Search

To better constrain Ca isotopic fractionation during volcanic magma differentiation, we report Ca isotope data for a series of volcanic glasses from the eastern Manus Basin that represent a continuum of compositions comprising basaltic andesite, andesite, dacite, and rhyolite. Our results show that these volcanic glasses exhibit ~0.25‰ variation in Ca isotopic compositions, with δ44/40Ca ranging from 0.81 ± 0.04 to 1.05 ± 0.04‰. The similar trace element patterns and Sr-Nd isotopic compositions exclude source heterogeneity as the cause of the varied Ca isotopic compositions. Due to their low K/Ca ratios and young age, the Ca isotopic variations cannot be attributed to 40K decay. Instead, the observed Ca isotopic variations are primarily controlled by their lithology. Specifically, δ44/40Ca of basaltic andesite (0.81 ± 0.04) and andesite (0.82 ± 0.01) are similar to the averages of reported mid-ocean ridge basalts (MORBs; 0.84 ± 0.10; 2SD, N = 26) and back-arc basin basalts (BABBs; 0.80 ± 0.08; 2SD, N = 21) in literature, but do not correlate with SiO2 or MgO contents, indicating the Ca isotopic fractionation is insignificant during intermediate-mafic magma evolution. As magma differentiation continues, δ44/40Ca of dacite and rhyolite evolves from 0.85 ± 0.01 to 1.05 ± 0.04‰. Moreover, δ44/40Ca of felsic glasses are negatively correlated with Eu/Eu*, which suggests that the increasing δ44/40Ca of felsic glasses is most likely related to plagioclase (Pl) fractional crystallization. Therefore, we conclude that Ca isotopes can behave differently at different stages of volcanic magma differentiation. Modeling of fractional crystallization using rhyolite-MELTS shows that the relative Ca fractions in crystallized minerals (e.g., Pl vs. Cpx) and the Ca isotopic fractionation factor between the residual melt and crystallized Pl (αmelt-Pl, which is likely related to the An number of Pl, magma temperature and kinetic effects during crystallization), provide a plausible explanation for the behavior of Ca isotopes during volcanic magma differentiation.

Increasing evidence shows that the mantle contributes (directly or indirectly) to Sn-bearing granites worldwide. However, the specific role of mantle in the formation of tin granites and related mineralization remains poorly understood. In the world-class Dachang district, South China, tin mineralization is related to the Longxianggai equigranular/porphyritic biotite granites and tin orebodies are cut by granite porphyry dykes hosting mafic microgranular enclaves (MMEs). A combination of zircon U-Pb dating and Hf-O isotopes, mineral chemistry, and whole-rock elemental and Sr-Nd isotopic compositions—for granitic rocks and MMEs, is employed to constrain the petrogenesis and to unravel the link between tin fertility and mantle upwelling. Laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) zircon U-Pb dating indicates that the biotite granites were emplaced at ca. 93 Ma, and the granite porphyry dykes and MMEs were formed at ca. 86 Ma. The biotite granites are silica- and alkali-enriched with A/CNK ratios of 1.04–1.36, and exhibit elevated concentrations of Li, F, P, Rb, Cs, Ta, Sn, W, and U, showing affinities with highly fractionated S-type granites. Whole-rock geochemical and Nd isotopic (εNd(t) = −10.0 to −7.8) data, and in situ zircon Hf-O (εHf(t) = −9.9 to −3.9, δ18O = 6.2–8.9‰) isotopes indicate that the biotite granites were formed by partial melting of metasedimentary rocks at relatively high temperatures (≥782 °C), possibly with minor input of mantle material. Likewise, the post-ore granite porphyry dykes have similar chemical and mineralogical characteristics as fractionated S-type granites. Zircon Hf-O isotopes (εHf(t) = −9.0 to −4.9, δ18O = 6.5–8.2‰) and whole-rock geochemical data suggest they were derived from a similar source as the biotite granites, whereas elevated εNd(t) values of −5.0 to −3.3 for granite porphyry dykes relative to biotite granites reveal an increasing mantle input. Distinct εNd(t) (−0.4 and −0.3) and zircon Hf-O (εHf(t) = 1.5–5.0, δ18O = 6.5–7.2‰) isotopes of the MMEs, suggest that the mafic melt could be sourced from the asthenospheric mantle, contaminated by subcontinental lithospheric mantle/continental crust during magma ascent, and hybridized by felsic melt at emplacement-level. The magmatic sequence in the Dachang district is indicative of an extensional tectonic setting where mantle-derived magmas are predicted to migrate to shallower crustal levels as the crust progressively becomes thinner and hotter. High-temperature partial melting of mature metasedimentary crust triggered by heat input from the upwelled mantle, may contribute to biotite breakdown, which is important for concentrating tin in melts. Fractional crystallization of initially Sn-rich felsic melts under reduced conditions makes further tin enrichment and produces Sn-bearing granites (the Longxianggai pluton). Prolonged mantle upwelling results in distinct magma mixing and the formation of granite porphyry dykes and MMEs. These dykes are highly fractionated with elevated Sn and W contents, which show great potential to form hydrothermal Sn-W mineralization.

