Search

Ancient mantle domains with unradiogenic Os isotopes have been well documented in the Yarlung-Tsangpo ophiolites (YTO), southern Tibet. Nonetheless, previous Os isotope studies have yielded relatively young Re -depletion model (T RD) ages of <1.0 Ga for mantle peridotites of the Zedong ophiolite in the eastern segment of the YTO. In this study, a comprehensive petrological and geochemical study has been conducted on a suite of mantle samples including harzburgites, lherzolites and one pyroxenite. Our results indicate that the Zedong harzburgites have been subjected to low to moderate degrees (∼6–20 %) of partial melting and late-stage melt-rock interaction with basaltic melts. In contrast, both lherzolites and pyroxenite were generated by melt refertilization. The Zedong lherzolites display flat patterns of highly siderophile elements (HSE; Os, Ir, Ru, Pt, Pd, Re), whereas the harzburgites show depletion in Pt, Pd and Re relative to Os, Ir and Ru. The Re -Os isotopes of the Zedong lherzolites have been modified by melt refertilization and thus cannot provide reliable age information. Three Zedong harzburgites with the lowest Pd N /Ir N and Re N /Ir N ratios have the least radiogenic 187Os/188Os of 0.11497–0.11537, giving the old T RD ages of 1.89–1.98 Ga. These ages are considerably older than previously reported T RD ages for the Zedong peridotites (< 1.0 Ga), but comparable to the oldest T RD age for the Luobusa peridotites. This supports the occurrence of similarly old mantle domains in both Zedong and Luobusa ophiolites in the eastern part of the YTO. The Zedong pyroxenite has a radiogenic 187Os/188Os ratio of 0.158 that is similar to the 187Os/188Os ratios (0.156–0.180) of previously reported for the YTO pyroxenites, which can be explained by selective assimilation of interstitial sulfides during the percolation of migrating melts through the mantle lithosphere. Grain boundary sulfides may play a significant role in the high 187Os/188Os of the YTO pyroxenites. • We report HSE and Re-Os isotopes for Zedong ophiolitic mantle peridotite and pyroxenite. • Ancient mantle domains (> 1.8 Ga) occur in Zedong ophiolite in Yarlung-Tsangpo Suture Zone (YTSZ). • Grain boundary sulfides may play an important role in the high 187Os/188Os of the YTSZ pyroxenites. [ABSTRACT FROM AUTHOR]

The heavier iron isotopic composition of some oceanic island basalts (OIB) relative to primary mid-ocean ridge basalts (MORB) has commonly been interpreted to reflect the heavy bulk isotopic compositions of their sources (e.g., of a pyroxenite-dominated mantle region). Alternatively, comparable Fe isotopic signatures observed in continental intraplate basalts (CIB) may record enhanced isotope fractionation during partial melting of an oxidized source, offering valuable insights into mantle oxidation induced by subducted carbonate. To distinguish between these two independent contributions, here we present high-precision Fe isotopic data (δ57Fe) and Fe3+/∑Fe ratios for a suite of Cenozoic intraplate basalts from Inner Mongolia, East Asia. These basalts are distributed along a northwest-southeast (NW–SE) profile and associated with earlier subduction of the Paleo-Asian oceanic slab. Variations in Fe/Mn ratios and zinc isotopic compositions (δ66Zn) of these basalts reveal that altered oceanic crust has always contributed to their mantle sources but the amounts of recycled carbonate gradually vary. Together with existing data, we demonstrate that all of these basalts generally exhibit heavier δ57Fe (0.07 ‰ to 0.31 ‰), higher Fe3+/∑Fe melt (0.08 to 0.39) and Fe/Mn (> 65) relative to primary MORB (δ57Fe ∼0.10 ‰; Fe3+/∑Fe melt ∼0.10; Fe/Mn ∼60), pointing to a common origin of heavy Fe isotopes in intraplate basalts that may reflect those of pyroxenite sources (e.g., recycled crust). However, these basalts containing larger amounts of recycled carbonate in their sources (i.e., higher δ66Zn) have heavier δ57Fe at similar Fe/Mn ratios and proportions of pyroxenite melt. Such Fe isotopic signatures exceed the exclusive role of pyroxenite and are best attributed to partial melting at oxidizing conditions (i.e., high Fe3+/∑Fe source) with an inferred enlarged fractionation factor. This oxidized mantle source formed through the reduction of recycled carbonate to graphite/diamond at depths greater than 150 km. A global dataset of Fe isotopes in intraplate basalts is compiled in order to examine this interpretation further. Compared with OIB typically from a pyroxenite/eclogite-dominant source, CIB with similarly heavy δ57Fe and formed by similar melting degrees are commonly more silica-undersaturated, indicating a widespread carbonated mantle source. Thus, these observations highlight that heavy Fe isotopes can record carbonate-induced mantle oxidation in addition to source lithological heterogeneity for intraplate basalts. • Pyroxenite sources are a common origin of heavy Fe isotopes in intraplate basalts. • Basalts with higher δ57Fe have higher δ66Zn at similar proportions of pyroxenite melt. • Heavy Fe isotopes in intraplate basalts can record carbonate-induced mantle oxidation. [ABSTRACT FROM AUTHOR]

The lithospheric mantle affected by metasomatism is an important reservoir for halogens and often preserves significant heterogeneity in halogen abundances. However, the halogen compositions of the non-metasomatized mantle rocks with depleted geochemical features have not been well studied, impeding a comprehensive understanding of the halogen distribution in the mantle. The Balmuccia (BM) and Baldissero (BD) peridotite massifs in the Ivrea-Verbano Zone, Italian Alps, represent fragments of the sub-continental lithospheric mantle (SCLM) that are fresh and devoid of subduction-related mantle metasomatism. Here we examine the heavy halogen contents of well-characterized spinel-facies mantle peridotites (n = 20) and pyroxenites (n = 16) from the BM and BD massifs. The results indicate significant variation in whole rock halogen contents of peridotites and pyroxenites, with chlorine (Cl) ranging from <30 to 244 μg/g, bromine (Br) from <0.10 to 0.94 μg/g, and iodine (I) from <0.015 to 0.12 μg/g. Some peridotites and pyroxenites show rather low halogen contents (below the detection limit), consistent with their depleted features. However, a large number of samples display halogen contents far higher than the depleted MORB mantle (i.e., Cl: 3–10 μg/g) and peridotite xenoliths from the strongly metasomatized mantle, although ∼30 % of the halogens were leached out by ultrapure water or dilute nitric acid. The data suggest variable halogen enrichment in many mantle rocks. The halogen contents of different types of mantle rocks do not vary with major elements and incompatible trace elements (e.g., Nb), indicating that halogen compositions are not controlled by magmatic processes such as melting and melt-peridotite reaction. Instead, secondary fluid inclusions are abundant and mainly distributed as trails within olivine and pyroxene in both peridotites and pyroxenites, which were likely trapped during the exhumation of these mantle rocks. Their elevated halogen/Nb ratios compared with MORB mantle and strong correlations with Ba/Nb support the effect of fluids. Micro-Raman analyses identify hydrous serpentines within fluid inclusions. Moreover, mass balance calculation indicates that halogen enrichment is mainly caused by secondary fluids (> 90 %). Together, our findings highlight that the fluid percolation in the mantle rocks during exhumation can intensely modify the halogens, to a similar degree as mantle metasomatism, suggesting the sensitivity of halogens to secondary modification. • Halogen data for peridotites and pyroxenites from Balmuccia and Baldissero massifs • Halogens display low contents to variable enrichment compared with depleted MORB mantle. • More than 90 % of the halogens are preserved in the fluid inclusions of constituent minerals. • Crustal fluid percolation during exhumation is the main cause of halogen enrichment. [ABSTRACT FROM AUTHOR]

Continental intraplate basalts usually exhibit diverse compositions varying from silica-deficient alkaline basalts to silica-excess tholeiites. However, the mechanism of compositional diversity remains controversial. Here, we conducted comprehensive petrology, mineral chemistry, whole-rock geochemistry, and Sr–Nd–Hf isotopic compositions, as well as thermodynamic modeling for the coexisting Datong alkaline and tholeiitic basalts, North China, to put constraints on the source characteristics and melting conditions of these basalts. The alkaline basalts and tholeiites have contrasting major elemental and Sr–Nd–Hf isotopic compositions (87Sr/86Sr, 0.703431–0.703603 vs. 0.703912–0.704010; Ɛ Nd , 5.5–7.0 vs. 2.9–4.8; Ɛ Hf , 9.5–10.6 vs. 8.2–9.6) and trace-element patterns. The observed compositional differences could be explained by neither magmatic processes (e.g., assimilation and/or fractional crystallization), varying melting degrees due to lithospheric thickness variations, nor the melt-peridotite interaction. Instead, lithological heterogeneous mantle sources containing different types of pyroxenite should be considered to account for the compositional diversity, as indicated by olivine (high Ni, Zn/Fe, and low Mn/Zn) and whole-rock chemistry (low CaO, high Fe/Mn, FC3MS, and FCKANTMS ratios). The incompatible trace element-based thermodynamic modeling was carried out to put quantitative constraints on the mantle source characteristics and melting conditions. The modeling results suggest that the source of alkaline basalts contains 30% silica-deficient pyroxenite +70% lherzolite with melting pressure and temperatures of 2 GPa, 1370 °C, whereas the tholeiites have 15% silica-excess pyroxenite +85% harzburgite in the source with melting conditions of 1.5 GPa, 1425 °C. The estimations, together with the thermobarometric calculations using major-element compositions, suggest that the tholeiitic and alkaline basalts were generated by melting of lithological heterogeneous mantle sources under lithosphere with different thicknesses. The lithological heterogeneity originated from recycled components from historical peripheral subducted plates, and melting of such lithological heterogeneity is likely related to asthenosphere upwelling and convection induced by the deeply subducted Pacific slab. Our study emphasizes the fundamental role of source lithological heterogeneity in the generation of co-existing Datong alkaline and tholeiitic basalts and may provide insights into the origin of compositional diversity of other continental intraplate basalts. • Melting conditions of the Datong basalts was constrained using thermodynamic methods. • Silica-deficient and silica-excess pyroxenite is in the source of Datong alkaline and tholeiitic basalts, respectively. • Source lithological heterogeneity is important in the compositional diversity of continental intraplate basalts. [ABSTRACT FROM AUTHOR]

To investigate the effects of melt metasomatism on Ca isotope fractionation in the upper mantle, we report data for xenoliths of mantle rocks including composite peridotites (lherzolites and coexisting pyroxenite veins), wehrlites as well as eclogites and their minerals. The δ44/40Ca in clinopyroxenes (cpx) are ~0.3‰ lower than in coexisting garnets in eclogites from the Udachnaya kimberlite in the central Siberian craton, likely due to equilibrium isotope fractionation controlled by mineral structure. The δ44/40Ca in the pyroxenite veins (cumulates from mafic melts) and their host peridotites from Tariat in Mongolia range from 0.88‰ to 1.05‰ and are close to the Bulk Silicate Earth (BSE) value (0.95 ± 0.05‰). Calcium-Fe-rich low-orthopyroxene lherzolites and wehrlites (LW series) from Tok in the SE Siberian craton were formed by reaction of lherzolites and harzburgites with percolating silica-undersaturated melts. The δ44/40Ca (0.65‰–0.87‰) in the Tok peridotites are negatively correlated with Ca/Al, (Gd/Yb) N and positively with Mg# (indices of melt-rock reaction degrees). The δ44/40Ca ratios in the LW series rocks are lighter than the BSE estimate, and similar to δ44/40Ca in their host basalt and an olivine-cpx cumulate (0.70‰–0.76‰). These observations on the Tok suite can be explained by reaction of refractory peridotite protoliths with melts derived from mixed eclogite and peridotite sources. Overall, the δ44/40Ca variations in eclogites, wehrlites, and various rocks related to silicate melts in the mantle, demonstrate that melt metasomatism plays an important role in producing mantle Ca isotope heterogeneity. The data for coexisting garnet and cpx in eclogites outline a potential for garnet to affect Ca isotope inter-mineral and solid-melt fractionation in the mantle. [ABSTRACT FROM AUTHOR]

