TY - JOUR
AU1 - J, Ver Hoeve, Thomas
AU2 - S, Scoates, James
AU3 - J, Wall, Corey
AU4 - Dominique, Weis,
AU5 - Marghaleray, Amini,
AB - Abstract The near-solidus crystallization history of the Paleoproterozoic Bushveld Complex, the world’s largest layered intrusion, has been investigated using the in situ trace element geochemistry (LA-ICP-MS) of accessory minerals that crystallized from late, highly fractionated pockets of interstitial melt in layered cumulates and from granitic magmas in felsic roof rocks. Zircon with simple to complex sector zoning occurs in mafic–ultramafic rocks in interstitial pockets that contain quartz–biotite–plagioclase and local granophyric intergrowths. Chondrite-normalized rare earth element patterns are typical of igneous zircon and Ti is negatively correlated with Hf in most samples. Ti-in-zircon thermometry of the cumulates (T = 950–730°C) records the onset of zircon saturation through to the solidus, with notably cooler temperatures determined for Upper Zone and roof rock zircon (T = 875–690°C). Forward modelling of proposed Bushveld parental magmas using rhyolite-MELTS consistently yields similar temperatures for zircon saturation (800–740°C) from highly fractionated melts (∼5–20% remaining melt) with late-stage, near-solidus mineral assemblages similar to those observed in the rocks. Anomalously high and variable Th/U (2–77) in zircon from orthopyroxenites in the Critical Zone, including those associated with the PGE-rich UG2 chromitite and Merensky Reef in the Upper Critical Zone, can be related to U loss from the fractionated interstitial melt during exsolution of late, oxidized Cl-rich fluids. In addition to zircon, rutile occurs throughout the Critical Zone of the Bushveld Complex in two different textural settings, as interstitial grains with quartz and zircon and with chromite, each with distinctive chemistry. Euhedral rutile needles found in interstitial melt pockets have relatively high HFSE concentrations (Nb = 1000–20 000 ppm; Ta = 100–1760 ppm), high Zr-in-rutile temperatures (1000–800°C), and are magmatic in origin. Rutile associated with chromite, either as rims or inclusions, is strongly depleted in HFSE (Nb <1000 ppm; Ta <100 ppm) and in Cr and Sc relative to magmatic rutile, and represents a sub-solidus exsolution product of Ti from chromite. Exploring the near-solidus evolution of mafic layered intrusions such as the Bushveld Complex using the trace element chemistry of accessory minerals provides a novel approach to constraining the late stages of crystallization from highly fractionated interstitial melts in these petrologically significant intrusions. INTRODUCTION Layered intrusions preserve rock records of the processes by which mafic–ultramafic magmas crystallize in the continental crust and they provide ideal natural laboratories for studying magmatic evolution and differentiation in time and space (Wager & Brown, 1968; Parsons, 1987; Cawthorn, 1996; Irvine et al., 1998; Maier et al., 2001; Charlier et al., 2015). The petrology and geochemistry of layered intrusions has typically focused on the textural setting and major element chemistry of the primary cumulus mineral assemblage formed at high temperature (i.e. primocrysts: olivine, pyroxene, spinel, plagioclase), however, there is much to be learned from examining the final stages of crystallization of cumulates (e.g. Meurer & Meurer, 2006; Holness et al., 2011; Yudovskaya et al., 2013; Scoates & Wall, 2015). These late-stage processes, which can include textural maturity (Holness et al., 2007; Holness & Vernon, 2015), accessory mineral crystallization (Scoates & Chamberlain, 1995; Scoates & Wall, 2015; Zeh et al., 2015), isotopic mineral disequilibrium (Tepley & Davidson, 2003; Chutas et al., 2012), and liquid immiscibility (VanTongeren & Mathez, 2012; Veksler & Charlier, 2015), provide critical insight into the magmatic history of these remarkable intrusions. The mineral zircon (ZrSiO4) is now recognized as a relatively common accessory mineral in mafic and ultramafic rocks of layered intrusions (Scoates & Chamberlain, 1995;,Scoates & Friedman, 2008; Wotzlaw et al., 2012; Scoates & Wall, 2015; Zeh et al., 2015; Mungall et al., 2016; Wall et al., 2018). Zircon is widely used in U–Th–Pb geochronology (e.g. Silver & Deutsch, 1963; Schoene, 2013; Barboni & Schoene, 2014), and increasingly used in Hf isotope geochemistry (e.g. Griffin et al., 2002; Belousova et al., 2006, 2010) and O isotope geochemistry (e.g. Valley, 2003; Hawkesworth & Kemp, 2006). The emergence of zircon as a powerful tool for identifying and fingerprinting magmatic sources and processes (e.g. Belousova et al., 2002; Grimes et al., 2009, 2015; Claiborne et al., 2010) is due to the wide range of incompatible trace elements that can be readily incorporated into its crystal lattice. Rare earth element concentrations and ratios in zircon can be used to monitor crystallization processes and magmatic conditions (e.g. Belousova et al., 2006; Trail et al., 2012; Burnham & Berry, 2014). Additionally, the temperature-dependent substitution of Ti4+ in its structure when it is in equilibrium with quartz (SiO2) and rutile (TiO2) allows zircon to serve as a robust thermometer (Ti-in-zircon thermometry) under a range of geologic conditions (Watson & Harrison, 2005; Ferry & Watson, 2007). Other accessory minerals reported from mafic layered intrusions with applications to U–Pb geochronology and trace element-isotopic geochemistry include baddeleyite (ZrO2), apatite (Ca5(PO4)3(F, Cl, OH)), rutile (TiO2), and titanite (CaTiSiO5) (Scoates & Wall, 2015). The trace element geochemistry of accessory zircon and rutile has been investigated using samples that span nearly the entire 8 km-thick stratigraphic sequence of the Paleoproterozoic Bushveld Complex, South Africa, including samples from the overlying granites and granophyres. The textural setting of zircon and rutile was established by combined petrography-scanning electron microscopy (SEM), and the internal structure and zoning of individual grains was imaged by SEM-cathodoluminescence (SEM-CL). Their trace element concentrations were determined in situ by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS; n = 397 analyses). Trace element variations were used to evaluate petrologic relationships, including thermometry (e.g. Ti-in-zircon, Zr-in-rutile), between the major zones of the intrusion and to constrain the late-stage crystallization history of fractionated interstitial melt in the cumulates and granites. A forward geochemical model has been developed for proposed Bushveld parental magmas that tracks the evolution of late-stage fractionated melt to temperatures approaching the solidus of these rocks, with implications for the consolidation history of mafic layered intrusions in general. GEOLOGIC SETTING OF THE BUSHVELD COMPLEX South Africa’s Bushveld Complex is the world’s largest mafic–ultramafic layered intrusion and was emplaced into the Kaapvaal Craton at c.2·06 Ga (Fig. 1) (Eales et al., 1993; Eales & Cawthorn, 1996; Cawthorn et al., 2009; Maier et al., 2013; Cawthorn, 2015; Scoates & Wall, 2015; Zeh et al., 2015; Mungall et al., 2016). Covering more than 95 000 km2 (Finn et al., 2015), the Bushveld Complex has a number of associated satellite intrusions (e.g. Uitkomst, Losberg, Moloto) and overlying volcanic sequences within the broader Bushveld Magmatic Province. The Bushveld Complex includes the Rustenburg Layered Suite, an up to 8 km-thick sequence of stratified mafic and ultramafic cumulates that hosts world-class deposits of chromium, platinum group elements (PGE) and vanadium (Eales & Cawthorn, 1996; Maier, 2005; Cawthorn et al., 2009; Maier et al., 2013; Cawthorn, 2015) (Fig. 2). In the Eastern Limb, the Upper Zone is in direct contact with contemporaneous volcanic rocks of the Rooiberg Group, the Stavoren Granophyre of the Rashoop Granophyre Suite, and the Nebo Granite of the Lebowa Granite Suite, all of which show complex field and petrogenetic relationships with the Rustenburg Layered Suite (Twist & French, 1983; Walraven, 1985; VanTongeren et al., 2010; Mathez et al., 2013; VanTongeren & Mathez, 2015) (Fig. 3). Fig. 1 View largeDownload slide Generalized geologic map of the Bushveld Complex in South Africa showing the locations of the analysed samples containing zircon and rutile used in this study (sample locations noted by orange stars). The location of the PGE-rich Merensky Reef is indicated by the white dashed line. The purple unit includes the Marginal Zone, Lower Zone, and Critical Zone, and the green unit represents the stratigraphically higher Main Zone and Upper Zone. Felsic roof rocks of the Rashoop Granophyre Suite and Lebowa Granite Suite are shown in pink and red, respectively. Map modified from Scoates & Friedman (2008) and based on Kinnaird et al. (2005). TML, Thabazimbi-Murchison Lineament. Fig. 1 View largeDownload slide Generalized geologic map of the Bushveld Complex in South Africa showing the locations of the analysed samples containing zircon and rutile used in this study (sample locations noted by orange stars). The location of the PGE-rich Merensky Reef is indicated by the white dashed line. The purple unit includes the Marginal Zone, Lower Zone, and Critical Zone, and the green unit represents the stratigraphically higher Main Zone and Upper Zone. Felsic roof rocks of the Rashoop Granophyre Suite and Lebowa Granite Suite are shown in pink and red, respectively. Map modified from Scoates & Friedman (2008) and based on Kinnaird et al. (2005). TML, Thabazimbi-Murchison Lineament. Fig. 2 View largeDownload slide Schematic stratigraphic sections of the Bushveld Complex showing the samples examined in this study. Sample locations are indicated by stars; pink stars represent samples that yielded zircon for study and yellow stars indicate samples that did not contain zircon. The left-hand column shows a complete section of the Rustenburg Layered Suite and roof rocks with the maximum unit thickness in metres from the Western Limb (regular font) and Eastern Limb (bold font) taken from Cawthorn et al. (2009). The white dashed line indicates the PGE-rich Merensky Reef in the Upper Critical Zone. Magnetitite layers in the Upper Zone are indicated by black dashed lines. B1, B2, and B3 refer to marginal rocks from Barnes et al. (2010). UUMZ, Upper and Upper Main Zone from VanTongeren et al. (2010). The right-hand column is a detailed section of the Lower Zone and Critical Zone where most of the samples are located. Abbreviations: M, Magnetitite layer; MML, Main Magnetitite Layer; M21, Magnetitite layer 21. Fig. 2 View largeDownload slide Schematic stratigraphic sections of the Bushveld Complex showing the samples examined in this study. Sample locations are indicated by stars; pink stars represent samples that yielded zircon for study and yellow stars indicate samples that did not contain zircon. The left-hand column shows a complete section of the Rustenburg Layered Suite and roof rocks with the maximum unit thickness in metres from the Western Limb (regular font) and Eastern Limb (bold font) taken from Cawthorn et al. (2009). The white dashed line indicates the PGE-rich Merensky Reef in the Upper Critical Zone. Magnetitite layers in the Upper Zone are indicated by black dashed lines. B1, B2, and B3 refer to marginal rocks from Barnes et al. (2010). UUMZ, Upper and Upper Main Zone from VanTongeren et al. (2010). The right-hand column is a detailed section of the Lower Zone and Critical Zone where most of the samples are located. Abbreviations: M, Magnetitite layer; MML, Main Magnetitite Layer; M21, Magnetitite layer 21. Fig. 3 View largeDownload slide Geologic map of the Eastern Limb of the Bushveld Complex modified from Molyneaux (2008). Note the discontinuous nature of the Marginal Zone, the restricted area of the Critical Zone, the major trough structure in the Burgersfort area dominated by rocks of the Lower Zone, the thick Main Zone, and the complex geologic relationships between the upper part of the Rustenburg Layered Suite and the overlying felsic rocks (‘leptite’, Stavoren Granophyre, Nebo Granite, Rooiberg felsite). Fig. 3 View largeDownload slide Geologic map of the Eastern Limb of the Bushveld Complex modified from Molyneaux (2008). Note the discontinuous nature of the Marginal Zone, the restricted area of the Critical Zone, the major trough structure in the Burgersfort area dominated by rocks of the Lower Zone, the thick Main Zone, and the complex geologic relationships between the upper part of the Rustenburg Layered Suite and the overlying felsic rocks (‘leptite’, Stavoren Granophyre, Nebo Granite, Rooiberg felsite). The Bushveld Complex outcrops in five major limbs—Eastern, Western, Far Western, Northern and Southern—spanning nearly 400 km from east to west, with the limbs dipping gently inward to form a bowl-shaped intrusion (Cawthorn & Webb, 2001; Webb et al., 2011). The limbs appear to be (inter)connected at depth based on geophysical evidence, stratigraphic relations, and xenolith studies (Hall, 1932; Cawthorn & Walraven, 1998; Webb et al., 2011). The Eastern Limb of the Bushveld Complex provides the best surficial exposures due to relatively rugged topography and lack of vegetation compared to the Western and Northern limbs and, as a result, has been mapped in the most detail (von Gruenewaldt, 1973; Cameron, 1978, 1980; ,Sharpe, 1981; Harmer & Sharpe, 1985; Molyneaux, 2008) (Fig. 3). Geology of the Rustenburg Layered Suite Rocks of the Rustenburg Layered Suite are divided from base to top into the Marginal, Lower, Critical, Main, and Upper zones, with most contacts determined by changes in cumulus or primocryst mineralogy (Kruger et al., 1987; Eales & Cawthorn, 1996; Cawthorn et al., 2009). Wilson (2015) reported a newly discovered Basal Ultramafic Sequence, including a basal chill sequence, beneath the Marginal Zone, with in excess of 800 m of harzburgites and feldspathic peridotites in the Clapham Compartment of the Eastern Limb. The Lower Zone has a maximum thickness of 1·4 km in the Eastern Limb and is composed primarily of adcumulate dunites and harzburgites at the base and top with lesser volumes of orthopyroxenite in the middle of the unit (Cameron, 1978; Maier et al., 2013; Wilson, 2015). The Critical Zone has been subdivided into the Lower Critical Zone and Upper Critical Zone, and is characterized by abundant orthopyroxenites and numerous cyclic units (Eales et al., 1990). The Upper Critical Zone is host to chromium and PGE deposits found within laterally continuous horizons of chromitite and sulphide-bearing ‘reefs’ (e.g. PGE-rich Merensky Reef) (Cameron, 1980; Naldrett et al., 1986, 2012; Eales et al., 1990; Mondal & Mathez, 2006). The >2 km thick Main Zone of the Bushveld Complex includes mafic rocks of noritic to gabbronoritic composition (Tegner et al., 2006; Maier & Barnes, 2010). The Main Zone extends from the top of the Bastard Reef, which lies stratigraphically above the Merensky Reef, to the Pyroxenite Marker, which represents a major compositional reversal from massive noritic rocks to layered orthopyroxenites (von Gruenewaldt, 1973; Sharpe, 1985). The <2 km thick Upper Zone, from the top of the Pyroxenite Marker to the roof of the Rustenburg Layered Suite, consists of gabbronorites and diorites that contain layers (up to 21) of massive to semi-massive magnetite, the most significant being the Main Magnetite Layer (MML) (VanTongeren et al., 2010; Cawthorn, 2013a). In the southeastern section of the Bushveld Complex, a sequence of quartz hornblende monzonites referred to as the ‘Residual Zone’ (Cawthorn, 2013b) lies above the Upper Zone and beneath the Rooiberg Group. Parental magmas have been proposed for the four major zones of the Bushveld Complex based on the chemistry of marginal rocks and sills found in contact with the Rustenburg Layered Suite (Fig. 2) (Sharpe, 1981, 1985; Tegner et al., 2006; Barnes et al., 2010). Where present, these marginal rocks, which typically consist of fine-grained norites to gabbronorites, are up to a few hundred metres thick and are in contact with the Lower Zone, Critical Zone and Main Zone in the Eastern Limb (Sharpe, 1981; Maier et al., 2013). Some marginal rocks and sills are considered to represent Bushveld Complex parental melts based on trace element concentrations, platinum group element ratios and modelled crystallization sequences (Barnes et al., 2010). In the Burgersfort area of the Eastern Limb (Fig. 3), marginal rocks associated with the Lower Zone are relatively Mg-rich tholeiites (Mg# = 77) and are referred to as B1 (Sharpe, 1981, 1982; Barnes et al., 2010). Marginal rocks in contact with the Lower Critical Zone are comparable to the Mg-rich B1 compositions of the Lower Zone. Upper Critical Zone marginal rocks consist of a more fractionated tholeiitic basaltic composition (Mg# = 55) than the B1 rocks and are referred to as B2 (Sharpe, 1981; Barnes et al., 2010). The Main Zone parental magma is considered to be the B3 composition, which is similar to the B2 tholeiite, but with higher Mg# = 62 (Sharpe, 1981; Barnes et al., 2010). The Upper Zone is interpreted to represent crystallization of the final phase of Bushveld Complex magmatism as represented by the proposed mafic UUMZ parental magma (UUMZ = Upper Zone Upper Main Zone) (VanTongeren et al., 2010). Geology of the felsic roof rocks Bushveld Complex roof rocks that overlie the Rustenburg Layered Suite comprise a series of felsic igneous rocks dominated by granites and granophyres (e.g. Walraven, 1987; Kleemann & Twist, 1989; Cawthorn et al., 2009; Fourie & Harris, 2011; Mathez et al., 2013). In the Eastern Bushveld Complex, the Upper Zone is immediately overlain by the Rashoop Granophyre Suite and includes the Stavoren Granophyre (Fig. 3), which is defined by micrographic intergrowths of quartz and alkali feldspar (Walraven, 1985, 1987). Sheets of Stavoren Granophyre are in direct contact with the underlying Rustenburg Layered Suite and are found between the Nebo Granite of the extensive Lebowa Granite Suite and the overlying volcanic rocks of the Rooiberg Group (e.g. Walraven, 1987; Mathez et al., 2013). The 3–5 km thick Rooiberg Group consists of thick volcanic sequences (up to 400 m each) that range from basaltic to rhyolitic in composition (Twist & French, 1983; Schweitzer et al., 1995; Buchanan et al., 1999, 2002). In the Rooiberg Group, the Damwal, Kwaggasnek, and Schrikkloof formations stratigraphically overlie the Rustenburg Layered Suite, whereas the lowermost basaltic–dacitic Dullstrom Formation is found both below and above the Rustenburg Layered Suite (Twist & French, 1983; Schweitzer et al., 1995; Buchanan et al., 1999). Previous work on zircon and rutile in the Bushveld Complex Zircon in the Bushveld Complex has been the focus of a number of geochronological and trace element studies. Zircon was recovered from a pegmatitic orthopyroxenite of the Merensky Reef in the Western Limb by Scoates & Friedman (2008), who recognized the interstitial textural setting of this U–Th–Pb-bearing accessory mineral and published the first high-precision U–Pb date from the Bushveld Complex using the chemical abrasion-ID-TIMS or CA-TIMS technique (weighted mean 207Pb/206Pb date = 2054·4 ± 1·3 Ma, 2σ). Scoates & Wall (2015) re-analysed zircon from this sample and reported a revised age of 2057·04 ± 0·55 Ma, as well as a 2056·88 ± 0·41 Ma date for zircon from the Merensky Reef in the Eastern Limb, thus confirming synchronous crystallization of this horizon across >300 km in the Bushveld Complex. As part of a broader study of the effects of downhole fractionation corrections in LA-ICP-MS U–Pb geochronology, Ver Hoeve et al. (2018) also investigated in situ dates for zircon grains from the Merensky Reef that were variably untreated, annealed, and annealed. Zeh et al. (2015) presented CA-TIMS U–Pb zircon results for nine samples spanning the stratigraphic sequence of the Rustenburg Layered Suite and proposed that the entire stack of mafic–ultramafic rocks crystallized within about one million years at c.2055–2056 Ma. Mungall et al. (2016) focused on samples from the Critical Zone and found a restricted range of CA-TIMS U–Pb zircon dates of c.2056 Ma for eight samples, that based in part on statistically significant differences between them, indicate an out-of-sequence emplacement of sills in this part of the intrusion. The trace element geochemistry of zircon from the mafic–ultramafic rocks of the Bushveld Complex, mainly focused on samples from the Critical Zone, has been examined by Yudovskaya et al. (2013) and Zeh et al. (2015). Yudovskaya et al. (2013) analysed zircon from chromitites of the Critical Zone in the Eastern, Western, and Northern limbs for trace elements by LA-ICP-MS. They identified two types of zircon, each with distinctive trace element characteristics. Rare metamict or partially metamict cores with dark cathodoluminescence (CL) response are enriched in rare earth elements (REE), Y, Th, and U compared to abundant transparent zircon with bright CL response. Yudovskaya et al. (2013) proposed that the metamict, CL-dark cores represent a rarely preserved high-temperature crystallization stage of zircon and that the common CL-bright grains and rims crystallized from fractionated interstitial melt following new magma inputs and mixing of magmas with different proportions of crystals and melt. Ti-in-zircon thermometry for the chromitites in the Critical Zone yielded a range of temperatures from 930–760°C (Yudovskaya et al., 2013). Zeh et al. (2015) recognized recrystallized melt inclusions in zircon consisting of quartz–K-feldspar intergrowths, which were also noted by Yudovskaya et al. (2013), and identified systematic variations in CL intensity coupled with compositional changes in zircon from the Upper Critical Zone and lowermost Main Zone. Their Ti-in-zircon thermometry results indicated crystallization temperatures from 940–670°C. Zeh et al. (2015) also found significant Th/U variation in zircon, ranging from typical values of ∼0·5–1·5 in zircon from the Upper, Marginal, and Lower Critical zones to highly variable values (0·6 to 10) in zircon from the Upper Critical Zone and the base of the Main Zone. They related the variable and high Th/U in zircon to fractionation of accessory minerals with zircon, including rutile, apatite, and thorite. The presence of rutile in the Bushveld Complex has long been recognized (Cameron & Emerson, 1959). Cameron (1979) described rutile in samples from the Critical Zone and recognized seven distinct settings and morphologies, including associations with chromite, either as rims or as inclusions, and as independent crystals within interstitial minerals (e.g. plagioclase, biotite). For the rutile intergrown with chromite, Cameron (1979) proposed an origin by subsolidus oxidation–exsolution from chromite during cooling. Scoates & Friedman (2008), and subsequently Scoates & Wall (2015), identified rutile in samples from the Merensky Reef, present as overgrowths on chromite or as discrete grains adjacent to chromite and as acicular needles in interstitial plagioclase and biotite. U–Pb dating of rutile from the Merensky Reef in the Western Limb and >300 km to the east in the Eastern Limb yielded ages of 2053·00 ± 2·74 Ma and 2052·96 ± 0·61 Ma, respectively (Scoates & Friedman, 