Hydrous minerals in the hybridized zone at the slab-mantle interface play a critical role in retaining and transporting fluid mobile elements in subduction zones. The observed high B and As concentrations and high δ11B values in arc magmas were previously thought to have been resulted from the mechanical transport of forearc serpentinites down to subarc depths by mantle corner flows or through subduction channels because B, As, and Sb are mostly released during early stages of subduction. This study provides new insights into the recycling of B, As, and Sb in subduction zones via an in-situ study on a set of carbonate-bearing serpentinite samples from the Mianlue melange complex in the Qinling Orogen. Petrological observations and bulk-rock Sr-Pb isotopes reveal that the Jianchaling serpentinites were formed by interactions between forearc peridotites and sediment-derived fluids at a shallow depth in the subduction channel. Some Jianchaling serpentinites experienced the lizardite to antigorite transition at 300–400 °C with infiltrations of CO2-rich fluids. In contrast, the Liangyazi antigorite serpentinites were produced at a greater depth. Antigorite grains in the Jianchaling serpentinites have lower Feapfu than those in the Liangyazi serpentinites and most antigorite with geochemical data in the literature. Meanwhile, B, As, and Sb concentrations of Fe-poor antigorite from the Jianchaling serpentines (385–653 ppm B, 198–334 ppm As, and 6.24–8.01 ppm Sb) are not only higher than those of the Fe-rich antigorite from the Liangyazi serpentinites but also 2–3.5 times higher than those of the surrounding lizardite. The lower modal percentage of spinel and ferritchromite in antigorite veinlets, together with the lower bulk-rock FeO content of antigorite-lizardite serpentinites than that of lizardite serpentinites from the Jianchaling region, indicates that some FeO might be lost during the lizardite to antigorite transition. The presence of carbonate minerals in the antigorite veinlets and the remarkably high B, As, and Sb concentrations in the Jianchaling antigorite indicate that the penetrating fluids during the lizardite to antigorite transition are most likely derived from sediment and metabasite decarbonations at

The Circum-Pacific subduction zone is a famous gold metallogenic domain in the world, with two important gold metallogenic provinces, the North China Craton and Nevada, which are related to the destruction of the North China Craton and the Wyoming Craton, respectively. Their ore-forming fluids were possibly derived from the stagnant slab in the mantle transition zone. The oceanic lithospheric mantle usually contains serpentine layers up to thousands of meters thick. During plate subduction, serpentine is dehydrated at depths of 300 km. The overlying big mantle wedge is hydrated during the breakdown of these hydrous facies in the mantle transition zone. The dehydration of the subducted slab in the big mantle wedge releases sulfur-rich fluid, which extracts gold and other chalcophile elements in the surrounding rocks, forming gold-rich fluid. Because the cratonic geotherm is lower than the water-saturated solidus line of lherzolite, the fluid cannot trigger partial melting. Instead, it induces metasomatism and forms pargasite and other water-bearing minerals when it migrates upward to depths of less than 100 km in the cratonic lithospheric mantle, resulting in a water- and gold-rich weak layer. During the destruction of craton, the weak layer is destabilized, releasing gold-bearing fluids that accelerate the destruction. The ore-forming fluids migrate along the shallow weak zone and are accumulated at shallow depths, and subsequently escape along deep faults during major tectonic events, leading to explosive gold mineralization. The ore-forming fluids are rich in ferrous iron, which releases hydrogen at low pressure through iron hydrolysis. Therefore, decratonic gold deposits are often reduced deposits.

The geological evolution of the Proto-Tethys Ocean remains vague. The Tianshuihai Terrane (TSHT), a subterrane of the West Kunlun, distributed to the south of the Proto-Tethys Ocean during the Earl...

The Dongguashan deposit is the largest and deepest Cu (Au) deposit in the Tongling region, eastern China. The presence of skarn and stratiform ore bodies in the same system of the Dongguashan deposit has caused much debate as to whether the mineralization was associated with magmatic-hydrothermal or sedimentary origin. In this study, in situ (LA-ICP-MS and EPMA) elemental compositions of medium-to coarse-grained euhedral pyrite from the stratiform and skarn ore bodies, and pyritization marble are investigated in the Dongguashan deposit. The results of the studied pyrite samples have relatively consistent trace element composition characteristics and several genetically significant elemental combinations. Gold and Ag are mainly of lattice bound appearance in pyrite. The Pb, Cu, Sb and Tl elements are mainly incorporated into the pyrite structure as substitutions. The distinctly trace element distribution correlations (e.g., Mn, Se and Tl) of the studied pyrites indicate that they were formed in a high temperature and oxidizing environment. Low Au (1) and Se/Tl (>10) ratios of the studied pyrites imply that the Dongguashan deposit is most likely to be a magmatic-hydrothermal origin.

The late Cambrian tonalite-trondhjemite association in the eastern segment of North Qilian suture zone (E. NQSZ) is important for understanding of the sub-arc magmatic process and tectonic evolutio...

The Duolong mineral district in western Tibet is one of the largest porphyry Cu–Au deposit fields with significant metallogenic potential in China. Its tectonic environment relevant to Early Cretaceous Cu–Au mineralization remains controversial. Here we report new whole-rock major and trace element, and Sr-Nd-Hf-Pb isotopic data for the newly discovered basalt in the Nadun area, Duolong mineral district, to decipher their genesis and further constrain the tectonic environment. A contemporaneous rhyolite sample interbedded with the basalt in the lower part of the volcanic section in the Nadun area yields an LA-ICP-MS zircon U–Pb age of 122.5 ± 1.2 Ma. The basalt samples exhibit high-K calc-alkaline/shoshonite properties and are enriched in high field strength elements, e.g., high TiO2 (1.43–1.79 wt.%) and Nb (14.6–19.5 ppm) contents, with high Nb/La ratios (0.4–0.6), which are compositionally comparable to those of Nb-enriched arc basalts (NEABs). The (87Sr/86Sr)i ratios of 0.7052 to 0.7056, negative eNd(t) (− 0.7 to − 0.2) and eHf(t) values (+ 6.0 to + 6.5), and high (206Pb/204Pb)i, (207Pb/204Pb)i, (208Pb/204Pb)i and ratios (18.522 to 18.561, 15.641 to 15.645 and 38.679 to 38.730, respectively) suggest that the Nadun NEABs are more enriched than those of the island arc basalts (IABs) in the area. The slightly enriched radiogenic isotopes for the Nadun NEABs indicate that the subducting sediments play an important role in the source. Furthermore, their high Nb, Ti, and Cu contents indicate that the source mantle wedge was metasomatized by slab melts. The Nadun NEAB and other coeval magmatic rocks in the Duolong mineral district, including adakite, OIB-like basalt, MORB-type basalt, A-type rhyolite, and common IAB, are typical rock assemblages of ridge subduction. We infer that the Duolong mineral district were formed by ridge subduction in the Early Cretaceous.

