Search

Water is critical in the evolution of the mantle due to its strong influence on the physicochemical properties of mantle rocks. Mid-ocean ridge basalts (MORBs) are commonly used to study the compositional characteristics of the convecting upper mantle. However, there remains abundant samples in the global MORB data sets without direct measurements of water contents. The commonly observed good correlation between H₂O and other incompatible trace components, such as Ce, has been applied to quantify water contents of MORBs. However, this approach assumes constant H₂O/Ce in the target samples, which is not always true as the H₂O/Ce ratios of MORBs could be rather heterogeneous even in some short ridge segments. Utilizing the present compositional data of global MORB glasses with measured water contents (n = 1,467), we construct a Random Forest Regression model based on machine learning, which can predict water concentrations of samples based on selected major and trace element data, without assuming a ratio between H₂O and other trace elements. This model allows us to precisely recover water contents for MORBs with comparable accuracy with traditional analytical methods. The predicted results of MORB glasses from this model (n = 1,931) expand the water content database of global MORBs and indicate a broad existence of high-H₂O MORBs. This new approach allows us to investigate the water content of MORBs from some ridges lacking previous water content measurements (e.g., the Chile Ridge and the Pacific-Antarctic Ridge) and infer changes in the water content of MORB sources through applying the model to transform fault samples.

The Early Cretaceous Jehol Biota evolution has remarkable spatiotemporal correlation with the destruction of the North China craton though the coupling mechanism remains enigmatic. The craton destruction was accompanied by intense magmatic activity and the released volatiles and nutrients might have had climatic and environmental impacts on the biotic evolution. In this study, we investigated the mentioned hypothetical causal link by determining concentrations and total emissions of volatile elements (S, F, Cl) and bulk-rock P contents of volcanic rocks that were erupted during the pre-flourishing, flourishing and post-flourishing stages of the Jehol Biota. Our results show that the volcanism near the flourishing stage has lower S (1,083–2,370 ppm), Cl (1,277–5,608 ppm) and higher P₂O₅ contents (0.48–0.84 wt.%) than that in the non-flourishing stages with S of 1,991–3,288 ppm, Cl of 7,915–12,315 ppm and P₂O₅ of 0.17–0.23 wt.%. Fluorine contents in the three stages vary from 893 to 3,746 ppm. The total volatile emissions are minor in the flourishing stage (1–14 Gt S, 0.6–10 Gt Cl, 0.6–9 Gt F) but elevated in the non-flourishing stages (2–766 Gt S, 4–1,168 Gt Cl, 1–175 Gt F). Our data suggest that regional climatic and environmental impacts of volcanism in the non-flourishing stages probably hindered the species diversification. The high P flux released from lithospheric mantle-derived lavas during the peak time of craton destruction might have enhanced primary productivity and contributed to the flourishing of the Jehol Biota. Our study provides insights into the relationship between the biosphere and deep geodynamic processes driven by volcanism.

Initiation of large igneous provinces (LIPs) and temporal and spatial variations in their chemical compositions provide unique opportunities to understand mantle dynamics. It has been recently revealed that water-rich reservoirs in the mantle sources have played a significant role in the production of LIPs. However, the dominant causes for the chemical variations in primitive magmas of LIPs remain debated. In the Emeishan LIP (ELIP), there are several places where picrites accompany more abundant flood basalts. These picrites span a large range of Ti/Y, Sm/Yb, and TiO₂/Al₂O₃. In this work, we have determined the water contents of picrites with Ti/Y > 500 from the ELIP, based on the water content of clinopyroxene phenocrysts and water partitioning coefficients. We have also calculated the water content of the mantle sources by applying batch and fractional partial melting models. We show that the chemical compositions of picrites in the ELIP closely correlate with the calculated water content in their mantle sources. Our model suggests that the variation in the chemical composition of the primary magmas of the ELIP both on a local and regional scale can be mainly explained by the complementary roles of relatively dry recycled pyroxenite and hydrated components in the mantle plume, without requiring catastrophic thinning of the lithospheric mantle. Our model thus implies that both the initiation and chemical variation of the ELIP are affected by water from the deep mantle.