2008; Scoates & Wall, 2015), reflecting cooling through the closure temperature for Pb diffusion in rutile (400–450°C: Schmitz & Bowring, 2003). In their study of microstructures in chromite grains from the Merensky Reef, Vukmanovic et al. (2013) observed rutile with morphologies similar to those described by Cameron (1979) that they also interpreted as an exsolution product from chromite. SAMPLES AND ANALYTICAL TECHNIQUES All samples examined in this study are from the Eastern Limb of the Bushveld Complex, with the exception of SA04–13, which was collected from the West Mine (Townlands shaft; now known as the Khuselaka platinum mine, Khuseleka 1 shaft) near Rustenburg in the Western Limb (Table 1). Samples with the prefix ‘SA04-’ were collected in May 2004 and all other samples are from the collection at the American Museum of Natural History (New York). A total of 20 samples, including surficial, underground and drill-core samples were processed through mineral separation, of which 13 yielded zircon grains and six contained rutile. The sample set includes one each from the Lower Zone (LZ10–02) and Lower Critical Zone (TW477–661), with the majority of samples coming from the Upper Critical Zone, including two UG2 chromitite horizon samples (DT28–912, B00–1-6), three samples from the PGE-rich Merensky Reef (B90–7, SA04–08, SA04–13) and one sample from the (uneconomic) Bastard Reef (MP24D2). Main Zone and Upper Zone rocks proved to be relatively poor targets for zircon recovery and only one sample from the base of the Main Zone (B90–1, Tennis Ball Marker) and one sample from near the top of the Upper Zone (B07–040) yielded zircon. The felsic roof rocks of the Bushveld Complex are represented by three samples. Sample B07–051 is a ‘leptite’, a granophyre with graphic intergrowths of quartz, plagioclase and alkali feldspar (Iannello, 1971), and has been interpreted by VanTongeren & Mathez (2015) and VanTongeren et al. (2016) to represent the molten equivalent of highly thermally metamorphosed Rooiberg Group lava. Sample B10–054 is a Stavoren granophyre and sample B10–056 is a Nebo granite; both have been interpreted as differentiates of Upper Zone magma (Mathez et al., 2013; Zirakparvar et al., 2014; VanTongeren & Mathez, 2015). Hf isotope results determined by laser ablation-MC-ICP-MS have been reported for zircon from samples B07–040, B10–056, and B10–056 in Zirakparvar et al. (2014) and from sample B07–051 by VanTongeren et al. (2016). Table 1 Summary of samples examined from the Bushveld Complex for zircon and rutile trace element geochemistry Sample Stratigraphic Position Subunit Locality Rock Type Latitude Longitude Reference B10-056 Roof Rocks Lebowa (Nebo) Granite Mesekete River Section Granite 29°50·636'E 25°00·256'S Mathez et al. (2013) B10-054 Roof Rocks Rashoop (Stavoren) Granophyre Mesekete River Section Granophyre 29°50·634'E 25°00·214'S Mathez et al. (2013) B07-051 Roof Rocks Rashoop Droogehoek River Section Granodiorite 29°54·287'E 24°51·669'S VanTongeren et al. (2016) B07-040 Upper Zone Droogehoek River Section Diorite 29°54·465'E 24°51·794'S VanTongeren et al. (2010) B90-1 Main Zone Tennis Ball Marker NE of Stoffberg Norite 29°59·935'E 24°17·767'S Zirakparvar et al. (2014) MP24D21 Upper Critical Zone Bastard Reef Atok Mine Pyroxenite 29°51'50·90”E 24°17'46·05”S Mathez et al. (1997) B90-7(0)1 Upper Critical Zone Merensky Reef Atok Mine Pyroxenite 29°51'50·90”E 24°17'46·05”S Mathez (1995) SA04-132 Upper Critical Zone Merensky Reef West Mine (Western Limb) Pyroxenite 27°15’31·9”E 25°37’29·1”S Scoates & Friedman (2008) SA04-08 Upper Critical Zone Merensky Reef Farm Driekop Pyroxenite 30°05’11·4”E 24°31’11”S Scoates & Wall (2015) B00-1-62 Upper Critical Zone UG2 Middelpunt Mine Pyroxenite Middlepunt Mine Mathez & Mey (2005) DT28-9121 Upper Critical Zone UG2 Diamand 422 KS Pyroxenite 29°49'22·63”E 24°17'08·16”S Mondal & Mathez (2006) SA04-06 Upper Critical Zone MG3 Surface Chromitite 29°53·293E 24°16·215'S This study TW477 661.151 Lower Critical Zone Cameron's Section Twickenham 114 KT Pyroxenite 30°00'46·65”E 24°23'40·08”S Cameron (1980), Mondal &  Mathez (2006) LZ10-02 Lower Zone Burgersfort Harzburgite 30°11·114'E 24°38·260'S Zirakparvar (2015) Sample Stratigraphic Position Subunit Locality Rock Type Latitude Longitude Reference B10-056 Roof Rocks Lebowa (Nebo) Granite Mesekete River Section Granite 29°50·636'E 25°00·256'S Mathez et al. (2013) B10-054 Roof Rocks Rashoop (Stavoren) Granophyre Mesekete River Section Granophyre 29°50·634'E 25°00·214'S Mathez et al. (2013) B07-051 Roof Rocks Rashoop Droogehoek River Section Granodiorite 29°54·287'E 24°51·669'S VanTongeren et al. (2016) B07-040 Upper Zone Droogehoek River Section Diorite 29°54·465'E 24°51·794'S VanTongeren et al. (2010) B90-1 Main Zone Tennis Ball Marker NE of Stoffberg Norite 29°59·935'E 24°17·767'S Zirakparvar et al. (2014) MP24D21 Upper Critical Zone Bastard Reef Atok Mine Pyroxenite 29°51'50·90”E 24°17'46·05”S Mathez et al. (1997) B90-7(0)1 Upper Critical Zone Merensky Reef Atok Mine Pyroxenite 29°51'50·90”E 24°17'46·05”S Mathez (1995) SA04-132 Upper Critical Zone Merensky Reef West Mine (Western Limb) Pyroxenite 27°15’31·9”E 25°37’29·1”S Scoates & Friedman (2008) SA04-08 Upper Critical Zone Merensky Reef Farm Driekop Pyroxenite 30°05’11·4”E 24°31’11”S Scoates & Wall (2015) B00-1-62 Upper Critical Zone UG2 Middelpunt Mine Pyroxenite Middlepunt Mine Mathez & Mey (2005) DT28-9121 Upper Critical Zone UG2 Diamand 422 KS Pyroxenite 29°49'22·63”E 24°17'08·16”S Mondal & Mathez (2006) SA04-06 Upper Critical Zone MG3 Surface Chromitite 29°53·293E 24°16·215'S This study TW477 661.151 Lower Critical Zone Cameron's Section Twickenham 114 KT Pyroxenite 30°00'46·65”E 24°23'40·08”S Cameron (1980), Mondal &  Mathez (2006) LZ10-02 Lower Zone Burgersfort Harzburgite 30°11·114'E 24°38·260'S Zirakparvar (2015) All samples collected from outcrop, except (1) drill core sample and (2) underground sample. Table 1 Summary of samples examined from the Bushveld Complex for zircon and rutile trace element geochemistry Sample Stratigraphic Position Subunit Locality Rock Type Latitude Longitude Reference B10-056 Roof Rocks Lebowa (Nebo) Granite Mesekete River Section Granite 29°50·636'E 25°00·256'S Mathez et al. (2013) B10-054 Roof Rocks Rashoop (Stavoren) Granophyre Mesekete River Section Granophyre 29°50·634'E 25°00·214'S Mathez et al. (2013) B07-051 Roof Rocks Rashoop Droogehoek River Section Granodiorite 29°54·287'E 24°51·669'S VanTongeren et al. (2016) B07-040 Upper Zone Droogehoek River Section Diorite 29°54·465'E 24°51·794'S VanTongeren et al. (2010) B90-1 Main Zone Tennis Ball Marker NE of Stoffberg Norite 29°59·935'E 24°17·767'S Zirakparvar et al. (2014) MP24D21 Upper Critical Zone Bastard Reef Atok Mine Pyroxenite 29°51'50·90”E 24°17'46·05”S Mathez et al. (1997) B90-7(0)1 Upper Critical Zone Merensky Reef Atok Mine Pyroxenite 29°51'50·90”E 24°17'46·05”S Mathez (1995) SA04-132 Upper Critical Zone Merensky Reef West Mine (Western Limb) Pyroxenite 27°15’31·9”E 25°37’29·1”S Scoates & Friedman (2008) SA04-08 Upper Critical Zone Merensky Reef Farm Driekop Pyroxenite 30°05’11·4”E 24°31’11”S Scoates & Wall (2015) B00-1-62 Upper Critical Zone UG2 Middelpunt Mine Pyroxenite Middlepunt Mine Mathez & Mey (2005) DT28-9121 Upper Critical Zone UG2 Diamand 422 KS Pyroxenite 29°49'22·63”E 24°17'08·16”S Mondal & Mathez (2006) SA04-06 Upper Critical Zone MG3 Surface Chromitite 29°53·293E 24°16·215'S This study TW477 661.151 Lower Critical Zone Cameron's Section Twickenham 114 KT Pyroxenite 30°00'46·65”E 24°23'40·08”S Cameron (1980), Mondal &  Mathez (2006) LZ10-02 Lower Zone Burgersfort Harzburgite 30°11·114'E 24°38·260'S Zirakparvar (2015) Sample Stratigraphic Position Subunit Locality Rock Type Latitude Longitude Reference B10-056 Roof Rocks Lebowa (Nebo) Granite Mesekete River Section Granite 29°50·636'E 25°00·256'S Mathez et al. (2013) B10-054 Roof Rocks Rashoop (Stavoren) Granophyre Mesekete River Section Granophyre 29°50·634'E 25°00·214'S Mathez et al. (2013) B07-051 Roof Rocks Rashoop Droogehoek River Section Granodiorite 29°54·287'E 24°51·669'S VanTongeren et al. (2016) B07-040 Upper Zone Droogehoek River Section Diorite 29°54·465'E 24°51·794'S VanTongeren et al. (2010) B90-1 Main Zone Tennis Ball Marker NE of Stoffberg Norite 29°59·935'E 24°17·767'S Zirakparvar et al. (2014) MP24D21 Upper Critical Zone Bastard Reef Atok Mine Pyroxenite 29°51'50·90”E 24°17'46·05”S Mathez et al. (1997) B90-7(0)1 Upper Critical Zone Merensky Reef Atok Mine Pyroxenite 29°51'50·90”E 24°17'46·05”S Mathez (1995) SA04-132 Upper Critical Zone Merensky Reef West Mine (Western Limb) Pyroxenite 27°15’31·9”E 25°37’29·1”S Scoates & Friedman (2008) SA04-08 Upper Critical Zone Merensky Reef Farm Driekop Pyroxenite 30°05’11·4”E 24°31’11”S Scoates & Wall (2015) B00-1-62 Upper Critical Zone UG2 Middelpunt Mine Pyroxenite Middlepunt Mine Mathez & Mey (2005) DT28-9121 Upper Critical Zone UG2 Diamand 422 KS Pyroxenite 29°49'22·63”E 24°17'08·16”S Mondal & Mathez (2006) SA04-06 Upper Critical Zone MG3 Surface Chromitite 29°53·293E 24°16·215'S This study TW477 661.151 Lower Critical Zone Cameron's Section Twickenham 114 KT Pyroxenite 30°00'46·65”E 24°23'40·08”S Cameron (1980), Mondal &  Mathez (2006) LZ10-02 Lower Zone Burgersfort Harzburgite 30°11·114'E 24°38·260'S Zirakparvar (2015) All samples collected from outcrop, except (1) drill core sample and (2) underground sample. All samples were crushed and separated using standard heavy liquid and magnetic separation techniques at the Pacific Centre for Isotopic and Geochemical Research at the University of British Columbia (UBC). Zircon grains were concentrated into the N2 (least magnetic) and N2/M5 magnetic splits and selected under a binocular microscope on the basis of clarity, size, external morphology and lack of visible inclusions or discoloration. Rutile was most commonly found in the N2/M5 magnetic split, with some grains sourced from the N2 split. Grains of each mineral were selected and mounted in epoxy set in 2·5 cm diameter pucks that were subsequently ground and polished to their approximate mid-sections. Mounted grains of zircon and rutile were imaged for internal structure at the Electron Microbeam/X-Ray Diffraction Facility (EMXDF) at UBC, Vancouver, using a Philips Xl-30 scanning electron microscope (SEM). Zircon grains were imaged using a Robinson cathodoluminescence (CL) detector, and rutile grains, along with a subset of zircon grains from the roof rock samples, were analysed by backscattered electron imaging (BSE) using a Bruker Quanta 200 energy-dispersion X-ray microanalysis system with XFlash 6010 SDD detector at a voltage of 15 kV. To verify the stoichiometry of zircon for use as an internal standard during analysis by LA-ICP-MS, a random sampling of three to five grains per sample (total analyses = 64) were analysed by electron probe micro-analysis (EPMA) at the EMXDF, with full results provided in Supplementary Data Electronic Appendix 1; Supplementary Data are available for downloading at http://www.petrology.oxfordjournals.org. Analyses were carried out using a fully automated CAMECA SX-50 instrument operating in wavelength-dispersion mode with an excitation voltage of 15 kV and a beam current of 20 nA. Concentrations of Si, Zr, and Hf were measured using spot diameters of 5 μm and peak count times of 20 seconds, 10 seconds, and 100 seconds, respectively. The following standards, X-ray lines, and crystals were used for the elements considered: zircon, SiKα, TAP; zircon, ZrLα, PET; Hf element, HfLα, LIF; diopside, CaKα, PET. Spots were mostly placed in the centers of grains; fractures or slopes on the grain surfaces were avoided. Both Zr and Si are stoichiometric with an average value of 48·6 wt % and 14·9 wt %, respectively (Fig. 4); Hf contents are more variable (0·93–1·41 wt %). Fig. 4 View largeDownload slide Major and minor element oxide variations in Bushveld Complex zircon measured by EPMA. (a) Zr/Hf vs HfO2 ; (b) SiO2 vs HfO2; (c) ZrO2 vs HfO2; (d) ZrO2 vs SiO2. The analyses define the expected decreasing trend in Zr/Hf with increasing Hf content and show minimal variation in ZrO2 and SiO2 (i.e. stoichiometric Zr and Si) of zircon over the entire range of HfO2 from the Bushveld Complex. Abbreviations in legend: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ-MR, Upper Critical Zone-Merensky Reef; UCZ, Upper Critical Zone; MZ, Main Zone; UZ, Upper Zone. Fig. 4 View largeDownload slide Major and minor element oxide variations in Bushveld Complex zircon measured by EPMA. (a) Zr/Hf vs HfO2 ; (b) SiO2 vs HfO2; (c) ZrO2 vs HfO2; (d) ZrO2 vs SiO2. The analyses define the expected decreasing trend in Zr/Hf with increasing Hf content and show minimal variation in ZrO2 and SiO2 (i.e. stoichiometric Zr and Si) of zircon over the entire range of HfO2 from the Bushveld Complex. Abbreviations in legend: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ-MR, Upper Critical Zone-Merensky Reef; UCZ, Upper Critical Zone; MZ, Main Zone; UZ, Upper Zone. Trace element concentrations were determined in zircon and rutile by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) at the Pacific Centre for Isotopic and Geochemical Research (PCIGR) at UBC using a Resonetics (now ASI) RESOlution M-50-LR Class I laser ablation system coupled to an Agilent 7700x quadrupole ICP-MS (Table 2). Ablations were carried out with the 193 nm excimer laser, using a beam energy of 100 mJ and a pit diameter of 34 μm, with a pulse rate of 5 Hz for a duration of 40 seconds followed by 20 seconds of gas blank. Spots were chosen based on grain-scale variations observed in CL images and grain size, with typically two analyses per grain. For zircon, measurements were conducted on the masses 7Li, 29Si, 43Ca, 45Sc, 49Ti, 85Rb, 88Sr, 89Y, 90Zr, 93Nb, 139La, 140Ce, 141Pr, 146Nd, 147Sm, 153Eu, 157Gd, 159Tb, 163Dy, 165Ho, 166Er, 169 Tm, 172Yb, 175Lu, 177Hf, 188Ta, 202Hg (as a monitor for Pb), Pb (204Pb, 206Pb, 207Pb, 208Pb), 232Th, and U (235U and 238U); these masses were chosen based on high relative isotopic abundances and the absence of interferences. For rutile, 53Cr, 57Fe, and 182W were also analysed. Sample-standard bracketing was carried out with spot analyses of synthetic glasses NIST 612 and NIST 610 between zircon and rutile analyses and the natural zircon reference materials 91 500, Plešovice, and FC-1 were also analysed. Table 2 Procedure for LA-ICP-MS analysis of zircon and rutile Parameter Value Zircon Rutile Resonetics RESOlution M-50 LR Class I Laser  Wavelength 193 nm  Energy density 5 J cm-2  Repetition rate 5 Hz  Spot diameter 34 μm  Helium flow 0·75 L min-1  High-purity  N2 gas flow 0·002 L min-1  Ablation time /  gas blank time 40 s / 20 s Agilent 7700x Quadrupole ICP-MS  Argon carrier gas flow 0·53 L min-1 0·53 L min-1  Masses measured 34 37  Mass sweep time 705 ms 606 ms  ThO+/Th+ (NIST612) <0·4 wt %  238U/232Th sensitivity (NIST 612) 102-120% Standardization and Reduction  Trace element standard NIST 612  Reference materials NIST 610, 91 500, Plešovice, FC-1 NIST 610  Data reduction Iolite 2·5 extensiona for Igor Pro, Trace_Elements _IS reduction scheme  Internal standard Zr ∼ 49·3 wt· % (EPMA)b Ti = 59·5 wt· % (stoichiometric) Parameter Value Zircon Rutile Resonetics RESOlution M-50 LR Class I Laser  Wavelength 193 nm  Energy density 5 J cm-2  Repetition rate 5 Hz  Spot diameter 34 μm  Helium flow 0·75 L min-1  High-purity  N2 gas flow 0·002 L min-1  Ablation time /  gas blank time 40 s / 20 s Agilent 7700x Quadrupole ICP-MS  Argon carrier gas flow 0·53 L min-1 0·53 L min-1  Masses measured 34 37  Mass sweep time 705 ms 606 ms  ThO+/Th+ (NIST612) <0·4 wt %  238U/232Th sensitivity (NIST 612) 102-120% Standardization and Reduction  Trace element standard NIST 612  Reference materials NIST 610, 91 500, Plešovice, FC-1 NIST 610  Data reduction Iolite 2·5 extensiona for Igor Pro, Trace_Elements _IS reduction scheme  Internal standard Zr ∼ 49·3 wt· % (EPMA)b Ti = 59·5 wt· % (stoichiometric) a See Paton et al. 2011. b Internal standard values measured by EPMA (Fig. 4). Table 2 Procedure for LA-ICP-MS analysis of zircon and rutile Parameter Value Zircon Rutile Resonetics RESOlution M-50 LR Class I Laser  Wavelength 193 nm  Energy density 5 J cm-2  Repetition rate 5 Hz  Spot diameter 34 μm  Helium flow 0·75 L min-1  High-purity  N2 gas flow 0·002 L min-1  Ablation time /  gas blank time 40 s / 20 s Agilent 7700x Quadrupole ICP-MS  Argon carrier gas flow 0·53 L min-1 0·53 L min-1  Masses measured 34 37  Mass sweep time 705 ms 606 ms  ThO+/Th+ (NIST612) <0·4 wt %  238U/232Th sensitivity (NIST 612) 102-120% Standardization and Reduction  Trace element standard NIST 612  Reference materials NIST 610, 91 500, Plešovice, FC-1 NIST 610  Data reduction Iolite 2·5 extensiona for Igor Pro, Trace_Elements _IS reduction scheme  Internal standard Zr ∼ 49·3 wt· % (EPMA)b Ti = 59·5 wt· % (stoichiometric) Parameter Value Zircon Rutile Resonetics RESOlution M-50 LR Class I Laser  Wavelength 193 nm  Energy density 5 J cm-2  Repetition rate 5 Hz  Spot diameter 34 μm  Helium flow 0·75 L min-1  High-purity  N2 gas flow 0·002 L min-1  Ablation time /  gas blank time 40 s / 20 s Agilent 7700x Quadrupole ICP-MS  Argon carrier gas flow 0·53 L min-1 0·53 L min-1  Masses measured 34 37  Mass sweep time 705 ms 606 ms  ThO+/Th+ (NIST612) <0·4 wt %  238U/232Th sensitivity (NIST 612) 102-120% Standardization and Reduction  Trace element standard NIST 612  Reference materials NIST 610, 91 500, Plešovice, FC-1 NIST 610  Data reduction Iolite 2·5 extensiona for Igor Pro, Trace_Elements _IS reduction scheme  Internal standard Zr ∼ 49·3 wt· % (EPMA)b Ti = 59·5 wt· % (stoichiometric) a See Paton et al. 2011. b Internal standard values measured by EPMA (Fig. 4). Data were reduced using Iolite 2.5 software running within the Igor Pro environment using the Trace Elements IS data reduction scheme (Paton et al., 2011). For zircon, NIST 612 was used as the standard and the average Zr contents (by sample) determined by EPMA were employed as an internal standard value for the unknowns. For rutile, the same reduction scheme was initially employed with a stoichiometric value of 59·5 wt % Ti as the internal standard. This method resulted in trace element concentrations in the NIST 610 reference material that were ∼30% below values reported in Pearce et al. (1997), likely due to the refractory nature of Ti and the physical and compositional differences between NIST 612 and NIST 610. Use of a semi-quantitative reduction scheme (no internal standard) yielded precise and accurate trace element concentrations in the NIST 610 reference material. The trace element concentrations and uncertainties for the rutile analyses were nearly identical for the two reduction methods; therefore, results from the semi-quantitative reduction method are utilized for this study. All concentrations are reported in ppm with uncertainties of 2σ. RESULTS Textural setting and internal structure of zircon and rutile in the Bushveld Complex Zircon occurs as a relatively common accessory mineral in ultramafic–mafic cumulates of the Bushveld Complex and is associated with interstitial pockets containing quartz, plagioclase, and locally alkali feldspar, which are commonly observed as granophyric and myrmekitic intergrowths (Figs 5, 6). This textural setting is consistent with crystallization of zircon from fractionated pockets of interstitial melt (e.g. Yudovskaya et al., 2013; Scoates & Wall, 2015; Zeh et al., 2015; Wall & Scoates, 2016). SEM-CL images of the analysed grains from the Bushveld Complex demonstrate the absence of embayed cores and truncated growth zones, with no evidence for xenocrysts, inheritance of older grains, or significant changes in crystallization conditions that could lead to dissolution and regrowth (e.g. Miller & Wooden, 2004) (Fig. 7). Fig. 5 View largeDownload slide Photomicrographs showing the textural setting of zircon and rutile in rocks from the Rustenburg Layered Suite (all scale bars are 200 μm). (a) Euhedral needle of rutile in contact with chromite in interstitial plagioclase (TW477–661, XPL). (b) Irregular zircon grain in small interstitial pocket consisting of clinopyroxene, quartz, and biotite (TW477–661, XPL). (c) Two different settings of rutile, including a euhedral needle within interstitial plagioclase and a larger anhedral rutile grain on a chromite rim that shares a grain boundary with cumulus orthopyroxene (B00–1-6, PPL). (d) Two zircon grains in interstitial pockets associated with quartz and biotite (SA04–08, XPL). (e) Zircon and rutile in the core of a chromite crystal that in turn is hosted in an interstitial pocket of plagioclase. Note biotite rimming chromite (SA04–13, XPL). (f) Zircon and rutile in the core of a large chromite grain associated with interstitial plagioclase and clinopyroxene (SA04–13, XPL). (g) Interstitial quartz containing a rutile needle (MPD24D2, XPL). (h) Elongate zircon grains rimming chromite in contact with sericitized plagioclase. The chromite grain in the lower left is in contact with biotite and quartz (B90–1, XPL). (i) Large euhedral zircon associated with quartz and feldspar in small interstitial pockets in gabbro (plagioclase, clinopyroxene, orthopyoxene) (B07–040, XPL). Abbreviations: PPL, plane-polarized light; XPL, cross-polarized light; opx, orthopyroxene; plag, plagioclase; chr, chromite; cpx, clinopyroxene; ksp, alkali feldspar; qtz, quartz; bt, biotite. Fig. 5 View largeDownload slide Photomicrographs showing the textural setting of zircon and rutile in rocks from the Rustenburg Layered Suite (all scale bars are 200 μm). (a) Euhedral needle of rutile in contact with chromite in interstitial plagioclase (TW477–661, XPL). (b) Irregular zircon grain in small interstitial pocket consisting of clinopyroxene, quartz, and biotite (TW477–661, XPL). (c) Two different settings of rutile, including a euhedral needle within interstitial plagioclase and a larger anhedral rutile grain on a chromite rim that shares a grain boundary with cumulus orthopyroxene (B00–1-6, PPL). (d) Two zircon grains in interstitial pockets associated with quartz and biotite (SA04–08, XPL). (e) Zircon and rutile in the core of a chromite crystal that in turn is hosted in an interstitial pocket of plagioclase. Note biotite rimming chromite (SA04–13, XPL). (f) Zircon and rutile in the core of a large chromite grain associated with interstitial plagioclase and clinopyroxene (SA04–13, XPL). (g) Interstitial quartz containing a rutile needle (MPD24D2, XPL). (h) Elongate zircon grains rimming chromite in contact with sericitized plagioclase. The chromite grain in the lower left is in contact with biotite and quartz (B90–1, XPL). (i) Large euhedral zircon associated with quartz and feldspar in small interstitial pockets in gabbro (plagioclase, clinopyroxene, orthopyoxene) (B07–040, XPL). Abbreviations: PPL, plane-polarized light; XPL, cross-polarized light; opx, orthopyroxene; plag, plagioclase; chr, chromite; cpx, clinopyroxene; ksp, alkali feldspar; qtz, quartz; bt, biotite. Fig. 6 View largeDownload slide Thin section scan and photomicrographs showing the distribution, shape, and mineralogy of interstitial pockets in Lower Critical Zone sample TW477–661. (a) Thin section scan (4 x 2·5 cm) in transmitted light shows cumulus orthopyroxene (grey) with abundant interstitial