Search

International audience; The syenitic rocks of South Rallier du Baty Intrusive Complex (SRBIC) represent in- trusions into the oceanic plateau basalts of the south-western Kerguelen Islands. The SRBIC was previously interpreted as a typical ring complex due to magma emplace- ment with cauldron subsidence. Our new structural and geochronological data reveal that it is a laccolith built between 11.6 and 7.9 Ma by successive injections of magma sheets around the crust–mantle boundary, with an average injection rate between 0.8 and 1.4 × 10−4 km3/year. These results establish strong similarities between the SRBIC, the only recorded example of a felsic laccolith in an oceanic intraplate setting, and many continental plutons emplaced in various geodynamic setting. The SRBIC thus has the characteristics of a continental plutonic complex emplaced in an oceanic plateau crust. We postulate the critical parameter relevant to causing such similari- ties and plutonic magmatism is crustal thickness.

This study analyzes the microstructures and deformational characteristics of spinel peridotite xenoliths from the Nógrád-Gömör Volcanic Field (NGVF), located on the northern margin of a young extensional basin presently affected by compression. The xenoliths show a wide range of microstructures, bearing the imprints of heterogeneous deformation and variable degrees of subsequent annealing. Olivine crystal preferred orientations (CPOs) have dominantly [010]-fiber and orthorhombic patterns. Orthopyroxene CPOs indicate coeval deformation with olivine. Olivine J indices correlate positively with equilibration temperatures, suggesting that the CPO strength increases with depth. In contrast, the intensity of intragranular deformation in olivine varies as a function of the sampling locality. We interpret the microstructures and CPO patterns as recording deformation by dislocation creep in a transpressional regime, which is consistent with recent tectonic evolution in the Carpathian-Pannonian region due to the convergence between the Adria microplate and the European platform. Postkinematic annealing is probably linked to percolation of metasomatism by mafic melts through the upper mantle of the NGVF prior to the eruption of the host alkali basalt. Elevated equilibration temperatures in xenoliths from the central part of the volcanic field are interpreted to be associated with the last metasomatic event, which only shortly preceded the ascent of the host magma. Despite well-developed olivine CPOs in the xenoliths, which imply a strong seismic anisotropy, the lithospheric mantle alone cannot account for the shear wave splitting delay times measured in the NGVF, indicating that deformation in both the lithosphere and the asthenosphere contributes to the observed shear wave splitting., The authors would like to thank the people who contributed to this work. We owe thanks to Fabrice Barou and David Adams for their help with EBSD-SEM analyses at Geosciences Montpellier and at CCFS Macquarie University, respectively. L?szl? Aradi is acknowledged for his help with field work, petrography and MTEX application. We are grateful for the constructive criticism of Sandra Piazolo and Jacques Pr?cigout and two anonymous reviewers, as well as for the editorial handling of Stephen Parman. Our research received financial support from a Marie Curie International Reintegration Grant (grant NAMS-230937), a postdoctoral grant (grant PD101683) of the Hungarian Scientific Research Fund (OTKA), and a Bolyai J?nos Postdoctoral Research Fellowship of the Hungarian Academy of Sciences to I. J. K., as well as from the Lend?let Pannon LitH2Oscope Research Group (Research Centre for Astronomy and Earth Sciences, Hungarian Academy of Sciences). N. L. received support from Macquarie University international PhD scholarship and project and travel funding from ARC Centre of Excellence for Core to Crust Fluid Systems (CCFS). K. H. acknowledges funding from Ministry of Economy, Industry and Competitiveness (MINECO, Spain) and the State Research Agency (AEI, Spain; grants FPDI-2013-16253 and CGL2016-81085-R). Instruments used at Macquarie University are funded by DEST Systemic Infrastructure Grants, ARC LIEF, NCRIS/AuScope, industry partners, and Macquarie University. The data used in this paper are listed in the references, tables, and supporting information. The raw EBSD data are available from the corresponding author upon request. This is the 92nd publication of the Lithosphere Fluid Research Lab (LRG), contribution 1361 from the ARC Centre of Excellence for Core to Crust Fluid Systems (www.ccfs.mq.edu.au) and 1320 from the GEMOC Key Centre (www.gemoc.mq.edu.au).

The interaction of hydrous fluids and melts with dry rocks of the lithospheric mantle inevitably modifies their viscoelastic and chemical properties due to the formation of compositionally distinct secondary phases. In addition, melt percolation and the associated metasomatic alteration of mantle rocks may also facilitate modification of the pre–existing rock texture and olivine crystallographic preferred orientation (CPO) and thus seismic properties. Here we explore the relationship between mantle metasomatism, deformation and seismic anisotropy using subduction–related mantle xenoliths from the Penghu Islands, western Taiwan. The investigated xenoliths have equilibrated at upper lithospheric mantle conditions (879 °C to 1127 °C) based on pyroxene geothermometry and show distinct variations in clinopyroxene chemical composition, texture and olivine CPO allowing for the classification of two distinct groups. Group 1 xenoliths contain rare earth element (REE) depleted clinopyroxene, show a porphyroclastic texture and olivine grains are mostly characterized by [100]–axial pattern symmetries. In contrast, REE-enriched clinopyroxene from Group 2 xenoliths occur in a fine–grained equigranular texture and coexisting olivine frequently displays [010]–axial pattern symmetries. The clinopyroxene compositions are indicative of cryptic and modal to stealth metasomatic alteration of Group 1 and Group 2 xenoliths, respectively. Furthermore, the observed olivine [100]–axial pattern of Group 1 xenoliths reflects deformation by dislocation creep at high temperature, low pressure and dry conditions, whereas olivine [010]–axial patterns of Group 2 xenoliths imply activation of olivine [001] glide planes along preferentially wet (010) grain boundaries. This correlation indicates that the variation in olivine CPO symmetry from [100]– to [010]–axial pattern in Penghu xenoliths results from deformation and intra-crystalline recovery by subgrain rotation during metasomatic alteration induced by melt percolation. The microstructural observations and olivine CPO combined with petrological and geochemical data suggest that Group 1 xenoliths preserve microstructural and chemical characteristics of an old, probably Proterozoic lithosphere, while Group 2 xenoliths record localized Miocene deformation associated with wall–rock heating and metasomatism related to melt circulation. Furthermore, the observed transition of olivine CPO from [100]–axial pattern to [010]–axial pattern by deformation in the presence of variable melt fractions and associated metasomatic alteration can be inferred to modify the physical properties of mantle rocks.

We have compiled whole-rock analyses in mantle rocks to investigate the abundance of gold in the lithospheric mantle, and how this distribution can be modified. We show that while gold is heterogeneously distributed in peridotites from the upper mantle, the concentrations in most geographical regions cluster around the global median value of 1.2 ppb. This means Au concentration in the modern lithospheric mantle is similar to estimates for the crust and primitive upper mantle, indicating that Au has not been substantially lost from the mantle into the crust, or that it has been effectively recycled between the crust and the mantle, over Earth's history. The only region that shows substantial deviation from the typical lithospheric mantle values is the Eastern Block of the North China Craton, which shows elevated Au contents (median 3.5 ppb) across multiple studies from seven different sampling sites. Au shows no systematic relationship with the LILE, and does not appear to be enriched by hydrous fluids in the mantle, which strongly contrasts with its behaviour in the crust. There is likewise no evidence that Au is enriched by carbonatitic metasomatism in the mantle. A weak correlation with Ta and Hf, along with significantly elevated Ta concentrations in samples with Au concentrations >5 ppb is observed, which possibly indicates the transport of Au by silicate melts in the mantle. We also show that the Au content of pyroxenites is commonly elevated relative to that of peridotites in the lithospheric mantle (median Aupyroxenite = 2.0 ppb; median Auperidotite = 1.2 ppb). The data indicate that the melts from which these pyroxenites have crystallised may be effective metasomatic agents within the lithospheric mantle. Furthermore, these gold-rich pyroxenites may play an important role in mineralisation processes.