Crustal recycling is an important cause of mantle heterogeneity and can have significant control on basalt compositions. Recycled components from the subducted (Paleo-) Pacific slab have frequently been recognized in Cenozoic basalts from the eastern North China Craton (NCC). However, it still remains unclear if the subducted Paleo-Asian oceanic slab contributed to intraplate basalts in this Craton. In a search of evidence for the recycled components from this slab, we have studied the Ar Ar age and elemental and Sr-Nd-Pb isotope compositions of newly-discovered basalts from the Langshan area and compiled a regional synthesis of Cenozoic alkali basalts from the northwest NCC. This region is far from the Pacific domain but near the suture zone of the Paleo-Asian Ocean. With a Late Cretaceous eruption age (~89 Ma), Langshan basalts have low silica and high FeO, MgO and alkali contents, high incompatible elemental concentrations, positive Sr, Eu, Ba, Nb, Ta and negative Pb and Ti anomalies with high Ce/Pb, Nb/U, Ba/Rb and low Rb/Sr ratios, superchondritic Zr/Hf ratios, and uniform radiogenic isotopes (ɛ Nd(t) = 2.20–2.51, 87Sr/86Sr i = 0.703807–0.704348, 206Pb/204Pb i = 18.200–18.384, 207Pb/204Pb i = 15.470–15.515, 208Pb/204Pb i = 38.007–39.430). These compositional characteristics are shared by the compiled Cenozoic alkali basalts from northwest NCC. The low silica and high MgO, FeO and alkali contents together with the positive Nb, Ta and negative Ti anomalies were probably controlled by silica-deficient garnet pyroxenite in the mantle source. The partial melting is estimated to have occurred in the asthenosphere under an average mantle potential temperature of 1300–1450 °C and a pressure of ~2.5 GPa. The positive Sr, Eu, Ba, Nb, Ta anomalies and the canonical indices (high Ce/Pb, Nb/U, Ba/Rb and low Rb/Sr ratios) indicate the involvement of a subducted oceanic igneous slab. Considering 1) the particular tectonic locality of the study region, 2) the inferred northward increase of silica-deficient pyroxenite in the mantle source of Cenozoic alkali basalts, 3) the evidence for strong lithospheric modification beneath the northwest NCC induced by slab-derived components, and 4) the longevity of subducted slabs in convecting mantle, the subducted Paleo-Asian oceanic materials could have introduced ubiquitous mantle heterogeneity beneath the northwest NCC and played a significant role in the generation of the Late Cretaceous to Cenozoic intraplate alkali basalts there. [ABSTRACT FROM AUTHOR]

Partial melting of mantle peridotite from which considerable amounts of melt have been extracted during prior melting episodes generates melts characterized by low incompatible element contents and very low ratios of highly to moderately incompatible elements, so-called 'ultra-depleted' melts. Reaction of peridotite with percolating ultra-depleted melts has been inferred from petrological-geochemical studies of abyssal peridotites and ophiolites. But so far, direct evidence for the existence of ultra-depleted melts, only comes from rare melt inclusions. Here, we show that a pyroxenite layer within abyssal peridotite from the Mid Atlantic Ridge (8°N, Doldrums Fracture Zone) formed by crystallization of a segregated melt that is highly depleted in incompatible elements, at >27 km-depth beneath the ridge axis (at T ∼ 1250 °C), and with little or no modification by interaction with the host harzburgite. During exhumation, the pyroxenite experienced decompression and partial re-equilibration under plagioclase-facies conditions (∼1060 °C and ∼ 15 km depth). The high Hf isotope ratio (εHf = 40.3) of the pyroxenite clinopyroxene is inherited from a melt sourced from an ultra-depleted peridotite that evolved with high Lu/Hf. The associated MORB-like Nd isotope ratios (εNd = 10.6), however, imply a long-term evolution of the source peridotite with composition moderately depleted in incompatible elements (rather low Sm/Nd). These compositions are different from the host harzburgite, but typical for peridotites that have melted and partially reacted with migrating melts in ancient times. Hence, the pyroxenite investigated here is a partial melt from an ultra-depleted peridotite that has become re-enriched in incompatible elements and clinopyroxene. This melt crystallized in the oceanic lithosphere, and partially re-equilibrated at low pressure during exhumation. Overall, our results show that renewed melting of ultra-depleted peridotites with a complex history of prior melting and melt-rock reaction occurs, and that such melts migrate through the sub-ridge mantle, and can thus contribute to mid ocean ridge magmatism, although to a still unknown extent. • Abyssal pyroxenite hold the isotopic signature of the parental melt mantle source. • Doldrums pyroxenite records melting of ancient ultra-depleted, refertilized mantle. • Refertilization enables melting of otherwise ancient refractory mantle. • Segregation shortly after melt generation preserves trace of ultra-depleted melts. [ABSTRACT FROM AUTHOR]

Abstract Constant concentration ratios of trace elements of similar incompatibility in oceanic basalts have been used to determine the abundances and ratios of incompatible elements in mantle sources. However, the Mo content in the depleted mantle estimated from Mo/Ce of oceanic basalts is much lower than the estimates based on Mo contents of mantle peridotites (25 ± 7 ng/g versus 113–180 ng/g). This discrepancy partly reflects uncertainties about the geochemical behavior of Mo in the upper mantle. New Mo concentration data obtained on unserpentinized spinel-facies lherzolites, harzburgites, dunites and pyroxenites (n = 47) from the Balmuccia and Baldissero mantle tectonites (Ivrea-Verbano Zone, Italian Alps) provide new insights into the magmatic behavior of Mo and its content in the depleted mantle. The peridotites and pyroxenites display variable and very low Mo contents (4–16 and 3–11 ng/g, respectively). A slightly serpentinized peridotite shows the highest Mo content of 37 ng/g. The Mo contents show no systematic variation with other incompatible or compatible elements. Chromium spinel in the mantle rocks likely does not control the behavior of Mo because complete or incomplete digestion of spinel did not lead to a noticeable change in Mo contents. The data indicate incorporation of little Mo into olivine, pyroxene and spinel, consistent with the predominant occurrence of Mo6+ and the experimentally-determined very low partition coefficients at typical upper mantle conditions. Bulk rock Mo contents from these mantle rocks, upper limits of Mo contents in sulfides and the lack of variations of Mo concentrations with varying sulfide fractions indicate a minor influence of mantle sulfides on the bulk Mo budget. The incorporation of little Mo into pyroxenes of the mantle rocks contrasts with Ce which is mainly controlled by clinopyroxene. This contrasting behavior suggests a higher incompatibility of Mo than Ce during mantle melting. Modelling polybaric melting of mantle peridotites indicates that melts formed at low degrees of partial melting (e.g., <2–3%) would not retain the Mo/Ce of the mantle sources. In contrast, >5% melting, as in tholeiites and komatiites, leads to extraction of nearly the complete inventory of Mo and Ce into the melts and thus Mo/Ce of the melts should be representative of mean mantle source compositions, even if Mo and Ce show different bulk partition coefficients. Molybdenum contents of mantle rocks from the Ivrea Zone and the data on oceanic basalts and komatiites indicate that the depleted upper mantle has a low Mo content of <30 ng/g. Highlights • Mantle peridotites and pyroxenites from the Ivrea Zone have very low Mo contents. • Little Mo is incorporated into olivine, pyroxene and spinel in mantle rocks, consistent with experimental constraints. • Mantle sulfides play a limited role for the bulk Mo budget of peridotites and pyroxenites. • Mo is more incompatible than Ce in the depleted mantle. • The depleted mantle has a low Mo content of <30 ng/g. [ABSTRACT FROM AUTHOR]

For the last three decades recycled oceanic lithosphere has been proposed as the principal mantle source for mafic magma erupting on oceanic islands. However, the composition of this lithosphere, as well as how and what part of it melts, is a matter of debate. Several melting experiments show that a mafic pyroxenite/eclogite lithology melts at higher pressure than that of ultramafic peridotite/lherzolite. In addition, the role of carbonate recycling and a carbon-rich fluxing agent producing carbonated silicate melt is increasingly being discussed. The historical Timanfaya eruption (1730–1736) at Lanzarote, Canary Islands, produced a range of mafic magmas that evolved with time from basanite trough alkali basalt to silica-saturated tholeiite, whereas only basanite was erupted in the smaller 1824 eruption. Mantle nodules are common and most lavas and tephra are characterised by abundant olivine macrocrysts. Variable trace element ratios of whole-rock tephra (Rb/Th: 3.3–7.7; La/Yb: 12–44; Nb/Ta: 16–21), and more especially those of olivine-bearing melt inclusions (Rb/Th: 2.5–12; La/Yb: 9–77; Nb/Ta: 14–22) suggest rapid melt extraction from a heterogeneous mantle source beneath the island. Increasing (Nb,Ta)/Th, Zr/Hf and decreasing Nb/Ta from basanite through alkali basalt to tholeiite reflect decreasing carbon-rich fluxing on a time scale of less than a few years. Increasing K, Rb, Cs, and Ba over Th with time during the Timanfaya eruption most likely result from diminishing phlogopite control during the production of increasingly silica-saturated magma. Highly variable melt inclusion compositions are compatible with fractional melting of carbonated pyroxenite, with the whole-rock composition most likely representing the amalgamation of these fractional melts at the asthenosphere-lithosphere boundary beneath the eastern Canary Islands. [ABSTRACT FROM AUTHOR]

A detailed study of the Sr- and O- isotopic composition of minerals in the Upper Group 2 (UG2) unit of the eastern Bushveld Complex revealed significant variations in the pegmatoidal footwall pyroxenite and the UG2 chromitite layer itself. Initial 87 Sr/ 86 Sr ratios (Sr i ) of interstitial plagioclase in the footwall rocks are significantly higher than in adjacent rocks of the Critical Zone. The δ 18 O values of plagioclase-pyroxene and plagioclase-chromite mineral pairs in footwall pyroxenite and chromitite are consistent with oxygen isotope exchange from crystallization temperatures down to 700 °C and 800 °C, respectively, whereas in other parts of the UG2, plagioclase and pyroxene δ 18 O values are consistent with O isotope equilibration at liquidus crystallization temperatures of approximately 1150 °C. We propose that the higher initial 87 Sr/ 86 Sr of plagioclase in the UG2 chromitite and the footwall pyroxenite reflects infiltration of fluid derived from dehydrating country rocks beneath the Bushveld Complex. Fluids could have been introduced either late or during chromitite crystallization. The chromitite layer and its pegmatoidal footwall could have acted as barriers to entrap late stage crustal fluids. Alternatively, reaction of fluids with the chromite-bearing crystallizing mush layer would induce hydration melting of pyroxene and plagioclase supplying Cr and Al to the melt, thus invoking further chromite formation. This dissolution-reaction releases Cr from pyroxene, which along with pre-existing disseminated chromite could refine the chromitite layers. The pegmatoidal UG2 footwall constitutes the reaction front-line of this process, which recrystallized to a coarse-grained rock upon cooling. This study emphasizes the role of crustal fluids and late- and post-magmatic processes in layered intrusions taking place at subsolidus temperatures, whereby isotopic and chemical compositions may be significantly changed. [ABSTRACT FROM AUTHOR]

