clino- and orthoamphiboles in the system Na2O-CaO-FeO-MgO-Al2O3-SiO2-H2O-O

    J. Metamorph. Geol.

    (2007)
  • H.V. Eales et al.

    Crustally contaminated komatiite: primary source of the chromitites and marginal, lower, and critical zone magmas in a staging chamber beneath the Bushveld Complex

    Econ. Geol.

    (2012)
  • J.M. Ferry et al.

    New thermodynamic models and revised calibrations for the Ti-in-zircon and Zr-in-rutile thermometers

    Contrib. Mineral. Petrol.

    (2007)
  • B. Fu et al.

    Ti-in-zircon thermometry: applications and limitations

    Contrib. Mineral. Petrol.

    (2008)
  • M.L. Fuhrman et al.

    Ternary-feldspar modeling and thermometry

    Am. Mineral.

    (1988)
  • T. Geisler et al.

    Re-equilibration of Zircon in Aqueous Fluids and Melts

    Elements

    (2007)
  • M.S. Ghiorso et al.

    An H2O–CO2 mixed fluid saturation model compatible with rhyolite-MELTS

    Contrib. Mineral. Petrol.

    (2015)
  • C.B. Grimes et al.

    On the occurrence, trace element geochemistry, and crystallization history of zircon from in situ ocean lithosphere

    Contrib. Mineral. Petrol.

    (2009)
  • C.B. Grimes et al.

    “Fingerprinting” tectono-magmatic provenance using trace elements in igneous zircon

    Contrib. Mineral. Petrol.

    (2015)
  • G.A.R. Gualda et al.

    Rhyolite-MELTS: a modified calibration of MELTS optimized for silica-rich, fluid-bearing magmatic systems

    J. Petrol.

    (2012)
  • T. Günther et al.

    Mantle sources and magma evolution of the Rooiberg lavas, Bushveld Large Igneous Province, South Africa

    Contrib. Mineral. Petrol.

    (2018)
  • A.L. Hall

    The Bushveld igneous complex in the central Transvaal

    Geological Survey of the Union of South Africa

    (1932)
  • M. Hill et al.

    Geochemical Characteristics and Origin of the Lebowa Granite Suite, Bushveld complex

    Int. Geol. Rev.

    (1996)
  • T.J.B. Holland et al.

    Thermodynamics of order-disorder in minerals. 2. Symmetric formalism applied to solid solutions

    Am. Mineral.

    (1996)
  • Cited by (18)

    • Archean crustal growth and reworking revealed by combined U-Pb-Hf-O isotope and trace element data of detrital zircons from ancient and modern river sediments of the eastern Kaapvaal Craton

      2022, Geochimica et Cosmochimica Acta
      Citation Excerpt :

      Detritus from a river system draining a large area may provide an average sample of the exposed crustal basement (Rudnick and Gao, 2003; Gehrels, 2014; Liu et al., 2017), and thus detrital zircons recovered from clastic sediments and sedimentary rocks provide valuable information about the evolution of the continental crust during early Earth via detailed microbeam U-Pb dating, and trace element and Hf-O isotopic analyses (Pietranik et al., 2008; Zeh et al., 2008; Belousova et al., 2010; Payne et al., 2016; Ranjan et al., 2020). Nevertheless, care must be taken when using detrital zircon grains for provenance analyses, as these grains can be seriously altered by metamorphic and hydrothermal processes, causing age reset, decoupling of the U-Pb and Lu-Hf isotope systems (e.g., Zeh et al., 2008, 2010; Gerdes and Zeh, 2009; Lenting et al., 2010; Zeh and Gerdes, 2014), and bias of the primary trace element budget used to derive petrogenetic parameters (e.g., Zeh et al., 2008; Gudelius et al., 2020a and references therein). The Kaapvaal Craton is one of the oldest and best-preserved Archean continental fragments on Earth (Kröner et al., 1996; Anhaeusser, 2014).

    • Combining detrital zircon shape and U–Pb–Hf isotope analyses for provenance studies – An example from the Aquiri region, Amazon Craton, Brazil

      2021, Precambrian Research
      Citation Excerpt :

      Furthermore, those zircon grains from the Congo Craton do not show combinations with the R type, characterised by {3 0 1} pyramids (T, OC or V types of Pupin, 1980), which are the most ubiquitous type in sample ACAD-04 (76% of the zircon population), as well as in samples ACAD-20 (16%) and AGLD-11 (34%), comprising zircon grains of all age groups from 3600 to 2800 Ma (Fig. 10, Supplementary Table S4). However, zircon grains showing high-temperature typologies and {3 0 1} pyramids have been described from A-type granites, granophyres and felsites with tholeiitic affinity, forming the roof of the Bushveld Complex in southern Africa (Gudelius et al., 2020a; Zeh et al., 2020). Based on this, it is likely that the zircon grains with {3 0 1} pyramids were not derived from TTG rocks, but instead from Archaean granites, perhaps those of A-type affinity in the Amazon Craton.

    View all citing articles on Scopus
    View full text