In order to investigate the behavior of potassium (K) isotopes during metamorphic dehydration and to further constrain its implications for K recycling in the subduction zone, we measured K isotopic compositions of whole-rocks and separated minerals (phengite, omphacite, amphibole) from Sumdo eclogites, Tibet, using a recently established high-precision analysis method. Our data reveal that the δ41K (‰) of the whole-rock eclogites (−1.64 to −0.24) displays dramatically lower values than observed in fresh mid-ocean ridge basalts (MORB) (−0.43 ± 0.17) and altered MORB (−0.76 to −0.11). In addition, the δ41K values of eclogites show a positive correlation with both K2O contents and K/Nb ratios, which suggests that the low δ41K values were most likely caused by dehydration during subduction. Thus, isotopically heavy K may be released into the mantle wedge, while the light component is subducted into the deep mantle. Therefore, the K isotope systems have the potential to trace subducted crustal materials and to create heterogeneity within the mantle. Mineral separates from Sumdo eclogites are highly heterogeneous in K isotopic compositions, ranging from −0.98 to +0.23 in phengite, from −1.25 to −1.10 in omphacite and from −1.16 to −0.09 in amphibole. The K isotope fractionations between amphibole and phengite and between amphibole and omphacite vary from −0.30 to +0.25 and from −0.04 to +0.63, respectively, indicating K isotopic disequilibrium between amphibole and omphacite/phengite, which might result from the multistage growth of minerals during subduction metamorphism. However, generally phengites show heavier K isotopic compositions than the coexisting omphacite (Δ41Kphengite-amphibole = +0.25) and amphibole (Δ41Kphengite-amphibole = +0.03 ∼ +0.30, except TB193 is −0.25), which may imply that the K isotopic fractionation is controlled by the difference in coordination numbers of K between phengite (6) and omphacite (7 to 8)/amphibole (8).

Calcium isotopes can potentially be used to trace many geological processes. Characterizing the Ca isotopic compositions of marine sediments is a precondition for using Ca isotopes to trace the recycling of marine sediments into the mantle and mantle-derived rocks. The best way to achieve this goal is to investigate different types of marine sediments that are adjacent to subduction zones and would undergo subduction in the future. Here, we present high-precision Ca isotopic data for marine sediments from the Deep Sea Drilling Project (DSDP) Sites 543 and 144 along the Lesser Antilles forearc region, to constrain the Ca isotope signatures of marine sediments with various components and assess their possible effects on δ44/40Ca of the local mantle and arc lavas via subduction. Our results show that the δ44/40Ca values of these marine sediments vary from 0.10 to 1.21‰ and depend on lithology. Carbonate-rich sediments from Site 144 exhibit δ44/40Ca values in the range of 0.10–0.58‰ with a weighted average value of 0.36‰, which is ∼0.6‰ lower than that of the mantle. In comparison, δ44/40Ca of clay-rich sediments from Site 543 range from 0.79 to 1.21‰ with a mantle-like mean value of 0.96‰. The data obtained here suggest that carbonate- and clay-rich marine sediments differ significantly both in Ca isotopic compositions and CaO contents. Recycling of these sediments will result in different effects on the isotopic compositions of the local mantle and probably affect the local arc lavas. Specifically, carbonate-rich sediments from Site 144 with high CaO contents and light Ca isotopic compositions can alter the local mantle to lower δ44/40Ca signatures, which is expected in the southern Lesser Antilles arc lavas.

Granitoids in active continental margins are probes of sub-arc magmatic processes and are crucial for the understanding of crust–mantle interaction during subduction. Here we report data for the mi...

Subduction related metasomatism has been proposed to be one of the most important processes that may cause Ca isotopic heterogeneities in the mantle. Back-arc basin basalts (BABBs) affected by different contributions of subducted materials have various geochemical compositions ranging from MORB-like to arc-like, so that they are ideal media to investigate the behavior of Ca isotopes during plate subduction. Here we report Ca isotopic compositions of a series of BABBs from the Lau Basin, Woodlark Basin and Coriolis Troughs in the southwestern Pacific. Our results show that δ44/40Ca of all the BABBs exhibit a narrow range of 0.73 ± 0.07 to 0.89 ± 0.06‰ with an average of 0.80 ± 0.08‰ (2SD, N = 21), which are lower than the estimated value of the Bulk Silicate Earth (BSE, 0.94 ± 0.05‰), but are similar to that of MORBs reported in the literature (0.83 ± 0.11‰, 2SD, N = 21). Even though the mantle sources of BABBs studied here have been influenced by various amounts of slab-derived fluids and/or sediments, there is no systematic variation of δ44/40Ca with indicators of subduction contributions (e.g., Ce/Pb, Nb/U, Ba/La, Th/Yb, 87Sr/86Sr and 143Nd/144Nd). This suggests that either there is no Ca isotopic discrepancy between subducted components and mantle, and/or the amount of Ca that is transferred from the subducting Pacific slab to the studied BABB mantle sources is insignificant. Water depths at which subducting sediments were deposited control the components of marine sediments that may be subducted, which in turn determine how subduction affects Ca isotopic composition of the mantle. Because the southwestern Pacific has deeper water depths than the carbonate compensation depth, pelagic clays dominate the contribution from the subducted slab rather than carbonates, which leads to the negligible Ca isotopic fractionation during subduction along these trenches. The lack of correlations of δ44/40Ca with geochemical indicators of magma differentiation (e.g., MgO, CaO/Al2O3, Sr/Nd, Eu/Eu*) suggests that intermediate-mafic magma differentiation has insignificant effects on Ca isotopic compositions of BABBs. Therefore, we conclude that the observed offset of δ44/40Ca (∼0.14‰) between the mantle and BABBs studied here should be primarily controlled by Ca isotopic fractionation during partial melting without significant contributions from subducted materials.