Mineral water contents, together with the major and trace element compositions of minerals and whole-rock, were determined for garnet pyroxenites enclosed by ultrahigh-pressure (UHP) metamorphic gneiss at Hujialin in the Sulu orogen. The garnet pyroxenites have low SiO 2 contents of 40.25 to 46.68 wt% and MgO contents of 10.99 to 14.79 wt%. They are characterized by enrichment in LREE and LILE (Ba, Sr, Pb) but depletion in HFSE (Nb, Zr) and HREE. They were generated in the Triassic by metasomatic reaction of the mantle wedge peridotite with hydrous felsic melts derived from partial melting of the deeply subducted continental crust. Measured water contents vary from 523 to 1213 ppm for clinopyroxene, and 55 to 1476 ppm for garnet. These mineral water contents are not only correlated with mineral major and trace element abundances but also relatively homogenous within single mineral grains. Such features preclude significant disturbance of the mineral water contents during pyroxenite exhumation from the mantle depth to the surface and thus indicate preservation of the primary water contents for the UHP metasomatites. The garnet pyroxenites are estimated to have bulk water contents of 424–660 ppm, which are higher than those for the MORB source, similar to or higher than those for the OIB sources and close to the lower limit for the arc magma source. The relationships between contents of mineral water and some elements suggest that the high water contents of garnet pyroxenites are primarily determined by the abundance of water-rich clinopyroxene. Garnet also has the high water contents, suggesting its importance in hosting water at mantle depths. Calculated whole-rock H 2 O/Ce ratios are 63–145, higher than those for Hawaiian garnet pyroxenites and SWIR abyssal pyroxenites. These observations suggest that metasomatic pyroxene-rich lithologies have the capacity to contribute high H 2 O concentrations and variable H 2 O/Ce ratios to the mantle. This lends support to the interpretation that the source of some intraplate basalts may be a heterogeneous mixture of peridotite and pyroxenite. On the other hand, the high water contents of garnet pyroxenites suggest that the presence of ultramafic metasomatites in the mantle wedge would enhance its water storage and thus reduce the water transport into deeper mantle by subduction. [ABSTRACT FROM AUTHOR]

Nickel contents and Mn/Fe in olivine phenocrysts have been suggested to reflect the mineral composition of the mantle source of the host magma. This hypothesis is tested here against a well-characterized suite of meimechitic (or Ti-rich komatiitic) dikes from the Antarctic extension of the Jurassic ~ 180 Ma Karoo large igneous province. The presented trace element data on Fo 82–92 olivines show relatively high Ni (2430–3570 ppm) and low 100 ∗ Mn/Fe (1.32–1.5; Mn = 890–1570 ppm), compatible with pyroxenite-rich sources ( X px = 37–75%). Many other mantle source indicators (parental melt MgO and whole-rock Zn/Fe, MgO/CaO, FC3MS, Zr/Y vs. Nb/Y, and radiogenic isotope compositions) suggest dominantly or solely peridotitic mantle sources, however. Therefore, the measured high Ni and low Mn/Fe are likely to reflect high temperatures and pressures of melting and possibly high water contents in such peridotite sources. We recommend considerable caution when using Ni and Mn contents of olivine as source indicators, as they may only serve for qualitative comparison of primitive volcanic rocks that originated under fairly similar mantle conditions. [ABSTRACT FROM AUTHOR]

Mantle-derived eclogite and pyroxenite xenoliths from the Jurassic Voyageur kimberlite on the northern Slave craton in Arctic Canada were studied for garnet and clinopyroxene major and trace element compositions, clinopyroxene Pb and garnet O isotopic compositions, and garnet Fe 3 + /ΣFe contents. The Voyageur xenoliths record a wide range of pressures, but are cooler compared to mantle xenoliths derived from the nearby, coeval Jericho kimberlite. The CaO, TiO 2 and Zr contents of Voyageur eclogites increase with depth, which is also observed in northern Slave peridotite xenoliths, demonstrating ‘bottom-up’ metasomatic processes within cratonic mantle lithosphere. The Voyageur eclogites have positive Eu anomalies, flat HREE N patterns, and major element compositions that are consistent with ultimate origins from basaltic and gabbroic protoliths within oceanic lithosphere. Clinopyroxene Pb isotope ratios intercept the Stacey-Kramers two-stage terrestrial Pb evolution curve at ca. 2.1 Ga, and form an array towards the host kimberlite, indicating isotopic mixing. The 2.1 Ga eclogite formation age broadly overlaps with known Paleoproterozoic subduction and collision events that occurred along the western margin of the Slave craton. Unlike the eclogites, the Voyageur pyroxenites contain garnet with distinctive fractionated HREE N , sinusoidal REE patterns of calculated bulk rocks, and clinopyroxene with 206 Pb/ 204 Pb ratios that intercept the Stacey-Kramers curve at 1.8 Ga. This suggests a distinct origin as Paleoproterozoic high-pressure mantle cumulates. However, the pyroxenite Pb isotope ratios fall within the eclogite array and could also be explained by protoliths formation at ca. 2.1 Ga followed by minor isotopic mixing during mantle metasomatism. Thus, an alternative scenario involves pyroxenite formation within the mantle section of Paleoproterozoic oceanic lithosphere followed by variable metasomatism after incorporation into cratonic mantle lithosphere. This model allows for a linked petrogenesis of the Voyageur eclogites (crust) and pyroxenites (mantle) as part of the same subducting oceanic slab. Oxygen fugacity determinations for one pyroxenite and ten eclogite xenoliths show a range of 3 log units, from − 4.6 to − 1.6 ΔFMQ, similar to the range observed for nearby Jericho and Muskox eclogites (ΔFMQ − 4.2 to − 1.5). Importantly, the northern Slave eclogite and pyroxenite mantle components are highly heterogeneous in terms of redox state provided that they range from reduced to oxidized relative to Slave peridotite xenoliths. Moreover, the Voyageur eclogites do not exhibit any trend between oxidation state and equilibration depth, which contrasts with the downward decrease in f O 2 shown by Slave and worldwide cratonic peridotite xenoliths. Our investigation of mantle eclogite and pyroxenite f O 2 reinforces the important influence of recycled mafic components in upper mantle processes, because their high and variable redox buffering capacity strongly controls volatile speciation and melting relations under upper mantle conditions. [ABSTRACT FROM AUTHOR]

Magmatic underplates, which are characterized by high seismic velocity and anomalously high density in comparison to normal lower crust, are among the most significant compositional heterogeneities identified in the rift lithosphere. The interaction of a magma with the lithospheric mantle to form pyroxenites is considered a viable mechanism for the creation of an underplate, though the details of such interaction are difficult to constrain with geophysical techniques alone. Here we present a petrographic and geochemical study of pyroxenite xenoliths erupted on the western Ethiopian plateau from along the Yerer Tullul Wellel Volcano-Tectonic lineament (YTVL). The equilibration pressures recorded by the xenoliths (0.88–1.49 GPa) are consistent with an origin within a seismically and gravitationally defined underplate that exists beneath the western Ethiopian plateau. Three groups of pyroxenite have been identified: Group I – characterized by a relative enrichment of orthopyroxene, and clinopyroxene with unfractionated heavy rare elements (HREE; Tb N /Yb N = 1.2–1.6); Group II – characterized by a relative enrichment in clinopyroxene, which have fractionated HREE (Tb N /Yb N = 1.6–2.0); Group III – transitional xenoliths that contain pyroxene characteristic of Group I but also contain clinopyroxene characteristic of Group II. The magma responsible for the formation of Group I xenoliths was likely of sub-alkaline composition and derived from a spinel peridotite on the basis of relatively unfractionated REE profiles. This magma may have been the Oligocene Low Titanium (LT)-type flood basalts associated with the Afar plume, or subduction-related magmatism associated with island arc activity during the construction of the regional lithosphere during the Neoproterozoic Pan African Orogeny. The magma responsible for the formation of Group II xenoliths was likely of alkaline composition and derived from a source with some garnet on the basis of fractionated HREE. This magma may have been the Oligocene High Titanium (HT)-type flood basalts associated with the Afar plume or Miocene-recent magmas associated with the reactivation of the YTVL during rift development. Our results show that the lithospheric mantle beneath the YTVL was impacted by at least two metasomatic events and exhibits characteristics more similar to the xenolith suites from the rift valley than elsewhere on the western Ethiopian plateau. We propose that the magmatic underplate beneath the western Ethiopian plateau was formed by multiple episodes of magma-lithosphere interaction, likely developing over time rather than in a single event. [ABSTRACT FROM AUTHOR]

Oceanic island basalts are thought to originate from partial melting of lithologically heterogeneous mantle. Sobolev and coworkers proposed the use of fractionation-corrected Ni concentration Ni × (FeO/MgO) and Fe/Mn measured in olivine as indicators of pyroxenite-derived component in mixtures of primary melts. Trace element concentrations were measured in olivine from two historical eruptions in Lanzarote (Canary Islands), which erupted mafic and mantle nodule bearing magmas, ranging in composition from highly silica-undersaturated basanites through alkali basalts to tholeiites. Tephra from each of the five eruption phases of the 1730–36 CE Timanfaya eruption contains olivine phenocrysts with cores with restricted Fo variation (~ 3 mol%), whereas trace element concentrations and ratios vary by a factor of approximately two. The largest variation is observed in early basanites with later eruption phases producing olivine macrocrysts containing trace elements within that range. Ni × (FeO/MgO) and Fe/Mn of olivine in the basanites span the whole range of olivine in MORB and OIB worldwide, whereas the olivine of subsequent phases has values akin to those crystallized from pyroxenite melts. The Fo content decreases systematically with time during the 1730–36 eruption and the proportion of silica-saturated primary melt increased in the parental magma mixture with time. At the end of the eruption, tholeiite magmas crystallized olivine with lower Fo content, whereas the concentrations of Mn and Ca increased together with Ca/Al at relatively uniform Ni × (FeO/MgO) and Fe/Mn, all of which is readily explained by increased decompression melting at lower temperature. The basanite from the eruption that took place in 1824 CE has olivine with even higher Fo content and trace element variability similar to that of the Timanfaya basanite. The fact that the Lanzarote basanites contain olivine with trace element systematic spanning that of MORB and pyroxenite melt can be explained by CO 2 -flux melting of a lithologically heterogeneous source, generating the diverse compositions. Alternatively, reactive porous flow through depleted oceanic lithosphere and equilibration with dunitic restite caused Ni depletion of percolating pyroxenite melt from which olivine of the basanites crystallized. The fact that olivine compositions and basanite magma were reproduced approximately a century later may reflect episodic carbonatic fluxing in the slowly uprising Canarian mantle plume. [ABSTRACT FROM AUTHOR]

Mantle heterogeneity has a first-order control on the petrological and geochemical differences of erupted mafic lavas worldwide. Because of their contrasted distributions between mantle phases, First Row Transition Elements (FRTEs) have been considered potential lithological tracers. Using a combination of published data on natural and experimental samples and new high-precision analyses on a variety of pyroxenite samples, we investigate the parameters that control FRTE partition (D s) and exchange (K D s) coefficients between common mantle minerals. We demonstrate that mineral-clinopyroxene exchange coefficients are independent of composition and temperature and that coefficients obtained from natural samples can be accurate as long a sufficiently high number of compositionally diverse samples are considered, making them reliable input parameters in mantle melting models. As a proof of concept, we use the exchange coefficients determined from natural mantle lithologies in this study, along with published experimental clinopyroxene/melt partitioning coefficients, to perform simple inverse modeling on two basalt suites from the Western Volcanic Zone in Iceland and Samoa, selected for their contrasted Mn/Fe and Zn/Fe ratios. Our results show that a given FRTE ratio in basalt can be explained by a large range of modal proportions in the source. However, when combined, FRTE ratios become a powerful tool to constrain the nature of the source. [Display omitted] • Natural samples can be used to obtain K D s and complement experimental data. • The number and diversity of samples control the accuracy of the K D values. • K D FRTEs mnl / cpx are independent of composition and temperature. • A FRTE ratio in basalt can result from a range of modal proportions in the source. • When combined, FRTE ratios become a powerful tool to characterize magma sources. [ABSTRACT FROM AUTHOR]