The origins of the Mesozoic low-δ18O A-type granites in eastern China are quite controversial. A new “low-δ18O” granite, the Taohuadao alkaline (A1-type) granite from the Zhoushan archipelago, offshore from Zhejiang, is reported in this study. Integrated in situ analyses of Hf-O-Li isotopes in zircon are used to provide insights on the petrogenesis and evolution of this granite, and low-δ18O granites in general. Three types of zircon have been recognized in Taohuadao samples: zircon with sector/oscillatary zoning (Type 1), disturbed zoning (Type 2) and extremely high-U (Type 3) grains. Data are carefully screened for radiation damage and post-magmatic alteration and only Type 1 zircon represents a magmatic origin. In Type 1 zircons, Type 1a with normal mantle δ18Ozrn value of 5.23‰ has a positive δ7Li value (2.0 ± 3.3‰), while Type 1b has a δ18O value of ∼4.81‰ and highly variable negative δ7Li values (−31.9‰ to −0.3‰). In contrast to the O-Li isotopes, zircon is quite homogeneous in terms of Hf isotopes (eHf(t) = −3.5 ± 1.7), suggesting that the Taohuadao A1-type granites were derived from a simple source. Our results show that the magmatic zircon δ18Ozrn values are negatively correlated with Hf content but positively correlated with Zr/Hf ratio in zircon, which indicates that fractional crystallization plays an important role in producing the slightly low δ18O signature. The anhydrous nature of the A1-type granites suppressed the crystallization of biotite and hornblende, while low oxygen fugacity is responsible for insufficient precipitation of magnetite, both of which drive the later differentiates towards slight depletion in 18O. Therefore, we suggest a ferrogabbro-type fractional crystallization may have played a major role in producing the chemical and oxygen isotopic variations of the Taohuadao A1-type granite and accounts for ∼0.5‰ 18O-depletion. The extremely depleted δ7Li in Type 1b zircons is most likely a result of diffusion-driven fractionation. Mantle-like O-Li isotopes and very low Li abundance in zircon suggest a predominantly mantle origin with little, if any, involvement of continental crustal materials in the granite. Thus the Taohuadao A1-type granite was most likely produced by intensive fractionation of reduced, anhydrous basaltic melts, rather than by mixing of mantle- and crustally-derived melts as previously thought.

Magnesian ilmenite is a common kimberlite indicator mineral, although its use in diamond exploration is still controversial. Complex crystallisation and replacement processes have been invoked to explain the wide compositional and textural ranges of ilmenite found in kimberlites. This work aims to shed light on these processes, as well as their implications for diamond exploration. Petrographic studies were combined for the first time with both major- and trace-element analyses to characterise the ilmenite populations found in xenoliths and xenocrysts in two Angolan kimberlites (Congo-Kasai craton). A multi-stage model describes the evolution of ilmenite in these pipes involving: i) crystallisation of ferric and Mg-rich ilmenite either as metasomatic phases or as megacrysts, both in crustal and in metasomatised mantle domains; ii) kimberlite entrainment and xenolith disaggregation producing at least two populations of ilmenite nodules differing in composition; iii) interaction of both types with the kimberlitic magma during eruption, leading to widespread replacement by Mg-rich ilmenite along grain boundaries and fractures. This process produced similar major-element compositions in ilmenites regardless of their primary (i.e., pre-kimberlitic) origin, although the original enrichment in HFSE (Zr, Hf, Ta, Nb) observed in Fe3 +-rich xenocrysts is preserved. Finally (iv) formation of secondary Mn-ilmenite by interaction with a fluid of carbonatitic affinity or by infiltration of a late hydrothermal fluid, followed in some cases by subsolidus alteration in an oxidising environment. The complexities of ilmenite genesis may lead to misinterpretation of the diamond potential of a kimberlite during the exploration stage if textural and trace-element information is disregarded. Secondary Mg-enrichment of ilmenite xenocrysts is common and is unrelated to reducing conditions that could favour diamond formation/preservation in the mantle. Similarly, Mn-rich ilmenite should be disregarded as a diamond indicator mineral, unless textural studies can prove its primary origin.

Chromitites enclosed within metasomatised Finero phlogopite peridotite (FPP) contain accessory platinum-group minerals, base metal (BM) sulfides, baddeleyite, zircon, zirconolite, uraninite and thorianite. To provide new insights into mantle-crustal interaction in the Finero lithosphere this study evaluates (1) the mineral chemistry and Os-isotope composition of laurite, (2) the crystal morphology, internal structure, in-situ U-Pb, trace-element and Hf-isotope data of zircon from two chromitite localities at Alpe Polunia and Rio Creves. The osmium isotope results reveal a resticted range of ‘unradiogenic’ 187Os/188Os values for laurite at Alpe Polunia (0.1247–0.1251, mean 0.1249 ± 0.0001). Re-Os model ages (TRD) of laurite reflect an Early Paleozoic partial melting event (ca 450 Ma or older), presumably before the Variscan orogeny. The Os isotopic composition of laurite/chromitite probably preserves their mantle signature and was not affected by later metasomatic processes. U-Pb and Hf-isotope data show that the Finero chromitites have distinct zircon populations with peculiar morphology, internal cathodoluminescence textures, trace-element composition and an overall U-Pb age span from ∼310 Ma to 190 Ma. Three age peaks at Rio Creves (220 ± 4 Ma, 234.2 ± 4.5 Ma and 277.5 ± 3.2 Ma) are consistent with a prolonged formation and multistage zircon growth, in contrast to the common assumption of a single metasomatic event during chromitite formation. The trace-element signatures of zircons are comparable with those of mantle-derived zircons from alkaline ultramafic rocks, supporting the carbonatitic nature of the metasomatism. Hf-isotope compositions of the Finero zircons, with eHf values ranging mainly from −3 to +1, are consistent with crustal input during metasomatism and could indicate that the parental melts/fluids were derived from a relatively old source; the minimum estimates for Hf model ages are 0.8–1.0 Ga. Our findings imply that mantle rocks and metasomatic events at Finero have a far more complex geological history than is commonly assumed.

Ultrahigh-pressure ( UHP ) materials ( e.g. , diamond, high-pressure polymorph of chromite) and super-reduced (SuR) phases ( e.g. , carbides, nitrides, silicides and native metals) have been identified in chromitites and peridotites of the Tibetan and Polar-Urals ophiolites. These unusual assemblages suggest previously unrecognized fluid- or melt-related processes in the Earth’s mantle. However, the origin of the SuR phases, and in particular their relationships with the UHP materials in the ophiolites, are still enigmatic. Studies of a recently recognized SuR mineral system from Cretaceous volcanics on Mt Carmel, Israel, suggest an alternative genesis for the ophiolitic SuR phases. The Mt Carmel SuR mineral system (associated with Ti-rich corundum xenocrysts) appears to reflect the local interaction of mantle-derived CH 4 ± H 2 fluids with basaltic magmas in the shallow lithosphere (depths of ∼30–100 km). These interactions produced desilication of the magma, supersaturation in Al 2 O 3 leading to rapid growth of corundum, and phase assemblages requiring local oxygen fugacity ( f O 2 ) gradually dropping to ∼11 log units below the iron–wustite (IW) buffer. The strong similarities between this system and the SuR phases and associated Ti-rich corundum in the Tibetan and Polar-Urals ophiolites suggest that the ophiolitic SuR suite probably formed by local influx of CH 4 ± H 2 fluids within previously subducted peridotites (and included chromitites) during their rapid exhumation from the deep upper mantle to lithospheric levels. In the final stages of their ascent, the recycled peridotites and chromitites were overprinted by a shallow magmatic system similar to that observed at Mt Carmel, producing most of the SuR phases and eventually preserving them within the Tibetan and Polar-Urals ophiolites.