material – quartz, Na-plagioclase (white), biotite (brown), and chromite (black). Black outlined boxes indicate areas shown in detail in the panels. (b-g) Photomicrographs illustrating the mineralogy and textures of individual interstitial pockets including zircon. Scale bars are 500 microns in all panels. Abbreviations: opx, orthopyroxene; plag, plagioclase; chr, chromite; bt, biotite; grano, granophyre; qtz, quartz; cpx, clinopyroxene; rt, rutile; ap, apatite; zr, zircon. Fig. 6 View largeDownload slide Thin section scan and photomicrographs showing the distribution, shape, and mineralogy of interstitial pockets in Lower Critical Zone sample TW477–661. (a) Thin section scan (4 x 2·5 cm) in transmitted light shows cumulus orthopyroxene (grey) with abundant interstitial material – quartz, Na-plagioclase (white), biotite (brown), and chromite (black). Black outlined boxes indicate areas shown in detail in the panels. (b-g) Photomicrographs illustrating the mineralogy and textures of individual interstitial pockets including zircon. Scale bars are 500 microns in all panels. Abbreviations: opx, orthopyroxene; plag, plagioclase; chr, chromite; bt, biotite; grano, granophyre; qtz, quartz; cpx, clinopyroxene; rt, rutile; ap, apatite; zr, zircon. Fig. 7 View largeDownload slide Representative scanning electron microscope-cathodoluminescence images of zircon from the Bushveld Complex. Yellow circles indicate locations of spot analyses by LA-ICP-MS (circle diameter is 34 μm) with values of Hf and U ppm with 2σ indicated and Th/U shown. Temperatures (in circles) are Ti-in-zircon temperatures calculated using aSiO2 = 1 and aTiO2 = 1 (Ferry & Watson, 2007). The exceptions are samples from the Upper Zone and roof granites where aTiO2 = 0·7. (a) LZ10–02, Lower Zone; (b) TW477–661, Lower Critical Zone; (c) DT28–912, UG2; (d) B00–1-6, UG2; (e) B90–7, Merensky Reef (Eastern Limb); (f) SA04–08, Merensky Reef (Eastern Limb) (g) MP24D2–26, Bastard Reef; (h) B90–1, base of Main Zone – Tennis Ball Marker; (i) B07–040, Upper Zone; (j) B07–051, Rashoop granophyre; (k) B10–054, Stavoren Granophyre (Rashoop); (l) B10–056, Nebo Granite (Lebowa). Scales as indicated by the white bars on each panel. Fig. 7 View largeDownload slide Representative scanning electron microscope-cathodoluminescence images of zircon from the Bushveld Complex. Yellow circles indicate locations of spot analyses by LA-ICP-MS (circle diameter is 34 μm) with values of Hf and U ppm with 2σ indicated and Th/U shown. Temperatures (in circles) are Ti-in-zircon temperatures calculated using aSiO2 = 1 and aTiO2 = 1 (Ferry & Watson, 2007). The exceptions are samples from the Upper Zone and roof granites where aTiO2 = 0·7. (a) LZ10–02, Lower Zone; (b) TW477–661, Lower Critical Zone; (c) DT28–912, UG2; (d) B00–1-6, UG2; (e) B90–7, Merensky Reef (Eastern Limb); (f) SA04–08, Merensky Reef (Eastern Limb) (g) MP24D2–26, Bastard Reef; (h) B90–1, base of Main Zone – Tennis Ball Marker; (i) B07–040, Upper Zone; (j) B07–051, Rashoop granophyre; (k) B10–054, Stavoren Granophyre (Rashoop); (l) B10–056, Nebo Granite (Lebowa). Scales as indicated by the white bars on each panel. In the mafic–ultramafic cumulates of the Rustenburg Layered Suite, zircon grains (generally) range from 100–250 μm in their longest dimension, whereas those from the felsic roof rocks are mostly smaller than 100 μm. Zircon from the Lower Zone (LZ10–02) is irregular (anhedral) and shows complex sector zoning (Fig. 7a); some grains appear featureless in CL imaging. Zircon from the Critical Zone is highly variable, both between samples and within grains from a single sample (Fig. 7b–g). Most grains are distinguished by sector zoning and some grains reveal convolute zoning with internal oscillatory bands (e.g. Fig. 7c). Zircon from the UG2 chromitite and Merensky Reef in the Critical Zone displays a range of features dominated by anhedral, sector-zoned grains that commonly contain fine oscillatory zoning within sectors (Fig. 7b–f). Zircon from the Main Zone (B90–1) is characterized by poor CL response and few zoning features, with many grains appearing flat in CL or displaying minor CL gradients or banding (Fig. 7h). Zircon in the Upper Zone sample is coarser grained and euhedral compared with those from the Main Zone, Critical Zone, and Lower Zone. Zircon from the Upper Zone has a brighter CL response and commonly displays micron-scale oscillatory zoning within growth sectors (Fig. 7i). Zircon from the three different felsic roof units is typically euhedral and small, ranging from 50–150 microns in length. Oscillatory zoning is present in nearly all zircon from these samples, with more complex zoning locally present in some grains (Fig. 7j–l). Small inclusions within zircon observed in BSE images from the roof rock samples are identified as apatite, sodic plagioclase, and potassium feldspar. Rutile, which occurs in Lower Zone and Main Zone samples and in all samples from the Critical Zone, typically ranges from 60–250 μm in length, with many grains being ∼150 × 100 μm in size. Most rutile is found as acicular crystals within interstitial pockets containing quartz, plagioclase, alkali feldspar, biotite, and zircon (Fig. 5a, d, i). In the Critical Zone, sub-equant to rounded rutile is also found intimately associated with chromite, either on the rims or in the cores of chromite grains (Fig. 5d, f, g, h) as documented by Cameron (1979) and Vukmanovic et al. (2013). In some samples (e.g. B00–1-6 from UG2), both associations are present within millimetres of each other (Fig. 5d). BSE images reveal minimal internal structure in rutile (Fig. 8); however, some grains display fine sharp bright lines that represent exsolution lamellae of ilmenite (Fig. 8c and d). Fig. 8 View largeDownload slide Backscattered electron (BSE) images of representative rutile grains from the Critical Zone of the Bushveld Complex. Circles indicate locations of spot analyses by LA-ICP-MS (circle diameter is 34 μm) with concentrations of Hf, Nb, and Ta (ppm) shown. Temperatures (in circles) are Zr-in-rutile temperatures calculated assuming aSiO2 = 1 for primary magmatic rutile (yellow circles: b, c, d, e, i) and aSiO2 = 0·5 for grains exsolved from chromite (white circles: a, f, g, h) (Ferry & Watson, 2007). (a-b) TW477–661 (Lower Critical Zone) rutile characterized by fractured subhedral grains with no obvious zonation in BSE. (c) DT28–912 (UG2 – Upper Critical Zone) rutile with euhedral prismatic habit. (d) DT28–912 (UG2 – Upper Critical Zone) rutile with anhedral habit and exsolved ilmenite (bright lines). Small inclusions and fractures are present. (e-f). SA04–08 (Merensky Reef – Upper Critical Zone) rutile grains are small and irregular showing no significant BSE variation. (g) SA04–13 (Merensky Reef – Upper Critical Zone, Western Limb) rutile with angular–subhedral morphology and minimal BSE variation. (h) SA04–13 (Merensky Reef – Upper Critical Zone, Western Limb) rutile with prismatic appearance, no major BSE variation, and minor fracturing. (i) MP24D2 (Bastard Reef – Upper Critical Zone) very small rutile (<50 μm). Scales as indicated by the white bars on each panel. Fig. 8 View largeDownload slide Backscattered electron (BSE) images of representative rutile grains from the Critical Zone of the Bushveld Complex. Circles indicate locations of spot analyses by LA-ICP-MS (circle diameter is 34 μm) with concentrations of Hf, Nb, and Ta (ppm) shown. Temperatures (in circles) are Zr-in-rutile temperatures calculated assuming aSiO2 = 1 for primary magmatic rutile (yellow circles: b, c, d, e, i) and aSiO2 = 0·5 for grains exsolved from chromite (white circles: a, f, g, h) (Ferry & Watson, 2007). (a-b) TW477–661 (Lower Critical Zone) rutile characterized by fractured subhedral grains with no obvious zonation in BSE. (c) DT28–912 (UG2 – Upper Critical Zone) rutile with euhedral prismatic habit. (d) DT28–912 (UG2 – Upper Critical Zone) rutile with anhedral habit and exsolved ilmenite (bright lines). Small inclusions and fractures are present. (e-f). SA04–08 (Merensky Reef – Upper Critical Zone) rutile grains are small and irregular showing no significant BSE variation. (g) SA04–13 (Merensky Reef – Upper Critical Zone, Western Limb) rutile with angular–subhedral morphology and minimal BSE variation. (h) SA04–13 (Merensky Reef – Upper Critical Zone, Western Limb) rutile with prismatic appearance, no major BSE variation, and minor fracturing. (i) MP24D2 (Bastard Reef – Upper Critical Zone) very small rutile (<50 μm). Scales as indicated by the white bars on each panel. Trace element geochemistry of zircon Table 3 contains a summary of trace element concentrations in zircon; complete LA-ICP-MS results are listed in Supplementary Data Electronic Appendix 2. In the following section, Hf–Ti and U–Th variations in zircon are addressed by comparing the results of the entire dataset (Fig. 9) and by examining spatial variations within individual grains (Fig. 10). Variations in the REE abundances and patterns of zircon from the different rocks in the Bushveld Complex are also evaluated (Fig. 11). Table 3 Summary of ranges of trace element concentrations in zircon from the Bushveld Complex Element Li Sc Ti Y Nb La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Pb Th U Th/U Tzrc (°C) LZ10-02 - Lower Zone (n = 19) Q1 0·50 411 21·1 377 3·28 0·03 14·9 0·16 2·10 2·61 0·76 10·9 3·08 36·4 12·2 50·7 9·38 75·3 14·7 8340 0·81 230 184 170 1·0 821 Median 1·05 419 22·9 499 3·51 0·04 16·3 0·21 3·08 4·33 1·29 16·8 4·59 49·6 16·3 66·4 11·94 94·1 19·2 8493 1·17 280 220 207 1·1 830 Q3 2·81 431 25·2 589 4·58 0·07 16·9 0·42 5·07 5·19 1·44 19·7 5·38 58·2 19·0 79·1 14·28 116 22·6 8848 1·36 375 310 288 1·1 840 TW477-661 - Lower Critical Zone (n = 18) Q1 0·79 362 13·3 451 2·73 0·00 11·4 0·06 1·27 2·64 0·60 13·4 3·94 42·8 14·5 58·6 10·80 85·1 16·9 9659 0·55 82 69·3 59·3 1·0 773 Median 1·89 375 15·2 697 3·15 0·02 12·7 0·17 2·65 4·45 1·09 17·5 5·08 59·1 21·5 95·6 18·29 147 29·0 10855 0·80 191 149 81·8 3·0 786 Q3 3·42 397 27·3 1420 4·10 0·05 23·8 0·37 5·00 7·09 1·73 36·3 11·0 131 45·8 195 36·01 290 55·9 11284 1·28 539 446 122 4·4 849 DT28-912 - Upper Critical Zone, UG2 (n = 24) Q1 1·34 270 18·3 380 0·90 0·00 3·4 0·05 1·01 2·31 0·36 11·4 3·24 34·7 11·7 49·7 9·45 77·5 15·1 10048 0·33 130 131 5·35 13 806 Median 2·45 281 25·6 647 1·13 0·00 3·8 0·17 2·87 4·29 0·79 18·5 5·40 58·8 20·1 86·6 15·62 125 24·0 10360 0·40 192 205 6·78 20 842 Q3 3·85 287 29·2 771 1·62 0·01 5·5 0·25 4·61 7·07 0·97 26·2 6·77 72·8 23·9 101 18·47 141 27·0 10878 0·84 243 255 18·8 30 858 B00-1-6 - Upper Critical Zone, UG2 (n = 31) Q1 1·08 232 14·2 322 0·73 0·00 3·2 0·04 0·82 1·87 0·33 8·70 2·73 29·6 9·8 43·1 8·22 68·8 13·0 9125 0·26 118 101 3·54 8·0 779 Median 1·49 240 20·5 505 0·86 0·00 3·8 0·07 1·70 4·36 0·58 15·6 4·36 48·8 15·9 63·1 11·90 95·8 17·8 9550 0·34 152 135 4·48 25 818 Q3 3·98 246 23·0 654 1·47 0·00 5·9 0·18 3·12 6·04 0·83 21·7 5·91 63·3 20·3 83·4 15·45 124 24·1 10110 0·71 219 195 25 33 830 B90-7 - Upper Critical Zone, Merensky Reef (n = 20) Q1 0·25 389 22·5 358 2·23 0·00 2·8 0·03 0·74 1·60 0·29 7·64 2·48 31·0 11·2 51·5 10·62 92·3 18·6 9988 0·29 266 155 10·2 10 828 Median 0·38 395 24·0 620 2·35 0·00 3·2 0·14 2·39 4·32 0·54 17·5 5·17 58·7 19·8 83·4 15·54 128 24·8 10510 0·38 332 199 13·5 11 835 Q3 0·69 410 26·4 724 2·57 0·01 3·8 0·25 3·98 5·60 0·75 22·3 6·11 68·2 23·2 98·2 18·24 150 29·1 11270 0·47 518 252 18·8 19 846 SA04-08 Upper Critical Zone, Merensky Reef (n = 25) Q1 0·41 239 17·7 229 1·16 0·00 3·2 0·02 0·45 0·84 0·19 5·3 1·60 20·6 7·4 33·6 6·34 53·3 10·7 7212 0·38 106 99·0 101 0·84 802 Median 0·63 249 20·4 272 1·34 0·00 3·9 0·03 0·65 1·16 0·26 6·1 1·89 23·4 8·7 39·3 7·67 65·0 13·1 7455 0·48 186 149 114 1·0 817 Q3 0·98 256 21·9 448 1·58 0·03 4·8 0·14 2·21 2·83 0·46 12·9 3·53 42·1 14·7 62·1 11·54 93·3 18·1 7853 0·69 243 187 155 1·3 825 SA04-13 Upper Critical Zone, Merensky Reef (n = 22) Q1 0·17 227 17·4 329 1·14 0·00 3·7 0·03 0·53 1·52 0·23 8·02 2·51 30·9 10·5 43·5 8·05 65·1 12·0 8752 0·54 99·5 77·4 30·2 2·0 801 Median 0·27 233 20·8 406 1·38 0·00 5·9 0·06 1·01 2·19 0·34 10·7 3·25 37·7 12·8 53·0 9·96 82·3 15·2 8995 0·64 149 142 44·0 2·8 820 Q3 0·72 237 24·9 626 1·69 0·01 6·6 0·11 1·44 3·90 0·58 19·2 5·64 63·3 20·2 82·5 14·71 115 21·4 9741 0·77 195 168 56·5 3·7 839 MP24D2 - Upper Critical Zone, Bastard Reef (n = 15) Q1 0·45 281 22·9 314 1·36 0·00 5·4 0·08 1·26 2·26 0·50 11·3 3·01 32·4 10·5 41·5 7·74 57·9 11·0 8950 0·50 423 105 26·5 3·1 830 Median 1·11 290 33·8 348 1·69 0·01 7·2 0·13 2·30 3·12 0·65 12·6 3·32 36·0 11·2 45·1 8·14 65·0 12·1 9565 0·86 1507 147 44·6 3·6 875 Q3 1·56 295 46·8 530 2·62 0·01 7·8 0·23 3·57 4·73 0·96 19·1 5·09 54·2 17·2 68·3 12·27 94·3 17·7 10088 0·90 3158 155 47·9 6·9 915 B90-1 - Main Zone, Tennis Ball Marker (n = 20) Q1 1·05 333 24·9 981 2·78 0·09 21·3 0·36 4·38 6·00 0·71 25·0 7·76 90·6 31·6 142 27·93 236 45·8 9195 0·72 369 170 190 0·84 839 Median 2·73 346 28·9 1400 3·08 0·21 24·2 0·72 8·75 8·10 0·99 35·3 10·8 133 46·0 201 39·65 332 62·4 9611 1·11 622 227 263 1·0 856 Q3 30·5 351 35·2 1725 4·61 1·89 39·5 1·38 10·5 12·4 1·72 49·0 13·7 165 56·5 247 45·92 369 71·2 10418 1·62 935 475 643 1·1 879 B07-040 - Upper Zone (n = 32) Q1 0·60 235 5·90 470 0·39 0·00 2·6 0·02 0·37 1·14 0·22 7·40 2·74 37·5 15·1 74·9 15·95 140 27·0 8835 0·17 137 56·6 132 0·42 698 Median 1·58 239 6·95 767 0·43 0·00 3·1 0·05 1·17 2·82 0·54 16·2 5·17 66·6 24·8 118 23·55 200 38·9 9375 0·19 164 78·8 163 0·50 713 Q3 2·69 242 8·40 1089 0·51 0·01 3·3 0·10 2·13 4·50 0·86 25·1 7·68 97·0 36·2 165 32·82 273 50·9 9644 0·22 225 105 189 0·57 730 B07-051 - Felsic Roof Rocks, microgranite (n = 15) Q1 1·07 169 7·80 890 2·09 0·08 7·1 0·25 2·58 4·65 0·83 22·4 6·87 83·5 30·0 136 26·75 227 43·3 6835 0·81 89·7 88·2 159 0·56 723 Median 2·29 170 8·50 1140 2·29 0·39 10·4 0·38 5·33 7·13 1·14 30·2 9·25 111 37·0 161 31·40 257 48·9 6954 0·88 101 96·1 165 0·57 731 Q3 3·58 171 10·7 1572 3·09 7·45 26·5 2·55 13·6 8·78 1·26 37·1 11·3 138 48·4 214 40·20 330 62·0 7560 1·11 112 101 179 0·60 751 B10-054 - Stavoren Granophyre (n = 16) Q1 0·64 192 10·3 1130 2·41 0·06 7·3 0·21 3·22 5·32 0·91 25·7 8·47 107 38·2 174 33·53 284 54·2 7646 1·01 79·8 84·2 153 0·54 749 Median 1·16 194 10·9 1274 2·48 0·59 9·5 0·43 5·70 6·69 1·02 33·0 9·74 120 42·6 190 36·60 302 58·4 7750 1·04 98·3 92·9 165 0·67 754 Q3 2·57 195 11·3 1479 2·86 3·68 19·0 1·43 8·88 9·45 1·26 40·3 12·3 145 50·0 216 40·70 336 64·0 7811 1·13 133 109 185 0·59 758 B10-056 - Nebo Granite (n = 20) Q1 0·54 169 7·05 728 2·16 0·22 8·6 0·41 4·58 4·64 0·78 15·5 5·25 67·7 25·1 117 23·35 199 38·3 7305 0·88 78·9 73·7 156 0·44 714 Median 0·90 169 7·85 980 2·88 2·50 13·9 1·02 7·05 8·01 1·15 36·9 9·84 100 32·4 137 26·52 224 42·7 7510 0·98 89·2 77·6 170 0·52 723 Q3 1·93 170 11·4 1595 3·53 23·4 71·3 7·78 38·8 14·1 3·27 41·9 12·2 141 49·5 219 41·68 341 64·5 8158 1·29 111 104 188 0·57 757 Element Li Sc Ti Y Nb La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Pb Th U Th/U Tzrc (°C) LZ10-02 - Lower Zone (n = 19) Q1 0·50 411 21·1 377 3·28 0·03 14·9 0·16 2·10 2·61 0·76 10·9 3·08 36·4 12·2 50·7 9·38 75·3 14·7 8340 0·81 230 184 170 1·0 821 Median 1·05 419 22·9 499 3·51 0·04 16·3 0·21 3·08 4·33 1·29 16·8 4·59 49·6 16·3 66·4 11·94 94·1 19·2 8493 1·17 280 220 207 1·1 830 Q3 2·81 431 25·2 589 4·58 0·07 16·9 0·42 5·07 5·19 1·44 19·7 5·38 58·2 19·0 79·1 14·28 116 22·6 8848 1·36 375 310 288 1·1 840 TW477-661 - Lower Critical Zone (n = 18) Q1 0·79 362 13·3 451 2·73 0·00 11·4 0·06 1·27 2·64 0·60 13·4 3·94 42·8 14·5 58·6 10·80 85·1 16·9 9659 0·55 82 69·3 59·3 1·0 773 Median 1·89 375 15·2 697 3·15 0·02 12·7 0·17 2·65 4·45 1·09 17·5 5·08 59·1 21·5 95·6 18·29 147 29·0 10855 0·80 191 149 81·8 3·0 786 Q3 3·42 397 27·3 1420 4·10 0·05 23·8 0·37 5·00 7·09 1·73 36·3 11·0 131 45·8 195 36·01 290 55·9 11284 1·28 539 446 122 4·4 849 DT28-912 - Upper Critical Zone, UG2 (n = 24) Q1 1·34 270 18·3 380 0·90 0·00 3·4 0·05 1·01 2·31 0·36 11·4 3·24 34·7 11·7 49·7 9·45 77·5 15·1 10048 0·33 130 131 5·35 13 806 Median 2·45 281 25·6 647 1·13 0·00 3·8 0·17 2·87 4·29 0·79 18·5 5·40 58·8 20·1 86·6 15·62 125 24·0 10360 0·40 192 205 6·78 20 842 Q3 3·85 287 29·2 771 1·62 0·01 5·5 0·25 4·61 7·07 0·97 26·2 6·77 72·8 23·9 101 18·47 141 27·0 10878 0·84 243 255 18·8 30 858 B00-1-6 - Upper Critical Zone, UG2 (n = 31) Q1 1·08 232 14·2 322 0·73 0·00 3·2 0·04 0·82 1·87 0·33 8·70 2·73 29·6 9·8 43·1 8·22 68·8 13·0 9125 0·26 118 101 3·54 8·0 779 Median 1·49 240 20·5 505 0·86 0·00 3·8 0·07 1·70 4·36 0·58 15·6 4·36 48·8 15·9 63·1 11·90 95·8 17·8 9550 0·34 152 135 4·48 25 818 Q3 3·98 246 23·0 654 1·47 0·00 5·9 0·18 3·12 6·04 0·83 21·7 5·91 63·3 20·3 83·4 15·45 124 24·1 10110 0·71 219 195 25 33 830 B90-7 - Upper Critical Zone, Merensky Reef (n = 20) Q1 0·25 389 22·5 358 2·23 0·00 2·8 0·03 0·74 1·60 0·29 7·64 2·48 31·0 11·2 51·5 10·62 92·3 18·6 9988 0·29 266 155 10·2 10 828 Median 0·38 395 24·0 620 2·35 0·00 3·2 0·14 2·39 4·32 0·54 17·5 5·17 58·7 19·8 83·4 15·54 128 24·8 10510 0·38 332 199 13·5 11 835 Q3 0·69 410 26·4 724 2·57 0·01 3·8 0·25 3·98 5·60 0·75 22·3 6·11 68·2 23·2 98·2 18·24 150 29·1 11270 0·47 518 252 18·8 19 846 SA04-08 Upper Critical Zone, Merensky Reef (n = 25) Q1 0·41 239 17·7 229 1·16 0·00 3·2 0·02 0·45 0·84 0·19 5·3 1·60 20·6 7·4 33·6 6·34 53·3 10·7 7212 0·38 106 99·0 101 0·84 802 Median 0·63 249 20·4 272 1·34 0·00 3·9 0·03 0·65 1·16 0·26 6·1 1·89 23·4 8·7 39·3 7·67 65·0 13·1 7455 0·48 186 149 114 1·0 817 Q3 0·98 256 21·9 448 1·58 0·03 4·8 0·14 2·21 2·83 0·46 12·9 3·53 42·1 14·7 62·1 11·54 93·3 18·1 7853 0·69 243 187 155 1·3 825 SA04-13 Upper Critical Zone, Merensky Reef (n = 22) Q1 0·17 227 17·4 329 1·14 0·00 3·7 0·03 0·53 1·52 0·23 8·02 2·51 30·9 10·5 43·5 8·05 65·1 12·0 8752 0·54 99·5 77·4 30·2 2·0 801 Median 0·27 233 20·8 406 1·38 0·00 5·9 0·06 1·01 2·19 0·34 10·7 3·25 37·7 12·8 53·0 9·96 82·3 15·2 8995 0·64 149 142 44·0 2·8 820 Q3 0·72 237 24·9 626 1·69 0·01 6·6 0·11 1·44 3·90 0·58 19·2 5·64 63·3 20·2 82·5 14·71 115 21·4 9741 0·77 195 168 56·5 3·7 839 MP24D2 - Upper Critical Zone, Bastard Reef (n = 15) Q1 0·45 281 22·9 314 1·36 0·00 5·4 0·08 1·26 2·26 0·50 11·3 3·01 32·4 10·5 41·5 7·74 57·9 11·0 8950 0·50 423 105 26·5 3·1 830 Median 1·11 290 33·8 348 1·69 0·01 7·2 0·13 2·30 3·12 0·65 12·6 3·32 36·0 11·2 45·1 8·14 65·0 12·1 9565 0·86 1507 147 44·6 3·6 875 Q3 1·56 295 46·8 530 2·62 0·01 7·8 0·23 3·57 4·73 0·96 19·1 5·09 54·2 17·2 68·3 12·27 94·3 17·7 10088 0·90 3158 155 47·9 6·9 915 B90-1 - Main Zone, Tennis Ball Marker (n = 20) Q1 1·05 333 24·9 981 2·78 0·09 21·3 0·36 4·38 6·00 0·71 25·0 7·76 90·6 31·6 142 27·93 236 45·8 9195 0·72 369 170 190 0·84 839 Median 2·73 346 28·9 1400 3·08 0·21 24·2 0·72 8·75 8·10 0·99 35·3 10·8 133 46·0 201 39·65 332 62·4 9611 1·11 622 227 263 1·0 856 Q3 30·5 351 35·2 1725 4·61 1·89 39·5 1·38 10·5 12·4 1·72 49·0 13·7 165 56·5 247 45·92 369 71·2 10418 1·62 935 475 643 1·1 879 B07-040 - Upper Zone (n = 32) Q1 0·60 235 5·90 470 0·39 0·00 2·6 0·02 0·37 1·14 0·22 7·40 2·74 37·5 15·1 74·9 15·95 140 27·0 8835 0·17 137 56·6 132 0·42 698 Median 1·58 239 6·95 767 0·43 0·00 3·1 0·05 1·17 2·82 0·54 16·2 5·17 66·6 24·8 118 23·55 200 38·9 9375 0·19 164 78·8 163 0·50 713 Q3 2·69 242 8·40 1089 0·51 0·01 3·3 0·10 2·13 4·50 0·86 25·1 7·68 97·0 36·2 165 32·82 273 50·9 9644 0·22 225 105 189 0·57 730 B07-051 - Felsic Roof Rocks, microgranite (n = 15) Q1 1·07 169 7·80 890 2·09 0·08 7·1 0·25 2·58 4·65 0·83 22·4 6·87 83·5 30·0 136 26·75 227 43·3 6835 0·81 89·7 88·2 159 0·56 723 Median 2·29 170 8·50 1140 2·29 0·39 10·4 0·38 5·33 7·13 1·14 30·2 9·25 111 37·0 161 31·40 257 48·9 6954 0·88 101 96·1 165 0·57 731 Q3 3·58 171 10·7 1572 3·09 7·45 26·5 2·55 13·6 8·78 1·26 37·1 11·3 138 48·4 214 40·20 330 62·0 7560 1·11 112 101 179 0·60 751 B10-054 - Stavoren Granophyre (n = 16) Q1 0·64 192 10·3 1130 2·41 0·06 7·3 0·21 3·22 5·32 0·91 25·7 8·47 107 38·2 174 33·53 284 54·2 7646 1·01 79·8 84·2 153 0·54 749 Median 1·16 194 10·9 1274 2·48 0·59 9·5 0·43 5·70 6·69 1·02 33·0 9·74 120 42·6 190 36·60 302 58·4 7750 1·04 98·3 92·9 165 0·67 754 Q3 2·57 195 11·3 1479 2·86 3·68 19·0 1·43 8·88 9·45 1·26 40·3 12·3 145 50·0 216 40·70 336 64·0 7811 1·13 133 109 185 0·59 758 B10-056 - Nebo Granite (n = 20) Q1 0·54 169 7·05 728 2·16 0·22 8·6 0·41 4·58 4·64 0·78 15·5 5·25 67·7 25·1 117 23·35 199 38·3 7305 0·88 78·9 73·7 156 0·44 714 Median 0·90 169 7·85 980 2·88 2·50 13·9 1·02 7·05 8·01 1·15 36·9 9·84 100 32·4 137 26·52 224 42·7 7510 0·98 89·2 77·6 170 0·52 723 Q3 1·93 170 11·4 1595 3·53 23·4 71·3 7·78 38·8 14·1 3·27 41·9 12·2 141 49·5 219 41·68 341 64·5 8158 1·29 111 104 188 0·57 757 All concentrations reported in parts per million (ppm). Raw LA-ICP-MS data reduced using Iolite 2.5 trace elements reduction scheme with Zr concentration determined by EPMA as internal standard. Q1 = quartile 1, Q3 = quartile 3. Q1-Q3 represents the range of the middle 75% of analyses for each sample. Tzrc (°C) is the Ti-in-zircon temperature calculated using the method of Ferry and Watson (2007). Table 3 Summary of ranges of trace element concentrations in zircon from the Bushveld Complex Element Li Sc Ti Y Nb La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Pb Th U Th/U Tzrc (°C) LZ10-02 - Lower Zone (n = 19) Q1 0·50 411 21·1 377 3·28 0·03 14·9 0·16 2·10 2·61 0·76 10·9 3·08 36·4 12·2 50·7 9·38 75·3 14·7 8340 0·81 230 184 170 1·0 821 Median 1·05 419 22·9 499 3·51 0·04 16·3 0·21 3·08 4·33 1·29 16·8 4·59 49·6 16·3 66·4 11·94 94·1 19·2 8493 1·17 280 220 207 1·1 830 Q3 2·81 431 25·2 589 4·58 0·07 16·9 0·42 5·07 5·19 1·44 19·7 5·38 58·2 19·0 79·1 14·28 116 22·6 8848 1·36 375 310 288 1·1 840 TW477-661 - Lower Critical Zone (n = 18) Q1 0·79 362 13·3 451 2·73 0·00 11·4 0·06 1·27 2·64 0·60 13·4 3·94 42·8 14·5 58·6 10·80 85·1 16·9 9659 0·55 82 69·3 59·3 1·0 773 Median 1·89 375 15·2 697 3·15 0·02 12·7 0·17 2·65 4·45 1·09 17·5 5·08 59·1 21·5 95·6 18·29 147 29·0 10855 0·80 191 149 81·8 3·0 786 Q3 3·42 397 27·3 1420 4·10 0·05 23·8 0·37 5·00 7·09 1·73 36·3 11·0 131 45·8 195 36·01 290 55·9 11284 1·28 539 446 122 4·4 849 DT28-912 - Upper Critical Zone, UG2 (n = 24) Q1 1·34 270 18·3 380 0·90 0·00 3·4 0·05 1·01 2·31 0·36 11·4 3·24 34·7 11·7 49·7 9·45 77·5 15·1 10048 0·33 130 131 5·35 13 806 Median 2·45 281 25·6 647 1·13 0·00 3·8 0·17 2·87 4·29 0·79 18·5 5·40 58·8 20·1 86·6 15·62 125 24·0 10360 0·40 192 205 6·78 20 842 Q3 3·85 287 29·2 771 1·62 0·01 5·5 0·25 4·61 7·07 0·97 26·2 6·77 72·8 23·9 101 18·47 141 27·0 10878 0·84 243 255 18·8 30 858 B00-1-6 - Upper Critical Zone, UG2 (n = 31) Q1 1·08 232 14·2 322 0·73 0·00 3·2 0·04 0·82 1·87 0·33 8·70 2·73 29·6 9·8 43·1 8·22 68·8 13·0 9125 0·26 118 101 3·54 8·0 779 Median 1·49 240 20·5 505 0·86 0·00 3·8 0·07 1·70 4·36 0·58 15·6 4·36 48·8 15·9 63·1 11·90 95·8 17·8 9550 0·34 152 135 4·48 25 818 Q3 3·98 246 23·0 654 1·47 0·00 5·9 0·18 3·12 6·04 0·83 21·7 5·91 63·3 20·3 83·4 15·45 124 24·1 10110 0·71 219 195 25 33 830 B90-7 - Upper Critical Zone, Merensky Reef (n = 20) Q1 0·25 389 22·5 358 2·23 0·00 2·8 0·03 0·74 1·60 0·29 7·64 2·48 31·0 11·2 51·5 10·62 92·3 18·6 9988 0·29 266 155 10·2 10 828 Median 0·38 395 24·0 620 2·35 0·00 3·2 0·14 2·39 4·32 0·54 17·5 5·17 58·7 19·8 83·4 15·54 128 24·8 10510 0·38 332 199 13·5 11 835 Q3 0·69 410 26·4 724 2·57 0·01 3·8 0·25 3·98 5·60 0·75 22·3 6·11 68·2 23·2 98·2 18·24 150 29·1 11270 0·47 518 252 18·8 19 846 SA04-08 Upper Critical Zone, Merensky Reef (n = 25) Q1 0·41 239 17·7 229 1·16 0·00 3·2 0·02 0·45 0·84 0·19 5·3 1·60 20·6 7·4 33·6 6·34 53·3 10·7 7212 0·38 106 99·0 101 0·84 802 Median 0·63 249 20·4 272 1·34 0·00 3·9 0·03 0·65 1·16 0·26 6·1 