The origin of ferroan A-type granites in anorogenic tectonic settings remains a long-standing petrological puzzle. The proposed models range from extreme fractional crystallization of mantle-derived magmas to partial melting of crustal rocks, or involve combination of both. In this study, we apply whole-rock chemical and Sm-Nd isotopic compositions and thermodynamically constrained modeling (Magma Chamber Simulator, MCS) to decipher the genesis of a suite of A1-type peralkaline to peraluminous granites and associated intermediate rocks (monzodiorite-monzonite, syenite) from the southwestern margin of the Archean Karelia craton, central Finland, Fennoscandian Shield. These plutonic rocks were emplaced at ca. 2.05 Ga during an early stage of the break-up of the Karelia craton along its western margin and show trace element affinities to ocean island basalt-type magmas. The intermediate rocks show positive εNd(2050 Ma) values (+1.3 to +2.6), which are only slightly lower than the estimated contemporaneous depleted mantle value (+3.4), but much higher than average εNd(2050 Ma) of Archean TTGs (–10) in the surrounding bedrock, indicating that these rocks were essentially derived from a mantle source. The εNd(2050 Ma) values of the peralkaline and peraluminous granite samples overlap (–0.9 to +0.6 and –3.2 to +0.9, respectively) and are somewhat lower than those in the intermediate rocks, suggesting that the mafic magmas parental to granite must have assimilated some amount of older Archean continental crust during their fractionation, which is consistent with the continental crust-like trace element signatures of the granite members. The MCS modeling indicates that fractional crystallization of mantle-derived magmas can explain the major element characteristics of the intermediate rocks. The generation of the granites requires further fractional crystallization of these magmas coupled with assimilation of Archean crust. These processes took place in the middle t

Mantle plume-generated basalts have long been utilized to assess the large-scale heterogeneity of the deep mantle. Their chemical and isotopic enrichment has been thought to result from the incorporation of recycled sediments or oceanic crust in their sources. Variable extents of dehydration or rehydration of recycled materials during subduction can induce different H₂O/Ce ratios in plume-related basalts. The Hainan plume has been regarded as the cause of the massive Cenozoic basaltic volcanism in southeastern Asia. In this study, we measured bulk-rock major and trace element compositions of basalts from a volcano at Heishanling Hill, northern Hainan Island, and made FTIR analyses of water in clinopyroxene phenocrysts to determine magmatic water contents. In addition, we determined Sr-Nd isotope compositions of basalts at Heishanling Hill and other three localities. Water contents of Cpx phenocrysts vary from 20 to 254 ppm, which were used to calculate the water contents in the corresponding equilibrated melts, yielding values from 0.22 to 2.42 wt% with an average of 1.07 ± 0.47 wt%. Considering the effect of magma differentiation on water contents, the primary melts of Heishanling basalts were estimated to have a range of 0.67–1.40 wt% for individual Heishanling basalts, with the average of 1.03 ± 0.33 wt%. The calculated water contents in the mantle source of the Heishanling basalts vary from 279 to 582 ppm, falling in the range of global OIBs. Combined with the results previously reported, H₂O/Ce ratios of Hainan basalts were estimated to vary from 50 to 425, covering the reported range from EM-type OIBs to NMORBs. The enriched Sr-Nd isotopic compositions of the studied samples resemble those of other Hainan plume-related basalts, and their ¹⁴³Nd/¹⁴⁴Nd ratios are well correlated with incompatible trace element ratios (e.g., Th/La, Ba/Nb). Bulk-rock compositions are consistent with a contribution of pyroxenite in the mantle source of the Hainan basalts an