The mechanism of lithospheric removal and destruction of the North China Craton (NCC) has been hotly debated for decades. It is now generally accepted that the subduction of the (Paleo)-Pacific plate played an important role in this process. However, how the plate subduction contributed to the craton destruction remains unclear. Here we report high oxygen fugacity (fO2) characteristics of the Yunmengshan granite, e.g., hematite–magnetite intergrowth supported by zircon Ce4+/Ce3+ ratios and apatite Mn oxygen fugacity indicator. High fO2 magmas are widely discovered in Late Mesozoic (160–130 Ma) adakitic rocks in central NCC. The origin of high fO2 magma is likely related to the input of the “oxidized mantle components”, which shows a close connection between plate subduction and destruction of the craton. The research area is ~ 1500 km away from the current Pacific subduction zone. Considering the back-arc extension of Japan Sea since the Cretaceous, this distance may be shortened to ~ 800 km, which is still too far for normal plate subduction. Ridge subduction is the best candidate that was responsible for the large scale magmatism and the destruction of the NCC. Massive slab-derived fluids and/or melts were liberated into an overlying mantle wedge and modified the lithospheric mantle. Rollback of the subducting plate induced the large-scale upwelling of asthenospheric mantle and triggered the formation of extensive high fO2 intraplate magmas.

The region of southern Anhui Province (SAP) is one of Late Mesozoic magmatic belts and copper-molybdenum-gold-tungsten polymetallic ore districts in China, but the relationship of polymetallic mineralization and related granodiorite remains unclear. Previous studies indicate that apatite can be an indicator for magmatic evolution and mineral exploration. In this study, apatites from the SAP ore-bearing granodiorites were investigated using observations from the electron microscope, electron probe microanalysis (EPMA), laser-ablation inductively-coupled plasma-mass spectrometry (LA-ICP-MS). The studied apatite samples identified as fluorapatite, show high F (2.69–4.13 wt.%) and low Cl contents (mainly

Mesozoic A-type granites are widely distributed in the Lower Yangtze River Belt (LYRB) in China, but their petrogenesis and geodynamic settings are unresolved. Here, we describe geochronological an...

Significance Scientists have been working on the dating of atmospheric oxygenation in Earth’s history for decades. However, most previous studies relied on evidence from sedimentary rocks. Here, we show that igneous rocks can also be linked with surface oxidation by a key geological process: plate subduction. We here make an attempt to apply the Th/U of worldwide arc igneous rocks as an indicator for the timing of atmospheric oxygenation over the Earth’s history. The results are coincident with the previously defined Great Oxidation Event and Neoproterozoic Oxygenation Event., Atmospheric oxygen has evolved from negligible levels in the Archean to the current level of about 21% through 2 major step rises: The Great Oxidation Event (GOE) in the early Proterozoic and the Neoproterozoic Oxygenation Event (NOE) during the late Proterozoic. However, most previous methods for constraining the time of atmospheric oxygenation have relied on evidence from sedimentary rocks. Here, we investigate the temporal variations of the Th/U of arc igneous rocks since 3.0 billion y ago (Ga) and show that 2 major Th/U decreases are recorded at ca. 2.35 Ga and ca. 0.75 Ga, coincident with the beginning of the GOE and NOE. The decoupling of U from Th is predominantly caused by the significant rise of atmospheric oxygen. Under an increasingly oxidized atmosphere condition, more uranium in the surface environment became oxidized from the water-insoluble U4+ to the water-soluble U6+ valance and incorporated in the sea water and altered oceanic crust. Eventually, the subduction of this altered oceanic crust produced the low-Th/U signature of arc igneous rocks. Therefore, the sharp decrease of Th/U in global arc igneous rocks may provide strong evidence for the rise of atmospheric oxygen. We suggest that the secular Th/U evolution of arc igneous rocks could be an effective geochemical indicator recording the global-scale atmospheric oxygen variation.

Enrichments in light REE without concomitant enrichments in high-field-strength elements in mantle peridotites are usually attributed to inputs from carbonate-rich melts and referred to as ‘carbonatite’ metasomatism as opposed to interaction with evolved silicate melts. Alternatively, both enrichment types are ascribed to percolating volatile-bearing mafic liquids whose chemical signatures evolve from ‘silicate’ to ‘carbonatite’. Here we compare these models for peridotites in which these enrichment types are combined, as may be common in the mantle. We report new Ca-Sr-Nd isotope and chemical data for lherzolite and harzburgite xenoliths from Spitsbergen that were metasomatized, first, by silicate, then by carbonate-rich melts that formed carbonate-bearing pockets replacing earlier minerals. Seven crushed samples were treated with acetic acid that dissolved carbonates formed in the latest event, but not silicates. The leachates (acid-removed carbonates making up 0.6–1.4% of total sample mass) contain much more Sr than the residues after leaching (277–2923 vs. 16–60 ppm), have a greater overall 87Sr/86Sr range (0.7049–0.7141 vs. 0.7036–0.7055) and higher 87Sr/86Sr in each sample than the residues. The leachates have lower δ44/40Ca range (0.17–0.68‰) than the residues (0.78–1.00‰), as well as lower δ44/40Ca than the residues in all samples but one. By and large, the carbonates are out of Ca-Sr isotope equilibrium with the host peridotites implying that the older silicate and younger carbonatite metasomatism were produced by different parental melts, thus supporting the existence of distinctive carbonate-rich metasomatic media in the lithospheric mantle, possibly including recycled materials. The δ44/40Ca in the leachates (i.e. carbonates, 0.17–0.68‰) are well below bulk silicate Earth (BSE) estimates (0.94 ± 0.05‰) and δ44/40Ca in non-metasomatized melt-depleted mantle. Yet, δ44/40Ca in the non-leached whole rock (WR) carbonate-bearing samples (0.75–0.95‰) fall within, or are only slightly lower than, the BSE range. The 87Sr/86Sr range in these WR samples (0.7030–0.7112) includes very high values for peridotites with large aggregates of dolomite and Mg-calcite. It appears that both carbonatite and silicate metasomatism may produce δ44/40Ca values lower than the BSE such that Ca-isotope data cannot robustly tell apart these two enrichment types, yet carbonatite metasomatism may yield the lowest δ44/40Ca. Carbonates, even at small mass fractions, are significant hosts of Sr in the WR Spitsbergen peridotites (8–51 wt.% of Sr mass) because of very high Sr concentrations, but add little to WR Ca balance (3–12 wt.%). As a result, high Sr content and 87Sr/86Sr ratios may be indices (though not definitive proofs) of carbonatite metasomatism in mantle rocks.