The Bushveld Complex contains the largest platinum-group element (PGE) deposits of the world that are represented by persistent stratiform reefs highly enriched with PGE with respect to underlying and overlying rocks. New Re-Os isotope and elemental LA MC-ICPMS data on platinum-group minerals (PGM) from the mineralized reefs are presented with implications to correlation between the different segments of the Bushveld Complex and a role of superimposed processes at the reef formation. We analyzed laurite (RuS 2 ), hollingworthite (RhAsS), sperrylite (PtAs 2 ) and Pt-Fe alloys from the Merensky Reef, Pseudoreef and the PGE reef of the Platreef. The measured 187 Os/ 188 Os value for Platreef laurite is 0.1751 ± 0.0004 whereas the ratios for sperrylite and hollingworthite range to slightly higher values (0.1713–0.1818 and 0.1744–0.1835 respectively). The observed textures of the analyzed PGM, such as Pt-Fe symplectites in base metal sulfides (BMS), laurite inclusions in chromite and sperrylite rims around sulfide-silicate aggregates, are interpreted as features of primary magmatic crystallization whereas hollingworthite overgrowths and exsolutions in sperrylite are likely to have originated from later solid state transformation or metasomatic processes. The Platreef is a composite sill-like body in the northern limb correlative to the Critical Zone in terms of stratigraphic position, whole-rock geochemistry and isotope characteristics. The pristine magmatic character of sulfides and PGM in the stratiform reefs at the top of the Platreef strongly resembles the style of Merensky Reef mineralization. However, the basal part of the Platreef pyroxenitic sequence is variably contaminated and mineralized with a significant hydrothermal overprint. Sulfides from underlying Lower Zone peridotite yield δ 34 S values varying from +9‰ to +14.2‰ that are much higher than the values for the overlying Platreef and are a consequence of sulfur assimilation from sedimentary sulfates. The same homogeneous mantle-like S isotope compositions in the high-grade PGE reef of the Platreef and the Merensky Reef can be explained by enhanced S isotope exchange through the orthomagmatic process of sulfide upgrading. The similar values and the limited variations of the initial Os isotopic ratios of all PGMs in both reefs also support their magmatic crystallization and origin from a common source.

The origin and evolution of the Yarlung Zangbo ophiolites (South Tibet, China) is the key to the tectonics of the Neo-Tethyan Ocean between Greater India and Asia and the underlying upper mantle dynamics. This study presents a detailed investigation of the Zedang ultramafic body (comprising a harzburgitic and a lherzolitic domain) in the eastern Yarlung Zangbo Suture. Major-element compositions and Ti, Y, and HREE concentrations in peridotites and their minerals indicate that the harzburgites experienced higher degrees of melting than the lherzolites (∼13–19% versus ∼7–12%). The overall enrichment of LREE, Zr, and Sr in harzburgites and their clinopyroxenes suggest that the harzburgites were pervasively metasomatized (cryptically) by silicate melts. The harzburgites also record local strong metasomatism close to melt channels. Nd isotopes indicate that both metasomatic agents were derived from forearc basaltic magmas that intruded the harzburgites at ∼130–120 Ma. The lherzolites did not experience such metasomatism. Thermometry shows that the harzburgites experienced a thorough, lower-temperature reequilibration process in lithosphere, while the lherzolites rapidly accreted from the asthenosphere and preserved high equilibration temperatures (up to ∼1320°C). Comparable enrichment in fluid-mobile elements and radiogenic Sr-isotope compositions in both harzburgitic and lherzolitic pyroxenes reflect slab-fluid infiltration into both mantle domains. All the evidence and the presence of subduction-related chromitites in the harzburgites suggest that the Zedang harzburgites formed in a possibly Jurassic mature subduction system, while the lherzolites accreted later in an early Cretaceous forearc during subduction initiation. The two-layered lithospheric mantle reflects the episodic subduction of the Tethyan slabs.

Kimberlites are not only the most economically important source of diamonds; they also carry unique information encapsulated in rock fragments entrained as the magma traverses the whole thickness of the lithosphere. The Nurbinskaya pipe in the Siberian kimberlite province (Russia) is one of several intruded along the Vilyui Rift, a major terrane boundary. The pipe contains three populations of mantle-derived zircon xenocrysts: Archean (mean age 2709 ± 9 Ma), Devonian (mean age 371 ± 2.3 Ma), and a subset of grains with evidence of brittle deformation and rehealing, and a range of ages between 370 and 450 Ma. The Hf-isotope, O-isotope and trace-element signatures of the last group provide a link between the Archean and Devonian events, indicating at least three episodes of magmatic activity and zircon crystallization in the lithosphere beneath the pipe. The emplacement of the Nurbinskaya pipe ca 370 Ma ago was only the youngest activity in a magma plumbing system that has been periodically reactivated over at least 2.7 billion years, controlled by the lithosphere-scale structure of the Vilyui Rift.

Widespread Miocene (24–8 Ma) ultrapotassic rocks and their entrained xenoliths provide information on the composition, structure, and thermal state of the sub-continental lithospheric mantle in southern Tibet during the India-Asia continental collision. The ultrapotassic rocks along the Lhasa block delineate two distinct lithospheric domains with different histories of depletion and enrichment. The eastern ultrapotassic rocks (89°E–92°E) reveal a depleted, young, and fertile lithospheric mantle (87Sr/86Srt = 0.704–0.707 [t is eruption time]; Hf depleted-mantle model age [TDM] = 377–653 Ma). The western ultrapotassic rocks (79°E–89°E) and their peridotite xenoliths (81°E) reflect a refractory harzburgitic mantle refertilized by ancient metasomatism (lavas: 87Sr/86Srt = 0.714–0.734; peridotites: 87Sr/86Srt = 0.709–0.716). These data integrated with seismic tomography suggest that upwelling asthenosphere was diverted away from the deep continental root beneath the western Lhasa block, but rose to shallower depths beneath a thinner lithosphere in the eastern part. Heating of the lithospheric mantle by the rising asthenosphere ultimately generated the ultrapotassic rocks with regionally distinct geochemical signatures reflecting the different nature of the lithospheric mantle.

Magnesium and oxygen are critical elements in the solid Earth and hydrosphere. A better understanding of the combined behavior of Mg and O isotopes will refine their use as a tracer of geochemical processes and Earth evolution. In this study, the Mg-isotope compositions of garnet and omphacite separated from well-characterized xenolithic eclogites from the Roberts Victor kimberlite pipe (South Africa) have been measured by solution multi-collector ICP-MS. The reconstructed whole-rock δ 26 Mg values of Type I (metasomatized) eclogites range from − 0.61‰ to − 0.20‰ (Type IA) and from − 0.60‰ to − 0.30‰ (Type IB) (mean − 0.43‰ ± 0.12‰), while δ 26 Mg of Type IIA (fresh, least metasomatized) eclogites ranges from − 1.09‰ to − 0.17‰ (mean − 0.69‰ ± 0.41‰); a Type IIB (fresh, least metasomatized) has δ 26 Mg of − 0.37‰. Oxygen-isotope compositions of garnet were analyzed in situ by SIMS (CAMECA 1280) and cross-checked by laser fluorination. Garnets have δ 18 O of 6.53‰ to 9.08‰ in Type IA, 6.14‰ to 6.65‰ in Type IB, and 2.34‰ to 2.91‰ in Type IIB. The variation of δ 26 Mg and δ 18 O in Type IA and IB eclogites is consistent with the previously proposed model for the evolution of these samples, based on major and trace elements and radiogenic isotopes. In this model, the protoliths (Type II eclogites) were metasomatized by carbonatitic to kimberlitic melts/fluids to produce first Type IA eclogites and then Type IB. Metasomatism has changed the O-isotope compositions, but the Mg-isotope compositions of Type IA are mainly controlled by the protoliths; those of Type IB eclogites reflect mixing between the protoliths and the kimberlitic melt/fluid. The combination of a large range of δ 26 Mg and low δ 18 O in Type II eclogites cannot be explained easily by seawater alteration of oceanic crust, interaction of carbonate/silicate sediments with oceanic crust, or partial melting of mafic rocks.