The Otanmäki area in central Finland hosts two occurrences of Nb-Zr-REE mineralization, Katajakangas and Kontioaho, within a suite of ca. 2.04–2.06 Ga (A1-type) gneissic granites, syenite and monzonite-monzodiorite. They exhibit trace element characteristics and whole-rock εNd(2050 Ma) values (from +2.6 to -1.3) consistent with derivation by differentiation of mantle-derived mafic magmas with variable interaction with crustal material. The mineralization is localized in 0.1- to 1.4-m-thick dikes (Kataja-kangas) and a 30- to 50-m-thick sheet-like body (Kontioaho), containing allanite-(Ce), zircon, titanite, and Nb-REE-Th-U oxides. Their wall rocks are composed of ca. 2.06 Ga peraluminous monzogranite, which is genetically unrelated to the mineralized rock units, as evidenced by whole-rock chemical and Sm-Nd isotope data and zircon U-Pb geochronology. Instead, the mineralization is linked to the nearby peralkaline to metaluminous alkali feldspar granite magmatism dated at ca. 2.04–2.05 Ga. The development of REE-HFSE enrichment in the mineralized rock units required extensive crystallization of a peralkaline granite magma, producing residual metaluminous, high-silica melts enriched in REE-HFSE, Ca, and Fe relative to Na, K, and Al. The REE-HFSEand Ca enrichment was further promoted by volatile complexing with dissolved F-, CO₃²⁻ and SO₄²⁻. These highly evolved melts were parental to the mineralized dikes and the sheet-like intrusive body, which were emplaced into the monzogranite capping the intrusions of peralkaline granite.

Major and trace element compositions of iron oxides from Ni-Cu-PGE sulphide deposits hosted by mafic-ultramafic rocks in northern Fennoscandia (mainly Finland) have been analysed. Samples used were from the Ruossakero Ni-(Cu) deposit; Tulppio dunite and related Ni-PGE mineralisation; Hietaharju, Vaara, and Tainiovaara Ni-(Cu-PGE) deposits; and the Lomalampi PGE-(Ni-Cu) deposit. Reference samples from Ni-Cu-PGE sulphide deposits in Russia, China, and Finland, and a barren komatiite sequence fomr Finland were also studied. Magnetite and Cr-magnetite showed a wide range of trace element compositions as a result of the variation of silicate and sulphide melt compositions and their post-magmatic modification history. The Ni content in oxide showed a positive correlation with the Ni tenor of the sulphide phase in equilibrium with magnetite, regardless of whether the sulphide assemblage was magmatic or post-magmatic in origin. The massive sulphide samples contained an oxide phase varying in composition from Cr-magnetite to magnetite, indicating that Cr-magnetite can crystallise directly from sulphide liquid. The Mg concentration of magnetites in massive sulphide samples was lowest among the samples analysed, which can be regarded as a diagnostic feature of an oxide phase crystallised together with primitive Fe-rich monosulphide solid solution. The results show that magnetite geochemistry together with petrographical observations could be used as an indicator of potential Ni-(Cu-PGE) mineralisation., Major and trace element compositions of iron oxides from Ni-Cu-PGE sulphide deposits hosted by mafic-ultramafic rocks in northern Fennoscandia (mainly Finland) have been analysed. Samples used were from the Ruossakero Ni-(Cu) deposit; Tulppio dunite and related Ni-PGE mineralisation; Hietaharju, Vaara, and Tainiovaara Ni-(Cu-PGE) deposits; and the Lomalampi PGE-(Ni-Cu) deposit. Reference samples from Ni-Cu-PGE sulphide deposits in Russia, China, and Finland, and a barren komatiite sequence fomr Finland were also studied. Magnetite and Cr-magnetite showed a wide range of trace element compositions as a result of the variation of silicate and sulphide melt compositions and their post-magmatic modification history. The Ni content in oxide showed a positive correlation with the Ni tenor of the sulphide phase in equilibrium with magnetite, regardless of whether the sulphide assemblage was magmatic or post-magmatic in origin. The massive sulphide samples contained an oxide phase varying in composition from Cr-magnetite to magnetite, indicating that Cr-magnetite can crystallise directly from sulphide liquid. The Mg concentration of magnetites in massive sulphide samples was lowest among the samples analysed, which can be regarded as a diagnostic feature of an oxide phase crystallised together with primitive Fe-rich monosulphide solid solution. The results show that magnetite geochemistry together with petrographical observations could be used as an indicator of potential Ni-(Cu-PGE) mineralisation.