1·89 23·4 8·7 39·3 7·67 65·0 13·1 7455 0·48 186 149 114 1·0 817 Q3 0·98 256 21·9 448 1·58 0·03 4·8 0·14 2·21 2·83 0·46 12·9 3·53 42·1 14·7 62·1 11·54 93·3 18·1 7853 0·69 243 187 155 1·3 825 SA04-13 Upper Critical Zone, Merensky Reef (n = 22) Q1 0·17 227 17·4 329 1·14 0·00 3·7 0·03 0·53 1·52 0·23 8·02 2·51 30·9 10·5 43·5 8·05 65·1 12·0 8752 0·54 99·5 77·4 30·2 2·0 801 Median 0·27 233 20·8 406 1·38 0·00 5·9 0·06 1·01 2·19 0·34 10·7 3·25 37·7 12·8 53·0 9·96 82·3 15·2 8995 0·64 149 142 44·0 2·8 820 Q3 0·72 237 24·9 626 1·69 0·01 6·6 0·11 1·44 3·90 0·58 19·2 5·64 63·3 20·2 82·5 14·71 115 21·4 9741 0·77 195 168 56·5 3·7 839 MP24D2 - Upper Critical Zone, Bastard Reef (n = 15) Q1 0·45 281 22·9 314 1·36 0·00 5·4 0·08 1·26 2·26 0·50 11·3 3·01 32·4 10·5 41·5 7·74 57·9 11·0 8950 0·50 423 105 26·5 3·1 830 Median 1·11 290 33·8 348 1·69 0·01 7·2 0·13 2·30 3·12 0·65 12·6 3·32 36·0 11·2 45·1 8·14 65·0 12·1 9565 0·86 1507 147 44·6 3·6 875 Q3 1·56 295 46·8 530 2·62 0·01 7·8 0·23 3·57 4·73 0·96 19·1 5·09 54·2 17·2 68·3 12·27 94·3 17·7 10088 0·90 3158 155 47·9 6·9 915 B90-1 - Main Zone, Tennis Ball Marker (n = 20) Q1 1·05 333 24·9 981 2·78 0·09 21·3 0·36 4·38 6·00 0·71 25·0 7·76 90·6 31·6 142 27·93 236 45·8 9195 0·72 369 170 190 0·84 839 Median 2·73 346 28·9 1400 3·08 0·21 24·2 0·72 8·75 8·10 0·99 35·3 10·8 133 46·0 201 39·65 332 62·4 9611 1·11 622 227 263 1·0 856 Q3 30·5 351 35·2 1725 4·61 1·89 39·5 1·38 10·5 12·4 1·72 49·0 13·7 165 56·5 247 45·92 369 71·2 10418 1·62 935 475 643 1·1 879 B07-040 - Upper Zone (n = 32) Q1 0·60 235 5·90 470 0·39 0·00 2·6 0·02 0·37 1·14 0·22 7·40 2·74 37·5 15·1 74·9 15·95 140 27·0 8835 0·17 137 56·6 132 0·42 698 Median 1·58 239 6·95 767 0·43 0·00 3·1 0·05 1·17 2·82 0·54 16·2 5·17 66·6 24·8 118 23·55 200 38·9 9375 0·19 164 78·8 163 0·50 713 Q3 2·69 242 8·40 1089 0·51 0·01 3·3 0·10 2·13 4·50 0·86 25·1 7·68 97·0 36·2 165 32·82 273 50·9 9644 0·22 225 105 189 0·57 730 B07-051 - Felsic Roof Rocks, microgranite (n = 15) Q1 1·07 169 7·80 890 2·09 0·08 7·1 0·25 2·58 4·65 0·83 22·4 6·87 83·5 30·0 136 26·75 227 43·3 6835 0·81 89·7 88·2 159 0·56 723 Median 2·29 170 8·50 1140 2·29 0·39 10·4 0·38 5·33 7·13 1·14 30·2 9·25 111 37·0 161 31·40 257 48·9 6954 0·88 101 96·1 165 0·57 731 Q3 3·58 171 10·7 1572 3·09 7·45 26·5 2·55 13·6 8·78 1·26 37·1 11·3 138 48·4 214 40·20 330 62·0 7560 1·11 112 101 179 0·60 751 B10-054 - Stavoren Granophyre (n = 16) Q1 0·64 192 10·3 1130 2·41 0·06 7·3 0·21 3·22 5·32 0·91 25·7 8·47 107 38·2 174 33·53 284 54·2 7646 1·01 79·8 84·2 153 0·54 749 Median 1·16 194 10·9 1274 2·48 0·59 9·5 0·43 5·70 6·69 1·02 33·0 9·74 120 42·6 190 36·60 302 58·4 7750 1·04 98·3 92·9 165 0·67 754 Q3 2·57 195 11·3 1479 2·86 3·68 19·0 1·43 8·88 9·45 1·26 40·3 12·3 145 50·0 216 40·70 336 64·0 7811 1·13 133 109 185 0·59 758 B10-056 - Nebo Granite (n = 20) Q1 0·54 169 7·05 728 2·16 0·22 8·6 0·41 4·58 4·64 0·78 15·5 5·25 67·7 25·1 117 23·35 199 38·3 7305 0·88 78·9 73·7 156 0·44 714 Median 0·90 169 7·85 980 2·88 2·50 13·9 1·02 7·05 8·01 1·15 36·9 9·84 100 32·4 137 26·52 224 42·7 7510 0·98 89·2 77·6 170 0·52 723 Q3 1·93 170 11·4 1595 3·53 23·4 71·3 7·78 38·8 14·1 3·27 41·9 12·2 141 49·5 219 41·68 341 64·5 8158 1·29 111 104 188 0·57 757 Element Li Sc Ti Y Nb La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Pb Th U Th/U Tzrc (°C) LZ10-02 - Lower Zone (n = 19) Q1 0·50 411 21·1 377 3·28 0·03 14·9 0·16 2·10 2·61 0·76 10·9 3·08 36·4 12·2 50·7 9·38 75·3 14·7 8340 0·81 230 184 170 1·0 821 Median 1·05 419 22·9 499 3·51 0·04 16·3 0·21 3·08 4·33 1·29 16·8 4·59 49·6 16·3 66·4 11·94 94·1 19·2 8493 1·17 280 220 207 1·1 830 Q3 2·81 431 25·2 589 4·58 0·07 16·9 0·42 5·07 5·19 1·44 19·7 5·38 58·2 19·0 79·1 14·28 116 22·6 8848 1·36 375 310 288 1·1 840 TW477-661 - Lower Critical Zone (n = 18) Q1 0·79 362 13·3 451 2·73 0·00 11·4 0·06 1·27 2·64 0·60 13·4 3·94 42·8 14·5 58·6 10·80 85·1 16·9 9659 0·55 82 69·3 59·3 1·0 773 Median 1·89 375 15·2 697 3·15 0·02 12·7 0·17 2·65 4·45 1·09 17·5 5·08 59·1 21·5 95·6 18·29 147 29·0 10855 0·80 191 149 81·8 3·0 786 Q3 3·42 397 27·3 1420 4·10 0·05 23·8 0·37 5·00 7·09 1·73 36·3 11·0 131 45·8 195 36·01 290 55·9 11284 1·28 539 446 122 4·4 849 DT28-912 - Upper Critical Zone, UG2 (n = 24) Q1 1·34 270 18·3 380 0·90 0·00 3·4 0·05 1·01 2·31 0·36 11·4 3·24 34·7 11·7 49·7 9·45 77·5 15·1 10048 0·33 130 131 5·35 13 806 Median 2·45 281 25·6 647 1·13 0·00 3·8 0·17 2·87 4·29 0·79 18·5 5·40 58·8 20·1 86·6 15·62 125 24·0 10360 0·40 192 205 6·78 20 842 Q3 3·85 287 29·2 771 1·62 0·01 5·5 0·25 4·61 7·07 0·97 26·2 6·77 72·8 23·9 101 18·47 141 27·0 10878 0·84 243 255 18·8 30 858 B00-1-6 - Upper Critical Zone, UG2 (n = 31) Q1 1·08 232 14·2 322 0·73 0·00 3·2 0·04 0·82 1·87 0·33 8·70 2·73 29·6 9·8 43·1 8·22 68·8 13·0 9125 0·26 118 101 3·54 8·0 779 Median 1·49 240 20·5 505 0·86 0·00 3·8 0·07 1·70 4·36 0·58 15·6 4·36 48·8 15·9 63·1 11·90 95·8 17·8 9550 0·34 152 135 4·48 25 818 Q3 3·98 246 23·0 654 1·47 0·00 5·9 0·18 3·12 6·04 0·83 21·7 5·91 63·3 20·3 83·4 15·45 124 24·1 10110 0·71 219 195 25 33 830 B90-7 - Upper Critical Zone, Merensky Reef (n = 20) Q1 0·25 389 22·5 358 2·23 0·00 2·8 0·03 0·74 1·60 0·29 7·64 2·48 31·0 11·2 51·5 10·62 92·3 18·6 9988 0·29 266 155 10·2 10 828 Median 0·38 395 24·0 620 2·35 0·00 3·2 0·14 2·39 4·32 0·54 17·5 5·17 58·7 19·8 83·4 15·54 128 24·8 10510 0·38 332 199 13·5 11 835 Q3 0·69 410 26·4 724 2·57 0·01 3·8 0·25 3·98 5·60 0·75 22·3 6·11 68·2 23·2 98·2 18·24 150 29·1 11270 0·47 518 252 18·8 19 846 SA04-08 Upper Critical Zone, Merensky Reef (n = 25) Q1 0·41 239 17·7 229 1·16 0·00 3·2 0·02 0·45 0·84 0·19 5·3 1·60 20·6 7·4 33·6 6·34 53·3 10·7 7212 0·38 106 99·0 101 0·84 802 Median 0·63 249 20·4 272 1·34 0·00 3·9 0·03 0·65 1·16 0·26 6·1 1·89 23·4 8·7 39·3 7·67 65·0 13·1 7455 0·48 186 149 114 1·0 817 Q3 0·98 256 21·9 448 1·58 0·03 4·8 0·14 2·21 2·83 0·46 12·9 3·53 42·1 14·7 62·1 11·54 93·3 18·1 7853 0·69 243 187 155 1·3 825 SA04-13 Upper Critical Zone, Merensky Reef (n = 22) Q1 0·17 227 17·4 329 1·14 0·00 3·7 0·03 0·53 1·52 0·23 8·02 2·51 30·9 10·5 43·5 8·05 65·1 12·0 8752 0·54 99·5 77·4 30·2 2·0 801 Median 0·27 233 20·8 406 1·38 0·00 5·9 0·06 1·01 2·19 0·34 10·7 3·25 37·7 12·8 53·0 9·96 82·3 15·2 8995 0·64 149 142 44·0 2·8 820 Q3 0·72 237 24·9 626 1·69 0·01 6·6 0·11 1·44 3·90 0·58 19·2 5·64 63·3 20·2 82·5 14·71 115 21·4 9741 0·77 195 168 56·5 3·7 839 MP24D2 - Upper Critical Zone, Bastard Reef (n = 15) Q1 0·45 281 22·9 314 1·36 0·00 5·4 0·08 1·26 2·26 0·50 11·3 3·01 32·4 10·5 41·5 7·74 57·9 11·0 8950 0·50 423 105 26·5 3·1 830 Median 1·11 290 33·8 348 1·69 0·01 7·2 0·13 2·30 3·12 0·65 12·6 3·32 36·0 11·2 45·1 8·14 65·0 12·1 9565 0·86 1507 147 44·6 3·6 875 Q3 1·56 295 46·8 530 2·62 0·01 7·8 0·23 3·57 4·73 0·96 19·1 5·09 54·2 17·2 68·3 12·27 94·3 17·7 10088 0·90 3158 155 47·9 6·9 915 B90-1 - Main Zone, Tennis Ball Marker (n = 20) Q1 1·05 333 24·9 981 2·78 0·09 21·3 0·36 4·38 6·00 0·71 25·0 7·76 90·6 31·6 142 27·93 236 45·8 9195 0·72 369 170 190 0·84 839 Median 2·73 346 28·9 1400 3·08 0·21 24·2 0·72 8·75 8·10 0·99 35·3 10·8 133 46·0 201 39·65 332 62·4 9611 1·11 622 227 263 1·0 856 Q3 30·5 351 35·2 1725 4·61 1·89 39·5 1·38 10·5 12·4 1·72 49·0 13·7 165 56·5 247 45·92 369 71·2 10418 1·62 935 475 643 1·1 879 B07-040 - Upper Zone (n = 32) Q1 0·60 235 5·90 470 0·39 0·00 2·6 0·02 0·37 1·14 0·22 7·40 2·74 37·5 15·1 74·9 15·95 140 27·0 8835 0·17 137 56·6 132 0·42 698 Median 1·58 239 6·95 767 0·43 0·00 3·1 0·05 1·17 2·82 0·54 16·2 5·17 66·6 24·8 118 23·55 200 38·9 9375 0·19 164 78·8 163 0·50 713 Q3 2·69 242 8·40 1089 0·51 0·01 3·3 0·10 2·13 4·50 0·86 25·1 7·68 97·0 36·2 165 32·82 273 50·9 9644 0·22 225 105 189 0·57 730 B07-051 - Felsic Roof Rocks, microgranite (n = 15) Q1 1·07 169 7·80 890 2·09 0·08 7·1 0·25 2·58 4·65 0·83 22·4 6·87 83·5 30·0 136 26·75 227 43·3 6835 0·81 89·7 88·2 159 0·56 723 Median 2·29 170 8·50 1140 2·29 0·39 10·4 0·38 5·33 7·13 1·14 30·2 9·25 111 37·0 161 31·40 257 48·9 6954 0·88 101 96·1 165 0·57 731 Q3 3·58 171 10·7 1572 3·09 7·45 26·5 2·55 13·6 8·78 1·26 37·1 11·3 138 48·4 214 40·20 330 62·0 7560 1·11 112 101 179 0·60 751 B10-054 - Stavoren Granophyre (n = 16) Q1 0·64 192 10·3 1130 2·41 0·06 7·3 0·21 3·22 5·32 0·91 25·7 8·47 107 38·2 174 33·53 284 54·2 7646 1·01 79·8 84·2 153 0·54 749 Median 1·16 194 10·9 1274 2·48 0·59 9·5 0·43 5·70 6·69 1·02 33·0 9·74 120 42·6 190 36·60 302 58·4 7750 1·04 98·3 92·9 165 0·67 754 Q3 2·57 195 11·3 1479 2·86 3·68 19·0 1·43 8·88 9·45 1·26 40·3 12·3 145 50·0 216 40·70 336 64·0 7811 1·13 133 109 185 0·59 758 B10-056 - Nebo Granite (n = 20) Q1 0·54 169 7·05 728 2·16 0·22 8·6 0·41 4·58 4·64 0·78 15·5 5·25 67·7 25·1 117 23·35 199 38·3 7305 0·88 78·9 73·7 156 0·44 714 Median 0·90 169 7·85 980 2·88 2·50 13·9 1·02 7·05 8·01 1·15 36·9 9·84 100 32·4 137 26·52 224 42·7 7510 0·98 89·2 77·6 170 0·52 723 Q3 1·93 170 11·4 1595 3·53 23·4 71·3 7·78 38·8 14·1 3·27 41·9 12·2 141 49·5 219 41·68 341 64·5 8158 1·29 111 104 188 0·57 757 All concentrations reported in parts per million (ppm). Raw LA-ICP-MS data reduced using Iolite 2.5 trace elements reduction scheme with Zr concentration determined by EPMA as internal standard. Q1 = quartile 1, Q3 = quartile 3. Q1-Q3 represents the range of the middle 75% of analyses for each sample. Tzrc (°C) is the Ti-in-zircon temperature calculated using the method of Ferry and Watson (2007). Fig. 9 View largeDownload slide Trace element variations of selected high field strength elements and actinides in zircon from the Bushveld Complex. (a) Ti vs Hf showing negative trends of decreasing Ti with increasing Hf concentration. Horizontal dashed lines in (a) and (b) are Ti-in-zircon temperatures calculated using aSiO2 = 1 and aTiO2 = 1 (Ferry & Watson, 2007), except for samples from the Upper Zone and roof granites where aTiO2 = 0·7 (see text for additional details). (b) Ti vs U. (c) Th vs U with lines of constant Th/U indicated. Three populations are defined: Upper Zone and roof rocks with Th/U ∼0·5, Lower Zone and Main Zone rocks with Th/U ∼1, and highly variable Th/U (2 to >40) of Critical Zone samples. (d) Th/U vs Hf showing the large variations in Th/U of the Critical Zone samples and variable Hf contents. Note logarithmic y-axis. Average 2σ uncertainty is indicated in each panel. Symbols: square, lherzolite/norite; diamond, orthopyroxenite; triangle, diorite; circle, granite/granophyre. Abbreviations in legend: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ-MR, Upper Critical Zone-Merensky Reef; UCZ, Upper Critical Zone; MZ, Main Zone; UZ, Upper Zone. Fig. 9 View largeDownload slide Trace element variations of selected high field strength elements and actinides in zircon from the Bushveld Complex. (a) Ti vs Hf showing negative trends of decreasing Ti with increasing Hf concentration. Horizontal dashed lines in (a) and (b) are Ti-in-zircon temperatures calculated using aSiO2 = 1 and aTiO2 = 1 (Ferry & Watson, 2007), except for samples from the Upper Zone and roof granites where aTiO2 = 0·7 (see text for additional details). (b) Ti vs U. (c) Th vs U with lines of constant Th/U indicated. Three populations are defined: Upper Zone and roof rocks with Th/U ∼0·5, Lower Zone and Main Zone rocks with Th/U ∼1, and highly variable Th/U (2 to >40) of Critical Zone samples. (d) Th/U vs Hf showing the large variations in Th/U of the Critical Zone samples and variable Hf contents. Note logarithmic y-axis. Average 2σ uncertainty is indicated in each panel. Symbols: square, lherzolite/norite; diamond, orthopyroxenite; triangle, diorite; circle, granite/granophyre. Abbreviations in legend: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ-MR, Upper Critical Zone-Merensky Reef; UCZ, Upper Critical Zone; MZ, Main Zone; UZ, Upper Zone. Fig. 10 View largeDownload slide Spatial variations of Hf, Ti, U, and Th/U within representative zircon grains from the Rustenburg Layered Suite of the Bushveld Complex. (a) LZ10–02 z15, Lower Zone. (b) TW477–661 z6, Lower Critical Zone. (c) DT28–912 z13, Upper Critical Zone – UG2. (d) B90–7 z7, Upper Critical Zone – Merensky Reef. (e) B90–1 z3, Main Zone. (f) B07–040 z8, Upper Zone. Each panel contains a SEM-CL image with yellow circles indicating locations of spot analyses by LA-ICP-MS (circle diameter = 34 μm). Each spot number corresponds to the analysis number in Supplementary Data Table S2. Above each image is a plot of U (green) and Th/U (red) and below each image is a plot of Hf (blue) and Ti (orange) concentration that corresponds to each analysis on the CL image in order from left to right. The variation in U concentration is not strongly correlated with variations in CL brightness; U concentrations decrease systematically from core to rim in the grain from the Upper Zone sample (B07–040). There are no systematic variations in Th/U from the core to the margins of individual grains for the high-Th/U zircon grains (b, c, d). Concentrations of Hf and Ti in most grains show a negative correlation, with decreasing Ti for increasing Hf. Variability within each grain is typically not spatially systematic or associated with observable variations in CL zoning. The exception is zircon from the Upper Zone sample (f). The 2σ analytical uncertainties are indicated for each analysis when larger than the symbol size. Fig. 10 View largeDownload slide Spatial variations of Hf, Ti, U, and Th/U within representative zircon grains from the Rustenburg Layered Suite of the Bushveld Complex. (a) LZ10–02 z15, Lower Zone. (b) TW477–661 z6, Lower Critical Zone. (c) DT28–912 z13, Upper Critical Zone – UG2. (d) B90–7 z7, Upper Critical Zone – Merensky Reef. (e) B90–1 z3, Main Zone. (f) B07–040 z8, Upper Zone. Each panel contains a SEM-CL image with yellow circles indicating locations of spot analyses by LA-ICP-MS (circle diameter = 34 μm). Each spot number corresponds to the analysis number in Supplementary Data Table S2. Above each image is a plot of U (green) and Th/U (red) and below each image is a plot of Hf (blue) and Ti (orange) concentration that corresponds to each analysis on the CL image in order from left to right. The variation in U concentration is not strongly correlated with variations in CL brightness; U concentrations decrease systematically from core to rim in the grain from the Upper Zone sample (B07–040). There are no systematic variations in Th/U from the core to the margins of individual grains for the high-Th/U zircon grains (b, c, d). Concentrations of Hf and Ti in most grains show a negative correlation, with decreasing Ti for increasing Hf. Variability within each grain is typically not spatially systematic or associated with observable variations in CL zoning. The exception is zircon from the Upper Zone sample (f). The 2σ analytical uncertainties are indicated for each analysis when larger than the symbol size. Fig. 11 View largeDownload slide Chondrite-normalized rare earth element patterns of zircon from the Bushveld Complex. Lines and symbols represent median values and shaded areas indicate the range of all analyses within a sample. All panels are at the same scale. (a) Lower Zone; (b) Lower Critical Zone and Bastard Reef (Upper Critical Zone); (c) UG2 (Upper Critical Zone); (d) Merensky Reef (Upper Critical Zone); (e) Main Zone and Upper Zone; (f) Felsic roof rocks. Normalization values from McDonough & Sun (1995). Symbols: square, lherzolite/norite; diamond, orthopyroxenite; triangle, diorite; circle, granite/granophyre. Abbreviations: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ, Upper Critical Zone; MZ, Main Zone; UZ, Upper Zone. Fig. 11 View largeDownload slide Chondrite-normalized rare earth element patterns of zircon from the Bushveld Complex. Lines and symbols represent median values and shaded areas indicate the range of all analyses within a sample. All panels are at the same scale. (a) Lower Zone; (b) Lower Critical Zone and Bastard Reef (Upper Critical Zone); (c) UG2 (Upper Critical Zone); (d) Merensky Reef (Upper Critical Zone); (e) Main Zone and Upper Zone; (f) Felsic roof rocks. Normalization values from McDonough & Sun (1995). Symbols: square, lherzolite/norite; diamond, orthopyroxenite; triangle, diorite; circle, granite/granophyre. Abbreviations: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ, Upper Critical Zone; MZ, Main Zone; UZ, Upper Zone. Hafnium and titanium Hf concentrations in zircon from the Bushveld mafic–ultramafic cumulates span a relatively wide range from 6903–13 350 ppm, with zircon from the felsic roof rocks showing less variation (6768–9690 ppm) (Fig. 9). Titanium (4–61 ppm) is negatively correlated with Hf in nearly all analysed samples and three general Ti–Hf trends distinguish zircon from the Bushveld Complex (Fig. 9a). Zircon from the Lower Zone and Main Zone is characterized by moderate Ti variations (10·1–52 ppm) over a wide range of Hf (7900–12 860 ppm). Zircon from the Critical Zone has the highest and most variable Ti concentrations (4–61 ppm). Zircon from the Upper Zone has low Ti concentrations that overlap the Ti concentrations from felsic roof rock samples (Fig. 9a and b). Multiple analyses of individual grains (Figs 7, 10) reveal a range of zoning patterns, varying from no apparent differences in Ti–Hf between cores and rims (Figs 7c, d, 10b, e) to prominent normal zonation (increasing Hf: Fig. 7g, i) and reversed zonation (decreasing Hf: Fig. 10f). Thorium and uranium There are three distinct populations of zircon from the Bushveld Complex based on Th–U concentrations (Fig. 9c and d). Zircon from the Lower Zone and Main Zone contain comparable Th (98–1092 ppm) and U (104–1580 ppm) with Th/U = 0·15–1·5. Zircon from the Upper Zone and felsic roof rock samples has lower Th (28–201 ppm) and U (80–351 ppm) concentrations and relatively consistent Th/U (0·35–0·74). Zircon from the Critical Zone, including the UG2 and Merensky Reef horizons, is characterized by Th (280–1002 ppm) that is comparable with zircon from other samples. However, these zircon grains are characterized by anomalously low U concentrations (2·3–457 ppm), yielding elevated and highly variable Th/U (0·60–77) (Fig. 9c), with the exception of zircon from sample SA04–08 (Merensky Reef, Western Limb). Zircon with similarly high Th/U (2–20) also occurs in pyroxenites from the Ultramafic and Basal series of the Stillwater Complex (Wall et al., 2018). Hf concentrations of the high-Th/U zircon in the Critical Zone are nominally higher than in other samples; however, there is no apparent correlation between increasing Hf and higher Th/U in the dataset (Fig. 9d). The Th/U variations in rocks of the Critical Zone are significantly wider than those observed by Zeh et al. (2015) and individual grains from each of the Critical Zone samples have distinctive Th/U. Zonation in Th/U from core to rim is not systematic for the high-Th/U zircons, with no significant zonation in some of the analysed grains (Fig. 10b, c) and weak zonation in others (decreasing Th/U: Figs 7b, d, e, 10d). Rare earth elements Chondrite-normalized rare earth element (REE) patterns of zircon from the Bushveld Complex display typical igneous zircon patterns with relatively depleted light REE (LREE) and relatively enriched heavy REE (HREE), sharp positive Ce anomalies, and prominent negative Eu anomalies (Fig. 11). Patterns in zircon from Lower Zone and Critical Zone mafic–ultramafic cumulates are broadly subparallel and display similar ranges within each sample (Fig. 11a–d). The LREE in Main Zone zircon are elevated by 1–2 orders of magnitude compared with zircon from the Lower Zone, Critical Zone, and Upper Zone, while displaying similar Ce and Eu anomalies (Fig. 11e). The cores of zircon grains from the Critical Zone do not show the elevated REE observed in some analyses reported by Yudovskaya et al. (2013). Upper Zone zircon patterns are typical of igneous zircon and are comparable with Critical Zone samples (Fig. 11e). Zircon from roof rock granites and granophyres display strong enrichment in LREE, with La values varying more than five orders of magnitude (0·03–1000 times chondritic values) in the Nebo Granite sample (Fig. 11f). This LREE-enrichment is characteristic of all three roof rock samples and, as a result, many REE patterns are relatively flat. Some analyses from each of the three felsic roof rocks display more typical zircon REE patterns without significant LREE enrichment (Fig. 11f). Anomalies in the chondrite-normalized REE patterns and ratios of the REE distinguish zircon from the different Bushveld Complex zones and rock packages. Europium anomalies in the sample set are mostly below 0·5 (Eu/Eu* = 0·12–0·92) and show negative trends with Hf concentrations for analyses from a single sample (Fig. 12a). In contrast, cerium anomalies are more variable (Ce/Ce* = 0·4–12·6) and trends are not apparent (Fig. 12c). Yb/Dy in zircon defines the slope of the HREE and Lu/Hf is related to the partitioning of the REE (e.g. Lu) relative to high field strength elements (e.g. Hf) (e.g. Samperton et al., 2015). Yb/Dy ranges from 1·6–4·2 in Bushveld zircon, with the lower and less variable values (generally) associated with samples from the Lower Zone and Critical Zone (Fig. 11b, d). Lu/Hf is significantly higher and shows a greater range (Lu/Hf = 21–114 × 104) in zircon from the Main Zone and Upper Zone, and in the felsic roof rocks, compared with (most) zircon from the Lower Zone and Critical Zone (Lu/Hf <40 × 104) (Fig. 9b). Fig. 12 View largeDownload slide Trace element ratio variations in Bushveld Complex zircon. (a) Europium anomaly (Eu/Eu*) vs Hf showing a (general) negative relationship of increasingly negative Eu anomaly with increasing Hf concentration for zircon. (b) Lu/Hf vs Yb/Dy showing two distinct populations. Zircon from the Main Zone, Upper Zone and roof rock samples defines a steeper negative trend over a large range of Lu/Hf, whereas zircon from the Lower Zone and Critical Zone samples has a shallower negative trend and a smaller range of Lu/Hf. (c) Cerium anomaly (Ce/Ce*) vs Hf showing highly variable Ce anomalies over the range of Hf concentrations. (d) Sc/Th vs Yb/Dy. Average 2σ uncertainty is indicated in each panel. Symbols: square, lherzolite/norite; diamond, orthopyroxenite; triangle, diorite; circle, granite/granophyre. Abbreviations: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ, Upper Critical Zone; UCZ-MR, Upper Critical Zone-Merensky Reef; MZ, Main Zone; UZ, Upper Zone. Fig. 12 View largeDownload slide Trace element ratio variations in Bushveld Complex zircon. (a) Europium anomaly (Eu/Eu*) vs Hf showing a (general) negative relationship of increasingly negative Eu anomaly with increasing Hf concentration for zircon. (b) Lu/Hf vs Yb/Dy showing two distinct populations. Zircon from the Main Zone, Upper Zone and roof rock samples defines a steeper negative trend over a large range of Lu/Hf, whereas zircon from the Lower Zone and Critical Zone samples has a shallower negative trend and a smaller range of Lu/Hf. (c) Cerium anomaly (Ce/Ce*) vs Hf showing highly variable Ce anomalies over the range of Hf concentrations. (d) Sc/Th vs Yb/Dy. Average 2σ uncertainty is indicated in each panel. Symbols: square, lherzolite/norite; diamond, orthopyroxenite; triangle, diorite; circle, granite/granophyre. Abbreviations: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ, Upper Critical Zone; UCZ-MR, Upper Critical Zone-Merensky Reef; MZ, Main Zone; UZ, Upper Zone. Trace element geochemistry of rutile A summary of rutile trace element chemistry is provided in Table 4 and all analytical results by LA-ICP-MS are reported in Supplementary Data Electronic Appendix 3. Rutile readily accepts substitution of high field strength elements, including Zr, Hf, Nb, and Ta, up to wt % concentrations (e.g. Zack et al., 2004a; Zack & Luvizottow, 2006; Luvizotto et al., 2009). Hafnium and Zr concentrations in rutile span a wide range from 2–303 ppm and from 43–8230 ppm, respectively (Fig. 13a). The relationship between Nb and Ta distinguishes two distinct geochemical groups, a population of rutile typified by high HFSE (1000–20 000 ppm Nb, 100–1760 ppm Ta) and high Cr (5000–14 000 ppm) and Sc (8–17 ppm), and an anomalous population characterized by exceptionally low HFSE (<1000 ppm Nb, <100 ppm Ta) and low Cr (2000–4000 ppm) and Sc (3–15) (Fig. 11c). Table 4 Summary of ranges of trace element concentrations of rutile from the Bushveld Complex Element Si Sc Cr Fe Y Zr Nb Hf Ta W Pb Th U T (°C) TW477-661 - Lower Critical Zone (n = 14 (5)) Q1 85 5.44 3850 39.6 - 499 72.2 11.4 0.958 17.1 12.0 - 5.73 685 Median 790 8.68 6020 292 - 2446 5595 90.6 326 315 16.0 - 8.63 851 Q3 1040 10 8100 419 - 3053 9670 112 565 455 43.0 0.009 24.4 878 DT28-912 - Upper Critical Zone, UG2 (n = 27 (15)) Q1 400 6.47 2630 12.5 - 327 65.8 14.9 3.44 6.10 1.48 - 0.594 650 Median 890 12.2 7510 800 - 4220 10640 149 806 566 5.14 - 2.44 921 Q3 1095 13.5 7850 2565 0.005 5025 12375 204 927 584 6.47 - 3.03 945 SA04-06 - Upper Critical Zone, Merensky Reef (n = 16 (16)) Q1 470 5.89 2630 3.15 - 319 89.4 25.8 4.58 59.8 1.73 0.0008 0.319 648 Median 905 8.36 2875 23.6 0.001 474 232 27.1 8.99 136 4.19 0.0084 0.935 682 Q3 1122 10.73 3160 72.8 0.029 1016 424 42.3 16.8 235 7.13 0.0633 3.75 751 SA04-08 - Upper Critical Zone, Merensky Reef (n = 24 (12)) Q1 252 7.69 5050 16.9 - 2588 6348 91.6 261 267 49.3 - 27.5 857 Median 565 11.4 6610 34.8 - 3309 8200 112 465. 