The minimum oxygen fugacity ( f O2) of Earth’s upper mantle probably is controlled by metal saturation, as defined by the iron-wustite (IW) buffer reaction (FeO → Fe + O). However, the widespread occurrence of moissanite (SiC) in kimberlites, and a suite of super-reduced minerals (SiC, alloys, native elements) in peridotites in Tibet and the Polar Urals (Russia), suggest that more reducing conditions ( f O2 = 6–8 log units below IW) must occur locally in the mantle. We describe pockets of melt trapped in aggregates of corundum crystals ejected from Cretaceous volcanoes in northern Israel which contain high-temperature mineral assemblages requiring extremely low f O2 (IW < −10). One abundant phase is tistarite (Ti2O3), previously known as a single grain in the Allende carbonaceous chondrite (Mexico) and believed to have formed during the early evolution of the solar nebula. It is associated with other reduced phases usually found in meteorites. The development of super-reducing conditions in Earth’s upper mantle may reflect the introduction of CH4 + H2 fluids from the deep mantle, specifically related to deep-seated volcanic plumbing systems at plate boundaries.

Perovskite (CaTiO 3 ) has become a very useful mineral for dating kimberlite eruptions, as well as for constraining the compositional evolution of a kimberlitic magma and its source. Despite the undeniable potential of such an approach, no similar study had been done in Angola, the fourth largest diamond producer in Africa. Here we present the first work of in situ U–Pb geochronology and Sr–Nd isotope analyses of perovskite in six Angolan kimberlites, supported by a detailed petrographic and geochemical study of their perovskite populations. Four types of perovskite were identified, differing in texture, major- and trace-element composition, zoning patterns, type of alteration and the presence or absence of inclusions. Primary groundmass perovskite is classified either as anhedral, Na-, Nb- and LREE-poor perovskite (Ia); or euhedral, strongly zoned, Na-, Nb- and LREE-rich perovskite (Ib). Secondary perovskite occurs as reaction rims on ilmenite (IIa) or as high Nb (up to 10.6 wt% Nb 2 O 5 ) perovskite rims on primary perovskite (IIb). The occurrence of these four types within the Mulepe kimberlites is interpreted as an evidence of a complex, multi-stage process that involved mingling of compositionally different melts. U–Pb dating of these perovskites yielded Lower Cretaceous ages for four of the studied kimberlites: Mulepe 1 (116.2 ± 6.5 Ma), Mulepe 2 (123.0 ± 3.6 Ma), Calonda (119.5 ± 4.3 Ma) and Cat115 (133 ± 10 Ma). Kimberlite magmatism occurred in NE Angola likely due to reactivation of deep-seated translithospheric faults (> 300 km) during the break-up of Gondwana. Sr–Nd isotope analyses of four of these kimberlites indicate that they are Group I kimberlites, which is consistent with the petrological observations.

The LA-ICP-MS U-(Th-)Pb geochronology international community has defined new standards for the determination of U-(Th-)Pb ages. A new workflow defines the appropriate propagation of uncertainties for these data, identifying random and systematic components. Only data with uncertainties relating to random error should be used in weighted mean calculations of population ages; uncertainty components for systematic errors are propagated after this stage, preventing their erroneous reduction. Following this improved uncertainty propagation protocol, data can be compared at different uncertainty levels to better resolve age differences. New reference values for commonly used zircon, monazite and titanite reference materials are defined (based on ID-TIMS) after removing corrections for common lead and the effects of excess 230Th. These values more accurately reflect the material sampled during the determination of calibration factors by LA-ICP-MS analysis. Recommendations are made to graphically represent data only with uncertainty ellipses at 2s and to submit or cite validation data with sample data when submitting data for publication. New data-reporting standards are defined to help improve the peer-review process. With these improvements, LA-ICP-MS U-(Th-)Pb data can be considered more robust, accurate, better documented and quantified, directly contributing to their improved scientific interpretation. La communauté internationale de la géochronologie LA-ICP-MS U-(Th-)Pb a défini les nouvelles normes pour la détermination des âges U-(Th-)Pb. Un nouveau flux de travail définit la propagation adéquate des incertitudes pour ces données, en identifiant les composants aléatoires et systématiques. Seules les données avec des incertitudes relatives à l'erreur aléatoire devraient être utilisées dans les calculs des moyennes pondérées des âges de la population; les composantes de l'incertitude pour les erreurs systématiques sont propagées après cette étape, empêchant leur réduction erronée. En suivant ce protocole amélioré pour la propagation de l'incertitude, les données peuvent être comparées à différents niveaux d'incertitude pour mieux résoudre les différences d’âge. Les nouvelles valeurs pour les matériaux de référence zircon, monazite et titanite sont définies (basée sur ID-TIMS) après les corrections du plomb commun et les effets de l'excès en 230Th. Ces valeurs reflètent plus fidèlement le matériau échantillonné lors de la détermination des facteurs d’étalonnage par analyse LA-ICP-MS. Des recommandations sont faites pour représenter graphiquement des données avec seulement des ellipses d'incertitude à 2s et pour soumettre ou pour citer des données de validation avec les données de l’échantillon lors de la soumission des données pour publication. Les nouvelles normes de déclaration des données sont définies pour améliorer le processus d'évaluation par les pairs. Grâce à ces améliorations, les données LA-ICP-MS U-(Th)-Pb peuvent être considérés comme plus solides, exactes, mieux documentées et quantifiées, contribuant ainsi à une meilleure interprétation scientifique.

The preservation of ultrahigh-pressure and super-reduced phases (diamond, moissanite, etc.) in the harzburgites and chromitites of the Yarlung Zangbo ophiolites (South Tibet, China) has major implications for mantle recycling and lithosphere evolution in the tectonic system related to the closing of the Neo-Tethyan Ocean. However, important aspects of the genesis of these enigmatic ophiolites and the related geodynamic evolution are still unclear. In the Zedang ophiolite of the eastern Yarlung Zangbo Suture, detailed mineral chemical data reveal that the harzburgite domain in the east [spinel Cr# (mole Cr3+/(Cr3+ + Al3+) = 0.62–0.33] is more depleted than the lherzolite domain in the west (spinel Cr# = 0.30–0.17) and shows much lower equilibration temperatures (by ∼250–150 °C) than the lherzolites. Clinopyroxene trace-element compositions indicate that the harzburgites underwent pervasive metasomatism after melt extraction, while the lherzolites did not. New zircon U–Pb ages show that the harzburgites were intruded by dolerite dykes with chilled margins at ∼130–128 Ma, consistent with the widespread mafic magmatism at ∼130–120 Ma in the Yarlung Zangbo ophiolites. Nd–Hf isotopic data indicate that the Zedang lherzolites subcreted the pre-emplaced harzburgites concurrently with the intrusion of the dolerite dykes into the harzburgites, and that the lherzolites and dolerites both were derived from upwelling asthenosphere with minor slab input. Available zircon geochronology and Hf-isotope data show that juvenile magmatism in the adjacent Gangdese Arc was almost completely interrupted from ∼130–120 Ma. We suggest that the extension of the overlying harzburgitic lithosphere, subcretion of lherzolites, intrusion of mafic dykes, and the waning of Gangdese-Arc magmatism all reflect a southward trench migration in the Neo-Tethyan subduction system from the Gangdese Arc to the oceanic forearc lithosphere. This magmatic relocation and tectonic linkage are inferred to be the far-field effects of plate reorganization related to the Lhasa-Qiangtang collision and the breakup of Gondwanaland in the early Cretaceous. This model provides a solution for the global “ophiolite conundrum”: supra-subduction-zone type ophiolites with mid-ocean-ridge features.

Mantle pyroxenites are the crystallised products of mafic silicate melts, which are commonly invoked as metasomatic agents in the upper mantle. This study has analysed the trace elements of sulfides, with a specific focus on gold, hosted in a suite of mantle pyroxenite xenoliths from Qilin in the Cathaysia Block, southeast China. These are compared with sulfides hosted in peridotite xenoliths from the same locality to assess the difference in the abundances of Au, and a suite of siderophile and chalcophile elements between the sulfides hosted in mobile melts in the upper mantle and their host “wall” rocks. Both the peridotite- and pyroxenite-hosted sulfides show a wide spectrum of trace element contents. The pyroxenite-hosted sulfides typically have PGE and Au concentrations that are an order of magnitude or more below those measured in the peridotite-hosted sulfides (lherzolite-hosted sulfides: total PGE = 95 ± 118 ppm, Au = 1.4 ± 2.6 ppm; pyroxenite-hosted sulfides: total PGE = 0.25 ± 0.70 ppm, Au = 0.14 ± 0.39 ppm). Furthermore, the Ir group PGE (Ir, Os and Ru) are present in lower concentrations than the Pd-group PGE (Pd, Pt and Rh). This may lead to a distinct signature if the melts from which these sulfides crystallise interact with lherzolitic sulfides. The overall low abundances of these elements within the pyroxenites suggests that the parent melts are an inefficient medium for enriching any of these elements in the upper mantle.