The history of the Hawaiian hotspot is of enduring interest in studies of plate motion and mantle flow, and has been investigated by many researchers using the detailed history of the Hawaiian-Emperor Seamount chain. One of the unexplained aspects of this history is the apparent offset of several Emperor seamounts from the Hawaii plume track. Here we show that the volcanic migration rates of the Emperor seamounts based on existing data are inconsistent with the drifting rate of the Pacific plate, and indicate northward and then southward “absolute movements” of the seamounts. Numerical modeling suggests that attraction and capture of the upper part of the plume by a moving spreading ridge led to variation in the location of the plume’s magmatic output at the surface. Flow of the plume material towards the ridge led to apparent southward movement of Meiji. Then, the upper part of the plume was carried northward until 65 Ma ago. After the ridge and the plume became sufficiently separated, magmatic output moved back to be centered over the plume stem. These changes are apparent in variations in the volume of seamounts along the plume track. Chemical and isotopic compositions of basalt from the Emperor Seamount chain changed from depleted (strong mid-ocean ridge affinity) in Meiji and Detroit to enriched (ocean island type), supporting declining influence from the ridge. Although its surface expression was modified by mantle flow and by plume-ridge interactions, the stem of the Hawaiian plume may have been essentially stationary during the Emperor period.

Kimberlite is characterized by explosive eruption powered by excess carbon dioxides (CO2)1 and water2. Given that diamond is the dominant stable phase of carbon in the upper mantle3, it is obscure where does the excess CO2 in kimberlite has come from. Here we show that ferric iron oxidizes diamond at 1900K, 20GPa and 2000K, 25GPa, forming CO2. The lower mantle is dominated by bridgmanite, which is rich in ferric iron4. Bridgmanite decomposes once it is brought to the upper mantle, releasing extra ferric iron. Therefore, the oxidation of diamond may have been popularly occurring at the base of the upper mantle, forming CO2-rich carbonated domains that are the main source of kimberlite. The rising kimberlitic magma reaches the lithosphere mantle of thick cratons before it crosses the solidus line of mantle peridotite, and thus keeps its volatile-rich nature that drives explosive eruptions. When the lithospheric mantle is thinner than ~140 km, kimberlite changes into much less explosive magmas due to partial melting of mantle peridotite, and, consequently, entrained diamond is mostly oxidized during the magma’s slower ascension.

Seamounts on the drifting oceanic crust are inevitably carried by plate motions and eventually accreted or subducted. However, the geochemical signatures of the subducted seamounts and the significance of seamount subduction are not well constrained. Hundreds of seamounts have subducted beneath the Philippine Sea Plate following the westward subduction of the Pacific Plate since the Eocene (~52 Ma). The subducted oceanic crust and seamount materials can be exhumed from the mantle depth to the seafloor in the Mariana forearc region by serpentinite mud volcanoes, providing exceptional opportunities to directly study the subducted oceanic crust and seamounts. The International Ocean Discovery Program (IODP) expedition 366 has recovered a few metamorphosed mafic clasts exhumed from the Mariana forearc serpentinite mud volcanoes, e.g., the Fantangisna and Asut Tesoru seamounts. These mafic clasts have tholeiitic to alkaline affinities with distinct trace elements and Nd-Hf isotopes characteristics, suggesting different provenances and mantle sources. The tholeiites from the Fantangisna Seamount have trace element characteristics typical of mid-ocean ridge basalt. The Pacific-type Hf-Nd isotopic compositions, combined with the greenschist metamorphism of these tholeiites further suggest that they came from the subducted Pacific oceanic crust. The alkali basalts-dolerites from the Fantangisna and Asut Tesoru seamounts show ocean island basalt (OIB)-like geochemical characteristics. The OIB-like geochemical signatures and the low-grade metamorphism of these alkali basalts-dolerites suggest they came from subducted seamounts that originally formed in an intraplate setting on the Pacific Plate. The Pacific Plate origin of these metabasites suggests they were formed in the Early Cretaceous or earlier. Two types of OIBs have been recognized from alkali metabasites, one of which is geochemically similar to the HIMU-EMI-type OIBs from the West Pacific Seamount Province, and another is similar to the EMII-type OIBs from the Samoa Island in southern Pacific, with negative Nb-Ta-Ti anomalies and enriched Nd-Hf isotopes. Generally, these alkali metabasites are sourced from the heterogeneous mantle sources that are similar to the present South Pacific Isotopic and Thermal Anomaly. This study provides direct evidence for seamount subduction in the Mariana convergent margins. We suggest seamount subduction is significant to element cycling, mantle heterogeneity, and mantle oxidation in subduction zones.

The magmatic character of early subduction zone and arc development is unlike mature systems. Low-Ti-K tholeiitic basalts and boninites dominate the early Izu-Bonin-Mariana (IBM) system. Basalts recovered from the Amami Sankaku Basin (ASB), underlying and located west of the IBM’s oldest remnant arc, erupted at ~49 Ma. This was 3 million years after subduction inception (51-52 Ma) represented by forearc basalt (FAB), at the tipping point between FAB-boninite and typical arc magmatism. We show ASB basalts are low-Ti-K, aluminous spinel-bearing tholeiites, distinct compared to mid-ocean ridge (MOR), backarc basin, island arc or ocean island basalts. Their upper mantle source was hot, reduced, refractory peridotite, indicating prior melt extraction. ASB basalts transferred rapidly from pressures (~0.7-2 GPa) at the plagioclase-spinel peridotite facies boundary to the surface. Vestiges of a polybaric-polythermal mineralogy are preserved in this basalt, and were not obliterated during persistent recharge-mix-tap-fractionate regimes typical of MOR or mature arcs., Magmatism associated with early growth of subduction zones is unlike that of mature island arc systems. Here, the authors find basalts with distinct mineralogical and geochemical characteristics were erupted during this early stage, and derived from extremely refractory, hot mantle sources.