397 66.5 - 35.1 888 Q3 1027 13.4 7875 58.9 0.006 4400 10728 155 705 481 88.5 0.023 46.8 927 SA04-13 - Upper Critical Zone, Merensky Reef (n = 30 (18)) Q1 507 6.02 2418 11.5 - 173 31.3 7.75 0.999 0.108 0.260 - 0.121 601 Median 800 6.59 2510 15.8 - 220 45.4 9.65 2.61 0.440 0.720 - 0.370 619 Q3 1042 7.42 2653 20.3 - 359 91.5 14.5 5.88 3.65 1.10 - 0.667 658 MP24D2 - Upper Critical Zone, Bastard Reef (n = 4 (0)) Q1 340 9.78 NA 81.8 - 3569 9360 132 778 444 4.89 - 2.55 898 Median 485 10.3 NA 93 0.15 4490 10625 165 823 450 8.56 0.125 4.27 928 Q3 910 12.9 NA 575 0.44 5385 15640 207 1072 606 16.7 0.490 8.86 955 Element Si Sc Cr Fe Y Zr Nb Hf Ta W Pb Th U T (°C) TW477-661 - Lower Critical Zone (n = 14 (5)) Q1 85 5.44 3850 39.6 - 499 72.2 11.4 0.958 17.1 12.0 - 5.73 685 Median 790 8.68 6020 292 - 2446 5595 90.6 326 315 16.0 - 8.63 851 Q3 1040 10 8100 419 - 3053 9670 112 565 455 43.0 0.009 24.4 878 DT28-912 - Upper Critical Zone, UG2 (n = 27 (15)) Q1 400 6.47 2630 12.5 - 327 65.8 14.9 3.44 6.10 1.48 - 0.594 650 Median 890 12.2 7510 800 - 4220 10640 149 806 566 5.14 - 2.44 921 Q3 1095 13.5 7850 2565 0.005 5025 12375 204 927 584 6.47 - 3.03 945 SA04-06 - Upper Critical Zone, Merensky Reef (n = 16 (16)) Q1 470 5.89 2630 3.15 - 319 89.4 25.8 4.58 59.8 1.73 0.0008 0.319 648 Median 905 8.36 2875 23.6 0.001 474 232 27.1 8.99 136 4.19 0.0084 0.935 682 Q3 1122 10.73 3160 72.8 0.029 1016 424 42.3 16.8 235 7.13 0.0633 3.75 751 SA04-08 - Upper Critical Zone, Merensky Reef (n = 24 (12)) Q1 252 7.69 5050 16.9 - 2588 6348 91.6 261 267 49.3 - 27.5 857 Median 565 11.4 6610 34.8 - 3309 8200 112 465. 397 66.5 - 35.1 888 Q3 1027 13.4 7875 58.9 0.006 4400 10728 155 705 481 88.5 0.023 46.8 927 SA04-13 - Upper Critical Zone, Merensky Reef (n = 30 (18)) Q1 507 6.02 2418 11.5 - 173 31.3 7.75 0.999 0.108 0.260 - 0.121 601 Median 800 6.59 2510 15.8 - 220 45.4 9.65 2.61 0.440 0.720 - 0.370 619 Q3 1042 7.42 2653 20.3 - 359 91.5 14.5 5.88 3.65 1.10 - 0.667 658 MP24D2 - Upper Critical Zone, Bastard Reef (n = 4 (0)) Q1 340 9.78 NA 81.8 - 3569 9360 132 778 444 4.89 - 2.55 898 Median 485 10.3 NA 93 0.15 4490 10625 165 823 450 8.56 0.125 4.27 928 Q3 910 12.9 NA 575 0.44 5385 15640 207 1072 606 16.7 0.490 8.86 955 All concentrations reported in parts per million (ppm). Raw LA-ICP-MS data reduced using Iolite 2.5 trace element reduction scheme with stoichiometric Ti of 59.5% used as internal standard. Q1 = quartile 1, Q3 = quartile 3. Q1-Q3 represents the range of the middle 75% of analyses for each sample. Rare earth element (La-Lu) concentrations were collected however values were all below detection limits and are not included. Numbers in secondary parentheses indicate number of analyses with measured Cr concentrations. T (°C) is the Zr-in-rutile temperature calculated using the method of Ferry and Watson (2007). NA, not analyzed; -, below detection limits. Table 4 Summary of ranges of trace element concentrations of rutile from the Bushveld Complex Element Si Sc Cr Fe Y Zr Nb Hf Ta W Pb Th U T (°C) TW477-661 - Lower Critical Zone (n = 14 (5)) Q1 85 5.44 3850 39.6 - 499 72.2 11.4 0.958 17.1 12.0 - 5.73 685 Median 790 8.68 6020 292 - 2446 5595 90.6 326 315 16.0 - 8.63 851 Q3 1040 10 8100 419 - 3053 9670 112 565 455 43.0 0.009 24.4 878 DT28-912 - Upper Critical Zone, UG2 (n = 27 (15)) Q1 400 6.47 2630 12.5 - 327 65.8 14.9 3.44 6.10 1.48 - 0.594 650 Median 890 12.2 7510 800 - 4220 10640 149 806 566 5.14 - 2.44 921 Q3 1095 13.5 7850 2565 0.005 5025 12375 204 927 584 6.47 - 3.03 945 SA04-06 - Upper Critical Zone, Merensky Reef (n = 16 (16)) Q1 470 5.89 2630 3.15 - 319 89.4 25.8 4.58 59.8 1.73 0.0008 0.319 648 Median 905 8.36 2875 23.6 0.001 474 232 27.1 8.99 136 4.19 0.0084 0.935 682 Q3 1122 10.73 3160 72.8 0.029 1016 424 42.3 16.8 235 7.13 0.0633 3.75 751 SA04-08 - Upper Critical Zone, Merensky Reef (n = 24 (12)) Q1 252 7.69 5050 16.9 - 2588 6348 91.6 261 267 49.3 - 27.5 857 Median 565 11.4 6610 34.8 - 3309 8200 112 465. 397 66.5 - 35.1 888 Q3 1027 13.4 7875 58.9 0.006 4400 10728 155 705 481 88.5 0.023 46.8 927 SA04-13 - Upper Critical Zone, Merensky Reef (n = 30 (18)) Q1 507 6.02 2418 11.5 - 173 31.3 7.75 0.999 0.108 0.260 - 0.121 601 Median 800 6.59 2510 15.8 - 220 45.4 9.65 2.61 0.440 0.720 - 0.370 619 Q3 1042 7.42 2653 20.3 - 359 91.5 14.5 5.88 3.65 1.10 - 0.667 658 MP24D2 - Upper Critical Zone, Bastard Reef (n = 4 (0)) Q1 340 9.78 NA 81.8 - 3569 9360 132 778 444 4.89 - 2.55 898 Median 485 10.3 NA 93 0.15 4490 10625 165 823 450 8.56 0.125 4.27 928 Q3 910 12.9 NA 575 0.44 5385 15640 207 1072 606 16.7 0.490 8.86 955 Element Si Sc Cr Fe Y Zr Nb Hf Ta W Pb Th U T (°C) TW477-661 - Lower Critical Zone (n = 14 (5)) Q1 85 5.44 3850 39.6 - 499 72.2 11.4 0.958 17.1 12.0 - 5.73 685 Median 790 8.68 6020 292 - 2446 5595 90.6 326 315 16.0 - 8.63 851 Q3 1040 10 8100 419 - 3053 9670 112 565 455 43.0 0.009 24.4 878 DT28-912 - Upper Critical Zone, UG2 (n = 27 (15)) Q1 400 6.47 2630 12.5 - 327 65.8 14.9 3.44 6.10 1.48 - 0.594 650 Median 890 12.2 7510 800 - 4220 10640 149 806 566 5.14 - 2.44 921 Q3 1095 13.5 7850 2565 0.005 5025 12375 204 927 584 6.47 - 3.03 945 SA04-06 - Upper Critical Zone, Merensky Reef (n = 16 (16)) Q1 470 5.89 2630 3.15 - 319 89.4 25.8 4.58 59.8 1.73 0.0008 0.319 648 Median 905 8.36 2875 23.6 0.001 474 232 27.1 8.99 136 4.19 0.0084 0.935 682 Q3 1122 10.73 3160 72.8 0.029 1016 424 42.3 16.8 235 7.13 0.0633 3.75 751 SA04-08 - Upper Critical Zone, Merensky Reef (n = 24 (12)) Q1 252 7.69 5050 16.9 - 2588 6348 91.6 261 267 49.3 - 27.5 857 Median 565 11.4 6610 34.8 - 3309 8200 112 465. 397 66.5 - 35.1 888 Q3 1027 13.4 7875 58.9 0.006 4400 10728 155 705 481 88.5 0.023 46.8 927 SA04-13 - Upper Critical Zone, Merensky Reef (n = 30 (18)) Q1 507 6.02 2418 11.5 - 173 31.3 7.75 0.999 0.108 0.260 - 0.121 601 Median 800 6.59 2510 15.8 - 220 45.4 9.65 2.61 0.440 0.720 - 0.370 619 Q3 1042 7.42 2653 20.3 - 359 91.5 14.5 5.88 3.65 1.10 - 0.667 658 MP24D2 - Upper Critical Zone, Bastard Reef (n = 4 (0)) Q1 340 9.78 NA 81.8 - 3569 9360 132 778 444 4.89 - 2.55 898 Median 485 10.3 NA 93 0.15 4490 10625 165 823 450 8.56 0.125 4.27 928 Q3 910 12.9 NA 575 0.44 5385 15640 207 1072 606 16.7 0.490 8.86 955 All concentrations reported in parts per million (ppm). Raw LA-ICP-MS data reduced using Iolite 2.5 trace element reduction scheme with stoichiometric Ti of 59.5% used as internal standard. Q1 = quartile 1, Q3 = quartile 3. Q1-Q3 represents the range of the middle 75% of analyses for each sample. Rare earth element (La-Lu) concentrations were collected however values were all below detection limits and are not included. Numbers in secondary parentheses indicate number of analyses with measured Cr concentrations. T (°C) is the Zr-in-rutile temperature calculated using the method of Ferry and Watson (2007). NA, not analyzed; -, below detection limits. Fig. 13 View largeDownload slide Trace element variations of Bushveld Complex rutile. (a) Zr vs Hf. (b) Zr vs Cr. (c) Ta vs Nb, note the logarithmic axes. (d) Sc vs Cr. Two distinct types of rutile are revealed on the basis of trace element geochemistry. The orange field highlights rutile that is interpreted to represent primary magmatic grains with elevated Zr, Hf, Ta, Nb, Cr, and Sc that crystallized from fractionated interstitial melt. The grey field outlines rutile that is interpreted to have exsolved from chromite and is characterized by low Zr and Hf, very low Ta and Nb, and anomalously low Cr. Average 2σ uncertainty is indicated in each panel. Abbreviations: LCZ, Lower Critical Zone; UG2, Upper Group 2 chromitite; MR, Merensky Reef; BR, Bastard Reef. Fig. 13 View largeDownload slide Trace element variations of Bushveld Complex rutile. (a) Zr vs Hf. (b) Zr vs Cr. (c) Ta vs Nb, note the logarithmic axes. (d) Sc vs Cr. Two distinct types of rutile are revealed on the basis of trace element geochemistry. The orange field highlights rutile that is interpreted to represent primary magmatic grains with elevated Zr, Hf, Ta, Nb, Cr, and Sc that crystallized from fractionated interstitial melt. The grey field outlines rutile that is interpreted to have exsolved from chromite and is characterized by low Zr and Hf, very low Ta and Nb, and anomalously low Cr. Average 2σ uncertainty is indicated in each panel. Abbreviations: LCZ, Lower Critical Zone; UG2, Upper Group 2 chromitite; MR, Merensky Reef; BR, Bastard Reef. DISCUSSION Trace elements and their ratios (e.g. Ti, Th/U, Lu/Hf, ∑REE, Sc/Yb) vary significantly with stratigraphic height in zircon from the Bushveld Complex from the Lower Zone through the Upper Critical Zone to the top of the Upper Zone and into the roof granites (Fig. 14). The trace element chemistry of Bushveld Complex zircon and rutile reflects the compositons of the parent magmas and variations in the crystallization path of fractionated interstitial melt in their host cumulates. The significance of trace element concentrations and ratios (e.g. Th/U) with respect to the processes that are recorded into their respective mineral systems are considered below. Using Ti-in-zircon and Zr-in-rutile thermometry, in conjunction with forward geochemical modelling of zircon saturation from proposed parental magmas, the crystallization pathways are constrained and their impact on the late, near-solidus processes of rocks from the Bushveld Complex is assessed. Fig. 14 View largeDownload slide Box-and-whisker diagrams showing trace element variations in zircon vs stratigraphic height in the Bushveld Complex. (a) Ti. (b) Th/U, note logarithmic scale. (c) Lu/Hf. (d) ∑REE = sum of the concentrations of the rare earth elements (La–Lu). (e) Sc/Yb. Each box represents the lower 25% to upper 75% of values for each sample and the whiskers extend to minimum and maximum values. The vertical line in each box shows the median value. Average 2σ uncertainty is indicated in each panel. The thick dashed red line indicates the location of the Merensky Reef for reference. Abbreviations in stratigraphic column: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ, Upper Critical Zone; MZ, Main Zone; UZ, Upper Zone. Units indicated on left side of column: LG, Lower Group chromitites; MG, Middle Group chromitites; UG, Upper Group chromitites; BB, Boulder Bed; MR, Merensky Reef; BR, Bastard Reef; MML, Main Magnetitite Layer; M21, Magnetitite layer 21. Fig. 14 View largeDownload slide Box-and-whisker diagrams showing trace element variations in zircon vs stratigraphic height in the Bushveld Complex. (a) Ti. (b) Th/U, note logarithmic scale. (c) Lu/Hf. (d) ∑REE = sum of the concentrations of the rare earth elements (La–Lu). (e) Sc/Yb. Each box represents the lower 25% to upper 75% of values for each sample and the whiskers extend to minimum and maximum values. The vertical line in each box shows the median value. Average 2σ uncertainty is indicated in each panel. The thick dashed red line indicates the location of the Merensky Reef for reference. Abbreviations in stratigraphic column: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ, Upper Critical Zone; MZ, Main Zone; UZ, Upper Zone. Units indicated on left side of column: LG, Lower Group chromitites; MG, Middle Group chromitites; UG, Upper Group chromitites; BB, Boulder Bed; MR, Merensky Reef; BR, Bastard Reef; MML, Main Magnetitite Layer; M21, Magnetitite layer 21. Mapping the solidus of Bushveld Complex cumulates and granites using Ti-in-zircon thermometry In the Bushveld Complex, zircon occurs in mafic–ultramafic rocks that are characterized either by heterogeneous textures, such as coarse grain size and the irregular distribution of minerals (e.g. UG2 pyroxenite, Merensky Reef, Bastard Reef), or by the presence of interstitial minerals, typically poikilitic plagioclase, as well as evidence of minerals that crystallized from highly fractionated melts (e.g. quartz, K-feldspar, biotite). The minimum temperature (Tzrc) at which zircon crystallized in the Bushveld cumulates and felsic roof rocks can be calculated following the method of Ferry & Watson (2007): log  (ppm Ti−in−zircon)=(5.711±0.072)–(4800±86)/Tzrc(K)– log aSiO2+log aTiO2 Interstitial quartz is present in all mafic–ultramafic cumulates and thus aSiO2 = 1 for all calculations. Rutile occurs in all Lower Zone, Critical Zone, and Main Zone samples, which fixes aTiO2 = 1; for the rutile-free Upper Zone and felsic roof rock samples, a value of aTiO2 = 0·7 was used (e.g. Hayden & Watson, 2007; Grimes et al., 2009). The Ferry & Watson (2007) thermometer was calibrated for a pressure of 1 GPa (10 kbars). They estimated a pressure dependence of -5°C/kbar for pressures below 10 kbar, which would lower the calculated temperatures for the Bushveld Complex (pressure of emplacement ∼3 kbar; Wallmach et al., 1995; VanTongeren et al., 2010) zircon by ∼35°C. In the following discussion, uncorrected Tzrc values are compared with published Ti-in-zircon results (Yudovskaya et al., 2013; Zeh et al., 2015) and with the Zr-in-rutile thermometry results (Ferry & Watson, 2007; Tomkins et al., 2007). Application of Ti-in-zircon thermometry yields temperatures that range from ∼950°C to 690°C (Figs 15, 16) in the Bushveld Complex, comparable to those previously reported for Bushveld zircon (930–760°C: Yudovskaya et al., 2013; 940–670°C: Zeh et al., 2015) and from a comprehensive study of zircon from the mafic–ultramafic rocks of the c.2709 Ma Stillwater Complex (990–720°C: Wall et al., 2018). The highest temperature for each sample is interpreted to mark the onset of zircon saturation in the fractionated interstitial melt and the lowest temperatures are assumed to represent solidus or near-solidus temperatures for each rock. In Lower Zone and Critical Zone ultramafic rocks, higher Ti concentrations in zircon (up to 61 ppm) yield higher temperatures and a range of Tzrc from ∼950–730°C, with the highest temperatures recorded in the Bastard Reef sample (MP24D2) from the top of the Upper Critical Zone. Main Zone zircon temperatures span a range comparable to that determined from the Critical Zone (Tzrc ∼930–750°C). In contrast, zircon from the more evolved Upper Zone diorite and overlying felsic roof rocks are characterized by notably lower temperatures (Tzrc ∼875–690°C) (Fig. 15j–m). Estimated temperature ranges for individual samples are ΔT = 182 °C (LZ10–02, Lower Zone) to ΔT = 83°C (B07–040, Upper Zone), ranges that are similar to those determined by Wall et al. (2018) for samples from the Stillwater Complex (ΔT = 50–150°C). Fig. 15 View largeDownload slide Ti vs Hf diagrams for zircon from all Bushveld Complex samples. The horizontal dashed lines indicate Ti-in-zircon temperatures (°C) calculated using aSiO2 = 1 and aTiO2 = 1 for all samples, except for the Upper Zone and roof granite samples where aTiO2 = 0·7 (Ferry & Watson, 2007). All plots are shown at the same scale for comparison. (a) LZ10–02. (b) TW477–661. (c) DT28–912. (d) B00–1-6. (e) B90–7(0). (f) SA04–08. (g) SA04–13. (h) MPD24D2. (i) B90–1. (j) B07–040. (k) B07–051. (l) B10–054. (m) B10–056. Most samples show distinct trends of decreasing Ti (and temperature) with increasing Hf that tracks the progressive crystallization of zircon from fractionated interstitial melt down to temperatures approaching the solidus for mafic–ultramafic rocks (800–750°C). The roof granites show more restricted variation suggesting crystallization at or near the eutectic. Symbols: square, lherzolite/norite; diamond, orthopyroxenite; triangle, diorite; circle, granite/granophyre. Abbreviations: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ, Upper Critical Zone; MZ, Main Zone; UZ, Upper Zone; MR, Merensky Reef; BR, Bastard Reef. Fig. 15 View largeDownload slide Ti vs Hf diagrams for zircon from all Bushveld Complex samples. The horizontal dashed lines indicate Ti-in-zircon temperatures (°C) calculated using aSiO2 = 1 and aTiO2 = 1 for all samples, except for the Upper Zone and roof granite samples where aTiO2 = 0·7 (Ferry & Watson, 2007). All plots are shown at the same scale for comparison. (a) LZ10–02. (b) TW477–661. (c) DT28–912. (d) B00–1-6. (e) B90–7(0). (f) SA04–08. (g) SA04–13. (h) MPD24D2. (i) B90–1. (j) B07–040. (k) B07–051. (l) B10–054. (m) B10–056. Most samples show distinct trends of decreasing Ti (and temperature) with increasing Hf that tracks the progressive crystallization of zircon from fractionated interstitial melt down to temperatures approaching the solidus for mafic–ultramafic rocks (800–750°C). The roof granites show more restricted variation suggesting crystallization at or near the eutectic. Symbols: square, lherzolite/norite; diamond, orthopyroxenite; triangle, diorite; circle, granite/granophyre. Abbreviations: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ, Upper Critical Zone; MZ, Main Zone; UZ, Upper Zone; MR, Merensky Reef; BR, Bastard Reef. Fig. 16 View largeDownload slide Summary of thermometry results for zircon from the Bushveld Complex. (a) Ti-in-zircon temperature vs Hf showing distinct negative trends within the dataset. Ti-in-zircon temperatures calculated using aSiO2 = 1 and aTiO2 = 1 for all samples, except for the Upper Zone and roof granite samples where aTiO2 = 0·7 (Ferry & Watson, 2007). (b) Histogram showing the distribution of calculated Ti-in-zircon temperatures ranging from ∼950°C down to 690°C. Average 2σ uncertainty in temperature is based on uncertainty in Ti concentration only. Uncertainty including error reported in the Ti-in-zircon thermometry equation is shown as grey band behind uncertainty in panel a. Symbols: square, lherzolite/norite; diamond, orthopyroxenite; triangle, diorite; circle, granite/granophyre. Abbreviations: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ-MR, Upper Critical Zone-Merensky Reef; UCZ, Upper Critical Zone; MZ, Main Zone; UZ, Upper Zone. Fig. 16 View largeDownload slide Summary of thermometry results for zircon from the Bushveld Complex. (a) Ti-in-zircon temperature vs Hf showing distinct negative trends within the dataset. Ti-in-zircon temperatures calculated using aSiO2 = 1 and aTiO2 = 1 for all samples, except for the Upper Zone and roof granite samples where aTiO2 = 0·7 (Ferry & Watson, 2007). (b) Histogram showing the distribution of calculated Ti-in-zircon temperatures ranging from ∼950°C down to 690°C. Average 2σ uncertainty in temperature is based on uncertainty in Ti concentration only. Uncertainty including error reported in the Ti-in-zircon thermometry equation is shown as grey band behind uncertainty in panel a. Symbols: square, lherzolite/norite; diamond, orthopyroxenite; triangle, diorite; circle, granite/granophyre. Abbreviations: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ-MR, Upper Critical Zone-Merensky Reef; UCZ, Upper Critical Zone; MZ, Main Zone; UZ, Upper Zone. Most Bushveld Complex samples show prominent negative Ti–Hf relationships that reflect crystallization of zircon from progressively fractionated interstitial melt (i.e. increasing Hf) as temperature decreased (i.e. decreasing Ti) (Fig. 15) (e.g. Grimes et al., 2009). The (relative) range in temperature and composition exhibited by each sample likely reflects the connectivity of interstitial melt pockets within a given cumulate mass during cooling, compaction, and accessory mineral saturation (e.g. Meurer & Meurer, 2006; Holness et al., 2011; Cawthorn, 2013a). Zircon from the Upper Zone sample and zircon from the three felsic roof rocks record relatively limited Ti–Hf variation, indicating that crystallization of zircon was likely near the eutectic temperature for these rocks. Forward modelling and significance of zircon saturation in fractionated interstitial melts during crystallization of the Bushveld Complex Forward geochemical modelling was carried out using rhyolite-MELTS v1.02 (Gualda et al., 2012) to investigate the conditions required for zircon saturation in Bushveld Complex mafic–ultramafic rocks, the predicted mineral assemblage at zircon saturation for comparison with the minerals observed in the late-crystallized interstitial pockets, and the thermometry results. The parental magma compositions used include the B1, B2, and B3 magmas from Barnes et al. (2010) and a 60:40 mixed composition of B1 and B2 magmas as a potential parent to the Upper Critical Zone (Barnes et al., 2010). The MELTS runs were carried out at 3 kilobars pressure in both equilibrium and fractional crystallization mode at 10°C temperature increments, starting from just above the liquidus for each composition down to 690°C, or until no melt remained. To reproduce the predominantly orthopyroxenite cumulates present in the Critical Zone, clinopyroxene crystallization was suppressed. The initial starting water contents were varied from 0·25–1·0 wt % H2O; 1·0 wt % H2O yielded phases that were most consistent with the observed late-stage mineral assemblages of quartz, biotite, and apatite. Amphibole crystallization was also suppressed due to its absence in the mafic–ultramafic cumulates. Distribution coefficients for Zr for the mineral phases present in the modelled crystallization sequences were taken from Bédard (2006, 2007). Bulk zircon saturation was calculated following the experimentally determined relationship of Watson (1979) with revised coefficients by Boehnke et al. (2013) that relate zircon saturation in silicate magmas to melt composition and temperature: lnDZr=(10108±32)/T(K)– 3(1.16±0.15) (M–1)–(1.48±0.09) where DZr = distribution coefficient, T(K) = temperature in Kelvin, and M = (Na + K + 2Ca)/(Al · Si), with each component representing the normalized molar ratio; calculation of Zrmelt requires division of DZr by the abundance of Zr in zircon (∼500 000 