For more than 20 years, the reported occurrence of diamonds in the chromites and peridotites of the Luobusa massif in Tibet (a complex described as an ophiolite) has been widely ignored by the diamond research community. This skepticism has persisted because the diamonds are similar in many respects to high-pressure high-temperature (HPHT) synthetic/industrial diamonds (grown from metal solvents), and the finding previously has not been independently replicated. We present a detailed examination of the Luobusa diamonds (recovered from both peridotites and chromitites), including morphology, size, color, impurity characteristics (by infrared spectroscopy), internal growth structures, trace-element patterns, and C and N isotopes. A detailed comparison with synthetic industrial diamonds shows many similarities. Cubo-octahedral morphology, yellow color due to unaggregated nitrogen (C centres only, Type Ib), metal–alloy inclusions and highly negative δ C 13 values are present in both sets of diamonds. The Tibetan diamonds ( n = 3 ) show an exceptionally large range in δ N 15 (−5.6 to + 28.7 ‰ ) within individual crystals, and inconsistent fractionation between {111} and {100} growth sectors. This in contrast to large synthetic HPHT diamonds grown by the temperature gradient method, which have with δ N 15 = 0 ‰ in {111} sectors and + 30 ‰ in {100} sectors, as reported in the literature. This comparison is limited by the small sample set combined with the fact the diamonds probably grew by different processes. However, the Tibetan diamonds do have generally higher concentrations and different ratios of trace elements; most inclusions are a NiMnCo alloy, but there are also some small REE-rich phases never seen in HPHT synthetics. These characteristics indicate that the Tibetan diamonds grew in contact with a C-saturated Ni–Mn–Co-rich melt in a highly reduced environment. The stable isotopes indicate a major subduction-related contribution to the chemical environment. The unaggregated nitrogen, combined with the lack of evidence for resorption or plastic deformation, suggests a short (geologically speaking) residence in the mantle. Previously published models to explain the occurrence of the diamonds, and other phases indicative of highly reduced conditions and very high pressures, have failed to take into account the characteristics of the diamonds and the implications for their formation. For these diamonds to be seriously considered as the result of a natural growth environment requires a new understanding of mantle conditions that could produce them.

In our paper, we suggested that the Junan granulite xenoliths and xenocrysts record evolution of the Precambrian lower crust beneath the southeastern North China Craton (NCC). Yuan and Xia (2015) disagree with us. However, they have not fully considered the evolutional histories of the NCC lithosphere, and geochemical and isotopic compositions of the Junan xenoliths. We also contend that they have misinterpreted the available geophysical data. Synthesizing the geochronological characteristics of the NCC lower crust, nature of the Junan granulite xenoliths, and reinterpretation of the resistivity profile, we again emphasize that the Junan granulite xenoliths are tectonically affiliated to the NCC lower crust, and the Junan zircon data could reflect the complex evolution of the lower crust beneath the southeastern NCC.

Melting and reactions between minerals and melts are important processes in the evolution of the lithospheric mantle, and are usually inferred from their geochemical fingerprints in mantle samples. However, a suite of mantle-derived peridotite xenoliths from the Xilong area, South China, records the reaction of successive silicate melts of different compositions with mineral assemblages in the mantle, preserved by quenching during entrainment. These xenoliths form two groups and record a compositionally layered mantle. Group 1 has olivine Mg# ~ 91, (and is thus relatively refractory), is derived from depths of ~ 50–65 km, and shows the trace-element geochemical signature of “old” carbonatitic metasomatism. Group 2 is more fertile with olivine Mg# mainly ~ 89–90, is derived from ~ 40 to 55 km and has ubiquitous modal spinel. Xenoliths of both groups then show sequential infiltration by two compositionally distinct melts (Na-rich and K-rich) not long before eruption. The Na-rich melts are enclosed in spongy clinopyroxene and spinel rims and are inferred to have triggered the reactions that formed the spongy rims, which have lower Al2O3, Na2O and Mg#, but higher FeO, TiO2 and Cr# than the primary phases. The undersaturated Na-rich mafic melts were probably formed in the asthenosphere by low-degree melting. The K-rich melts occur mainly in reaction zones around orthopyroxene and in reaction patches containing fine-grained secondary olivine, clinopyroxene and minor spinel. The melts have high contents of SiO2, K2O (mean 14.3 wt.%), Rb, Ba, and LREE but very low Na2O/K2O (0.01–0.29), positive anomalies in Eu and Sr, and variable HFSE anomalies. These compositional characteristics are consistent with an origin as low-degree partial melts of pre-existing phlogopite-bearing rocks. The K-rich melts also react with primary olivine, and the spongy-textured secondary clinopyroxene and spinel inferred to have formed by reaction with the Na-rich melts, yielding secondary olivine, vermicular clinopyroxene and spinel compositionally similar to that in the reaction patches. This observation suggests that the infiltration of K-rich melt occurred after the percolation of Na-rich melts. This study demonstrates the complexity of metasomatic processes in the lithospheric mantle, and emphasizes the caution required when interpreting the geochemical signatures of mantle-derived xenoliths in which multiple overprinting by compositionally diverse fluid and melt interactions with mantle minerals cannot be so clearly observed and characterized as in these samples.

Gold is a strongly chalcophile element and its concentration and behaviour in the mantle are intimately connected to the nature, abundance and mobility of sulfide phases. In this study Au has been analysed together with a suite of chalcophile and siderophile elements in sulfides hosted in 31 mantle-derived xenoliths from Spitsbergen in the Svalbard Archipelago. These data have been combined with analysis of silicate phases to assess the mobility of Au, and sulfides in general, during mantle processes. The 31 samples comprise lherzolites, harzburgites, dunites and a wehrlite. Several lherzolites contain amphibole and apatite, and two are cross-cut by amphibole veins. The REE patterns in clinopyroxene reveal three geochemically distinct groups of peridotites. Group I samples have flat or LREE depleted patterns ((La/Yb)n 1). The sulfides hosted in the Group I samples have very heterogeneous Au compositions (1.9 ± 1.4 ppm), whereas those hosted in Groups II and III samples have lower and more homogenous Au concentrations (Group II Au: 0.66 ± 0.60 ppm; Group III Au: 0.53 ± 0.57 ppm). A chromatographic metasomatic model is proposed to explain the variations between the three groups of samples. The LREE enrichment, and common inclusion of amphibole in the Group III samples is consistent with modification by a carbonate-rich melt. It is interpreted that these samples were located close to the metasomatic conduit. The sulfide homogeneity in these samples results from extensive equilibration with the metasomatic agent; Au has partitioned into the fluid phase, reducing the Au content of these sulfides. The Group I samples show little to no metasomatic alteration in the silicate phases, indicating they were located far from the conduit. The agents that have affected these samples have been modified by extensive fractionation and wall–rock interaction during percolation through the mantle, becoming enriched in Au. Group II samples are inferred to have been located at an intermediate distance from the conduit with silicate and sulfide characteristics that fall in between those from the other two groups.

The Os-isotope compositions of sulphides in mantle xenoliths hosted by Late Miocene alkali basalts from the Sviyaginsky volcano, Russian Far East, reveal the presence of Archaean–Proterozoic subcontinental lithospheric mantle beneath the Khanka massif. Their TMA and TRD model ages reveal similar peaks at 1.1 and 0.8 Ga suggesting later thermotectonic events in the subcontinental lithospheric mantle, whereas TRD model ages range back to 2.8 ± 0.5 (2σ) Ga. The events recognized in the subcontinental lithospheric mantle are consistent with those recorded in the crust of the Khanka massif. The sulphide Os-isotope data show that the subcontinental lithospheric mantle beneath the Khanka massif had formed at least by the Mesoproterozoic, and was subsequently metasomatized by juvenile crustal-growth events related to the evolution of the Altaids. The Khanka massif is further proposed to have tectonic affinity to the Siberia Craton and should originate from it accordingly.