Melt inclusions and their host olivine provide unique information about the nature, distribution and scale of mantle source heterogeneity. We present the first analyses of the chemical and Pb isotope compositions of melt inclusions and their host olivine from the Cenozoic Hannuoba basalts, North China Craton, which contains coexisting suites of alkali and tholeiitic basalts. There is limited variation in major and trace element composition, but significant Pb isotopic variation, in the tholeiitic samples. This contrasts with the substantial variation in major and trace elements, but limited Pb isotopic variation, in the alkali basalt samples. Based on the results of the major- and trace-element modeling, the compositional variation of the alkali basalts may be primarily attributed to garnet pyroxenite melting, with only a small input from garnet peridotite melts. The garnet–pyroxenite component involved in the genesis of the alkali basalts, of metasomatic origin, may have formed in the lithosphere. The wide range of isotopic composition of the tholeiitic basalts and remarkably limited variability of major and trace element composition is argued to be characteristic of the source region, which is a mixture of peridotite with small amounts of pyroxenite transformed from recycled ancient oceanic crust and sediment. The presence of recycled oceanic crust in the mantle source and the low–velocity anomaly observed beneath the Taihang Mountains, located 50 km to the southeast of the Hannuoba region, supports a plume model for their origin. [ABSTRACT FROM AUTHOR]

Mapping the compositional variability of the Earth's mantle is fundamental for understanding mantle dynamics, crustal recycling and melt generation. The geochemistry of intraplate oceanic basalts in particular points to a heterogeneous convecting mantle, often explained by variable proportions of fertile (pyroxenite, eclogite) and depleted (peridotitic) domains. A key parameter necessary to constrain the Earth's deep processes is water because it influences melting and affects the physical properties of the Earth's mantle. However, there are only a few direct water determinations on samples of the oceanic mantle. Here we report water concentrations by FTIR on garnet pyroxenite xenoliths from Salt Lake Crater vent, Oahu, Hawaii, in order to constrain the role of lithological variability in the distribution of water in the upper oceanic mantle. The clinopyroxenes have 260 to 576 ppm H 2 O and the orthopyroxenes about half these amounts. Curiously, garnets have water concentrations below detection limits (< 0.5 ppm H 2 O). Calculated melts in equilibrium with the pyroxenes have, on average, 2.3 ± 0.4 wt.% H 2 O and 243 ± 83 H 2 O/Ce. These H 2 O concentrations are similar to estimates for the H 2 O concentration of primary magmas of the rejuvenated Hawaiian stage volcanism, and 3 to 5 times higher than primary magmas of the shield stage. This supports earlier conclusions where the pyroxenites are interpreted as high pressure cumulates from alkali magmas similar to the rejuvenated stage magmas within the Pacific lithosphere. The reconstructed bulk pyroxenites have 211 to 467 ppm H 2 O, similar to estimates for the source of Hawaiian magmas and two to four times higher than estimates for the depleted MORB source (~ 100 ppm H 2 O). Despite their high water concentrations, however, the bulk pyroxenites have much lower H 2 O/Ce ratios than the MORB source (35–115 vs. 150–210, respectively) or Hawaiian magma sources (> 160). The discrepancy between the low H 2 O/Ce ratios in the bulk pyroxenites and high H 2 O/Ce in the equilibrium melts is consistent with experimental data that predicts 4 to 5 times higher partition coefficient for Ce than H in these pyroxenes. Therefore, the process of high-pressure crystallization in the oceanic lithosphere will create pyroxene-rich lithologies which are paradoxically, both “wet” (i.e., high H 2 O concentrations) and “dry” (low H 2 O/Ce ratios) compared to the source of their parental melts. Phlogopite is present as a trace phase in these rocks (< 0.4% modal) with relatively minor contribution on the bulk water contents. The coupled high H 2 O, low H 2 O/Ce ratios of the pyroxenites are similar to the inferred source of several Enriched Mantle (EM)-type Ocean Island Basalts (OIB), like the Samoa, Pitcairn, Society, and Kergulen hot spots, as well as the EM-1 type Koolau endmember of the Hawaiian magmas. We suggest that recycling of pyroxenite-bearing oceanic lithosphere can explain the relatively high H 2 O and low H 2 O/Ce ratios of some EM-type OIB. Our data suggests a link between lithological variability and heterogeneous water distribution in the upper mantle. [ABSTRACT FROM AUTHOR]

The 190 Pt– 186 Os system could be useful to study a number of different geologic processes. Over the last decade, this system has received considerable attention owing to its potential to detect core–mantle signatures in plume-derived lavas. Additionally, this system may also be useful as a proxy for melt depletion, examining the role of pyroxenites/sulfides in mantle melting and for dating of Platinum Group Element (PGE) based ores. The total natural variation of 186 Os/ 188 Os is limited (~ 150 ppm in the mantle), which thus requires extremely high levels of accuracy and precision to discern natural variations. 186 Os/ 188 Os ratios were measured in laboratory standards and natural samples to examine and quantify the sources of error affecting high precision 186 Os/ 188 Os measurements and a set of refined analytical procedures to minimize and reduce these various sources of error were recommended. Inter and intra-run variations in oxygen isotopic composition used in oxide correction of OsO 3 − peaks can produce systematic shifts in 186 Os/ 188 Os. Useful analysis time is wasted in measuring O isotopic composition line-by-line and it is sufficient to utilize pre- and post-run determination of O isotopic composition for performing oxide correction. Although the use of 189 Os/ 188 Os as a normalizing ratio for mass fractionation results in lower uncertainty in Os isotopes from oxide correction, this benefit is overwhelmed by the increased error propagation from mass fractionation. Instead, usage of the 192 Os/ 188 Os normalization generates the least propagated error due to mass fractionation and is recommended. Earlier studies have variably used different O isotopic compositions for oxide correction and different normalizations for mass fractionation correction, which can generate systematic shifts in the reduced data. Use of mutually consistent normalizing values for mass fractionation largely reduces the offsets between reduced data using different normalizing schemes, thus improving data comparison across laboratories. For the quantities of Os and beam intensities of 186 Os utilized in previous 186 Os/ 188 Os studies (several 10's of ng Os, 80–250 mV of 186 Os with a 10 11 Ω resistor), the largest source of analytical error is derived from Johnson noise on baseline integrations. Increasing the duration of baseline measurement significantly reduces the Johnson noise error, allowing to achieve maximum external precision for a given sample size. Several 186 Os/ 188 Os analyses reported in the literature seem to have been compromised by an interference, which may be WO 3 − , organics or some other unidentified species that affects both the 186 Os/ 188 Os and the 184 Os/ 188 Os ratios. This makes it extremely critical to monitor and correct PtO 2 − interferences, which could mask other potential interferences that directly affect 186 Os/ 188 Os measurements. [ABSTRACT FROM AUTHOR]

Forsterite (Fo)-rich olivine compositions from San Carlos (Arizona, SW USA) are commonly used as starting material in experimental petrology. In comparison to the San Carlos reference material USNM 111312/444, it has been shown that the major element variability of non-USNM San Carlos olivine is significant. We complement the characterization of the compositional variability of the non-USNM San Carlos olivine with new data, including minor and trace element analyses. High precision analyses reveal that selected minor elements (e.g., ~6% NiO, ~10% MnO, ~16% CaO, relative) and trace elements (e.g., ~75% Cr, ~120% Cu, ~160% P and Ti, relative) present significant concentration variations between grains. At the scale of the individual grain, however, San Carlos Fo-rich olivines appear homogeneous with no systematic core-rim variations. We also discuss the origin of olivine pyroxenites associated with the peridotite xenoliths and argue that they are derived by melt-rock reaction resulting in olivine dissolution and pyroxene precipitation from the peridotite host. Further segregation and in situ crystallization of the hybrid residual melt can produce Fe-rich olivine-poor pyroxenites. Finally, we discuss the origin of the P variability in San Carlos mantle olivine and suggest that P enrichment by metasomatism implies a highly reactive process with fast dissolution-reprecipitation of solid phases. • We report the major, minor & trace variability of non-USNM San Carlos olivine • Olivine trace elements can discriminate magmatic and mantle origins for xenoliths • SC olivine pyroxenites are formed by melt-rock reaction • Metasomatic events must result in fast crystal growth to explain P olivine enrichment [ABSTRACT FROM AUTHOR]

Abstract: We report high-precision Sr, Nd, Hf and Pb isotope compositions for twenty-one postshield-stage and two post-caldera lavas from Kahoolawe, a Loa-trend Hawaiian volcano. Kahoolawe postshield- and shield-stage lavas have overlapping though highly heterogeneous Sr, Nd, Hf and Pb isotope compositions, implying that the shield- and postshield-stage volcanism at Kahoolawe sampled the same isotopically heterogeneous mantle source. This differs from that of most other Hawaiian volcanoes, such as Haleakala, Mauna Kea, and Hualalai, whose shield-to-postshield transitions are characterized by shifts to lower 87Sr/86Sr and higher 143Nd/144Nd. There are correlations between CaO, Sc and V contents and radiogenic isotope compositions within Kahoolawe postshield-stage lavas. For example, Sc abundance is negatively correlated with 87Sr/86Sr, and positively correlated with εNd and εHf; V abundance is positively correlated with εNd, εHf, and 206Pb/204Pb. Element-isotope correlations are also observed in Mauna Kea postshield-stage lavas: Sc and V abundances are negatively correlated with εHf and 206Pb/204Pb, and positively correlated with εNd. These trends may be due to magma–magma mixing. That is, in addition to clinopyroxene fractionation to account for the low CaO, Sc and V contents in some postshield-stage lavas, partial melts of eclogite/garnet pyroxenite, characterized by low CaO, Sc and V contents, may also be part of the petrogenesis of Kahoolawe postshield-stage lavas. It is well established that lavas erupted at the geographically defined Loa- and Kea-trend volcanoes have different isotopic and geochemical compositions. Specifically, compared to the Kea-trend lavas, Loa-trend lavas have higher 208Pb/204Pb at a given 206Pb/204Pb. However, cases exist of both shield- and postshield-stage volcanism where Kea-type isotopic signatures are present in Loa-trend volcanoes and the reverse. We propose that Loa- and Kea-type source components are present beneath both Loa- and Kea-trend volcanoes in such a way that the average source compositions of Loa-trend volcanoes have a Loa-type isotopic signature, and that of the Kea-trend volcanoes have a Kea-type isotopic signature. When the size of the magma capture zone is much larger than that of the source components, the erupted lavas have the average compositions of the source. If the size of the magma capture zone is comparable to that of the source components, the erupted lavas could have either Loa- or Kea-type isotopic signatures. [Copyright &y& Elsevier]