We present high-precision iron isotope data for harzburgites and constituting minerals from the Yushigou supra-subduction zone ophiolite in the North Qilian orogen of northern Tibet to provide insights into iron isotope behavior in subduction zones. The Yushigou harzburgites represent typical melting residues of silica-enriched forearc mantle wedge with significant melt extraction and minimal melt percolation. The δ56Fe ranges from −0.096‰ to 0.110‰ in olivine, from −0.111‰ to 0.105‰ in orthopyroxene and from −0.141‰ to 0.054‰ in spinel. Coexisting minerals fall on a line with a slope of 1 in the δ-δ plot, indicating iron isotope equilibrium between mineral pairs. Iron isotopic fractionation between olivine and orthopyroxene is minor, which is consistent with theoretical predictions and previous observations. Good iron isotopic equilibrium between olivine and orthopyroxene is a further indicator that silica enrichment in arc peridotites is inherited from their mantle source. In contrast, significantly negative iron isotopic fractionation between spinel and olivine contradicts with most previous constraints, which may be ascribed to Cr substitution in spinel. Bulk rock harzburgites have δ56Fe ranging from −0.097‰ to 0.076‰, the average of which (0.002 ± 0.020‰) is similar to average δ56Fe determined for abyssal peridotites. Based on this inference, the lower δ56Fe of arc magmas than MORB should be due to smaller discrepancy in Fe3+/ΣFe between arc magmas and arc peridotites when compared with MORB and abyssal peridotites. Our modelling demonstrates that this can be achieved if arc magmas are produced by redox-buffered melting of silica-enriched mantle wedge while MORB are produced by redox-unbuffered melting of the depleted mantle.

Plate subduction is the largest natural factory that processes elements, which controls recycling and mineralization of a variety of elements. There are three major ore deposit belts in the world: the circum-Pacific, the central Asian, and the Tethys belts. All the three belts are closely associated with plate subductions, the mechanism remains obscure. We approached this problem from systematic studies on the behaviours of elements during geologic processes. This contribution summaries the recent progress of our research group. Our results suggest that porphyry Cu deposits form through partial melting of subducted young oceanic crust under oxygen fugacities higher than ΔFMQ ∼+1.5, which is promoted after the elevation of atmospheric oxygen at ca. 550 Ma. Tin deposits are associated with reducing magmatic rocks formed as a consequence of slab rollback. The Neo-Tethys tectonic regime hosts more than 60% of the world’s total Sn reserves. This is due to the reducing environment formed during the subduction of organic rich sediments. For the same reason, porphyry Cu deposits formed in the late stages during the closure of the Neo-Tethys Ocean. Tungsten deposits are also controlled by slab rollback, but is not so sensitive to oxygen fugacity. Subduction related W/Sn deposits are mostly accompanied by abundant accessory fluorites due to the breakdown of phengite and apatite. Decomposition of phengite is also significant for hard rock lithium deposits, whereas orogenic belt resulted from plate subduction promote the formation of Li brine deposits. Cretaceous red bed basins near the Nanling region are favorable for Li brines. Both Mo and Re are enriched in the oxidation-reduction cycle during surface processes, and may get further enriched once Mo-, Re-enriched sediments are subducted and involved in magmatism. During plate subduction, Mo and Re fractionate from each other. Molybdenum is mainly hosted in porphyry Mo deposits and to a less extent, porphyry Cu-Mo deposits, whereas Re is predominantly hosted in porphyry Cu-Mo deposits and sedimentary sulfide deposits.

Over the Earth’s evolutionary history, the style of plate subduction has evolved through time due to the secular cooling of the mantle. While continuous subduction is a typical feature of modern plate tectonics, a stagnant-lid tectonic regime with localized episodic subduction likely characterized the early Earth. The timing of the transition between these two subduction styles bears important insights into Earth’s cooling history. Here we apply a statistical analysis to a large geochemical dataset of mafic rocks spanning the last 3.5 Ga, which shows an increasing magnitude of alkali basaltic magmatism beginning at ca. 2.1 Ga. We propose that the rapid rise of continental alkali basalts correlates with an abruptly decreasing degree of mantle melting resulting from the enhanced cooling of the mantle at ca. 2.1 Ga. This might be a consequence of the initiation of continuous subduction, which recycled increasing volumes of cold oceanic crust into the mantle., The point in time in which the ancient stagnant-lid tectonic regime with episodic plate subduction changed to continuous plate subduction is still topic of scientific debate. Here, the authors present vast geochemical dataset covering the last 3.5 billion years of Earth’s history and date the onset of continuous subduction to about 2.1 billion years ago.

The Cu-Mo polymetallic deposits are widespread in the Chizhou region, the Middle and Lower Yangtze River metallogenic belt, eastern China, but the formation of Cu-Mo polymetallic deposits remains unclear. Geochronological, geochemical and Sr-Nd-Hf isotopic compositions of the Cu-Mo polymetallic deposits in the Chizhou region are systemically investigated to reveal the formation of Cu-Mo polymetallic deposits and petrogenesis of the granodiorite (porphyry). The geochronological dating results show the formation ages of Cu-Mo polymetallic deposits (molybdenite Re-Os age, 151–148 Ma) and the granodiorites (porphyries) (zircon U-Pb age, 150–146 Ma) in the Chizhou region are simultaneous (within error). The Chizhou granodiorites (porphyries) (SiO2 = 62.9–65.7 wt%) are high-K calc-alkaline peraluminous I-type granodiorites. They have high Mg# (mainly >40) values and low Na2O ( 0.2 wt%), and high log fO2 values (−28.3 to −6.22). The distinctly chemical and isotopic compositions indicate that the Chizhou granodiorites formed from metasomatized mantle-derived magmas mixing with Neoproterozoic crustal components. The results indicate that the formation of the Cu-Mo polymetallic deposits in the Chizhou region may depend on the slab-derived Cl-enriched fluids, an oxidizing environment and a mantle-crust mixed source.