Podiform chromitites occur in the meta-dunite horizon of the Dobromirtsi ultramafic massif, in the Central Rhodope metamorphic core complex (southern Bulgaria). Although these were deformed and metamorphosed together with their host rocks several chromite cores from massive chromitite samples still preserve compositions unaffected by metamorphism – at least in terms of major elements . In situ laser ablation ICP-MS analyses of these chromite cores show a bimodality in the composition of chromite between high-Cr and high-Al {Cr# = [Cr/(Cr + Al)], atomic ratio = 0.74–0.55} with Mg# between 0.69 and 0.60 [Mg# = Mg/(Mg + Fe 2 + ), atomic ratio]. These compositions of chromite and the whole rock enrichment of Os–Ir–Ru relative to Pt–Pd in the chromitites resemble that reported for podiform chromitites in ophiolites elsewhere. Magmatic platinum-group minerals (PGM), including laurite (RuS 2 ), Os–Ir alloys and sulfarsenides (irarsite, IrAsS), exhibit a wide range of 187 Os/ 188 Os (from 0.1097 to 0.1272) whereas 187 Re/ 188 Os is nearly zero. The distribution of the Os model ages calculated for these PGM ranges between 0.13 and 2.6 Ga, with age peaks at 0.25, 0.4, 0.7, 2.1 and 2.6 Ga. Two xenocrystic zircons were identified in the chromitites and yield concordant U–Pb age of 2257 ± 80 Ma (2σ) and 1952 ± 82 Ma (2σ). The oldest zircon exhibits a depleted-mantle Hf model age ( T DM ) of 2.77 Ga and a “crustal” Hf model age ( T C DM ) of 3.1 Ga, whereas the youngest has a T DM of 2.21 Ga and a T C DM of 2.35 Ga. We interpret the coexistence of both types of chromitites in a horizon of meta-dunites as the result of their precipitation from fractionating island arc tholeiite melts. The first reactions produced dunite; later the batches of magma were isolated from reaction with the peridotite wall-rocks and progressively fractionated in a network of dunite channels developed in the upper mantle beneath the back-arc spreading centre. The Os and Hf model ages suggest that the ultramafic protolith of the Dobromirtsi ophiolite was derived from an original Archean mantle at least 3 Ga old, which underlay a continental crust and was reworked in the Paleo-proterozoic and possibly even younger times.

We suggest a new explanation for the presence of crustally derived zircons in the upper-mantle rocks of ophiolitic complexes, as an alternative to subduction-related models. Integrated isotopic (U-Pb, Hf, and O isotopes) and trace-element data for zircons from the Tumut ophiolitic complex (southeast Australia) indicate that these grains are related to granitic magmatism and were introduced into the mantle rocks after their emplacement into the crust. These observations emphasize that a clear understanding of the origin of individual zircon populations and their relationship to the host rock is essential to interpretations of the tectonic history of upper-mantle rocks and the dynamics of crust-mantle interactions.

Peridotite xenoliths from the Nógrád–Gömör Volcanic Field (NGVF) record the geochemical evolution of the subcontinental lithospheric mantle beneath the northern margin of the Pannonian Basin. This study is focused on spinel lherzolites and presents petrography, and major and trace element geochemistry for 51 xenoliths selected from all xenolith-bearing localities of the NGVF. The xenoliths consist of olivine, orthopyroxene, clinopyroxene and spinel ± amphibole. No correlations between modal composition and textures were recognized; however, major and trace element geochemistry reveals several processes, which allow the distinction of xenolith groups with different geochemical evolution. The xenoliths have undergone varying degrees (∼7–25%) of partial melting with overprinting by different metasomatic processes. Based on their Mg#, the xenoliths can be subdivided into two major groups. Group I has olivine Mg# between 89 and 91, whereas Group II has Mg#, Journal of Petrology, 58 (6), ISSN:0022-3530, ISSN:1460-2415

The microstructures, major- and trace-element compositions of minerals and electron backscattered diffraction (EBSD) maps of high- and low-Cr# [spinel Cr# = Cr3+/(Cr3+ + Al3+)] chromitites and dunites from the Zedang ophiolite in the Yarlung Zangbo Suture (South Tibet) have been used to reveal their genesis and the related geodynamic processes in the Neo-Tethyan Ocean. The high-Cr# (0.77–0.80) chromitites (with or without diopside exsolution) have chromite compositions consistent with initial crystallization by interaction between boninitic magmas, harzburgite and reaction-produced magmas in a shallow, mature mantle wedge. Some high-Cr# chromitites show crystal-plastic deformation and grain growth on previous chromite relics that have exsolved needles of diopside. These features are similar to those of the Luobusa high-Cr# chromitites, possibly recycled from the deep upper mantle in a mature subduction system. In contrast, mineralogical, chemical and EBSD features of the Zedang low-Cr# (0.49–0.67) chromitites and dunites and the silicate inclusions in chromite indicate that they formed by rapid interaction between forearc basaltic magmas (MORB-like but with rare subduction input) and the Zedang harzburgites in a dynamically extended, incipient forearc lithosphere. The evidence implies that the high-Cr# chromitites were produced or emplaced in an earlier mature arc (possibly Jurassic), while the low-Cr# associations formed in an incipient forearc during the initiation of a new episode of Neo-Tethyan subduction at ~130–120 Ma. This two-episode subduction model can provide a new explanation for the coexistence of high- and low-Cr# chromitites in the same volume of ophiolitic mantle.

Several ultramafic–mafic complexes of assumed ophiolitic affinity that contain ore deposits occur mostly along the Pacific margin of Mexico and the boundaries of the Guerrero composite terrane, which have been described in association with various tectonomagmatic settings. These are (1) roots of volcanic arcs, (2) ophiolites, (3) oceanic basins, (4) continental fragments underlying a magmatic arc, and (5) subduction complexes or rift-transform magmatic suites. However, the petrogenesis of such ultramafic–mafic complexes and the associated ore deposits remains ill defined to date, as their comprehensive study virtually lacks an appropriate geochemical approach. Ultramafic–mafic complexes with known or estimated ages in Mexico are mostly of Paleozoic and Mesozoic ages, and very few of them formed once the crustal dynamics of the Mexican Pacific margin switched from generally extensional to compressional tectonic settings during the Late Cretaceous. A few deposits, aged Late Cretaceous to Paleocene, were interpreted to have formed in association with tectonomagmatic settings different from most deposits in the Guerrero composite terrane. To move forward in the understanding on how these mafic–ultramafic complexes and their associated ore deposits formed, the genesis of the chromitite deposits within the Loma Baya ultramafic–mafic complex (southernmost Guerrero composite terrane) is hereby endeavored. The combination of the geochemical study of both major and trace elements in chromite (including platinum-group elements; PGE), the mineralogy, PGE geochemistry and radiogenic isotopes (i.e., Re–Os) in platinum-group minerals (PGM), provides new assessment on the genesis of this ultramafic–mafic complex. From the obtained results it stems that the Loma Baya complex and the chromitite deposits associated with it are closely related with the formation of a back-arc basin in an active supra-subduction zone during the Lower Cretaceous, and the parental melts resulted from re-melting of an ultramafic protolith of Paleozoic (or older) age. The melts that originated in such back-arc basin were of tholeiitic affinity and produced Al-rich chromitites with a moderate enrichment in PGE, associated with rates of partial melting of a depleted mantle source of at least 20%. Such geological framework is in frank disagreement with previous work, thus highlighting the necessity of further research in other Mexican ultramafic–mafic complexes to better constrain their geodynamic setting of formation. The obtained Re–Os model ages in PGM of the Loma Baya chromitite deposits (ca. 300 and 130 Ma) suggest a possible correlation between these deposits and those in the Californias, in the sense that both areas record similar geological processes (mantle rock melting, re-melting, and chromitite emplacement) within similar tectonomagmatic environments and time frame for their occurrence. Such possibility further highlights the necessity to carry on similar studies in chromitite deposits elsewhere along the Pacific margin in Mexico and southwestern USA, in order to better reconstruct its tectonomagmatic evolution during the Mesozoic.