Abstract: Detailed mineralogical and geochemical studies have been carried out on a suite of pyroxenite xenoliths captured in a late Mesozoic basaltic dike from Junan, eastern Shandong, which is tectonically situated in the Sulu ultrahigh pressure orogenic belt. Two types of pyroxenites, namely websterite and garnet pyroxenite were identified according to their mineral assemblages. The equilibrium temperatures (828–935°C) of websterite and garnet pyroxenite xenoliths, falling within the temperature range of lower crustal xenoliths, suggest that they were derived from the lower crust rather than the lithospheric mantle. The websterite xenoliths are characterized by higher MgO and lower Al2O3; their convex upward REE patterns, along with lower concentrations of highly incompatible elements indicate that they were high pressure cumulates. Their extremely unradiogenic Nd and radiogenic Sr isotopic compositions imply the contribution of crustal materials to their sources. It is suggested that the precursor melts of websterite xenoliths were derived from a mantle source which had been intensely modified by the subducted lower continental crust of the Yangtze craton following the collision with the North China craton in the Triassic. The compositional features of garnet pyroxenites also suggest their cumulative origin, however, the sharp contrast in trace element concentrations and Sr–Nd isotopic compositions with those of websterite xenoliths indicate they have different precursor melts. Their Sr and Nd isotopic compositions are consistent with the suggestion that their precursor melts were derived from an enriched lithospheric mantle which had been metasomatized by melts released from the Proto-Tethyan oceanic crust which was subducted into the mantle prior to the subduction of continental crust of Yangtze craton. The occurrence of websterite and garnet pyroxenite xenoliths provide evidence that the lithospheric mantle of the North China craton had been significantly modified by the recycled oceanic and continental crust resulting from the collision between the Yangtze craton and the North China craton. [Copyright &y& Elsevier]

Abstract: Recent drilling in the Duke Island Complex shows that some olivine clinopyroxenites from Hall Cove contain low-S but high Pt+Pd. In contrast, dunites from the Duke Island Complex and the Yellow Hill intrusion on nearby Annette Island are low in Pt+Pd as well as S. New data from this study show that the low-S ultramafic rocks from Duke and Annette Islands are extreme cumulates of olivine and/or clinopyroxene and all contain trace to minor amounts of cumulus sulfide. As a result, the ratios of the least-mobile, highly chalcophile elements in these rocks mainly reflect those of original sulfide liquids which in turn reflect those of their parental magmas. Our data also show that all dunite samples from Annette Island contain Ni-depleted, primitive olivine with Fo contents >90mol% and Ni contents <1200ppm, indicating a primitive but significantly Ni-depleted parental magma for these rocks. Whole-rock compositions indicate that the parental magma was not depleted in platinum-group elements (PGE). The mantle-normalized PGE patterns of these samples are remarkably similar to dunites from some Ural-type ultramafic intrusions in Ural, Russia. They are all characterized by mantle-like Ru/Ir ratio plus significant Pt enrichment relative to other PGEs. Given their common occurrences in the world, we propose a Ni-depleted, Pt-enriched mantle component as the source for the parental magma of this type of dunite. A comparison of olivine and whole-rock compositions between different types of rocks from the Duke Island Complex reveals that significant Ir–Ru depletions relative to Pt–Pd in olivine clinopyroxenites from Hall Cove are most likely due to Ir-alloy and laurite segregation from their parental magma before it reached sulfide saturation. [Copyright &y& Elsevier]

Abstract: The extent to which the lithological heterogeneity of the mantle source controls mid-oceanic ridge basalt (MORB) compositions remains largely debated. The recycling of oceanic crust into the mantle is widely considered to be an important cause of mantle source heterogeneity. We provide data for major and trace elements and Sr–Nd–Pb isotopes of thirteen mid-ocean ridge basalt (MORB) samples from the East Pacific Rise (EPR) to investigate the mantle source heterogeneity introduced by recycled oceanic crust (ROC). Relationships of Sr–Nd–Pb isotopes and incompatible trace element ratios, such as Sm/Yb and Hf/Zr, and ratios of Zn/Fe, Co/Fe and (Pb/Pb⁎)N indicate a geochemically and lithologically heterogeneous mantle source. Mixing between components from depleted mantle (DMM) and enriched ROC-derived pyroxenite is required to explain these relationships. The relatively high Sm/Yb and the good correlation of 87Sr/86Sr vs. Sr/Nd are in concert with variable contributions from a pyroxenite component derived from recycled plagioclase-fractionated E-MORB. The narrow range of MgO contents (7.49wt.% to 8.58wt.%) and the correlations of major element contents with 87Sr/86Sr and Zn/Fe ratios indicate the control of a mixed pyroxenite–peridotite source on the major element compositions of these lavas. To further investigate the major element contribution from the lithologically heterogeneous mantle source (peridotite/pyroxenite), major element compositions were corrected for fractional crystallization both to a given MgO wt.% and a given liquidus temperature. After correction for fractional crystallization, the FeOT and Na2O increase and CaO decreases with increasing melt proportion from the enriched end-member. Based on the observations of this study and the results from partial melting experiments on pyroxenite and peridotite, a melting model, in which magma forms by mixing a deep and enriched melt (pyroxenite melt) with a shallow and depleted DMM melt (peridotite melt), provides a likely mechanism that can produce the major element systematics. [Copyright &y& Elsevier]

Abstract: We present SHRIMP U–Pb ages and REE data for zircon of two granulites, a garnet pyroxenite and a clinopyroxene amphibolite from the North Dabie complex. The North Dabie complex is a major fault-bounded petrotectonic unit in the Dabie continental collisional orogen between the South and North China Blocks. Zircon REE data are applied to link ages to specific petrological processes. In addition to SHRIMP zircon U–Pb dating, measurements were made on monazites from a HP–UHT metapsammitic granulite. Metamorphic zircon grains and coeval overgrowths on detrital cores from one granulite yielded a weighted mean 207Pb/206Pb age of 2042±7Ma, corresponding to peak HP/UHT metamorphism. The zircon cores have well-preserved oscillatory, magmatic growth zoning and generated a range of 207Pb/206Pb ages from 2766±8Ma to 3089±3Ma, representing the detrital source. Concordant analyses of monazites from the same granulite sample yielded a weighted mean 207Pb/206Pb age of 2001±6Ma, which we interpret as the time of exhumation to the middle crustal level. A magnetite–orthopyroxene–garnet quartzite associated with a banded iron formation has a metamorphic zircon age of 2293±7Ma (granulite-facies) with a detrital core age of 2435±28Ma. These results are interpreted to indicate a composite affinity to the South China and North China Blocks for the ancient Archean–Paleoproterozoic crust in the North Dabie metamorphic rocks. The garnet pyroxenite, known as a typical HP–UHP granulite, contains a multiple, latest Permian to Triassic metamorphic zircon population with 206Pb/238U ages of ca. 254–253, ca. 242–226, ca. 210, and ca. 203Ma. Along with published zircon and titanite U–Pb age data and mineralogical evidence, we suggest that the HP–UHP granulite evolved broadly synchronously with UHP eclogites during continental collision in the latest Permian to Triassic. In the garnet pyroxenite, there is also zircon inheritance of variable ages (ca. 291–2477Ma), indicative of a probable sedimentary origin of the protolith; there also occur young, mid-Jurassic to Cretaceous zircons ranging in age between 165 and 112Ma, signifying considerable crustal reworking in mid-Jurassic and Cretaceous times. Zircons from the clinopyroxene amphibolite have a weighted mean 206Pb/238U age of 780±6Ma representing the minimum time of protolith emplacement, and 133±2Ma for an anatectic event in the post-collisional stage. The age data of this amphibolite are similar to those of granitic gneisses/migmatites in the North Dabie complex. The new U–Pb zircon ages for the North Dabie rocks provide constraints on the affinity of collided crustal blocks, the metamorphic history of HP–UHP granulites, and post-collisional crustal reworking. [Copyright &y& Elsevier]

Abstract: The Jingbulake mafic–ultramafic intrusion in the South Tianshan orogenic belt, Xinjiang, NW China, is a zoned intrusive body composed of gabbro-diorite, olivine gabbro and wehrlite, locally intruded by pyroxenite, within a major W–E structure. Both olivine gabbro and wehrlite contain disseminated Ni–Cu sulfides, whereas pyroxenite hosts an ore body containing massive, net-textured and disseminated sulfide ores. Both silicate rocks and sulfide ores from the Jingbulake intrusion have low Pd/Ir ratios (10–63). The silicate rocks contain Ni-poor olivine (generally <0.1wt.% NiO) and have low Pt (0.48–2.3ppb) and Pd (0.52–4.8ppb) contents and high Cu/Pd ratios (128,000 to 200,000), indicative of significant chalcophile element depletion. The disseminated sulfide ores from No. 1 and 2 mineralization zones have 0.28–1.5ppb Pt and 2.8–8.1ppb Pd, and Cu/Pd ratios (200,000–300,000). The disseminated and net-textured sulfide ores from No. 8 mineralization zone have Pt (39–315ppb) and Pd (76–1100ppb), and intermediate to high Cu/Pd ratios (6850–138,000). The variably high Cu/Pd ratios indicate that the parental magma may have reached earlier sulfide saturation and removal. The low Os contents (0.0146 to 0.276ppb) and high Re/Os (29 to 292) and 187Os/188Os(i) ratios (0.277 to 0.916) of the silicate rocks are consistent with a major contribution of Os from crustal materials. On the other hand, the net-textured sulfide ores from No. 8 mineralization zone have high Os contents (15.1 to 32.6ppb), low Re/Os (2.0 to 2.3) and low 187Os/188Os(i) ratios (0.125 to 0.141), indicating a mantle-derived origin. Modeling indicates that >30% crustal contamination is required to explain the radiogenic Os isotopic composition if average upper crust is adopted as the crustal contaminant, or 5%–25% if sulfide-rich crustal rock is used as the crustal contaminant. However only 5–10% crustal contamination is required to explain the Nd and Sr isotopes. We propose that the decoupling of Os from Sr–Nd isotopes in the silicate rocks was due to selective crustal contamination. Addition of external sulfur from a crustal source may be the key factor triggering sulfide saturation. The silicate magma that underwent sulfide extraction before emplacement to shallow depth preserved a crustal Os isotope signature. Continued reaction of sulfide melts with new pulses of mantle-derived magma increased the Os content and decreased the 187Os/188Os(i) ratios, effectively masking the crustal Os contribution in the sulfides. [Copyright &y& Elsevier]

Abstract: We conducted hydrothermal experiments in olivine (Ol; Fo91)–H2O and orthopyroxenite (Opx; composed of 95% of orthopyroxene, En66)–H2O systems under conditions of 250 °C and vapor-saturated pressure (Psat) to examine the temporal evolution of the solution chemistry and products in runs of up to 1008h in duration. The maximal degree of hydration (i.e., H2O content in the solid sample) in the Ol–H2O experiments (3.6wt.%) was much higher than that in the Opx–H2O experiments (0.4wt.%). In the Ol–H2O experiments, Mg and Si in solution showed an initial increase (stage I) before decreasing (stage II) and finally attaining a steady state after 504h (stage III). Following a drop in silica activity toward the level of brucite stability field, the products also changed from serpentine+magnetite (stages I and II) to serpentine+brucite+magnetite (stage III). Serpentine minerals also changed from lizardite (stages I and II) to lizardite+chrysotile (stage III). The zoning pattern of the products around olivine at 1008h (brucite+serpentine at the olivine contact, and serpentine at the outermost rim) is consistent with the temporal changes in the mineralogy of the products, and is similar to the pattern observed in the mesh rims in partly serpentinized dunites. In the Opx–H2O experiments, chlorite formed after orthopyroxene grains, which differs from the formation of talc and serpentine after orthopyroxene (bastite), as observed in natural hydrated harzburgites. The Opx–H2O system maintained 10–103 times higher silica activity than Ol–H2O system, suggesting that brucite does not form after olivine during hydration of peridotites when the Ol–H2O system is linked to the Opx–H2O system. The progress of hydration reactions is affected by mechanical properties of host rocks. The hydration reactions observed in this study produced hierarchical fractures in the reactants, which became filled with reaction products, similar to mesh textures after olivine in natural peridotites. This reaction-induced fracturing produced new reaction surfaces and fluid pathways that enhanced the hydration reactions. The overall reaction producing serpentine+brucite in the Ol–H2O experiments showed the large volume expansion (V/V 0 =1.43 at stage III), whereas that producing only serpentine proceeded with near constant volume (V/V 0 =1.19 at stage I). The volume expansion is more difficult to occur in the oceanic lithosphere than in our experiments during serpentinization. Thus, in the case that volume expansion is prevented at reaction sites, one of the following outcomes occurs: (1) the hydration reaction stops until new fractures form, or (2) the reaction proceeds with low volume expansion (absence of brucite) by removing Mg from the system. These two outcomes would produce contrasting distributions or extent of hydration in oceanic lithosphere. [Copyright &y& Elsevier]