ppm: Watson, 1979). The results of the MELTS modelling and zircon saturation are shown in Figs 17 and 18. Fig. 17 View largeDownload slide Plots of Zr vs temperature (°C) summarizing zircon saturation modelling results using rhyolite-MELTS for four proposed parental Bushveld magmas. (a) B1 magma. (b) 60:40 mix of B1 and B2 magmas. (c) B2 magma. (d) B3 magma. Estimated parental magma compositions are from Barnes et al. (2010) and based on the whole rock chemistry of marginal sills – see the relative locations on the stratigraphic column in Fig. 2. The red lines track the evolution of Zr concentration in the melt with fractionation (decreasing temperature) and the blue lines indicate calculated Zr concentration for zircon saturation in the melt (Watson & Harrison, 1983; Boehnke et al., 2013). Zircon saturation occurs where the lines intersect and is denoted with a red star. The saturation temperature and the Zr concentration, SiO2 content, and M-value (M = [Na + K + 2Ca] / [Al · Si]) in the melt at zircon saturation, and initial Zr content of the magma, are indicated for each composition – see text for additional details and MELTS parameters. Note that all plots are at the same scale for comparison. Vertical dashed lines indicate percent melt remaining based on MELTS calculations. Fig. 17 View largeDownload slide Plots of Zr vs temperature (°C) summarizing zircon saturation modelling results using rhyolite-MELTS for four proposed parental Bushveld magmas. (a) B1 magma. (b) 60:40 mix of B1 and B2 magmas. (c) B2 magma. (d) B3 magma. Estimated parental magma compositions are from Barnes et al. (2010) and based on the whole rock chemistry of marginal sills – see the relative locations on the stratigraphic column in Fig. 2. The red lines track the evolution of Zr concentration in the melt with fractionation (decreasing temperature) and the blue lines indicate calculated Zr concentration for zircon saturation in the melt (Watson & Harrison, 1983; Boehnke et al., 2013). Zircon saturation occurs where the lines intersect and is denoted with a red star. The saturation temperature and the Zr concentration, SiO2 content, and M-value (M = [Na + K + 2Ca] / [Al · Si]) in the melt at zircon saturation, and initial Zr content of the magma, are indicated for each composition – see text for additional details and MELTS parameters. Note that all plots are at the same scale for comparison. Vertical dashed lines indicate percent melt remaining based on MELTS calculations. Fig. 18 View largeDownload slide Predicted temperature vs mineral abundance for equilibrium crystallization (solid lines) and fractional crystallization (dashed lines) as modeled with rhyolite-MELTS for Bushveld parental magma B1:B2 (60:40 mix). (a) Complete temperature range from 1300°C (liquidus) to 700°C (solidus) showing similar minerals and abundances for both models. (b) Detailed view of the final 20% of crystallization from 800–700°C showing the differences between the models with biotite, quartz, and titanite crystallization in the equilibrium model and magnetite, alkali feldspar, and quartz crystallization in the fractional model. The secondary y-axis shows the % melt remaining (thick black lines, solid and dashed). The red star indicates the point of bulk zircon saturation at ∼13% remaining melt. Clinopyroxene crystallization was suppressed in all models. Abbreviations: opx, orthopyroxene; plag, plagioclase; mt, magnetite; ilm, ilmenite; ksp, alkali feldspar. Fig. 18 View largeDownload slide Predicted temperature vs mineral abundance for equilibrium crystallization (solid lines) and fractional crystallization (dashed lines) as modeled with rhyolite-MELTS for Bushveld parental magma B1:B2 (60:40 mix). (a) Complete temperature range from 1300°C (liquidus) to 700°C (solidus) showing similar minerals and abundances for both models. (b) Detailed view of the final 20% of crystallization from 800–700°C showing the differences between the models with biotite, quartz, and titanite crystallization in the equilibrium model and magnetite, alkali feldspar, and quartz crystallization in the fractional model. The secondary y-axis shows the % melt remaining (thick black lines, solid and dashed). The red star indicates the point of bulk zircon saturation at ∼13% remaining melt. Clinopyroxene crystallization was suppressed in all models. Abbreviations: opx, orthopyroxene; plag, plagioclase; mt, magnetite; ilm, ilmenite; ksp, alkali feldspar. The MELTS modeling results confirm that zircon saturation can be achieved using all four of the proposed parental melts as starting compositions (Fig. 17). The concentration of Zr, an incompatible element, in the melt increases during crystallization of the major cumulus minerals (i.e. Zr-free minerals) and zircon saturation occurs with just over 20% remaining residual melt for the B1 parent down to as low as 6% remaining melt for B3 magmas (Fig. 17c and d) at temperatures ranging from 800°C down to 740°C. These temperatures are directly comparable with those determined by Ti-in-zircon thermometry (Fig. 16) and the predicted remaining melt volumes correspond to the observed rock textures (e.g. 5–20 vol.% interstitial material). The modelling results also reproduce the observed mineralogy in the mafic–ultramafic cumulates, forming predominantly orthopyroxene, followed by plagioclase and a late assemblage of minor and accessory minerals. At zircon saturation, the composition of the remaining liquid is intermediate-felsic (59–67 wt % SiO2) and the predicted stable phases include biotite, quartz, and alkali feldspar, with titanite appearing in equilibrium models at <800°C (Figs 17, 18). These predicted phases have been observed in the Bushveld samples themselves (e.g. Figs 5a, b, e, 6). Zircon will be a stable liquidus or autocrystic phase in fractionated melts within the mafic–ultramafic cumulates of layered intrusions provided that sufficient interstitial melt remains trapped within the matrix of the growing primocrysts to allow for crystallization of a Si–H2O-rich near-eutectic mineral assemblage (e.g. quartz, Na-plagioclase, K-feldspar, biotite), at temperatures approaching the solidus. Recognition that zircon can be extracted from mafic–ultramafic rocks in layered intrusions and associated Cr–Ni–Cu–PGE–V–Ti mineralization by targeting samples with heterogeneous textures (i.e. pegmatites) or macroscopic evidence for interstitial minerals that crystallized from highly fractionated melt (e.g. quartz, alkali feldspar, biotite) provides the petrologic community with new geochemical tools to decipher and quantify processes during their crystallization and consolidation (e.g. Scoates & Wall, 2015; Manor et al., 2017; Wall et al., 2018). The proposed temperature-composition framework for zircon crystallization in the Bushveld Complex also has important implications for U–Pb geochronologic studies of zircon from layered intrusions. The closure temperature for Pb diffusion in zircon is in excess of 950–1000°C (Cherniak & Watson, 2000, 2003; Cherniak, 2010), which is higher than the Ti-in-zircon and zircon saturation temperatures established for individual samples from the Bushveld Complex. High-precision U–Pb geochronology employing the chemical annealing and abrasion technique of Mattinson (2005) for individual zircon grains from layered intrusions that have been characterized for internal textural morphology by SEM-CL and trace element concentrations, including Ti-in-zircon thermometry, will yield dates that correspond to zircon crystallization within this relatively restricted temperature-time range at the latest stages of crystallization. With current CA-TIMS analytical precision of ≤0·1% on individual zircon analyses (2σ), the Bushveld Complex is sufficiently ancient (c.2·06 Ga) that it is not yet possible to distinguish the age of the earliest crystallized zircon (i.e. higher temperatures) from the last-crystallized zircon (i.e. lowest temperatures) in a single sample (Scoates & Friedman, 2008; Scoates & Wall, 2015; Zeh et al., 2015; Mungall et al., 2016). In contrast, identification of dispersed zircon dates (Δt = several hundred thousand years) within individual samples from Mesozoic-Cenozoic mafic to intermediate intrusions (e.g. Rioux et al., 2012; Samperton et al., 2015) offers the prospect of resolving integrated temperature-composition-time paths at the sample scale for young layered intrusions where zircon grains are fully characterized by SEM-CL and LA-ICP-MS prior to analysis. The significance of high-Th/U zircon in the Critical Zone The range of Th/U from ∼0·5 to 1–2 for most Bushveld Complex zircon is typical of magmatic values (e.g. Belousova et al., 2002; Xiang et al., 2011; Kirkland et al., 2014); however, zircon from Upper Critical Zone orthopyroxenites displays anomalously high values (Th/U = 2–77) (Figs 9a, 14). As the majority of the Th concentrations in these zircon grains are similar to those analysed from zircon in other samples from the Rustenburg Layered Suite and felsic roof rocks (Fig. 9a), the high-Th/U analyses appear to be related to a depletion in U relative to Th. High-Th/U zircon from the Bushveld Complex may have formed as a result of co-crystallization of zircon with other U-rich phases, or from fractionation of zircon alone (Yudovskaya et al., 2013; Zeh et al., 2015). Alternatively, U may have been lost from the fractionated interstitial melt during late-stage oxidation and fluid saturation, resulting in high-Th/U (and U-depleted) zircon. The relative effectiveness of these processes is evaluated below. The occurrence of co-crystallizing phases with zircon that have a greater affinity for U than Th (e.g. uraninite, zirconolite, baddeleyite) would lead to high Th/U in the fractionated interstitial melt and high Th/U in zircon that subsequently crystallized from that interstitial melt. Rare, U-bearing accessory minerals reported in the Bushveld Complex include baddeleyite (ZrO2) and loveringite ((Ca, Ce)(Ti, Fe, Cr, Mg)21O38)) (Yudovskaya et al., 2013; Mungall et al., 2016). However, they are present in such minute abundances that their crystallization would not have significantly affected the distribution of Th and U during crystallization from a mass balance perspective. Rutile and apatite may contain U and both are found in many Bushveld samples. Uranium concentrations in rutile are highly variable (0·01–74 ppm) (Table 4; Supplementary Data Electronic Appendix 2), reflecting, in part, the very low U concentrations of low-HFSE zircon. However, the highest U concentrations in rutile are from Merensky Reef sample SA04–08 (Eastern Limb), which has ‘normal’ Th/U zircon values (mean Th/U = 1·0; Table 4). This, combined with rutile being significantly less abundant than zircon in Bushveld rocks (i.e. by at least an order of magnitude), means that rutile fractionation in unlikely to produce high-Th/U residual melts. Finally, apatite analysed from the Merensky Reef is U-poor (1–3 ppm; Scoates & Wall, 2015) and thus is also a poor candidate for increasing Th/U during fractionation. The significant increase in Th/U required in the melt to produce the high-Th/U zircon typical of the Critical Zone cannot be explained by co-crystallization of other U-bearing phases with zircon. An increase in Th/U in both zircon and residual melt during crystallization of interstitial melt in the Bushveld Complex may result from zircon fractionation alone (Yudovskaya et al., 2013; Zeh et al., 2015). The viability of this process requires that during fractionation the zircon-melt distribution coefficient for U (DU) was higher than that for Th (DTh) such that the coefficient ratio DTh/DU decreased during cooling (e.g. Rubatto & Hermann, 2007). Recently, Kirkland et al. (2014) evaluated the evolution of Th/U in zircon and melt from a wide range of bulk-rock compositions (gabbro to alkali granite) based on 10 000 SIMS analyses of zircon. Ratios of Th/U are higher in mafic melts (gabbro Th/U ∼1) and decrease systematically with increasing silica content (granite Th/U ∼0·65). The ionic radius of Th4+ is approximately 4% larger than that of U4+ (Shannon, 1976) and, consequently, it is much less easily accommodated in the zircon crystal structure. The result is that although under equilibrium conditions there is a trend to higher Th/U in zircon with decreasing melt temperature due to lattice strain, fractionation will yield the opposite results and produce lower Th/U in zircon with decreasing temperature (Kirkland et al., 2014). An alternative process is required to explain the high-Th/U zircon in the Bushveld Complex. We speculate that a local change in the oxidation state of the fractionated interstitial melt resulted in crystallization of the high-Th/U zircon similar to the process proposed for high-Th/U zircon from pyroxenites in the Stillwater Complex (Wall et al., 2018). These high-Th/U compositions, characteristic of zircon from Upper Critical Zone orthopyroxenites, are due to partitioning of U6+ into a Cl-rich fluid phase (e.g. Keppler & Wyllie, 1991; Bacon et al., 2007). Apatite, a common interstitial mineral in the Bushveld Complex, has been used as a monitor of halogen variations of interstitial melts in mafic layered intrusions (Boudreau et al., 1986; Boudreau & McCallum, 1989). Apatite and biotite from rocks below the Bushveld Complex Main Zone are characterized by high ratios of Cl/F (i.e. chlorapatite) (Boudreau & McCallum, 1989; Willmore et al., 2000). To produce the high-Th/U zircon in the Upper Critical Zone orthopyroxenites, it is likely that zircon crystallized and grew from fractionated interstitial melt during and following exsolution of Cl-rich fluids. These fluids preferentially partitioned U6+, resulting in high Th/U in the remaining melt. This process must have been initiated at relatively high temperatures (∼950°C) based on the Ti-in-zircon thermometry (Fig. 16) and is consistent with evidence from the halogen geochemistry of apatite for fluid saturation in Critical Zone melt at temperatures >1000°C (Boudreau, 1999; Willmore et al., 2000). The variable Th/U in zircon from individual Critical Zone samples is a measure of the degree to which the interstitial melt pockets were interconnected and equilibrated with Cl-rich fluids. Origin of LREE-enriched signature of zircon from the granitic roof of the Bushveld Complex The LREE-enrichment (i.e. anomalously low (Sm/La)N) of zircon from the three felsic roof rocks and from the Main Zone sample is distinct from the normal magmatic REE patterns of zircon in Bushveld mafic–ultramafic cumulates (Fig. 11). Relative enrichment of LREE in zircon may result from a variety of processes, including co-crystallization of other accessory phases (e.g. xenotime YPO4, monazite CePO4, apatite Ca5(PO4)3(F, Cl, OH)) and subsequent incorporation as micro-inclusions, and hydrothermal alteration leading to solid-state partial recrystallization or re-equilibration with fluids during local dissolution-reprecipitation and formation of LREE-enriched micro-inclusions (Whitehouse & Kamber, 2002; Hoskin & Schaltegger, 2003; Hoskin, 2005; Geisler et al., 2007; Grimes et al., 2009). Although relatively large inclusions of apatite, alkali feldspar, and albite are found in some zircon in the roof granites, the analysed grains were devoid of these types of inclusions (Fig. 7) and do not show the porous or chaotic CL responses typical of zircon that has interacted with an aqueous fluid (Grimes et al., 2009). To evaluate the potential effect of ablating zircon with micro-inclusions (i.e. not resolvable by SEM), mixing calculations were carried out using an analysis from B10–056 (Nebo Granite) with the lowest La concentration (highest (Sm/La)N) as the starting composition and the most LREE-enriched analysis in the sample as the ‘contaminated’ end-member (Fig. 19). The effect of ablating micro-inclusions was simulated by using the average REE concentrations of apatite from the top of the Upper Zone (VanTongeren & Mathez, 2012) and monazite and xenotime compositions from Borai et al. (2002). The calculations demonstrate that addition of apatite could reproduce the LREE enrichment and the overall shape of the REE patterns, although this would require mixing in an unrealistic amount of apatite (up to 35 wt %). Mixing zircon with any quantity of xenotime yields increased REE concentrations due to the positive slope of the xenotime REE pattern, however, it does not flatten the LREE pattern as observed in the zircon patterns from the top of the Bushveld Complex. In contrast, mixing with as little as 0·01 wt % monazite could produce both the observed patterns and slopes. Significant LREE enrichment occurs with incorporation of only 0·001 wt % of monazite, which we assume represents the dominant micro-inclusion incorporated during growth of zircon in the felsic roof rock samples and in the Main Zone sample. As Y is notably more abundant in zircon with LREE-enriched compositions (range of median Y = 980–1400 ppm compared to 272–767 ppm for all other samples: Table 3), some proportion of the micro-inclusions is likely to be xenotime. Fig. 19 View largeDownload slide Rare earth element geochemistry and the effect of ablating micro-inclusions during zircon analysis in the felsic roof rocks. Modeling results show the effect of integrating accessory mineral inclusions on rare earth element (REE) concentrations during ablation of zircon for the Upper Zone sample (B10–056). (a-c) Chondrite-normalized REE patterns showing the effect of incorporating inclusions of apatite (a), xenotime (b), and monazite (c) during analysis. Yellow lines indicate zircon with a typical magmatic pattern from B10–056 (lower line) and an analysis with flat light REE (upper line). The Bushveld apatite pattern is from VanTongeren & Mathez (2010), and the xenotime and monazite compositions are from Borai et al. (2002). Mixing proportions next to the vertical arrows are in wt %. Fig. 19 View largeDownload slide Rare earth element geochemistry and the effect of ablating micro-inclusions during zircon analysis in the felsic roof rocks. Modeling results show the effect of integrating accessory mineral inclusions on rare earth element (REE) concentrations during ablation of zircon for the Upper Zone sample (B10–056). (a-c) Chondrite-normalized REE patterns showing the effect of incorporating inclusions of apatite (a), xenotime (b), and monazite (c) during analysis. Yellow lines indicate zircon with a typical magmatic pattern from B10–056 (lower line) and an analysis with flat light REE (upper line). The Bushveld apatite pattern is from VanTongeren & Mathez (2010), and the xenotime and monazite compositions are from Borai et al. (2002). Mixing proportions next to the vertical arrows are in wt %. Application of tectono-magmatic discrimination diagrams using trace elements in Bushveld zircon The trace element chemistry of zircon, especially the use of trace element ratios, is sensitive to rock type, crystallization environment, and tectono-magmatic setting (e.g. Hoskin & Ireland, 2000; Belousova et al., 2002; Grimes et al., 2007). Based on the classification diagrams of Belousova et al. (2002), the elemental geochemistry (e.g. Y, U, Th) of Bushveld zircon is similar to that of zircon that crystallized from granitic to syenitic magmas (Fig. 20a and b). The notable exception is the high-Th/U zircon from the Critical Zone that records the effects of late-stage oxidation, fluid saturation, and U loss from fractionated interstitial melt (Fig. 20a). This result is consistent with the predicted intermediate-felsic (59–67 wt % SiO2) melt compositions for zircon saturation derived from the MELTS modeling and with the associated phases in the rocks themselves, including biotite, quartz, and alkali feldspar. Fig. 20 View largeDownload slide Trace element distribution diagrams for Bushveld Complex zircon compared with classification schemes from global zircon datasets. (a-b) Logarithmic plots of Y vs U and Y vs Th compared to fields and rock types from Belousova et al. (2002). Note that the trend to low-U zircon defined mainly by samples from the Upper Critical Zone reflects U loss from interstitial melt during crystallization. (c-d) Logarithmic plots of U/Yb vs Nb/Yb and Sc/Yb vs Nb/Yb compared with fields established for various tectono-magmatic settings from Grimes et al. (2015). The majority of the samples, excluding the high-Th/U zircon, plot within the continental arc field (green-shaded field); the other fields include analyses of zircon from mid-ocean ridge gabbros (MOR-type) and ocean islands (OI-type). The prominent trends to low U/Yb values in (c) for zircon from the Critical Zone and Merensky Reef reflects U loss (high-Th/U zircon). The vectors indicate the effect of crystallization of zircon (zrn), apatite (ap), titanite (ttn), ilmenite (ilm), and amphibole (amph). Mean 2σ analytical uncertainties for each calculated ratio are smaller than the symbol size. Symbols: square, lherzolite/norite; diamond, orthopyroxenite; triangle, diorite; circle, granite/granophyre. Abbreviations: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ-MR, Upper Critical Zone-Merensky Reef; UCZ, Upper Critical Zone; MZ, Main Zone; UZ, Upper Zone. Fig. 20 View largeDownload slide Trace element distribution diagrams for Bushveld Complex zircon compared with classification schemes from global zircon datasets. (a-b) Logarithmic plots of Y vs U and Y vs Th compared to fields and rock types from Belousova et al. (2002). Note that the trend to low-U zircon defined mainly by samples from the Upper Critical Zone reflects U loss from interstitial melt during crystallization. (c-d) Logarithmic plots of U/Yb vs Nb/Yb and Sc/Yb vs Nb/Yb compared with fields established for various tectono-magmatic settings from Grimes et al. (2015). The majority of the samples, excluding the high-Th/U zircon, plot within the continental arc field (green-shaded field); the other fields include analyses of zircon from mid-ocean ridge gabbros (MOR-type) and ocean islands (OI-type). The prominent trends to low U/Yb values in (c) for zircon from the Critical Zone and Merensky Reef reflects U loss (high-Th/U zircon). The vectors indicate the effect of crystallization of zircon (zrn), apatite (ap), titanite (ttn), ilmenite (ilm), and amphibole (amph). Mean 2σ analytical uncertainties for each calculated ratio are smaller than the symbol size. Symbols: square, lherzolite/norite; diamond, orthopyroxenite; triangle, diorite; circle, granite/granophyre. Abbreviations: LZ, Lower Zone; LCZ, Lower Critical Zone; UCZ-MR, Upper Critical Zone-Merensky Reef; UCZ, Upper Critical Zone; MZ, Main Zone; UZ, Upper Zone. Grimes et al. (2015) proposed the use of select immobile trace element ratios (e.g. U/Yb, Nb/Yb, Sc/Nb) to effectively distinguish zircon from predominantly modern through to Mesozoic mid-ocean ridge, magmatic arc, and ocean island (+plume-influenced) settings based on a compilation of >5300 SHRIMP-RG analyses. Excluding the high-Th/U zircon from the Critical Zone and Merensky Reef (i.e. low U/Yb), most Bushveld Complex zircon plot above the mantle array (grey band) within the continental arc field on a U/Yb vs Nb/Yb diagram (Fig. 20c). In contrast to most other geochemical characteristics (Figs 9, 12, 14), zircon from the Upper Zone sample is distinguished from felsic roof rock zircon by relatively low