The Mustajarvi gold occurrence lies in the southern part of the Paleoproterozoic Central Lapland Greenstone Belt, next to a first-order transcrustal thrust fault system. The gold occurrence is structurally controlled by the second-order Mustajarvi shear zone, which occurs at the contact between siliciclastic metasedimentary and mafic to ultramafic metavolcanic rocks. The main mineralisation comprises a set of parallel veins and sulphidised rocks which are slightly oblique to the shear zone and are hosted by third-order shear structures. The gold-bearing rock at Mustajarvi is associated with pyrite which is present in 0.15 to 1 m wide quartz-pyrite-tourmaline veins and in zones of massive pyrite in the host rocks with thicknesses ranging from 1.15 to 2 m. In unweathered rock, hypogene gold is hosted by Au- and Au-Bi-telluride micro-inclusions in pyrite, whereas strong weathering at near surface levels has caused a remobilisation of gold, resulting in free gold deposited mainly in the cracks of oxidised pyrite. The geochemistry of both mineralisation styles is typical of orogenic gold systems. There is a strong enrichment of Ni and Co, unusual for orogenic gold deposits, which leads to the classification of Mustajarvi as an orogenic gold occurrence with atypical metal association., The Mustajarvi gold occurrence lies in the southern part of the Paleoproterozoic Central Lapland Greenstone Belt, next to a first-order transcrustal thrust fault system. The gold occurrence is structurally controlled by the second-order Mustajarvi shear zone, which occurs at the contact between siliciclastic metasedimentary and mafic to ultramafic metavolcanic rocks. The main mineralisation comprises a set of parallel veins and sulphidised rocks which are slightly oblique to the shear zone and are hosted by third-order shear structures. The gold-bearing rock at Mustajarvi is associated with pyrite which is present in 0.15 to 1 m wide quartz-pyrite-tourmaline veins and in zones of massive pyrite in the host rocks with thicknesses ranging from 1.15 to 2 m. In unweathered rock, hypogene gold is hosted by Au- and Au-Bi-telluride micro-inclusions in pyrite, whereas strong weathering at near surface levels has caused a remobilisation of gold, resulting in free gold deposited mainly in the cracks of oxidised pyrite. The geochemistry of both mineralisation styles is typical of orogenic gold systems. There is a strong enrichment of Ni and Co, unusual for orogenic gold deposits, which leads to the classification of Mustajarvi as an orogenic gold occurrence with atypical metal association.

The Kivilompolo molybdenite occurrence is located in the northern part of the Peräpoh jabelt, within the lithodemic Ylitornio nappe complex. It is hosted within a deformed porphyritic granite belonging to the pre-orogenic 1.99 Ga Kierovaara suite. The mineralization occurs mostly as coarse-grained molybdenite flakes in boudinaged quartz veins, with minor chalcopyrite, pyrite, magnetite, and ilmenite. In this study, we report new geochemical data from the host-rock granite and Re-Os dating results of molybdenite from the mineralization. For the whole-rock geochemistry, the mineralized granite is similar to the Kierovaara suite granites analyzed in previous studies. Also, the ca. 2.0 Ga Re-Os age for molybdenite is equal, within error, to the U-Pb zircon age of the Kierovaara suite granite. In addition, similar molybdenite and uraninite ages have been reported from the Rompas-Rajapalot Au-Co occurrence located 30 km NE of Kivilompolo. We propose that the magmatism at around 2.0 Ga ago initiated the hydrothermal circulation that was responsible for the formation of the molybdenite mineralization at Kivilompolo and the primary uranium mineralization associated with the Rompas-Rajapalot Au-Co occurrence or at least, the magmas provided heating, and in addition potentially saline magmatic fluids and metals from a large, cooling magmatic-hydrothermal system.