Abstract: Rhenium and Os concentrations and Os isotopic compositions were determined for a suite of previously characterized samples from a traverse across the Platreef at the Sandsloot–Mogolakwena mine of the Northern Limb of the Bushveld complex. The samples comprise five pyroxenites and five “parapyroxenites,” contact rocks representing variable mixtures of calc–silicate country rocks and magmatic rocks. The isotopic data from the pyroxenites define a Re–Os regression line with an age (2011±51Ma) that agrees within uncertainty with the accepted age for the Bushveld complex, whereas the parapyroxenite data scatter about this trend. The initial 187Os/188Os ratios of the pyroxenites (0.1820–0.1848) are similar to, but slightly more radiogenic than, previously published values for the Merensky Reef. Our preferred interpretation is that this section of the Platreef formed from a well mixed, crustally contaminated magma, genetically related to that which gave rise to the Merensky Reef. The Platreef parental magma may also have experienced minor additional contamination by wall rock interaction during emplacement. [Copyright &y& Elsevier]

Abstract: Study of the Maobei garnet peridotite and pyroxenite are significant, because these rocks were subjected to subduction-zone UHP metamorphism that is indicative of deep continental subduction. In order to better understand the origin of the Maobei peridotite and pyroxenite and the processes of continental subduction, we performed geochemical analyses on whole-rock and mineral samples. The studied samples were recovered in the main-hole of the Chinese Continental Scientific Drilling in the Sulu UHP terrane, east-central China. The garnet peridotites occur as lenses or layers in pyroxenite; they comprise garnet wehrlite and garnet dunite with minor interlayers of garnet websterite and garnet. Mini-bands, nodules or veins of eclogite and garnet clinopyroxenite (“internal”) also irregularly occur in garnet wehrlite. The peridotites are composed of olivine (Fo83–84), garnet (Prp51–56), ± diopside±enstatite (En86–87)±titanoclinochumite, and have lower Mg/(Mg+Fetotal) values of 0.79–0.83 and higher Al2O3 and CaO contents in comparison with the gneiss-hosted Sulu mantle peridotites. Garnet wehrlite displays variable REE patterns, from LREE-enriched to slightly LREE depleted relative to MREE and HREE; whereas, other ultramafic rocks show roughly flat REE patterns. In general, Rb, Ba and LREE increase, and compatible element (e.g. Cr, Co and Ni) abundances decrease from peridotite, pyroxenite to “internal eclogite” and “internal Grt-clinopyroxenite”. The calculated REE partition coefficients (DREE) between Cpx and Grt of the Maobei peridotite and pyroxenite decrease regularly from LREE (DCe =32–708) to HREE (DYb =0.01–0.10), which is comparable with natural mantle peridotite and eclogite crystallized at 800–1100°C. The geochemical data and occurrence of these rocks suggest that the protoliths of the Maobei peridotites are cumulates derived by differentiation of basaltic magma at the lower continental crust. Thus, the “Internal eclogite and Grt-clinopyroxenite” result from post-cumulus crystallization of the trapped melt. Ti-clinohumite orthopyroxenite may have formed by reaction between peridotitic olivine and indigenous melt or country rock-derived (“exotic”) TiO2-bearing fluids at high-P condition during continental subduction to mantle depths. The peak P−T estimates of 795–840°C and 5.3–6.8GPa of the Maobei peridotites and the available petrological data of country rocks indicate that the continental crust was subducted to a depth of ∼200km. [Copyright &y& Elsevier]

Abstract: Two dominant petrographic types of sulfides occur in Hawaiian garnet pyroxenite xenoliths — Type I occurs as globular, poikilitic inclusions only in clinopyroxenes, and Type II occurs interstitially between silicate phases. Chemically, the two types are essentially identical and their compositions are consistent with Fe–Ni monosulfide solid solution (Fe=55–57%, Ni=3–10%, Cu=0.5–2.0%, S=35–37% and Zn=0.01–0.5%). Both types of pyroxenitic sulfides have a factor of 10 to 1000 lower PGE contents than sulfides found in peridotites elsewhere (e.g. Os ∼1 to <0.01ppm), show fractionated PGE patterns (Pd( n )/Ir( n ) ∼1–35; n =chondrite-normalized) and very high Re( n )/Os( n ) ratios (∼10–400). The Ni and the total PGE (Os+Ir+Pt+Pd) contents of the sulfides correlate positively with the Mg# of clinopyroxene in the host rock, which points to a genetic link between the sulfides and silicates and a high-temperature mantle origin. The Hawaiian sulfides formed from an immiscible sulfide melt that separated from a silicate melt, which formed the host garnet pyroxenite. Experimental constraints place this sulfide–silicate melt immiscibility at 1530±100°C and 3.1±0.6GPa, i.e. near the base or slightly below the Pacific lithosphere. Melt/mineral partition calculations suggest that the parental silicate melts (prior to immiscibility) were similar to the Honolulu Volcanics (HV) alkali lavas that host the xenoliths, and that limited (0.1–0.3%) sulfide fractionation from a parental HV-type melt can account for the observed HSE variability in the sulfides. The relatively low Pt/Re ratios (<10) suggest that, upon “aging”, such sulfides cannot generate the coupled 186Os–187Os isotope enrichments observed in Hawaiian lavas and some komatiites. Therefore, recycling of mantle sulfides of pyroxenitic parentage is unlikely to explain the enriched Pt–Re–Os isotope systematics of plume-derived lavas. [Copyright &y& Elsevier]

Abstract: The genesis of intra-plate alkaline basalts remains controversial, and three sources have been proposed: silica-deficient eclogite–pyroxenite, hornblendite, and carbonated peridotite. Here, we assess these models by analyzing Cenozoic intra-continental alkaline basalts from Shandong province, North China. The Cenozoic basalts of Shandong province consist of an early sequence of weakly alkaline rocks (alkali olivine basalts) and a late sequence of strongly alkaline rocks (basanites and nephelinites). In comparison with the weakly alkaline rocks, the strongly alkaline rocks have lower concentrations of SiO2 (39.2–45.1wt.%) and Al2O3 (10.3–13.8wt.%), higher alkalis (Na2O+K2O=4.3–8.6wt.%) and CaO (8.0–12.6wt.%), higher concentrations of most incompatible elements, and higher values of Ca/Al (0.7–1.3), La/Yb (39.4–65.7), and Sm/Yb (6.1–9.9). On the whole, primitive-mantle normalized spidergrams reveal that the strongly alkaline rocks have stronger negative K, Zr, Hf, and Ti anomalies (Hf/Hf*=0.59–0.77, Ti/Ti*=0.46–0.71) than do the weakly alkaline rocks. All these rocks have superchondritic Zr/Hf ratios (>44). Inverse rare earth element (REE) modeling suggests that the strongly and weakly alkaline rocks represent melting amounts of <3% and 3–10%, respectively. Neither silica-deficient eclogite–pyroxenite melts nor hornblendite melts can satisfy all the features mentioned above. Here we prefer a carbonated mantle source because the main characteristics of the strongly alkaline rocks resemble those of carbonatites (e.g., enrichment of most incompatible elements, high Ca/Al ratios, superchondritic Zr/Hf ratios, and negative K, Zr, Hf, and Ti anomalies). Since dry peridotite has a higher solidus temperature than does carbonated peridotite in the mantle, an increased degree of melting under higher temperatures may result in the dilution of “carbonatitic fingerprints.” Although contributions from silica-deficient eclogite–pyroxenite or hornblendite cannot be ruled out, our observations suggest that carbonated peridotite is the main source for the Cenozoic strongly alkaline basalts of Shandong. [Copyright &y& Elsevier]

Abstract: An assemblage of MORB analyses (n =792 samples), including a suite of new, high-precision LA-ICP-MS measurements (n =79), has been critically compiled in order to provide a window into the chemical composition of these mantle-derived materials and their respective source region(s), commonly referred to as the depleted MORB mantle (DMM). This comprehensive MORB data set, which includes both “normal-type” (N-MORB, defined by (La/Sm)N <1.00) and “enriched-type” samples (E-MORB, (La/Sm)N ≥1.00), defines a global MORB composition that is more enriched in incompatible elements than previous models. A statistical evaluation of the true constancy of “canonical” trace element ratios using this data set reveals that during MORB genesis Ti/Eu, Y/Ho and Ce/Pb remain constant at the 95% confidence-level; thus, the ratios recorded in MORB (Ti/Eu=7060±1270, 2σ; Y/Ho=28.4±3.6, 2σ; Ce/Pb=22.2±9.7, 2σ) may reflect the composition of the DMM, presuming the degree of source heterogeneity, component mixing and conditions of melting/crystallization of the DMM are adequately recorded by global MORB. Conversely, Ba/Th, Nb/U, Zr/Hf, Nb/Ta, Sr/Nd, and Th/U are shown to fractionate as a function of MORB genesis, and thus these ratios do not faithfully record the composition of the DMM. Compared to samples from the Pacific and Indian Oceans, MORB derived from Atlantic ridge segments are characterized by statistically significant (≥95% confidence-level) enrichments in both highly incompatible elements (e.g., light REE, TITAN group elements, Sr, Ba, Pb, Th, and U) as well as less incompatible elements (e.g., heavy REE), indicating: i) a prominent recycled source component; ii) variable proportions of pyroxenite in the Atlantic source region; and/or, most likely iii) smaller degrees of melting and/or greater extents of fractional crystallization due to slower ridge spreading rates. Conversely, Pacific MORB has the most depleted regional signatures with regard to highly incompatible elements (e.g., Ba, Pb, Th, and U), likely due to faster ridge spreading rates. Indian Ocean MORB exhibit limited variation in incompatible element enrichments/depletions but are generally the most depleted in more compatible elements (e.g., Ti, Cr, Sc, and heavy REE), potentially due to distinct source characteristics or deep source melting in the garnet field. Atlantic, Pacific and Indian MORB can also be distinguished by trace element ratios, particularly Ce/Pb and Th/U, which is distinct at the >99% confidence-level. Global MORB, and by inference the DMM, are characterized by enrichments in Y/Ho and depletions in Th/U relative to the chondritic ratios, and are complementary to the continental crust. However, the median of global MORB and the bulk continental crust both have sub-chondritic Ti/Eu and Nb/Ta ratios, suggesting an under-represented Ti- and Nb-rich reservoir in the Earth, potentially refractory, rutile-bearing eclogite at depth in the mantle. [Copyright &y& Elsevier]