The early stages of magmatic processes operating at mantle depths beneath continental arcs are poorly known. The chemical compositions of minerals and rocks, mineral Sr^Nd^Hf^O isotopes and zircon U^Pb ages of garnet clinopyroxenite dykes from the Shenglikou peridotite massif (North Qaidam Orogen, NE Tibet, China) were studied to constrain their sources and genesis, and the dynamic processes that controlled pyroxenite formation beneath an early Paleozoic active continental margin. Major-element compositions of bulkrocks suggest that the pyroxenitic protoliths were cumulates segregated from a melt, which was extracted from a peridotite-dominated mantle source. Bulk-rock and mineral trace-element patterns show strong enrichment in fluid-mobile elements (e.g. Cs, Rb, Ba,Th, U, K, Pb and Li) and marked negative anomalies in the high field strength elements relative to rare earth elements, similar to the characteristics of melts derived from a volatile-rich subarc mantle. Enriched Sr and Nd initial isotopic compositions at 500Ma (Sr/Sr of 0 70919^0 71774 and eNd of 16 3 to 3 4) are in contrast to the highly radiogenic Hf isotope compositions (similar to those of the depleted-mantle reservoir) and to the uncontaminated upper-mantle dOV-SMOW (garnet: 5 6 0 3o, 2SD, n1⁄4 61; zircon: 5 9 0 3o, 2SD, n1⁄4 28). These decoupled isotopic signatures suggest that the melt source was located in a convective mantle wedge (controlling the Hf and O isotopes) that had been pervasively metasomatized by fluids from a subducted ProtoTethys oceanic slab (controlling the Sr^Nd isotopes and highly incompatible elements). Zircons with two groups of U^Pb ages (430 5Ma and 401 7Ma) were generated by recrystallization events, corresponding to UHP metamorphism and a major uplift stage during the North Qaidam orogeny, respectively.The combined evidence reveals a picture of continental arc magmatism at mantle depths and subsequent continental collision. The subduction of the Proto-Tethys oceanic slab beneath the southern Qilian margin triggered flux melting of the metasomatized convective mantle wedge and generated hydrous arc magmas. These primitive magmas intruded into the overlying lithospheric mantle and segregated the cumulates parental to the Shenglikou pyroxenites. Subsequent continental subduction incorporated fragments of the mantle-wedge peridotite (containing pyroxenite dykes) at 430Ma and carried them to shallow depths during exhumation at 400Ma.

The stable isotope systems of Li and B are useful tools for the understanding of ocean island basalt genesis. We have applied both systems to a comprehensive set of well-characterised samples across the Azores islands in order to further evaluate the nature of the mantle source. These ocean islands represent the surface expression of a low-buoyancy mantle plume. The variability of Li and B concentrations and their isotopic compositions across the archipelago builds on recent radiogenic isotope studies from the islands to the west of the Mid-Atlantic Ridge (Flores and Corvo). The results for both systems yield some values that are typical for ocean island basalts; however, the variability of both δ7Li and δ11B observed in primitive lavas is most extreme on the western islands of Flores and Corvo (δ7Li = + 3.5 to + 8.2‰ and δ11B = − 3.5 to + 11.8‰). The large spread in isotopic composition is most likely due to contamination of the magmas during ascent through hydrothermally altered oceanic crust. Models of assimilation fractional crystallisation best explain the variability observed on each island and allow for robust estimates of the mantle source. The implication of these observations is that the interpretation of the radiogenic isotopes and trace elements may have to be treated with care if the stable isotopes record contamination by assimilation of hydrothermally altered material and that the use of Li and B as mantle source tracers may be obscured by shallow level processes. Further, it seems unlikely that the large range observed in Sr and Nd isotopes can reflect variable contributions from recycled low-T materials such as sediment or altered MORB, since there is no correlation between the radiogenic and stable isotope data.

LIMA phases (LIndsleyite-MAthiasite: K-Ba-Zr-Cr titanates of the crichtonite series) in the Jagersfontein kimberlite (95 ± 5 Ma) are xenocrysts, related to the Phlogopite-K richterite-Peridotite (PKP) style of metasomatism that has affected the lithospheric mantle beneath the Kaapvaal Craton. The LIMA grains contain high levels of Sr, Hf, Nd, Pb and U, allowing a multi-isotope approach to the timing of metasomatism and the sources of metasomatic material. We provide a detailed discussion of the interference-correction and mass-bias issues involved in the in-situ isotopic analysis of these elements. Multiple in situ isotopic analyses of 11 grains by laser-ablation multi-collector ICPMS (and 2 by solution techniques) show significant within-grain and inter-grain variability in Sr and Pb isotopic composition, less variation in Nd-isotope composition, and very little variation in Hf-isotope composition. U-Pb dating, limited by high levels of initial Pb, gives a poorly constrained age of 162 ± 46 Ma, which may reflect an episode of PKP metasomatism. Back-calculation of the Pb-isotope data using measured 238U/204Pb values, and comparison with mantle evolution models, indicate a maximum age for the PKP metasomatism of 180 Ma, suggesting a link with the plume-related Karoo volcanism (179-183 Ma). The relationships between Pb, Sr, Hf and Nd isotopic compositions in LIMA, and published data on PKP–series xenoliths, suggest mixing between two components. One had high 87Sr/86Sr but low 206Pb/204Pb and 143Nd/144Nd, similar to the EM1 “reservoir.” This component probably represents lithospheric mantle that was enriched in Rb/Sr before the metasomatic episode that produced the LIMA phases. The other has isotopic characteristics similar to the “HIMU” reservoir, and may correspond to a plume-related fluid or melt responsible for at least the initiation of the metasomatism.

We integrate an updated worldwide compilation of U/Pb, Hf-isotope and trace-element data on zircon, and Re–Os model ages on sulfides and alloys in mantle-derived rocks and xenocrysts, to examine patterns of crustal evolution and crust–mantle interaction from 4.5 Ga to 2.4 Ga ago. The data suggest that during the period from 4.5 Ga to ca 3.4 Ga, Earth's crust was essentially stagnant and dominantly mafic in composition. Zircon crystallized mainly from intermediate melts, probably generated both by magmatic differentiation and by impact melting. This quiescent state was broken by pulses of juvenile magmatic activity at ca 4.2 Ga, 3.8 Ga and 3.3–3.4 Ga, which may represent mantle overturns or plume episodes. Between these pulses, there is evidence of reworking and resetting of U–Pb ages (by impact?) but no further generation of new juvenile crust. There is no evidence of plate-tectonic activity, as described for the Phanerozoic Earth, before ca 3.4 Ga, and previous modelling studies indicate that the early Earth may have been characterised by an episodic-overturn, or even stagnant-lid, regime. New thermodynamic modelling confirms that an initially hot Earth could have a stagnant lid for ca 300 Ma, and then would experience a series of massive overturns at intervals on the order of 150 Ma until the end of the EoArchean. The subcontinental lithospheric mantle (SCLM) sampled on Earth today did not exist before ca 3.5 Ga. A lull in crustal production around 3.0 Ga coincides with the rapid buildup of a highly depleted, buoyant SCLM, which peaked around 2.7–2.8 Ga; this pattern is consistent with one or more major mantle overturns. The generation of continental crust peaked later in two main pulses at ca 2.75 Ga and 2.5 Ga; the latter episode was larger and had a greater juvenile component. The age/Hf-isotope patterns of the crust generated from 3.0 to 2.4 Ga are similar to those in the internal orogens of the Gondwana supercontinent, and imply the existence of plate tectonics related to the assembly of the Kenorland (ca 2.5 Ga) supercontinent. There is a clear link in these data between the generation of the SCLM and the emergence of modern plate tectonics; we consider this link to be causal, as well as temporal. The production of both crust and SCLM declined toward a marked low point by ca 2.4 Ga. The data naturally divide the Archean into three periods: PaleoArchean (4.0–3.6 Ga), MesoArchean (3.6–3.0 Ga) and NeoArchean (3.0–2.4 Ga); we suggest that this scheme could usefully replace the current four-fold division of the Archean.