The Mustajärvi gold occurrence lies in the southern part of the Paleoproterozoic Central Lapland Greenstone Belt, in proximity to the first-order transcrustal Venejoki thrust fault system. The gold occurrence is structurally controlled by the second-order Mustajärvi shear zone, which is located at the contact between siliciclastic metasedimentary and mafic to ultramafic metavolcanic rocks. The main mineralization comprises a set of parallel veins and sulfidized rocks that are slightly oblique to the shear zone and are hosted by third-order structures likely representing Riedel R-type shears. The gold-mineralized rock at Mustajärvi is associated with pyrite that is present in 0.15- to 1-m-wide quartz-pyrite-tourmaline veins and in zones of massive pyrite in the host rocks with thicknesses ranging from 1.15 to 2 m. In unweathered rock, hypogene gold is hosted by Au- and Au-Bi-telluride micro-inclusions in pyrite, whereas strong weathering at near surface levels has caused a remobilization of gold, resulting in free gold deposited mainly in the cracks of oxidized pyrite. The geochemistry of both mineralization styles is typical of orogenic gold systems with strong enrichments comprising Au, B, Bi, CO₂, Te, and Se; and less consistent anomalous amounts of Ag, As, Sb, and W. Unusual for orogenic gold deposits is the strong enrichment of Ni and Co, which leads to the classification of Mustajärvi as orogenic gold occurrence with atypical metal association.

Using electron probe microanalyzer (EPMA) and laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS), we analyzed major and trace element compositions of iron oxides from Ni-Cu-PGE sulfide deposits hosted by mafic-ultramafic rocks in northern Fennoscandia, mostly focusing on Finland. The main research targets were the Archean Ruossakero Ni-(Cu) deposit; Tulppio dunite and related Ni-PGE mineralization; Hietaharju, Vaara, and Tainiovaara Ni-(Cu-PGE) deposits; and Paleoproterozoic Lomalampi PGE-(Ni-Cu) deposit. In addition, some reference samples from the Pechenga (Russia), Jinchuan (China), and Kevitsa (Finland) Ni-Cu-PGE sulfide deposits, and a barren komatiite sequence in the Kovero area (Finland) were studied. Magnetite and Cr-magnetite show a wide range of trace element compositions as a result of the variation of silicate and sulfide melt compositions and their post-magmatic modification history. Most importantly, the Ni content in oxide shows a positive correlation with the Ni tenor of the sulfide phase in equilibrium with magnetite, regardless of whether the sulfide assemblage is magmatic or post-magmatic in origin. The massive sulfide samples contain an oxide phase varying in composition from Cr-magnetite to magnetite, indicating that Cr-magnetite can crystallize directly from sulfide liquid. The Mg concentration of magnetites in massive sulfide samples is lowest among the samples analyzed, and this can be regarded as a diagnostic feature of an oxide phase crystallized together with primitive Fe-rich MSS (monosulfide solid solution). Our results show that magnetite geochemistry, plotted in appropriate discrimination diagrams, together with petrographical observations could be used as an indicator of potential Ni-(Cu-PGE) mineralization.