Abstract: In situ U–Pb and Hf-isotope systematics of zircons, mineral chemistry of silicates and whole-rock major-element, trace-element and platinum-group element (PGE) data are reported for five mafic xenoliths (granulites and pyroxenite) in the Hannuoba Cenozoic basalts, North China. Temperature estimates of 859–886 °C suggest that these xenoliths are derived from the lower crust. The whole-rock compositions of the xenoliths show weakly negative Eu anomalies (δEu=0.67–0.95), high SiO2/Al2O3 (5.80–13.8), low Ir (0.10–0.29 ppb), and wide ranges of Ni (4.5–235 ppm) and La/Nb (1.70–7.64). Most zircons in these samples are structureless and give Paleogene (44.5–47.3 Ma) U–Pb ages; minor populations have Early Mesozoic (210–220 Ma), Late Mesozoic (90 Ma) and Neogene (14 Ma) ages. The fractionation of pyroxene and plagioclase controlled the petrogenesis of the xenoliths, which are interpreted as the products of basaltic underplating and fractionation (cumulates) in Paleogene time (45–47 Ma), with assimilation of some older intermediate crustal components. The Paleogene underplating corresponds in time to lithosphere-scale extension in the North China Craton, leading to the widespread formation of sedimentary basins. Most Paleogene zircons have positive ε Hf (up to +13.2) with uniform T DM (0.46–0.50 Ga), whereas most Mesozoic zircons have negative ε Hf; this evolution suggests a temporal decrease in the degree of crustal assimilation during the extension of the lithosphere. In addition to complex modification in Early and Late Mesozoic times, the thermal event related to the eruption of the host magmas (ca 14 Ma) also reheated the lower crust beneath the North China Craton. [Copyright &y& Elsevier]

Abstract: Variscan, mantle-derived peridotites and associated pyroxenites occur as boudins in Moldanubian granulite near the town of Horní Bory in western Moravia. The peridotites consist of two compositionally distinct suites, one of Mg–lherzolite (Mg-number=89–91, 87Sr/86Sr=0.7046–0.7068, ε Nd =+4.1 to +5.3), and another of Fe–dunite/wehrlite (Mg-number=83–88, 87Sr/86Sr=0.7079–0.7087, ε Nd =−2.8 to −1.3). Modelling of Mg–Fe exchange between peridotite and Fe-rich melts reveals that the modal and chemical composition of the Fe–dunite/wehrlite suite can be produced by melt–rock reactions between lherzolite and SiO2-undersaturated melts of basaltic composition at melt/rock ratios ranging from 0.3 to 2. In such a model, pyroxenites represent the crystalline product (±trapped liquid) of melts migrating along conduits in peridotite. The Fe–dunite/wehrlite suite and pyroxenites are enriched in the LILE and depleted in the HFSE. The trace element and Sr–Nd isotopic compositions of Horní Bory peridotites and pyroxenites point to a significant component of crustal material in the invasive melts. The melt–rock reactions recorded in the Horní Bory ultramafic boudins are attributed to melt percolation in a mantle wedge above a Variscan subduction zone. [Copyright &y& Elsevier]

Abstract: We have determined the Sr, Nd, Pb and Hf isotopic compositions of clinopyroxene separated from mantle-derived ultramafic xenoliths (six spinel peridotites, two composite Cr-diopside pyroxenites, and one discrete Al-augite pyroxenite) hosted by Cenozoic alkali basalts at Hannuoba, North China, in order to understand the nature of the mantle source for this intraplate volcanism, and the petrologic history of the mantle lithosphere beneath North China Block, a crustal segment of the Sino-Korean Craton. Measured Sr, Nd, Pb and Hf isotopic compositions in the clinopyroxene grains separated from spinel peridotite and Cr-diopside pyroxenite (87Sr/86Sr=0.70265 to 0.70485; 206Pb/204Pb=17.75 to 19.15; ɛ Nd =0 to +11; ɛ Hf =+10 to +38) display mixing hyperbolas between mantle compositional end members DMM and EMII on the Sr–Pb and Nd–Pb isotope correlation diagrams. This is distinctly different from the host basalt data which show a mixture of DMM and EMI components on the diagrams. We interpret this to reflect infiltration by metasomatic agents, possibly silicate melts, having an EMII-like isotopic signature, which enriched a precursor time-integrated depleted mantle. An Al-augite pyroxenite, also hosted by these basalts, is characterized by highly enriched Sr, Nd, and Hf isotopic compositions (87Sr/86Sr=0.70733; ɛ Nd =−16; ɛ Hf =−18) with only moderately radiogenic Pb that has a 206Pb/204Pb value of 18.23. All of these data plot outside (1) the fields for oceanic basalts, and (2) the mixing arrays defined on the isotopic correlation diagrams by peridotites/Cr-diopside pyroxenite with their metasomatic agents, and by the host basalt. These observations suggest that (1) Al-augite pyroxenite is not cogenetic with the Cr-diopside pyroxenite, (2) parental melts of the pyroxenites are not likely to be the source for the metasomatism, and (3) the thermo/mechanically reactivated pyroxenite and/or spinel peridotite, is not likely to be the source for host basalt magmatism. The Cenozoic intraplate volcanism, therefore, must have originated in the asthenosphere. We observe that the relatively little-metasomatized Hannuoba peridotites define a Lu–Hf isochron of 2587±86 Ma (2σ). This value is, within error, indistinguishable from the Sm–Nd isochron age of the overlying granulite terrain. We suggest, therefore, that the Lu–Hf system can be used to constrain the timing of lithospheric mantle differentiation. Preservation of the Neoarchean mantle lithosphere beneath Hannuoba, despite the protracted tectono-magmatic reactivation during the Mesozoic and Cenozoic in this area, suggests that complete removal of the lithospheric mantle beneath East Asia by wholesale delamination is unlikely. [Copyright &y& Elsevier]

Initial 187Os/188Os (350 Ma) of cumulate garnet pyroxenites enclosed in high-temperature peridotite massifs from a Paleozoic convergent plate margin (lower Austria) is slightly suprachondritic to highly radiogenic (187Os/188Osi up to 0.686). The main factor controlling 187Os/188Os appears to be the extent of Os exchange of the parent melts with the host peridotites. The radiogenic 187Os/188Osi of the least reacted parent melts is most easily explained by an Os flux from the slab, although low Os concentrations suggest that the magnitude of this flux may be rather small. Sediment-like Sr–Nd–Pb isotopic compositions and negative Eu anomalies in plagioclase-free pyroxenites indicate that the Os was derived from the sediment section in the slab. Osmium abundances in the pyroxenites correlate well with Ni and Cr, indicating compatible behavior during pyroxene and garnet fractionation, likely because of coprecipitating sulfide. Rhenium abundances in the pyroxenites do not correlate with mildly incompatible elements such as Yb and are obviously not controlled by silicate phases. Except for two samples with high Re, Re appears to have partitioned preferentially into the melt, which, judging from low sulfur and scarce sulfides in most pyroxenites, may have been undersaturated in sulfur.An isotopic and chemical profile across a modally layered pyroxenite indicates substantial Os isotopic disequilibrium on the centimeter scale, while Nd and Sr isotopes are equilibrated or nearly equilibrated. This observation and the Os concentration gradient suggest that sulfide melts cannot be mobile in these lithologies, and under these conditions, otherwise isotopic and concentration gradients should have been erased by fast Os diffusion in sulfide melt. The Os isotopic profile is inconsistent with a simple diffusion model. Rather, it suggests that the pyroxenites formed from multiple batches of magma with differing 187Os/188Os. Reopening of pyroxenitic dikes likely plays an important role in the preservation of pristine slab signatures in melt and fluid that would be lost with extensive interaction with peridotite. [Copyright &y& Elsevier]

Spinel and garnet pyroxenite xenoliths in Cenozoic basalts from Hannuoba, North China show extremely heterogeneous chemical and isotopic compositions (&epsiv;Nd=−27 to +34). Most of these pyroxenites are relatively young, probably late Mesozoic in age, although a few Al-pyroxenites could be very old (∼2 Ga). While their texture and major element compositions suggest an origin of high pressure cumulates, the trace element and isotopic compositions of the Hannuoba pyroxenites require multiple segregation processes from different parental magmas. Strong LREE enrichment, ubiquitous HFSE depletion and some Eu anomalies of the Al- and Cr-pyroxenites indicate the involvement of crust components in their source. Their Sr–Nd isotopic ratios are negatively correlated and plot below the MORB–OIB–IAB–sediment trend, suggesting that the parental melts of the Cr- and Al-pyroxenites may have been derived from a mixture of asthenospheric melts and a long-term evolved continental crust. The garnet pyroxenites significantly deviate from the isotopic array defined by the Al-pyroxenites, due to their relatively high 87Sr/86Sr at given &epsiv;Nd. They thus more likely represent segregates from melts derived from partial melting of hydrothermally altered oceanic crust (basalts+marine sediment). If the crustal component involved in the Al-pyroxenites is subducted terrigenous sediments or other continental materials from the Archean Sino-Korean Craton, the Al-pyroxenites and garnet pyroxenites may have formed contemporaneously at a palaeo-convergent plate margin. This may be related to the subduction of the Mongol–Okhotsk plate beneath North China during the late Jurassic. Alternatively, if the delaminated lower crust was involved, it implies that most of the Al-pyroxenites are younger than the garnet pyroxenites, and their formation may be temporally correlated with lithospheric thinning during the Cretaceous. This model is attractive because the inferred tectonic evolution from a convergent setting to an extensional environment is consistent with the geologic record in the area. [Copyright &y& Elsevier]

Pyroxenite veining is widely preserved in peridotite massifs, and used to derive information on the origin and evolution of upper mantle domains. These lithospheric mantle sections can be isolated from the convecting mantle for >1 Ga or more, suffering a long history of melting and/or melt-rock reaction processes, which modify their original chemical and isotopic compositions. Here, we show the effect of ancient process of melt-rock reaction in the chemistry of garnet pyroxenites from Lanzo North Massif, an iconic lithospheric mantle section exhumed during the opening of the Jurassic Alpine Tethys. Selected pyroxenites are more than 10 cm thick, and embedded within peridotites that have textures and chemical compositions indicative of a complex history of interaction with migrating melts. Whole rock and clinopyroxene Nd-Hf isotopes of the pyroxenites consistently indicate that the first melt-rock reaction event occurred at ~400 Ma, likely in combination with exhumation from the garnet to the spinel-facies mantle conditions. Two samples still retain textural relicts and chemical evidence of precursor garnet and have high ε Nd (~12) for comparatively low ε Hf (~10), when recalculated at 400 Ma, which suggest that they were less affected by this ancient percolation process. The chemical evidence of such a long history of melt-rock reactions was preserved from 400 Ma until present. Finally, two pyroxenites located within plagioclase peridotites show evidence for an event of re-equilibration at plagioclase-facies conditions, likely triggered by infiltration of melt in the host rock. These samples reveal the coexistence of two internal Sm-Nd isochrones at 152 ± 30 Ma and 149 ± 13 Ma, thereby providing temporal constraints to the event of melt impregnation of the host peridotites as consequence of the opening of the Ligurian Tethys ocean. [ABSTRACT FROM AUTHOR]