Nb/Yb (Fig. 20c). Bushveld zircon also overlaps the field defined by ocean island zircon as represented by Iceland felsic volcanic rocks and Hawaiian trachytes (Grimes et al., 2015) (Fig. 20c). Zircon from the felsic roof rocks plots at the high end of the continental arc field in the Sc/Yb vs Nb/Yb diagram (Fig. 20d). In contrast, nearly all zircon from the mafic–ultramafic rocks has distinctly high Sc/Yb, plotting well above the continental arc field, and shows a strong signal of expected zircon fractionation (Fig. 20d). Zircon from the Upper Zone sample is again distinguished from the felsic roof rocks by relatively high-Sc/Yb. Grimes et al. (2015) highlighted the potential for inaccurate Sc analyses in zircon due to isobaric interferences caused by 90Zr++ during in situ analyses; special attention was paid to this interference during analysis. The only other example of high-Sc/Yb zircon that we have encountered from mafic–ultramafic intrusions is from the Stillwater Complex (Wall et al., 2018). Based on both occurrences, we speculate that high Sc/Yb-zircon may signal a distinctive source composition for these two layered intrusions. Rutile as a petrogenetic indicator in the Bushveld Complex and other mafic layered intrusions Rutile, the most common naturally occurring titanium dioxide polymorph, is found in a wide range of rocks as an accessory mineral, including granitoids, metamorphic rocks, mantle rocks, and meteorites (Meinhold, 2010). Based on its chemical variability (e.g. major host of Nb, Ta and other HFSE such as Zr, appreciable U contents), rutile can be used as a key petrogenetic indicator mineral to monitor a variety of geochemical and geochronological processes (e.g. Rudnick, 2000; Zack et al., 2004a, 2004b; Schmidt et al., 2009; Meinhold, 2010; Scoates & Wall, 2015). In igneous rocks, rutile is most commonly found in granites and associated quartz veins, pegmatites, carbonatites, kimberlites, and metallic ore deposits (Meinhold, 2010), and its presence is now increasingly recognized in mafic–ultramafic plutonic rocks, including layered intrusions such as the Bushveld Complex (Cameron, 1979; Scoates & Friedman, 2008; Vukmanovic et al., 2013; Yudovskaya et al., 2013; Scoates & Wall, 2015) and the Great Dyke (Oberthür et al., 2002), as well as in anorthosite-hosted Fe–Ti oxide ore deposits (Morisset et al., 2010, 2013). The textural setting and geochemistry of rutile in rocks of the Critical Zone from the Bushveld Complex reveal two distinct paths for its formation in the mafic–ultramafic cumulates. Magmatic rutile that is inferred to have crystallized directly from interstitial melt is present as small euhedral to subhedral needles and is associated with biotite or finely dispersed within quartz (e.g. Fig. 5d). It is readily identifiable by high HFSE concentrations typical of igneous rutile from other settings (Meinhold, 2010) and by notably high Sc and Cr contents (Fig. 13d). The strong temperature dependence on the partitioning of Zr into rutile coexisting with zircon or other Zr-rich phases allows for calibration of the Zr-in-rutile thermometer (Zack et al., 2004a; Watson et al., 2006; Ferry & Watson, 2007; Tomkins et al., 2007). The thermometer of Ferry & Watson (2007) expresses temperature as: log  (ppm Zr−in−rutile)=(7.420±0.105)–(4530±111)/T(K)– log aSiO2 Zr-in-rutile thermometry results for interstitial rutile yield a range of temperatures (∼1000–800°C) that are interpreted as crystallization temperatures and that broadly overlap the Ti-in-zircon temperatures from the same rocks (Fig. 21). In contrast, rutile that forms rims or sub-equant grains on chromite is interpreted as an exsolution product of chromite (e.g. Ghisler, 1970; Cameron, 1979; Ghosh & Konar, 2011; Vukmanovic et al., 2013). This rutile, which is depleted in HFSE compared to magmatic rutile, is very highly depleted in Ta and Nb, and contains significantly less Cr than the interstitial grains (Fig. 13b–d). The exsolved rutile inherits the chemical signal of its original chromite host, which contains very low abundances of the HFSE (Pagé et al., 2012) and Cr as an essential structural constituent. Application of the Zr-in-rutile thermometer to the exsolved rutile requires an assessment of silica activity and a value of aSiO2 = 0·5 is assumed (e.g. Hayden & Watson, 2007). The estimated temperatures (800–480°C) are significantly lower than those determined for the coexisting magmatic rutile (Fig. 21b) and are consistent with exsolution of the TiO2 component of chromian spinel (i.e. ulvöspinel, Fe2TiO4) at subsolidus conditions during cooling of the Bushveld Complex. Some samples contain only magmatic rutile (e.g. DT28–912, MP24D2), others are dominated by exsolved rutile (e.g. SA04–13, SA04–06), and still others have mixed populations of magmatic and exsolved rutile (e.g. SA04–08, TW477–661). A combination of careful petrography and imaging and trace element geochemistry can successfully resolve the type of rutile present in any given sample. Fig. 21 View largeDownload slide Summary of thermometry results for Bushveld Complex rutile. Temperatures are Zr-in-rutile temperatures calculated assuming aSiO2 = 1 for the primary magmatic rutile grains and aSiO2 = 0·5 for the rutile exsolved from chromite (Ferry & Watson, 2007). (a) Zr-in-rutile temperature vs Nb concentration, showing the high temperatures characteristic of magmatic rutile that crystallized from interstitial melt in the cumulates of the Upper Critical Zone and the anomalously low estimated temperatures for rutile that exsolved from chromite. Note the logarithmic scale for Nb. (b) Histogram showing the distribution of calculated Zr-in-rutile temperatures with a clear distinction between the two types of rutile. Fig. 21 View largeDownload slide Summary of thermometry results for Bushveld Complex rutile. Temperatures are Zr-in-rutile temperatures calculated assuming aSiO2 = 1 for the primary magmatic rutile grains and aSiO2 = 0·5 for the rutile exsolved from chromite (Ferry & Watson, 2007). (a) Zr-in-rutile temperature vs Nb concentration, showing the high temperatures characteristic of magmatic rutile that crystallized from interstitial melt in the cumulates of the Upper Critical Zone and the anomalously low estimated temperatures for rutile that exsolved from chromite. Note the logarithmic scale for Nb. (b) Histogram showing the distribution of calculated Zr-in-rutile temperatures with a clear distinction between the two types of rutile. The discovery of accessory rutile in many rocks of the Bushveld Complex opens up new possibilities for examining geochemical pathways during late-stage consolidation of mafic–ultramafic cumulates and the timescales of cooling and uplift in layered intrusions. As this study demonstrates, the distinctive trace element chemistry of rutile allows for discriminating between different types of rutile even within individual samples. Isotopic fingerprinting of rutile (Hf–O isotopes) from layered intrusions holds promise to complement existing whole-rock and mineral isotopic studies (e.g. Rb–Sr, Sm–Nd, Lu–Hf, Pb–Pb) that provide both source constraints and reveal processes involved in their crystallization and consolidation (DePaolo & Wasserburg, 1979 [Stillwater]; Stewart & DePaolo, 1990 [Skaergaard]; Chutas et al., 2012 [Bushveld]; VanTongeren et al., 2016 [Bushveld]). Rutile can also be exploited as a mineral chronometer in layered intrusions using combined U–Th–Pb geochronology and (U–Th)/He thermochronology to establish complete cooling histories for mafic layered intrusions. U–Pb dating of Bushveld Complex rutile, where the closure temperature for Pb diffusion is in the range of 400–600°C (Cherniak, 2000; Schmitz & Bowring, 2003) yields dates (c.2053 Ma) that are ∼4 million years younger than the U–Pb zircon dates (c.2057 Ma) from the same rocks, and, consequently, a cooling history for this major layered intrusion can be constructed (Scoates & Wall, 2015). For (U–Th)/He thermochronology, the closure temperature of He diffusion in rutile is much lower (∼100–200°C) and is dependent on grain size and cooling rate (Stockli et al., 2007; Cherniak & Watson, 2011). This low-temperature chronometer can be used to determine rates of exhumation and erosion of layered intrusions, an as yet uninvestigated research topic. Based on the results of this study, we predict that rutile is undoubtedly present in other major layered intrusions. Target rocks should include those that contain chromite and especially those that contain crystallized pockets of interstitial minerals that represent the final crystallization products of fractionated interstitial melt (e.g. quartz, alkali feldspar, biotite, apatite, zircon). CONCLUSIONS The trace element geochemistry of zircon and rutile determined by LA-ICP-MS from a suite of samples spanning nearly the entire magmatic stratigraphy of the Bushveld Complex establishes a temperature-composition framework for crystallization of fractionated interstitial melt in mafic–ultramafic cumulates and associated granitic magmas. Ti-in-zircon thermometry yields near-solidus temperatures (950–730°C) for the mafic–ultramafic cumulates of the Lower Zone, Critical Zone, and Main Zone, and notably cooler temperatures (875–690°C) for an Upper Zone diorite and the overlying felsic roof rocks. Forward modelling using rhyolite-MELTS for the crystallization of proposed Bushveld Complex parental magmas produces zircon saturation (800–740°C) from highly fractionated melts (∼5–20% remaining melt) and late-stage, near-solidus mineral assemblages that are the same as to those observed in the rocks (e.g. quartz, Na-plagioclase, K-feldspar, biotite). High-Th/U zircon from the Critical Zone appears to reflect U loss from the interstitial melt during exsolution of an oxidized Cl-rich fluid. Identification of two distinct morphological and geochemical types of rutile in the Bushveld Complex, needle-like magmatic rutile with typically high HFSE contents that crystallized from fractionated interstitial melt (Zr-in-rutile thermometry = 1000–800°C) along with zircon and rutile exsolved from chromite (800–480°C) with anomalously low HFSE, provides a new tool for evaluating the consolidation of mafic–ultramafic cumulates and the timescales of cooling and uplift in layered intrusions. Exploring the near-solidus evolution of mafic layered intrusions such as the Bushveld Complex using the trace element chemistry of accessory minerals is a novel approach to constraining the late stages of crystallization from highly fractionated interstitial melt in these petrologically important intrusions. ACKNOWLEDGEMENTS We are grateful to Drs. Ed Mathez and Jill VanTongeren for their careful field work and sampling of the Bushveld Complex and for providing many of the samples used in this study from the collection at the American Museum of Natural History (New York). Thanks to Hai Lin for help with sample processing (crushing, Wilfley table, heavy liquids, magnetic separation). Many thanks to Dr Mati Raudsepp and to Elisabetta Pani, Jenny Lai, Lan Kato, and Edith Czech from the Electron Microbeam and X-ray Diffraction Facility at UBC for assistance with the SEM and electron microprobe. We appreciate the useful comments from Mati Raudsepp on an earlier version of this manuscript and the detailed edits of Jon Scoates on the submitted version. We thank Steve Prevec and Christian Tegner for careful and constructive journal reviews of the manuscript and Marjorie Wilson for editorial handling and additional comments. FUNDING Funding for this project came from a Natural Sciences and Engineering Research Council (NSERC) Collaborative Research and Training Experience (CREATE) grant for salary to Tom Ver Hoeve through the Multidisciplinary Applied Geochemistry Network (MAGNET). Support for the research and analytical work was provided by Natural Sciences and Engineering Research Council (NSERC) Discovery Grants to James Scoates and Dominique Weis. REFERENCES Bacon C. R. , Sisson T. W. , Mazdab F. K. ( 2007 ). Young cumulate complex beneath Veniaminof caldera, Aleutian arc, dated by zircon in erupted plutonic blocks . Geology 35 , 491 – 494 . Google Scholar Crossref Search ADS Barboni M. , Schoene B. ( 2014 ). Short eruption window revealed by absolute crystal growth rates in a granitic magma . Nature Geoscience 7 , 524 – 528 . Google Scholar Crossref Search ADS Barnes S. , Maier W. , Curl E. ( 2010 ). 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 . Economic Geology 105 , 1491 – 1511 . Google Scholar Crossref Search ADS Bédard J. H. ( 2006 ). Trace element partitioning in plagioclase feldspar . Geochimica et Cosmochimica Acta 70 , 3717 – 3742 . Google Scholar Crossref Search ADS Bédard J. H. ( 2007 ). Trace element partitioning coefficients between silicate melts and orthopyroxene: parameterizations of D variations . Chemical Geology 244 , 263 – 303 . Google Scholar Crossref Search ADS Belousova E. A. , Griffin W. L. , O’Reilly S. Y. ( 2006 ). Zircon crystal morphology, trace element signatures and Hf isotope composition as a tool for petrogenetic modelling: examples from Eastern Australian granitoids . Journal of Petrology 47 , 329 – 353 . Google Scholar Crossref Search ADS Belousova E. , Griffin W. , O’Reilly S. Y. , Fisher N. ( 2002 ). Igneous zircon: trace element composition as an indicator of source rock type . Contributions to Mineralogy and Petrology 143 , 602 – 622 . Google Scholar Crossref Search ADS Belousova E. A. , Kostitsyn Y. A. , Griffin W. L. , Begg G. C. , O’Reilly S. Y. , Pearson N. J. ( 2010 ). The growth of the continental crust: constraints from zircon Hf-isotope data . Lithos 119 , 457 – 466 . Google Scholar Crossref Search ADS Boehnke P. , Watson E. , Trail D. , Harrison T. M. , Schmitt A. K. ( 2013 ). Zircon saturation re-revisited . Chemical Geology 351 , 324 – 334 . Google Scholar Crossref Search ADS Borai E. H. , Eid M. A. , Aly H. F. ( 2002 ). Determination of REEs distribution in monazite and xenotime minerals by ion chromatography and ICP-AES . Analytical and Bioanalytical Chemistry 372 , 537 – 541 . Google Scholar Crossref Search ADS PubMed Boudreau A. ( 1999 ). Fluid fluxing of cumulates: the J-M Reef and associated aocks of the Stillwater Complex, Montana . Journal of Petrology 40 , 755 – 772 . Google Scholar Crossref Search ADS Boudreau A. E. , McCallum I. S. ( 1989 ). Investigations of the Stillwater Complex: part V. apatites as indicators of evolving fluid composition . Contributions to Mineralogy and Petrology 102 , 138 – 153 . Google Scholar Crossref Search ADS Boudreau A. E. , Mathez E. A. , McCallum I. S. ( 1986 ). Halogen geochemistry of the Stillwater and Bushveld Complexes: evidence for transport of the platinum-group elements by Cl-rich fluids . Journal of Petrology 27 , 967 – 986 . Google Scholar Crossref Search ADS Buchanan P. C. , Koeberl C. , Reimold W. U. ( 1999 ). Petrogenesis of the Dullstroom Formation, Bushveld Magmatic Province, South Africa . Contributions to Mineralogy and Petrology 137 , 133 – 146 . Google Scholar Crossref Search ADS Buchanan P. , Reimold W. , Koeberl C. , Kruger F. ( 2002 ). Geochemistry of intermediate to siliceous volcanic rocks of the Rooiberg Group, Bushveld Magmatic Province, South Africa . Contributions to Mineralogy and Petrology 144 , 131 – 143 . Google Scholar Crossref Search ADS Burnham A. D. , Berry A. J. ( 2014 ). The effect of oxygen fugacity, melt composition, temperature and pressure on the oxidation state of cerium in silicate melts . Chemical Geology 366 , 52 – 60 . Google Scholar Crossref Search ADS Cameron E. N. ( 1978 ). The Lower Zone of the Eastern Bushveld Complex in the Olifants River trough . Journal of Petrology 19 , 437 – 462 . Google Scholar Crossref Search ADS Cameron E. N. ( 1979 ). Titanium-bearing oxide minerals of the Critical Zone of the Eastern Bushveld Complex . American Mineralogist 64 , 140 – 150 . Cameron E. N. ( 1980 ). Evolution of the Lower Critical Zone, Central Sector, Eastern Bushveld Complex, and its chromite deposits . Economic Geology 75 , 845 – 871 . Google Scholar Crossref Search ADS Cameron E. N. , Emerson M. E. ( 1959 ). The origin of certain chromite deposits of the eastern part of the Bushveld complex . Economic Geology 54 , 1151 – 1213 . Google Scholar Crossref Search ADS Cawthorn R. G. (ed.) ( 1996 ). Layered Intrusions . Amsterdam : Elsevier , p. 531 . Cawthorn R. G. ( 2013a ). Rare earth element abundances in apatite in the Bushveld Complex-a consequence of the trapped liquid shift effect . Geology 41 , 603 – 606 . Google Scholar Crossref Search ADS Cawthorn R. G. ( 2013b ). The Residual or Roof Zone of the Bushveld Complex, South Africa . Journal of Petrology 54 , 1875 – 1900 . Google Scholar Crossref Search ADS Cawthorn R. G. ( 2015 ). The Bushveld Complex, South Africa. In: Charlier B. , Namur O. , Latypov R. , Tegner C. (eds) Layered Intrusions . Dordrecht, Netherlands : Springer Science , pp. 517 – 587 . Cawthorn R. G. , Walraven F. ( 1998 ). Emplacement and crystallization time for the Bushveld Complex . Journal of Petrology 39 , 1669 – 1687 . Google Scholar Crossref Search ADS Cawthorn R. G. , Webb S. J. ( 2001 ). Connectivity between the western and eastern limbs of the Bushveld Complex . Tectonophysics 330 , 195 – 209 . Google Scholar Crossref Search ADS Cawthorn R. G. , Eales H. V. , Walraven F. , Uken R. , Watkeys M. K. ( 2009 ). The Bushveld Complex. In: Johnson M. R. , Anhaeusser C. R. , Thomas R. J. (eds) The Geology of South Africa . Pretoria: Geological Society of South Africa , pp. 261 – 272 . Charlier B. , Namur O. , Latypov R. , Tegner C. (eds) ( 2015 ). Layered Intrusions . Dordrecht : Springer Netherlands , p. 531 . Cherniak D. J. ( 2000 ). Pb diffusion in rutile . Contributions to Mineralogy and Petrology 139 , 198 – 207 . Google Scholar Crossref Search ADS Cherniak D. J. ( 2010 ). Diffusion in accessory minerals: zircon, titanite, apatite, monazite and xenotime. In: Zhang Y. , Cherniak D. J. (eds) Diffusion in Minerals and Melts . Reviews in Mineralogy and Geochemistry 72, 827 – 869 . Cherniak D. J. , Watson E. B. ( 2001 ). Pb diffusion in zircon . Chemical Geology 172 , 5 – 24 . Google Scholar Crossref Search ADS Cherniak D. J. , Watson E. B. ( 2003 ). Diffusion in zircon. In: Hanchar J. M. , Hoskin P. W. O. (eds) Zircon. Reviews in Mineralogy and Geochemistry 53, 113 – 143 . Cherniak D. J. , Watson E. B. ( 2011 ). Helium diffusion in rutile and titanite, and consideration of the origin and implications of diffusional anisotropy . Chemical Geology 288 , 149 – 161 . Google Scholar Crossref Search ADS Chutas N. I. , Bates E. , Prevec S. A. , Coleman D. S. , Boudreau A. E. ( 2012 ). Sr and Pb isotopic disequilibrium between coexisting plagioclase and orthopyroxene in the Bushveld Complex, South Africa: microdrilling and progressive leaching evidence for sub-liquidus contamination within a crystal mush . Contributions to Mineralogy and Petrology 163 , 653 – 668 . Google Scholar Crossref Search ADS Claiborne L. L. , Miller C. F. , Wooden J. L. ( 2010 ). Trace element composition of igneous zircon: a thermal and compositional record of the accumulation and evolution of a large silicic batholith, Spirit Mountain, Nevada . Contributions to Mineralogy and Petrology 160 , 511 – 531 . Google Scholar Crossref Search ADS DePaolo D. , Wasserburg G. ( 1979 ). Sm-Nd age of the Stillwater Complex and the mantle evolution curve for neodymium . Geochimica et Cosmochimica Acta 43 , 999 – 1008 . Google Scholar Crossref Search ADS Eales H. V. , Botha W. J. , Hattingh P. J. , de Klerk W. J. , Maier W. D. , Odgers A. T. R. ( 1993 ). The mafic rocks of the Bushveld Complex: a review of emplacement and crystallization history, and mineralization, in the light of recent data . Journal of African Earth Sciences 16 , 121 – 142 . Google Scholar Crossref Search ADS Eales H. V. , Cawthorn R. G. ( 1996 ). The Bushveld Complex. In: Cawthorn R. G. (ed.) Layered Intrusions . Amsterdam : Elsevier , pp. 181 – 229 . Eales H. V. , de Klerk W. J. , Teigler B. ( 1990 ). Evidence for magma mixing processes within the Critical and Lower Zones of the northwestern Bushveld Complex, South Africa . Chemical Geology 88 , 261 – 278 . Google Scholar Crossref Search ADS Ferry J. M. , Watson E. B. ( 2007 ). New thermodynamic models and revised calibrations for the Ti-in-zircon and Zr-in-rutile thermometers . Contributions to Mineralogy and Petrology 154 , 429 – 437 . Google Scholar Crossref Search ADS Finn C. A. , Bedrosian P. A. , Cole J. C. , Khoza T. D. , Webb S. J. ( 2015 ). Mapping the 3D extent of the Northern Lobe of the Bushveld layered mafic intrusion from geophysical data . Precambrian Research 268 , 279 – 294 . Google Scholar Crossref Search ADS Fourie D. S. , Harris C. ( 2011 ). O-isotope study of the Bushveld Complex granites and granophyres: constraints on source composition, and assimilation . Journal of Petrology 52 , 2221 – 2242 . Google Scholar Crossref Search ADS Geisler T. , Schaltegger U. , Tomaschek F. ( 2007 ). Re-equilibration of zircon in aqueous fluids and melts . Elements 3 , 43 – 50 . Google Scholar Crossref Search ADS Ghisler M. ( 1970 ). Pre-metamorphic folded chromite deposits of stratiform type in the early Precambrian of West Greenland . Mineralium Deposita 5 , 223 – 236 . Google Scholar Crossref Search ADS Ghosh B. , Konar R. ( 2011 ). Chromites from meta-anorthosites, Sittampundi layered igneous complex, Tamil Nadu, southern India . Journal of Asian Earth Sciences 42 , 1394 – 1402 . Google Scholar Crossref Search ADS Griffin W. L. , Wang X. , Jackson S. E. , Pearson N. J. , O’Reilly S. Y. , Xu X. , Zhou X. ( 2002 ). Zircon chemistry and magma mixing, SE China: in-situ analysis of Hf isotopes, Tonglu and Pingtan igneous complexes . Lithos 61 , 237 – 269 . Google Scholar Crossref Search ADS Grimes C. B. , John B. E. , Cheadle M. J. , Mazdab F. K. , Wooden J. L. , Swapp S. , Schwartz J. J. ( 2009 ). On the occurrence, trace element geochemistry, and crystallization history of zircon from in situ ocean lithosphere . Contributions to Mineralogy and Petrology 158 , 757 – 783 . Google Scholar Crossref Search ADS Grimes C. B. , John B. E. , Kelemen P. B. , Mazdab F. K. , Wooden J. L. , Cheadle M. J. , Hanghøj K. , Schwartz J. J. ( 2007 ). Trace element chemistry of zircons from oceanic crust: a method for distinguishing detrital zircon provenance . Geology 35 , 643 – 646 . Google Scholar Crossref Search ADS Grimes C. B. , Wooden J. L. , Cheadle M. J. , John B. E. ( 2015 ). “ Fingerprinting” tectono-magmatic provenance using trace elements in igneous zircon . Contributions to Mineralogy and Petrology 170 , 46 . Google Scholar Crossref Search ADS Gualda G. A. R. , Ghiorso M. S. , Lemons R. V. , Carley T. L. ( 2012 ). Rhyolite-MELTS: a modified calibration of MELTS optimized for silica-rich, fluid-bearing magmatic systems . Journal of Petrology 53 , 875 – 890 . Google Scholar Crossref Search ADS Hall A. L. ( 1932 ). The Bushveld igneous complex of the central Transvaal . South African Geological Survey Memoir 28 , 560 . Harmer R. E. , Sharpe M. R. ( 1985 ). Field relations and strontium isotope systematics of the marginal rocks of the eastern Bushveld complex . Economic Geology 80 , 813 – 837 . Google Scholar Crossref Search ADS Hawkesworth C. J. , Kemp A. I. S. ( 2006 ). Using hafnium and oxygen isotopes in zircons to unravel the record