Platinum-group minerals are the dominant control on the budget of platinum-group elements in ophiolitic chromitites. They vary widely in distribution, morphology, chemistry and Os-isotope compositions. Their distribution in different types of chromitites, and different microstructural sites within them may reflect the role of chromite as a physical “collector” of the platinum-group elements during crystallisation of both phases from melts. The platinum-group minerals show significant Os-isotope heterogeneity on the scale of single micrometric grains enclosed within individual chromite grains. This micro-scale isotopic heterogeneity, and their random spatial distribution within the chromites, argue strongly against an origin of the platinum-group minerals via exsolution from the chromite lattice during post-magmatic cooling. A re-interpretation of their microstructures and Os isotopes analysed in situ had led us to review the traditional models, and to propose several origins for the platinum-group minerals in chromitites. (1) The suites of platinum-group minerals mainly record crystallisation during mixing of multiple different melts and physical entrapment in chromite. (2) Some platinum-group minerals in chromitites may have been scavenged from wall–rock peridotite during migration of the parental melts of the chromitite. (3) Other platinum-group minerals may have precipitated from metasomatic fluid/melts that infiltrated existing chromitites. The recrystallization of pre-existing platinum-group minerals during polyphase metamorphism or “recycling” of the chromitite into deeper mantle levels can explain the presence of micrometric platinum-group minerals with distinct Os-isotope compositions.

Most eclogite xenoliths from the Roberts Victor kimberlite (South Africa) have been metasomatized by melts/fluids in the carbonatitic–kimberlitic spectrum, and these can be used to define a progressive metasomatic process. Several stages of this progressive metasomatism have been recognized within one sample (RV07-17), using petrography, major- and trace-element compositions, and Sr- and O-isotope data. Four zones in RV07-17 are distinguished in terms of the compositions of their garnets. From Zone 1 to Zone 4, the microstructure becomes less equilibrated; secondary minerals and fluid inclusions become abundant; the pyrope content of the garnets increases; and clinopyroxene shows progressive enrichment in MgO. The garnets of Zone 1 have flat REE patterns from Lu to Sm, with a strong depletion in the LREE. Toward Zone 4, the relative abundance of the MREE of the garnets drops significantly, giving smoother patterns. A large relict clinopyroxene grain in Zone 1 shows a strong depletion in the LREE, but the LREE/MREE of the recrystallized clinopyroxene increases from Zone 1 to Zone 4. From Zone 1 to 4, 87Sr/86Sr in clinopyroxene increases along with Sr content and δ18O of the garnet decreases from 8.2 to 5.7‰ as the MgO content increases. Trace-element data imply that the fluids that metasomatized RV07-17 were closer to carbonatitic than kimberlitic in composition. Aside from the one relict clinopyroxene, all original compositional information in this xenolith has been swept away, making it difficult to define the protolith. To know the origin of the xenolithic eclogites, and to use them as evidence for different geodynamic/tectonic scenarios, the least metasomatized samples must be studied; unfortunately these make up a tiny proportion of the widely studied xenolith population at Roberts Victor.

Abundant volcanic ash layers, largely altered to clay, characterize the Permian–Triassic boundary beds of the Daxiakou section, South China. Volcanism may have had an important role in the Permian–Triassic mass extinction, so it is important to understand the origin of these ash beds. The lithology of ash bed 260, as constrained by modes of crystal fragments (66.7% plagioclase, 7.2% alkali feldspar, and 26.1% quartz), suggests that the ash represents dacitic volcanism. The U–Pb ages, trace elements and Hf-isotope compositions of zircons from ten ash beds have been analyzed using LA-ICPMS and LA-MC-ICPMS. The zircons can be divided into two groups: magmatic zircons (~ 97%, 227–279 Ma) and inherited zircons (~ 3%, 678–2424 Ma). Magmatic zircons have Y, Hf, Th and U contents and Nb/Ta ratios typical of zircons from silicic calc-alkaline volcanism. e Hf (t) values of magmatic zircons vary from − 10.9 to + 5.3, implying magmatic mixing between juvenile crustal material (probably Neoproterozoic in age) and ancient crustal material (probably Archean). Beds 252–259-b have relatively depleted Hf-isotope compositions with e Hf (t) values of − 10.9–+ 5.3 and averages of − 5.0–− 2.5, while other beds have e Hf (t) values of − 10.8–+ 4.4 (mostly − 10.8–− 3.5) and averages of − 8.9–− 6.8, implying more input of juvenile crustal material in the volcanism of Beds 252–259-b. Integration of the Hf-isotope and trace-element compositions of magmatic zircons suggests that the volcanism producing the ashes took place along the convergent continent margins during the formation of the Pangea supercontinent. This intense silicic volcanism may have played an important role in causing the Permian–Triassic mass extinction in South China.

Mixed-habit diamonds are a special type of crystal that form by simultaneous growth by two different mechanisms; octahedral and cuboid growth. This type of crystal growth is known to preferentially partition nitrogen into the octahedral growth sectors during growth. To identify if this fractionation occurs for any other trace element, this study applies in situ laser ablation mass spectrometry to a well-studied collection of mixed-habit diamonds. The results show high concentrations of nickel (3.6–36.4 ppm) and cobalt (0.108–1.25 ppm) in the cuboid sectors, whereas the octahedral sectors contain only Due to the extremely low concentrations of trace elements in these gem quality diamonds, many of the elements are commonly below the limit of quantification (LOQ). While we are confident that Pr, Sr, Pb, Zr, V, Cr and Zn are not fractionated by the crystal growth mechanisms of mixed-habit diamonds that affect Ni and Co, this remains to be proven for the remaining trace elements analysed. Data on the fractionation of copper are inconclusive, with 75% of data from the cuboid sectors above the LOQ and all data from the octahedral sectors below the LOQ. However, the values of the LOQ for the octahedral data are often higher than those of the cuboid data, so further investigation of Cu is necessary to confirm its behaviour.

Trace-element and Sr–Nd–Hf isotopic compositions of garnet and clinopyroxene in mantle xenoliths from the diamondiferous Lac de Gras kimberlites, central Slave craton, constrain the nature, effects and timing of metasomatic modification of the underlying subcontinental lithospheric mantle (SCLM). Hypothetical carbonatite melts in equilibrium with peridotitic garnet from the > 3.3 Ga ultra-depleted shallow SCLM ( 145 km), which has normal LREE-depleted to mildly sinusoidal REE patterns and weak (or absent) negative Ti anomalies, match natural kimberlite compositions. These systematics translate to distinctly higher Ce/Yb, Sm/Er and Lu/Hf, and lower Zr/Sm and Sm/Nd in shallow peridotites, compared to deep peridotites. The latter show a continuum of trace-element ratios compatible with progressive interaction between a kimberlite-like melt and garnet-bearing mantle, leading to loss of Y and HREE and enrichment in LREE and Sr in the melt, increasing sinuosity of the REE, and lower Sm/Nd and Lu/Hf in the garnet, as the melts evolve to more volatile-rich compositions. Minerals in the majority of deep peridotites display a small range of 143Nd/144Nd (0.5123–0.5129) over a large range of 147Sm/144Nd (0.08 to 0.43) at the time of entrainment in the kimberlite, which indicates isotopic homogenisation less than 350 Ma ago. Combined with their variable 176Hf/177Hf (0.2825–0.2881), this can be modelled by mixing an ancient fluid-metasomatised cratonic mantle residue with ~ 1 to 10% kimberlite melt, which obliterated the original Nd-isotope compositions, but was less efficient at overprinting Hf isotopes. While samples from the shallow SCLM are few, this layer appears to have largely escaped pervasive kimberlite-related overprint: it retained radiogenic Hf (176Hf/177Hf = 0.2879) similar to that of undisturbed ancient depleted mantle, and highly unradiogenic Nd (143Nd/144Nd = 0.5099). This may reflect Proterozoic or earlier metasomatism by HFSE-poor, strongly LREE-enriched metasomatic fluids.