Pyroxenite-peridotite sequences from the External Liguride (EL) Jurassic ophiolites (Northern Apennines, Italy) consist of portions of fertile MORB mantle that were modified by deep melt infiltration and melt-peridotite reaction. They represent an excellent natural example of a MORB-like veined mantle including unmodified peridotite, pyroxenite layers and metasomatized peridotite. We carried out a spatially controlled Hf isotope study on these mantle sequences to investigate how the Nd and Hf isotopic systems are affected by pyroxenite emplacement and melt-peridotite interactions. Present-day Lu Hf isotopic compositions of these lithologies show a large range of 176Lu/177Hf and 176Hf/177Hf ratios that are correlated with their Nd isotopic compositions. Pyroxenite-free peridotites delineate a Hf Nd isotope array that corresponds to a Proterozoic age (> 1.5 Ga) which is likely related to the accretion to the subcontinental lithosphere of this mantle sector. Heterogeneous 176Hf/177Hf isotopic compositions in pyroxenites mostly correlate with the significant variations of 176Lu/177Hf ratios and reflect variable garnet abundance in the primary modal assemblage. Over time, the pyroxenites acquired a large range of εHf values, which encompass the global range of Hf Nd isotopes in ocean ridge basalts. Infiltration of pyroxenite-derived melts led the host peridotite to acquire low Lu/Hf ratios with the consequent development of 176Hf/177Hf ratios lower than in the unmodified peridotite, generating an equivalent of an enriched mantle component. This melt-peridotite interaction likely occurred during the pyroxenite emplacement 430 Ma ago, as confirmed by two Lu Hf local pyroxenite-peridotite isochrons. The chemical and isotopic changes produced, over time, a spread of Hf Nd isotopic signatures of the EL veined mantle, covering almost the entire range of published MORB compositions. Pyroxenite emplacement and local metasomatism of the host peridotites thus created Hf Nd enriched mantle domains, making the EL veined mantle the first reported natural example of an enriched MORB-like mantle that formed through the combined effect of deep emplacement of pyroxenite and pyroxenite-peridotite interaction. The structure and isotopic characteristics of the EL veined mantle were used to model the isotopic compositions of melts produced by decompression melting of three-component heterogeneous mantle sources, providing an additional scenario to the generation of EMORB erupted at mid-ocean ridge settings. Our results emphasize the potential role of deep pyroxenite infiltration in modifying the host peridotites by interaction with pyroxenite-derived melts and creating heterogeneous mantle domains. [ABSTRACT FROM AUTHOR]

The Adamawa Volcanic Massif of the Cameroon Volcanic Line (CVL) is dominated by silica-undersaturated basaltic lavas, which have been attributed to plume volcanism. However, the relative contributions of peridotite and pyroxenite sources have not been assessed. Here, we present mineral and whole-rock data for the lavas to consider this issue and to better understand their petrogenesis and source mineralogy. The lavas are porphyritic, with phenocrysts and micro-phenocrysts of olivine, and to a lesser extent, clinopyroxene, in a groundmass of plagioclase, oxides, and apatite. The lava compositions are all basanites. They are silica-undersaturated and are primitive, mantle-derived melts. The basanites display lower CaO and CaO/Al 2 O 3 ratio but higher Fe/Mn values than expected for melts of peridotite, consistent with an appreciable contribution from melting of pyroxenite in the source. Olivine phenocrysts are characterized by low CaO and NiO contents but higher Fe/Mn ratios than those of olivine in melts of peridotitic sources, supporting the contribution of pyroxenite to the primary melts, and that the main mantle source lithology for the basanites was a mixture of spinel-facies peridotite and silica-deficient pyroxenite. The presence of both pyroxenite and peridotite xenoliths in this area attests to the contribution of these two mantle lithologies to the primary melts, and their relative contribution varies on a very local scale. Significant contamination by crustal materials during magma ascent is ruled out. The Youkou basanites provide a key for improving the general understanding of melt production in lithologically heterogeneous mantle beneath the CVL. Beneath the Youkou volcano, partial melting of pyroxenite by decompression formed silica-deficient melts at 60-80 km depth, which modified the peridotite mantle source by a low degree melt enrichment process indicated by the enrichment in incompatible elements. This modified mantle source then underwent low-degree melting to produce the basanitic melts, which are concentrated in a linear zone focused by the combination of the northern edge of the Congo craton and the Central African Shear Zone. [Display omitted] • Youkou volcano, sampled basanites hosted mantle xenoliths from the uppermost mantle. • The lavas experienced magmatic overprinting. • Melts are SiO 2 -undersaturated silicate melts from both peridotite and pyroxenite sources. • Melts from pyroxenite sources are silica-deficient melts. [ABSTRACT FROM AUTHOR]

Quantifying water contents in the lithospheric mantle is key to our understanding of global geodynamics, mantle composition, and related physical properties. Most mantle lithologies (peridotite) contain little water (~50 ppm), but petrological heterogeneities such as pyroxenites are more hydrous (~300 ppm) relative to the mantle rocks. Pyroxenites also melt at lower temperatures than peridotites and are thus important to magma genesis. Thus, quantifying pyroxenite water contents provides new information on the distribution of water in the mantle. Here, we present phase-specific FTIR measurements of the water contents in pyroxenite mantle xenoliths from two continental lithospheric domains that experienced intense metasomatism: the French Massif Central (FMC, France) and the Adamawa Volcanic Plateau (AVP, Cameroon). The AVP garnet pyroxenites are more hydrated ([H 2 O] Clinopyroxene = 386–685 ppm; [H 2 O] Orthopyroxene = 124–155 ppm; [H 2 O] Garnet < 0.5 ppm) than FMC ones ([H 2 O] Clinopyroxene = 112–465 ppm; [H 2 O] Orthopyroxene = 61–104 ppm; [H 2 O] Garnet < 0.5 ppm). These water concentrations are homogenous at the grain and correlate with equilibrated major element concentrations, indicating that they are representative of lithospheric water, although the FMC pyroxenites were dehydrated during metasomatism by a carbonatitic fluid (based on the correlation between La N /Sm N and Ti/Eu ratios); the water contents of AVP pyroxenites were likely not affected by metasomatism. FMC pyroxenites show peculiar FTIR spectra that may reflect the preferential dehydration of specific sites in the pyroxene structure. In both regions, metasomatism modified the light rare Earth element contents (e.g., Ce) of the pyroxenites, resulting in highly variable H 2 O/Ce ratios. Therefore, we conclude that the utility of the H 2 O/Ce ratio to identify the involvement of pyroxenites in magmas genesis is limited. [ABSTRACT FROM AUTHOR]

Clarifying the genesis of backarc basin basalts (BABBs) is of great significance for understanding the cycling of subducted materials. Here we show MORB-like Fe isotopic signatures (δ56Fe) of basalts recovered from the South China Sea (SCS), a well-developed backarc basin, as an indicator to investigate the influence of subduction on the BABBs. Whole-rock chemical compositions and olivine chemistry indicate the presence of pyroxenitic components in the peridotitic mantle. Melting of a pyroxenite-bearing peridotite indicates an addition of ~10–40% of pyroxenitic melts to the peridotitic mantle in the SCS, which should result in higher δ56Fe values than that of global MORBs. It is inconsistent with the MORB-like δ56Fe values we observed in the SCS basalts. Therefore, a component with low δ56Fe signatures may exist in the source of the SCS basalts, which is most likely derived from subducted materials. Such a scenario can be well corroborated by mixing depleted MORB mantle with subducted fluids and pyroxenitic melts. This study emphasizes that the MORB-like Fe isotopic signatures in backarc basin basalts can be caused by multi-factors, especially in the case of heterogeneous mantle source. [ABSTRACT FROM AUTHOR]

• We carried out trace element profiles through pyroxenite-peridotite associations. • Peridotites are chemically modified by reaction with pyroxenite derived melts. • Melt percolation model well reproduced the REE gradient in modified peridotite. • Metasomatized peridotite has larger volume than the associated pyroxenite. • Melt-rock reaction is able to create enriched-MORB mantle sources. Deep melt intrusion and melt-peridotite interaction may introduce small-scale heterogeneity in the MORB mantle. These processes generate pyroxenite-bearing veined mantle that represent potential mantle sources of oceanic basalts. Natural proxies of such veined mantle are very rare and our understanding of mechanisms governing the chemical modification of mantle peridotite by MORB-type pyroxenite emplacement is very limited. We report the results of detailed spatially-controlled chemical profiles in pyroxenite-peridotite associations from the Northern Apennine ophiolitic mantle sequences (External Liguride Units, Italy), and investigate the extent and mechanism driving the local modification of peridotite by the interaction with pyroxenite-derived melt. Pyroxenites occur as cm-thick layers parallel to mantle tectonite foliation and show diffuse orthopyroxene-rich reaction rims along the pyroxenite-peridotite contact. Relative to distal unmodified peridotites, wall-rock peridotites show i) modal orthopyroxene enrichment at the expense of olivine, ii) higher Al, Ca, Si contents and slightly lower X Mg , iii) Al-richer spinel and lower-X Mg pyroxenes. Clinopyroxenes from wall-rock peridotites exhibit variable LREE-MREE fractionation, always resulting in Sm N /Nd N ratios lower than distal peridotites. From the contact with pyroxenite layers, peridotite clinopyroxenes record a REE compositional gradient up to about 15 cm marked by an overall REE increase away from the pyroxenite. Beyond 15 cm, and up to 23 cm, the MREE and HREE content decreases while the LREEs remain at nearly constant abundances. This REE gradient is well reproduced by a two-step numerical simulation of reactive melt percolation assuming variable amounts of olivine assimilation and pyroxene crystallization. Percolative reactive flow at decreasing melt mass and rather high instantaneous melt/peridotite ratio (initial porosity of 30%), combined with high extents of fractional crystallization (i.e. relatively low M a /M c ratio), accounts for the overall REE enrichment in the first 15 cm. Change of melt-rock reaction regime, mostly determined by the drastic decrease of porosity (Φ i = 0.01) due to increasing crystallization rates, results in more efficient chemical buffering of the host peridotite on the HREE composition of the differentiated liquids through ion-exchange chromatographic-type processes, determining the observed increase of the LREE/HREE ratio. Emplacement of thin (cm-sized) pyroxenite veins by deep melt infiltration is able to metasomatize a much larger volume of the host peridotite. Hybrid mantle domains made by pyroxenite, metasomatized peridotite and unmodified peridotite potentially represent mantle sources of E-MORB. Results of this work stress the key role of melt-peridotite reactions in modifying the upwelling mantle prior to oceanic basalts production. [ABSTRACT FROM AUTHOR]

Continental flood basalts (CFB) are amongst the most voluminous volcanic eruptions in Earth's history. They are rapidly emplaced, and in rare cases the thick lava piles are associated with primitive magmas that have high MgO contents. The compositions of these primitive melts are consistent with a deep-sourced, high-temperature mantle plume origin. Whilst the association of CFBs with impacting mantle plumes is widely accepted, the magnitude of the thermal anomaly is not yet resolved. The development of Al-in-olivine thermometry, however, allows the crystallisation temperature of (near-)liquidus olivine to be determined without knowing the composition of the co-existing melt. This provides both a robust minimum estimate of mantle temperature and a value from which potential temperature (T P) can be back-calculated. This technique has previously confirmed that crystallisation temperatures in CFB settings can be a few hundred degrees greater than those estimated for MORB, and the results hint at a diversity in crystallisation temperatures between different CFB settings. In this study, we re-assess the T P of the mantle source of the Paraná-Etendeka CFB province by applying the Al-in-olivine thermometer to olivine–spinel pairs from picrites and ferropicrites. We show that the mean crystallisation temperature of olivine with Fo >90 in the picrites is 1458 °C, with a maximum temperature of 1511 °C. Using the mean value, we calculate a preferred T P of 1623 °C, for an assumed lithospheric thickness of 50 km and magma emplacement pressure of 0.5 GPa. This represents a thermal anomaly of around +300 °C relative to ambient mantle, and confirms that the mantle source of the Paraná-Etendeka CFB is the second hottest known from Phanerozoic Large Igneous Provinces, after the Caribbean Large Igneous Province. The ferropicrites record a cooler mean olivine crystallisation temperature of 1296 °C. Given that these low-volume melts derive from deeper and earlier melting of mantle pyroxenite, their temperature is not directly comparable to that of the picrites but they appear to require a somewhat cooler mantle source – perhaps found at the front or edges of a rising plume head. [ABSTRACT FROM AUTHOR]