of crustal evolution . Chemical Geology 226 , 144 – 162 . Google Scholar Crossref Search ADS Hayden L. A. , Watson E. B. ( 2007 ). Rutile saturation in hydrous siliceous melts and its bearing on Ti-thermometry of quartz and zircon . Earth and Planetary Science Letters 258 , 561 – 568 . Google Scholar Crossref Search ADS Heaman L. M. , LeCheminant A. N. ( 1993 ). Paragenesis and U-Pb systematics of baddeleyite (ZrO2) . Chemical Geology 110 , 95 – 126 . Google Scholar Crossref Search ADS Holness M. B. , Vernon R. H. ( 2015 ). The influence of interfacial energies on igneous microstructures. In: Charlier B. , Namur O. , Latypov R. , Tegner C. (eds) Layered Intrusions . Dordrecht, Netherlands : Springer Science , pp. 183 – 228 . Holness M. B. , Nielsen T. F. D. , Tegner C. ( 2006 ). Textural maturity of cumulates: a record of chamber filling, liquidus assemblage, cooling rate and large-scale convection in mafic layered intrusions . Journal of Petrology 48 , 141 – 157 . Google Scholar Crossref Search ADS Holness M. B. , Stripp G. , Humphreys M. C. S. , Veksler I. V. , Nielsen T. F. D. , Tegner C. ( 2011 ). Silicate liquid immiscibility within the crystal mush: late-stage magmatic microstructures in the Skaergaard Intrusion, East Greenland . Journal of Petrology 52 , 175 – 222 . Google Scholar Crossref Search ADS Hoskin P. W. O. ( 2005 ). Trace-element composition of hydrothermal zircon and the alteration of Hadean zircon from the Jack Hills, Australia . Geochimica et Cosmochimica Acta 69 , 637 – 648 . Google Scholar Crossref Search ADS Hoskin P. W. O. , Ireland T. R. ( 2000 ). Rare earth element chemistry of zircon and its use as a provenance indicator . Geology 28 , 627 – 630 . Google Scholar Crossref Search ADS Hoskin P. W. O. , Schaltegger U. ( 2003 ). The composition of zircon and igneous and metamorphic petrogenesis. In: Hanchar J. M. , Hoskin P. W. O. (eds) Zircon, Reviews in Mineralogy and Geochemistry 53, 27 – 62 . Iannello P. ( 1971 ). The Bushveld granites around Rooiberg, Transvaal, South Africa . Geologische Rundschau 60 , 630 – 655 . Google Scholar Crossref Search ADS Irvine T. N. , Andersen J. C. Ø. , Brooks C. K. ( 1998 ). Included blocks (and blocks within blocks) in the Skaergaard intrusion: Geologic relations and the origins of rhythmic modally graded layers . Bulletin of the Geological Society of America 110 , 1398 – 1447 . Google Scholar Crossref Search ADS Keppler H. , Wyllie P. J. ( 1991 ). Partitioning of Cu, Sn, Mo, W, U, and Th between melt and aqueous fluid in the systems haplogranite-H2O-HCl and haplogranite-H2O-HF . Contributions to Mineralogy and Petrology 109 , 139 – 150 . Google Scholar Crossref Search ADS Kinnaird J. A. , Hutchinson D. , Schurmann L. , Nex P. A. M. , de Lange R. ( 2005 ). Petrology and mineralisation of the southern Platreef: northern limb of the Bushveld Complex, South Africa . Mineralium Deposita 40 , 576 – 597 . Google Scholar Crossref Search ADS Kirkland C. L. , Smithies R. H. , Taylor R. J. M. , Evans N. , McDonald B. ( 2015 ). Zircon Th/U ratios in magmatic environs . Lithos 212–215 , 397 – 414 . Google Scholar Crossref Search ADS Kleemann G. J. , Twist D. ( 1989 ). The compositionally-zoned sheet-like granite pluton of the Bushveld Complex—evidence bearing on the nature of A-type magmatism . Journal of Petrology 30 , 1383 – 1414 . Google Scholar Crossref Search ADS Kruger F. J. , Cawthorn R. G. , Walsh K. L. ( 1987 ). Strontium isotopic evidence against magma addition in the Upper Zone of the Bushveld Complex . Earth and Planetary Science Letters 84 , 51 – 58 . Google Scholar Crossref Search ADS Luvizotto G. L. , Zack T. , Meyer H. P. , Ludwig T. , Triebold S. , Kronz A. , Münker C. , Stockli D. F. , Prowatke S. , Klemme S. , Jacob D. E. , von Eynatten H. ( 2009 ). Rutile crystals as potential trace element and isotope mineral standards for microanalysis . Chemical Geology 261 , 346 – 369 . Google Scholar Crossref Search ADS Maier W. D. ( 2005 ). Platinum-group element (PGE) deposits and occurrences: mineralization styles, genetic concepts, and exploration criteria . Journal of African Earth Sciences 41 , 165 – 191 . Google Scholar Crossref Search ADS Maier W. D. , Barnes S. J. ( 2010 ). The petrogenesis of platinum-group element reefs in the Upper Main Zone of the Northern Lobe of the Bushveld Complex on the Farm Moorddrift, South Africa . Economic Geology 105 , 841 – 854 . Google Scholar Crossref Search ADS Maier W. D. , Barnes S.-J. , Groves D. I. ( 2013 ). The Bushveld Complex, South Africa: formation of platinum–palladium, chrome- and vanadium-rich layers via hydrodynamic sorting of a mobilized cumulate slurry in a large, relatively slowly cooling, subsiding magma chamber . Mineralium Deposita 48 , 1 – 56 . Google Scholar Crossref Search ADS Maier W. D. , Li C. , De Waal S. A. ( 2001 ). Why are there no major Ni-Cu sulfide deposits in large layered mafic-ultramafic intrusions? Canadian Mineralogist 39 , 547 – 556 . Google Scholar Crossref Search ADS Manor M. J. , Scoates J. S. , Wall C. J. , Nixon G. T. , Friedman R. M. , Amini M. , Ames D. E. ( 2017 ). Age of the Late Cretaceous ultramafic-hosted Giant Mascot Ni-Cu-PGE deposit, Southern Canadian Cordillera: integrating CA-ID-TIMS and LA-ICP-MS U-Pb geochronology and trace element geochemistry of zircon . Economic Geology 112 , 1395 – 1418 . Google Scholar Crossref Search ADS Mathez E. A. ( 1995 ). Magmatic metasomatism and formation of the Merensky reef, Bushveld Complex . Contributions to Mineralogy and Petrology 119 , 277 – 286 . Google Scholar Crossref Search ADS Mathez E. A. , Hunter R. H. , Kinzler R. ( 1997 ). Petrologic evolution of partially molten cumulate: the Atok section of the Bushveld Complex . Contributions to Mineralogy and Petrology 129 , 20 – 34 . Google Scholar Crossref Search ADS Mathez E. A. , Mey J. L. ( 2005 ). Character of the UG2 chromitite and host rocks and petrogenesis of its pegmatoidal footwall, northeastern Bushveld Complex . Economic Geology 100 , 1617 – 1630 . Google Scholar Crossref Search ADS Mathez E. A. , VanTongeren J. A. , Schweitzer J. ( 2013 ). On the relationships between the Bushveld Complex and its felsic roof rocks, part 1: petrogenesis of Rooiberg and related felsites . Contributions to Mineralogy and Petrology 166 , 435 – 449 . Google Scholar Crossref Search ADS Mattinson J. M. ( 2005 ). Zircon U-Pb chemical abrasion (“CA-TIMS”) method: combined annealing and multi-step partial dissolution analysis for improved precision and accuracy of zircon ages . Chemical Geology 220 , 47 – 66 . Google Scholar Crossref Search ADS McDonough W. F. , Sun S.-S. ( 1995 ). The composition of the Earth . Chemical Geology 120 , 223 – 253 . Google Scholar Crossref Search ADS Meinhold G. ( 2010 ). Rutile and its applications in earth sciences . Earth-Science Reviews 102 , 1 – 28 . Google Scholar Crossref Search ADS Meurer W. P. , Meurer M. E. S. ( 2006 ). Using apatite to dispel the “trapped liquid” concept and to understand the loss of interstitial liquid by compaction in mafic cumulates: an example from the Stillwater Complex, Montana . Contributions to Mineralogy and Petrology 151 , 187 – 201 . Google Scholar Crossref Search ADS Miller J. S. , Wooden J. L. ( 2004 ). Residence, resorption and recycling of zircons in Devils Kitchen Rhyolite, Coso Volcanic Field, California . Journal of Petrology 45 , 2155 – 2170 . Google Scholar Crossref Search ADS Molyneaux T. ( 2008 ). Compilation on a Scale of 1: 50000 of the Geology of the Eastern Compartment of the Bushveld Complex. Special Publication 1 Johannes Willemse Bushveld Intelligence Centre, Department of Geology, University of Pretoria, Pretoria. Mondal S. K. , Mathez E. A. ( 2006 ). Origin of the UG2 chromitite layer, Bushveld Complex . Journal of Petrology 48 , 495 – 510 . Google Scholar Crossref Search ADS Morisset C.-E. , Scoates J. S. , Weis D. , Rahier A. ( 2013 ). Methodology and application of hafnium isotopes in ilmenite and rutile by MC-ICP-MS . Geostandards and Geoanalytical Research 38 , 159 – 176 . Morisset C. -E. , Scoates J. S. , Weis D. , Sauvé M. , Stanaway K. J. ( 2010 ). Rutile-bearing ilmenite deposits associated with the Proterozoic Saint-Urbain and Lac Allard anorthosite massifs, Grenville Province, Quebec . Canadian Mineralogist 48 , 821 – 849 . Google Scholar Crossref Search ADS Mungall J. E. , Kamo S. L. , McQuade S. ( 2016 ). U–Pb geochronology documents out-of-sequence emplacement of ultramafic layers in the Bushveld Igneous Complex of South Africa . Nature Communications 7 , 13385 . Google Scholar Crossref Search ADS PubMed Naldrett A. J. , Gasparrini E. C. , Barnes S. J. , Von Gruenewaldt G. , Sharpe M. R. ( 1986 ). The Upper Critical Zone of the Bushveld complex and the origin of Merensky-type ores . Economic Geology 81 , 1105 – 1117 . Google Scholar Crossref Search ADS Naldrett A. J. , Wilson A. , Kinnaird J. , Yudovskaya M. , Chunnett G. ( 2012 ). The origin of chromitites and related PGE mineralization in the Bushveld Complex: new mineralogical and petrological constraints . Mineralium Deposita 47 , 209 – 232 . Google Scholar Crossref Search ADS Oberthür T. , Davis D. W. , Blenkinsop T. G. , Höhndorf A. ( 2002 ). Precise U-Pb mineral ages, Rb-Sr and Sm-Nd systematics for the Great Dyke, Zimbabwe—constraints on late Archean events in the Zimbabwe craton and Limpopo belt . Precambrian Research 113 , 293 – 305 . Google Scholar Crossref Search ADS Pagé P. , Barnes S. J. , Bédard J. H. , Zientek M. L. ( 2012 ). In situ determination of Os, Ir, and Ru in chromites formed from komatiite, tholeiite and boninite magmas: implications for chromite control of Os, Ir and Ru during partial melting and crystal fractionation . Chemical Geology 302–303 , 3 – 15 . Google Scholar Crossref Search ADS Parsons I . (ed .) ( 1987 ). Origins of Igneous Layering NATO ASI Series C 196 . Dordrecht, Holland : D. Reidel Publishing Company , p. 666 . Paton C. , Hellstrom J. , Paul B. , Woodhead J. , Hergt J. ( 2011 ). Iolite: freeware for the visualisation and processing of mass spectrometric data . Journal of Analytical Atomic Spectrometry 26 , 2508 – 2518 . Google Scholar Crossref Search ADS Pearce N. J. G. , Perkins W. T. , Westgate J. A. , Gorton M. P. , Jackson S. E. , Neal C. R. , Chenery S. P. ( 1997 ). A compilation of new and published major and trace element data for NIST SRM 610 and NIST SRM 612 glass reference materials . Geostandards and Geoanalytical Research 21 , 115 – 144 . Google Scholar Crossref Search ADS Rioux M. , Lissenberg C. J. , McLean N. M. , Bowring S. A. , MacLeod C. J. , Hellebrand E. , Shimizu N. ( 2012 ). Protracted timescales of lower crustal growth at the fast-spreading East Pacific Rise . Nature Geoscience 5 , 275 – 278 . Google Scholar Crossref Search ADS Rubatto D. , Hermann J. ( 2007 ). Experimental zircon/melt and zircon/garnet trace element partitioning and implications for the geochronology of crustal rocks . Chemical Geology 241 , 38 – 61 . Google Scholar Crossref Search ADS Rudnick R. L. ( 2000 ). Rutile-bearing refractory eclogites: missing link between continents and depleted mantle . Science 287 , 278 – 281 . Google Scholar Crossref Search ADS PubMed Samperton K. M. , Schoene B. , Cottle J. M. , Brenhin Keller C. , Crowley J. L. , Schmitz M. D. ( 2015 ). Magma emplacement, differentiation and cooling in the middle crust: integrated zircon geochronological–geochemical constraints from the Bergell Intrusion, Central Alps . Chemical Geology 417 , 322 – 340 . Google Scholar Crossref Search ADS Schmidt A. , Weyer S. , John T. , Brey G. P. ( 2009 ). HFSE systematics of rutile-bearing eclogites: new insights into subduction zone processes and implications for the earth’s HFSE budget . Geochimica et Cosmochimica Acta 73 , 455 – 468 . Google Scholar Crossref Search ADS Schmitz M. D. , Bowring S. A. ( 2003 ). Constraints on the thermal evolution of continental lithosphere from U-Pb accessory mineral thermochronometry of lower crustal xenoliths, southern Africa . Contributions to Mineralogy and Petrology 144 , 592 – 618 . Google Scholar Crossref Search ADS Schoene B. ( 2013 ). U-Th-Pb Geochronology. Treatise on Geochemistry: Second Edition . Amsterdam : Elsevier , pp. 341 – 378 . Schweitzer J. K. , Hatton C. J. , De Waal S. A. ( 1995 ). Regional lithochemical stratigraphy of the Rooiberg Group, upper Transvaal . South African Journal of Geology 98 , 245 – 255 . Schweitzer J. K. , Hatton C. J. , De Waal S. A. ( 1997 ). Link between the granitic and volcanic rocks of the Bushveld Complex, South Africa . Journal of African Earth Sciences 24 , 95 – 104 . Google Scholar Crossref Search ADS Scoates J. S. , Chamberlain K. R. ( 1995 ). Baddeleyite (ZrO2) and zircon (ZrSiO4) from anorthositic rocks of the Laramie anorthosite complex, Wyoming: petrologic consequences and U-Pb ages . American Mineralogist 80 , 1317 – 1327 . Google Scholar Crossref Search ADS Scoates J. S. , Friedman R. M. ( 2008 ). Precise age of the platiniferous Merensky Reef, Bushveld Complex, South Africa, by the U-Pb zircon chemical abrasion ID-TIMS technique . Economic Geology 103 , 465 – 471 . Google Scholar Crossref Search ADS Scoates J. S. , Wall C. J. ( 2015 ). Geochronology of layered intrusions. In: Charlier B. , Namur O. , Latypov R. , Tegner C. (eds) Layered Intrusions . Dordrecht, Netherlands : Springer Science , pp. 3 – 74 . Shannon R. D. ( 1976 ). Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides . Acta Crystallographica Section A 32 , 751 – 767 . Google Scholar Crossref Search ADS Sharpe M. R. ( 1981 ). The chronology of magma influxes to the eastern compartment of the Bushveld Complex as exemplified by its marginal border groups . Journal of the Geological Society 138 , 307 – 326 . Google Scholar Crossref Search ADS Sharpe M. R. ( 1982 ). Noble metals in the marginal rocks of the Bushveld Complex . Economic Geology 77 , 1286 – 1295 . Google Scholar Crossref Search ADS Sharpe M. R. ( 1985 ). Strontium isotope evidence for preserved density stratification in the main zone of the Bushveld Complex, South Africa . Nature 316 , 119 – 126 . Google Scholar Crossref Search ADS Silver L. T. , Deutsch S. ( 1963 ). Uranium-lead isotopic variations in zircons: a case study . The Journal of Geology 71 , 721 – 758 . Google Scholar Crossref Search ADS Stewart B. W. , DePaolo D. J. ( 1990 ). Isotopic studies of processes in mafic magma chambers: II. The Skaergaard Intrusion, East Greenland . Contributions to Mineralogy and Petrology 104 , 125 – 141 . Google Scholar Crossref Search ADS Stockli D. F. , Wolfe M. R. , Blackburn T. J. , Zack T. , Walker J. D. , Luvizotto G. L. ( 2007 ). He diffusion and (U–Th)/He thermochronometry of rutile. Transactions of the American Geophysical Union, Fall Meeting Supplement, Abstract V23C-1548, San Francisco. Tegner C. , Cawthorn R. G. , Kruger F. J. ( 2006 ). Cyclicity in the Main and Upper Zones of the Bushveld Complex, South Africa: crystallization from a zoned magma sheet . Journal of Petrology 47 , 2257 – 2279 . Google Scholar Crossref Search ADS Tepley F. J. , Davidson J. P. ( 2003 ). Mineral-scale Sr-isotope constraints on magma evolution and chamber dynamics in the Rum layered intrusion, Scotland . Contributions to Mineralogy and Petrology 145 , 628 – 641 . Google Scholar Crossref Search ADS Tomkins H. S. , Powell R. , Ellis D. J. ( 2007 ). The pressure dependence of the zirconium-in-rutile thermometer . Journal of Metamorphic Geology 25 , 703 – 713 . Google Scholar Crossref Search ADS Trail D. , Bruce Watson E. , Tailby N. D. ( 2012 ). Ce and Eu anomalies in zircon as proxies for the oxidation state of magmas . Geochimica et Cosmochimica Acta 97 , 70 – 87 . Google Scholar Crossref Search ADS Twist D. , French B. M. ( 1983 ). Voluminous acid volcanism in the Bushveld Complex: a review of the Rooiberg Felsite . Bulletin Volcanologique 46 , 225 – 242 . Google Scholar Crossref Search ADS Valley J. W. ( 2003 ). Oxygen isotopes in zircon. In: Hanchar J. M. , Hoskin P. W. O. (eds.) Zircon. Reviews in Mineralogy and Geochemistry 53, 343 – 385 . VanTongeren J. A. , Mathez E. A. ( 2012 ). Large-scale liquid immiscibility at the top of the Bushveld Complex, South Africa . Geology 40 , 491 – 494 . Google Scholar Crossref Search ADS VanTongeren J. A. , Mathez E. A. ( 2015 ). On the relationship between the Bushveld Complex and its felsic roof rocks, part 2: the immediate roof . Contributions to Mineralogy and Petrology 170 , 56 . Google Scholar Crossref Search ADS VanTongeren J. A. , Mathez E. A. , Kelemen P. B. ( 2010 ). A felsic end to Bushveld differentiation . Journal of Petrology 51 , 1891 – 1912 . Google Scholar Crossref Search ADS VanTongeren J. A. , Zirakparvar N. A. , Mathez E. A. ( 2016 ). Hf isotopic evidence for a cogenetic magma source for the Bushveld Complex and associated felsic magmas . Lithos 248–251 , 469 – 477 . Google Scholar Crossref Search ADS Veksler I. V. , Charlier B. ( 2015 ). Silicate liquid immiscibility in layered intrusions. In: Charlier B. , Namur O. , Latypov R. , Tegner C. (eds) Layered Intrusions . Dordrecht, Netherlands : Springer Science , pp. 229 – 258 . Ver Hoeve T. J. , Scoates J. S. , Wall C. J. , Weis D. , Amini M. ( 2018 ). Evaluating downhole fractionation corrections in LA-ICP-MS U-Pb zircon geochronology . Chemical Geology 483 , 201 – 217 . Google Scholar Crossref Search ADS von Gruenewaldt G. ( 1973 ). The Main and Upper zones of the Bushveld Complex in the Roossenekal area, eastern Transvaal . Transactions of the Geological Society of South Africa 76 , 207 – 227 . Vukmanovic Z. , Barnes S. J. , Reddy S. M. , Godel B. , Fiorentini M. L. ( 2013 ). Morphology and microstructure of chromite crystals in chromitites from the Merensky Reef (Bushveld Complex, South Africa) . Contributions to Mineralogy and Petrology 165 , 1031 – 1050 . Google Scholar Crossref Search ADS Wager L. , Brown G. ( 1968 ). Layered Igneous Rocks . Edinburgh : WH Freeman , p. 588 . Wall C. J. , Scoates J. S. ( 2016 ). High-precision U-Pb zircon-baddeleyite dating of the J-M Reef platinum group element deposit in the Stillwater Complex, Montana (USA) . Economic Geology 111 , 771 – 782 . Google Scholar Crossref Search ADS Wall C. J. , Scoates J. S. , Weis D. , Friedman R. M. , Amini M. , Meurer W. P. ( 2018 ). The Stillwater Complex: integrating zircon geochronological and geochemical constraints on the age, emplacement history and crystallization of a large, open-system layered intrusion . Journal of Petrology 59 , 153 – 190 . Google Scholar Crossref Search ADS Wallmach T. , Hatton C. J. , De Wall S. A. , Gibson R. L. ( 1995 ). Retrogressive hydration of calc-silicate xenoliths in the eastern Bushveld Complex: evidence for late magmatic fluid movement . Journal of African Earth Sciences 21 , 633 – 646 . Google Scholar Crossref Search ADS Walraven F. ( 1985 ). Genetic aspects of the granophyric rocks of the Bushveld complex . Economic Geology 80 , 1166 – 1180 . Google Scholar Crossref Search ADS Walraven F. ( 1987 ). Textural, geochemical, and genetic aspects of the Bushveld Complex . Geological Survey of South Africa Memoir 72 , 145 . Watson E. B. ( 1979 ). Zircon saturation in felsic liquids: experimental results and applications to trace element geochemistry . Contributions to Mineralogy and Petrology 70 , 407 – 419 . Google Scholar Crossref Search ADS Watson E. B. , Harrison T. M. ( 1983 ). Zircon saturation revisited: temperature and composition effects in a variety of crustal magma types . Earth and Planetary Science Letters 64 , 295 – 304 . Google Scholar Crossref Search ADS Watson E. B. , Harrison T. M. ( 2005 ). Zircon thermometer reveals minimum melting conditions on earliest Earth . Science 308 , 841 – 844 . Google Scholar Crossref Search ADS PubMed Watson E. B. , Wark D. A. , Thomas J. B. ( 2006 ). Crystallization thermometers for zircon and rutile . Contributions to Mineralogy and Petrology 151 , 413 – 433 . Google Scholar Crossref Search ADS Webb S. J. , Ashwal L. D. , Cawthorn R. G. ( 2011 ). Continuity between eastern and western Bushveld Complex, South Africa, confirmed by xenoliths from kimberlite . Contributions to Mineralogy and Petrology 162 , 101 – 107 . Google Scholar Crossref Search ADS Whitehouse M. J. , Kamber B. S. ( 2002 ). On the overabundance of light rare earth elements in terrestrial zircons and its implication for Earth’s earliest magmatic differentiation . Earth and Planetary Science Letters 204 , 333 – 346 . Google Scholar Crossref Search ADS Willmore C. C. , Boudreau A E. , Kruger F. J. ( 2000 ). The halogen geochemistry of the Bushveld Complex, Republic of South Africa: implications for chalcophile element distribution in the Lower and Critical zones . Journal of Petrology 41 , 1517 – 1539 . Google Scholar Crossref Search ADS Wilson A. H. ( 2015 ). The earliest stages of emplacement of the Eastern Bushveld Complex: development of the Lower Zone, Marginal Zone and basal ultramafic sequence . Journal of Petrology 56 , 347 – 388 . Google Scholar Crossref Search ADS Wotzlaw J.-F. , Bindeman I. N. , Schaltegger U. , Brooks C. K. , Naslund H. R. ( 2012 ). High-resolution insights into episodes of crystallization, hydrothermal alteration and remelting in the Skaergaard intrusive complex . Earth and Planetary Science Letters 355–356 , 199 – 212 . Google Scholar Crossref Search ADS Xiang W. , Griffin W. , Jie C. ( 2011 ). U and Th contents and Th/U ratios of zircon in felsic and mafic magmatic rocks: improved zircon-melt distribution coefficients . Acta Geologica Sinica - English Edition 85 , 164 – 174 . Google Scholar Crossref Search ADS Yudovskaya M. , Kinnaird J. , Naldrett A. J. , Rodionov N. , Antonov A. , Simakin S. , Kuzmin D. ( 2013 ). Trace-element study and age dating of zircon from chromitites of the Bushveld Complex (South Africa) . Mineralogy and Petrology 107 , 915 – 942 . Google Scholar Crossref Search ADS Zack T. , Luvizottow G. L. ( 2006 ). Application of rutile thermometry to eclogites . Mineralogy and Petrology 88 , 69 – 85 . Google Scholar Crossref Search ADS Zack T. , Moraes R. , Kronz A. ( 2004a ). Temperature dependence of Zr in rutile: empirical calibration of a rutile thermometer . Contributions to Mineralogy and Petrology 148 , 471 – 488 . Google Scholar Crossref Search ADS Zack T. , von Eynatten H. , Kronz A. ( 2004b ). Rutile geochemistry and its potential use in quantitative provenance studies . Sedimentary Geology 171 , 37 – 58 . Google Scholar Crossref Search ADS Zeh A. , Ovtcharova M. , Wilson A. H. , Schaltegger U. ( 2015 ). The Bushveld Complex was emplaced and cooled in less than one million years—results of zirconology, and geotectonic implications . Earth and Planetary Science Letters 418 , 103 – 114 . Google Scholar Crossref Search ADS Zirakparvar N. , Mathez E. A. , Scoates J. S. , Wall C. J. ( 2014 ). Zircon Hf isotope evidence for an enriched mantle source for the Bushveld Igneous Complex . Contributions to Mineralogy and Petrology 168 , 1050–1018. © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)
TI - A Temperature-Composition Framework for Crystallization of Fractionated Interstitial Melt in the Bushveld Complex from Trace Element Systematics of Zircon and Rutile
JF - Journal of Petrology
DO - 10.1093/petrology/egy066
DA - 2018-07-01
UR - https://www.deepdyve.com/lp/oxford-university-press/a-temperature-composition-framework-for-crystallization-of-Bkb1J8vKB6
SP - 1383
VL - 59
IS - 7
DP - DeepDyve
ER -