Scheelite has been analyzed from the Shizhuyuan and the Xianglushan world-class W deposits from the Nanling W–Sn region and Jiangnan W belt, respectively. The Shizhuyuan deposit consists of proximal skarn and greisen W–Sn–Mo–Bi and distal Pb–Zn–Ag veins. The Xianglushan deposit, contains layer-like sulfide–scheelite and skarn W orebodies on granite cupolas overprinted by W greisen veins. Scheelite in skarn ores from the Shizhuyuan contains higher concentrations of Mo than those in the sulfide–scheelite and skarn ores from the Xianglushan deposit, reflecting differences between oxidizing and reducing magmatic-hydrothermal fluids. Under oxidizing conditions, W is accompanied by Mo partitions into exsolved fluids to form W–Mo garnet skarns, whereas under reducing conditions, little Mo is carried by exsolved fluids to form W pyroxene skarns. Trace element patterns of scheelite from both deposits show negative Ba, Sr, Zr, and Ti, and positive Ta anomalies. Rare earth element (REE)patterns of scheelite within skarns from the Shizhuyuan deposit have negatively inclined and flat M-type tetrad patterns, and scheelite from the greisens displays flat and positively inclined M-type tetrad patterns. We infer that the fluids formed scheelite within the W skarns and greisens inherited parental magma trace element and REE characteristics (depleted Ba, Sr, Zr, and Ti, enriched Ta, negative Eu anomalies, and tetrad effects). Whereas, scheelite from sulfide–scheelite veins and skarns of the Xianglushan deposit also has W- and MW-type tetrad REE patterns. The W-type tetrad REE patterns are complementary to REE patterns from the Renjiashan granite, and the MW-type tetrad REE patterns occur during a single evolutionary stage within a complex hydrothermal environment. Sulfide mineralization can form after or before W skarns (the former like Shizhuyuan deposit and the latter like Xianglushan deposit). The formation conditions of the latter included reducing conditions and sulfide firstly supersaturated in the melt, resulting in sulfide drops which carried W aggregated on the cupolas. W skarns and greisens in both deposits underwent generally successive processes related to water supersaturation in the melt. Following a temperature decrease and crystallization, bubbles carried material changing from Si and metal to Si oxide complexes.

Geochemical, Sr Nd isotopic and geochronological methods were used to reveal the protoliths and geodynamic implications of the newly discovered Triassic Baqing eclogites, eastern Qiangtang terrane, central Tibet. We assessed the mobility of trace elements, and it turned out that high field strength elements (HFSEs) and rare earth elements (REEs) were immobile in the high-pressure (HP) metamorphic process and reliably employed to discriminate the protoliths of the Baqing eclogites. Whole-rock geochemical data (especially immobile elements) suggested both arc-related (negative Nb Ta anomalies, high light (L) REEs/heavy (H) REEs ratios) and mid-oceanic ridge basalt (MORB)-related (without negative Nb Ta anomalies, relatively low REEs contents) characteristics, indicating the contribution of different proportions of subduction material with a backarc basin origin. According to the Sr Nd isotopic ratios (initial 87Sr/86Sr ratios: 0.7078–0.7086; eNd(t) values: −0.58 to +1.96), the protoliths of the Baqing eclogites originated from mantle which experienced continental crustal contamination in the subduction zone. Compared with nearby subduction-related magmatic rocks from eastern Qiangtang terrane, backarc basin was the most likely tectonic environment for the Baqing eclogite protoliths. The Jinsha Paleo-Tethyan Ocean southward subduction underneath East Qiangtang block (EQB) and rollback led to the formation of this backarc basin between the Late Permian and Early Triassic, and subsequent northward subduction of the backarc basin formed the Baqing eclogites in early Late Triassic (227–221 Ma).

The Dongyuan deposit, in the recently identified world-class porphyry–skarn W metallogenic belt, east Jiangnan Orogen, China, is characterized by a set of nearly parallel and horizontal multi-cycle orebodies within cupolas of porphyritic granodiorite. Almost every orebody from the top to the bottom shows similar vertical changes in vein morphology and mineral assemblage from ≤1 cm quartz–feldspar–scheelite–calcite veins to 1–2 cm quartz–feldspar–muscovite–scheelite veins to 2–5 cm quartz–feldspar–muscovite–scheelite veins to ≥5 cm quartz veins. From the bottom to the top of each orebody in every cycle, the mole number of AlIV in K-feldspar decreases, the concentrations of Mn2+ in calcite increases, and the mole number of OH− in muscovite decreases. These changes reflect repetitive fluctuations in the concentrations of major elements that were involved in the precipitation of minerals as well as the physicochemical conditions of fluids during vein formation in the multi-cycle orebodies. The concentrations of Mo are obviously lower in pyrite from the middle orebodies than in the top and bottom orebodies. Our data indicate that the ore-forming fluids were sourced from multi-batch fluids that were exsolved from magma. The ΣLREE/ΣHREE ratios in the scheelite decrease from the upper to lower orebodies. Similar trends are found within a single orebody from the top to the bottom of veins. These trends were due to feldspar fractionation in the magmas and ore-forming fluids, respectively, because the crystallization of feldspar tends to capture LREEs. Therefore, the following model is proposed for the formation of the Dongyuan porphyry W deposit: the magma crystallized first at the roof of the intrusion, causing the underlying magmas to become oversaturated in H2O; the exsolved H2O then accumulated at the boundary between the crystallized upper part and the underlying magmas; as fluid pressures increased, cracks formed in the overlying crystallized granodiorite, and as a consequence the first cycle of vein networks was formed. This process was repeated as the interface between magma and crystallized rock moved downwards, thus forming the multi-cycle W orebodies that are embedded in the pluton.