We have studied the Re-Os isotope systematics of komatiite-related Ni-Cu-PGE sulfide deposits in Finland, including the Archean Ruossakero Ni-(Cu), Vaara Ni-(Cu-PGE), Tainiovaara Ni-(Cu-PGE) deposits and Ni-PGE sulfide mineralization related to a dunitic body at Tulppio, and the Paleoproterozoic Lomalampi PGE-(Ni-Cu) deposit. We analyzed chromite separates from both mineralized and non-mineralized komatiitic rocks, bulk-sulfide separates, and country-rock black schist from one deposit, and for comparison, chromite separates from non-mineralized Archean komatiitic rocks from the Kovero and Kuhmo greenstone belts. We have observed open-system behavior of sulfides in three deposits, with their Re-Os isotope compositions indicating resetting at ca. 1.83–1.87 Ga by metamorphic events related to the Svecofennian orogeny. Most of the chromite separates, especially those with >10 ppb Os, retain their primary isotope compositions with near-chondritic initial ¹⁸⁷Os/¹⁸⁸Os ratios, indicating that their Os was derived from a mantle source with a chondritic Re-Os isotope evolution. These include the Vaara deposit, which yielded a chromite Re-Os isochron age of 2719 ± 140 Ma and low initial γOs of −1.2 ± 0.02, in spite of the evidence for the presence of mass-independent S isotope fractionation in this deposit, suggesting a low-Os, sulfur-bearing contaminant. In the case of the Lomalampi deposit, chromites yield an isochron with an age of 2058 ± 93 Ma and initial γOs of +0.9 ± 0.02, consistent with earlier isotope data from Central Lapland komatiites. However, one chromite from a mineralized cumulate gave a γOs(2050 Ma) value of +24 ± 18, indicating a minor/moderate degree of crustal contamination involving material from country rock black schists. This is also supported by earlier S isotope data. In the Archean Tainiovaara deposit, chromite also suggests crustal contamination (γOs + 13 ± 1), but country rocks are Os-poor Archean gneisses, leaving the source of radiogenic

The Rompas-Rajapalot gold prospect is located in the northern part of the Paleoproterozoic Peräpohja belt. It covers an area of at least 10 × 10 km and comprises various styles of gold mineralization ranging from localized high-grade Au pockets in uraninite- and pyrobitumen-bearing calcsilicate-carbonate-quartz veins in mafic metavolcanic rocks (Rompas area) to disseminated gold grains in Fe-Mg-rich metasediments and quartz-tourmaline-sulfide-native gold veins (Palokas area). This study deals with the petrography and mineral chemistry of the gold mineralization at Palokas, which occurs in the eastern part of the Rompas-Rajapalot prospect. Major and trace element data and fluid inclusion characteristics of tourmaline are used to evaluate the origin and the pressure-temperature-fluid composition parameters of hydrothermal fluids. Whole-rock geochemical analyses are utilized to evaluate the nature of the protolith of the host rocks. Gold occurs in a native form in at least two different textural settings: 1) single, relatively coarse grains disseminated among the rock-forming silicates in cordierite-orthoamphibole rocks and 2) smaller grains occurring in fractures of tourmaline in quartz-sulfide-tourmaline breccias and in fractures of chloritized cordierite-orthoamphibole rocks adjacent to the tourmaline-rich breccias. Fracture-related gold is associated with Bi-Se-S-bearing tellurides, native Bi, molybdenite, chalcopyrite, and pyrrhotite. Coarser-grained disseminated gold were not found to be clearly associated with sulfides nor any fractures. Statistical correlations show that the Au concentration correlates strongly with Te, Cu, Co, Se, Bi, Mo, and Ag (ρ = 0.730–0.619) whereas Au correlates moderately with As, Fe, W (ρ = 0.523–0.511) and to a lesser extent with U, Pb, and Ni (ρ = 0.492–0.407). Gold has the strongest negative correlations with Sr and Ca. The chondrite-normalized REE patterns of tourmaline from the Au-mineralized rocks (both vein type and hos

Cobalt is a highly sought-after metal due to its economic importance in many hightech applications. It is mainly obtained from sedimentary-hosted Cu-Co deposits and magmatic Ni-Cu deposits as a by-product. We describe a recently discovered Co-enriched Cu-Ni deposit hosted by the Hietakero mafic intrusion in north-western Finnish Lapland. The intrusion contains gabbroic to pyroxenitic cumulates and was emplaced into a supracrustal strata composed mainly of mafic volcanic rocks, sulfur-bearing graphite schists and felsic volcanic interlayers. Hybrid rocks provide clear evidence for interaction of mafic magma with its country rocks. All these rocks have undergone intensive post-magmatic hydrothermal alteration by influx of H2O-, CO2- and Cl-bearing saline fluids, forming strongly scapolitized (±albitized) rocks and resulting in re-mobilization of sulfides and their metals. The Hietakero deposit is associated with pyroxenitic cumulates and hybrid rocks, with the sulfides (pyrrhotite, cobaltian pentlandite, chalcopyrite and pyrite) occurring as patches, brecciated to net-textured and vein to veinlet forms. Sulfur content is 6.1 wt.% on average, reaching 20 wt.% in some cases. Nickel and Cu tenors are low, but Co is elevated, resulting in very low Ni/Co of 2.6 compared to ‘classic’ Ni-Cu sulfide deposit globally. The average metal tenors are: 0.55 wt.% Ni, 1.07 wt.% Cu, and 0.23 wt.% Co. In-situ sulfur isotope analyses of sulfides from mineralized rocks yielded δ34S values from -2.0 to +4.5‰. In spite of strong post-magmatic alteration, we can conclude that the Hietakero cumulates were formed from a low-MgO and PGE-depleted basaltic magma, which developed in a staging chamber at a deeper crustal level. The magma produced compositionally different gabbroic to pyroxenitic rocks (i.e., high-Ti and low-Ti groups). Model calculations indicate that the relatively high Co tenor and low Ni/Co cannot be explained by an earlier phase of fractional crystallization or sulfid

The genesis of large igneous provinces (LIP) is controlled by multiple factors including anomalous mantle temperatures, the presence of fusible fertile components and volatiles in the mantle source, and the extent of decompression. The lack of a comprehensive examination of all these factors in one specific LIP makes the mantle plume model debatable. Here, we report estimates of the water content in picrites from the Emeishan LIP in southwestern China. Although these picrites display an island arc-like H2O content (up to 3.4 by weight percent), the trace element characteristics do not support a subduction zone setting but point to a hydrous reservoir in the deep mantle. Combining with previous studies, we propose that hydrous and hot plumes occasionally appeared in the Phanerozoic era to produce continental LIPs (e.g., Tarim, Siberian Trap, Karoo). The wide sampling of hydrous reservoirs in the deep mantle by mantle plumes thus indicates that the Earth’s interior is largely hydrated.

Recent experimental studies and in situ LA-ICP-MS analysis on natural Cr-spinel have shown that Rh and IPGEs (Ir-group platinum group elements: Ru, Ir, Os) are enriched in the lattice of Cr-spinel. However, the factors controlling the partitioning behaviour of these elements are not well constrained. In this study, we report the Rh, IPGE, and trace element contents in primitive Cr-spinel, measured by LA-ICP-MS, from nine volcanic suites covering various tectonic settings including island arc picrites, boninites, large igneous province picrites and mid-ocean ridge basalts. The aim is to understand the factors controlling the enrichment of Rh and IPGEs in Cr-spinels, to estimate empirical partition coefficients between Cr-spinel and silicate melts, and to investigate the role of Cr-spinel fractional crystallization on the PGE geochemistry of primitive magmas during the early stages of fractional crystallization. There are systematic differences in trace elements, Rh and IPGEs in Cr-spinels from arc-related magmas (Arc Group Cr-spinel), intraplate magmas (Intraplate Group Cr-spinel), and mid-ocean ridge magmas (MORB Group Cr-spinel). Arc Group Cr-spinels are systematically enriched in Sc, Co and Mn and depleted in Ni compared to the MORB Group Cr-spinels. Intraplate Group Cr-spinels are distinguished from the Arc Group Cr-spinels by their high Ni contents. Both the Arc and Intraplate Group Cr-spinels have total Rh and IPGE contents of 22–689 ppb whereas the MORB Group Cr-spinels are depleted in Rh and IPGE (total < 20 ppb). Palladium and Pt contents are below detection limit for all of the studied Cr-spinels (<1–5 ppb). The time-resolved spectra of LA-ICP-MS data for Cr-spinels mostly show constant count rates for trace element and Rh and IPGEs, suggesting homogeneous distribution of these elements in Cr-spinels. The PGE spikes observed in several Cr-spinels were interpreted to be PGE-bearing mineral inclusions and excluded from calculating the PGE conte