Apatite as an excellent tracer of the magmatic and metallogenic processes, can incorporate large amounts of trace elements sensitive to melting and/or ore-fluid evolution. This paper reports a combined study of in situ apatite geochemistry and whole-rock geochemistry for Early Cretaceous ore-bearing (Group I) and ore-barren porphyries (Group II) in the Liguo Fe-Cu-Au deposit for the deciphering of its petrogenesis and the control factors of mineralization. The apatite from both Group I and Group II is magmatic fluorapatite, characterized by Eu negative anomalies, an enrichment of LREEs, and a depletion of HREEs. Meanwhile, both groups have high Sr/Y and delta Eu, indicating porphyry adakitic characteristics in origin. Compared to isotopes of whole rocks, the variable Sr-87/Sr-86 (0.70250-0.71262) and delta O-18 (6.22-9.00) values of both groups of apatite imply contributions of mantle-, crust- and/or sediment-derived materials. Although Group I apatite and II apatite show some similar geochemical characteristics, Group I apatite precedes the crystallization of plagioclase with no Sr-(La/Yb)(N)/(La/ Sm)(N)/(Sm/Yb)(N) correlations, whereas Group II apatite coincides with plagioclase crystallization, displaying positive correlations. The geochemistry of the elements (delta Eu, delta Ce, MnO and V) of these apatite groups, which are sensitive to a redox state, shows high oxygen fugacity (between HM and NNO), and Group I apatite has systematically higher oxygen fugacity than Group II apatite. More importantly, these different trace elements and oxygen fugacity characteristics between Group I apatite and Group II apatite can be used as an indicator of mineralization, and allow the skarn Fe-Cu-Au mineralization range to be drawn for the first time. Furthermore, the estimated F and Cl contents of the host parent magma (F = 1300-2446 ppm, Cl = 140-4780 ppm) are higher than those in a primitive mantle and average continental crust, suggesting an enrichment process of F and Cl. According to the above adakitic characteristics, high oxygen fugacity, and high F and Cl contents, the Pacific Plate subduction may be the main dynamic mechanism of the diagenesis and mineralization found in Liguo.

The East Taiwan Ophiolite (ETO) occurs as blocks and thrust sheets associated with the Lichi Melange in the Coastal Range of eastern Taiwan. The blocks consist of serpentinized harzburgite, serpentinite breccia, gabbro, dikes of dolerite and plagiogranite, pillow basalts, and red clay within a mud-and serpentinite-rich melange matrix. New U-Pb zircon dating of a pegmatite gabbro yields a weighted mean age of 16.65 +/- 0.20 Ma. This age is earlier than the North Luzon Arc but overlaps with the late-stage spreading of the South China Sea. ETO glassy basalt has low K2O, MgO and high CaO contents, similar to MORB. REE and trace element patterns show both N-MORB patterns with LREE depletion and E-MORB patterns with slight LREE enrichment. A few samples show slight depletion in Nb-Ta and Ti and enrichment in Rb, Ba, U and Sr, indicating a hint of subduction influence. Most ETO basalt plots within the overlapping fields of N-MORB and BABB on Ti-V, Cr-Y, Nb/Yb-Th/Yb, and Hf/3-Th-Ta discrimination diagrams. These geochemical compositions are emblematic of mid-ocean ridge or back-arc lava, like South China Sea basalt. We interpret ETO basalt and gabbro as fragments of the subducted South China Sea basement that were scrapped off and accreted to the Luzon forearc during the process of subduction initiation along the Manila Trench. Blocks of mantle material in the melange may originate from the upper plate of the arc-continent collision and were mixed with lower plate crustal material in a subduction channel now represented by the Lichi Melange.

The Yangshan gold ore belt contains a series of sediment-hosted disseminated gold deposits in the southern subzone of the western Qinling orogen. The ore deposits formed in the Middle Devonian carbonaceous clastic-carbonate sequences that underwent strong deformation and low greenschist facies regional metamorphism. The shapes and distribution of the ore bodies are strictly confined by the E-W-trending Anchanghe-Guanyinba thrust fault zone. Pyrite and arsenopyrite are the major Au-carrying minerals. Microthermometric results show that the ore-forming fluids are CO2-rich aqueous solutions. The homogenization temperatures range from 221 to 303.5 °C, and salinities range from 2.0 to 7.2 wt% NaCl equiv. These characteristics are consistent with orogenic gold deposit. Noble gas isotopic measurement shows that 3He/4He ratios of pyrite and quartz grains inherited from the mineralized carbonaceous phyllite, ranging from 0.033 to 0.081 Ra, implying that the origins of ore fluids are restricted within the crustal regime. The 40Ar/36Ar ratios range from 434 to 863, indicating 34 to 68% crust-derived radiogenic 40Ar* in the ore fluids. Based on the backscattered electron (BSE) maps of pyrite, three generations (i.e. core, mantle, rim) are distinguished: pre-ore pyrite including framboidal pyrites and the cores of zoned pyrites both in the phyllite and plagiogranite dikes; ore stage pyrites including the middle zones of the three-zoned pyrites in the granitic dikes and the outer zones of the zoned pyrites in the phyllite; post-ore pyrites, i.e. the outermost rims of the zoned pyrites in the dikes. The arsenopyrite in the phyllite has similar texture to the pyrite, and is regarded as the peak of the mineralization due to its highest Au concentrations. High spatial resolution NanoSIMS multi-elemental maps of the pyrites reveal that (i) gold is distributed heterogeneously as lattice-bounded Au in the zoned pyrites with higher Au (As) in the middle zones and lower Au (As) in the inner and outermost zones. No nano-particles of native gold was identified in pyrite; (ii) gold mineralization occurred in a single event instead of multistage ore formation; (iii) gold precipitation event is followed by an overgrown zone with low Au and As concentrations. In situ sulfur isotopes of Au-bearing pyrite identified two types of sulfur sources: one is framboidal pyrite and the barren core of pyrite in the carbonaceous phyllite, which have very negative δ34S values ( Combining Au concentrations with the corresponding δ34S values in the pyrites reveals that the zones with high Au have lower δ34S values (and vice versa). This coupling relation between Au and S isotope has been also identified in other orogenic and Carlin-type gold deposits. The oxidation of ore fluid is taken as the key for gold precipitation during formation of orogenic gold deposit. However, the oxidation of ore fluids accompanied by the escape of reduced gas (e.g. H2S) should result in high Au coupled with high δ34S. Therefore, we propose that the Yangshan gold deposit forms through partitioning aqueous vapor with high Au, As, and lower δ34S from a deep magma chamber under the ore field, which leads to overgrowth of the pre-existing pyrites. Based on the geological features, texture of pyrite, and sulfur isotopic characteristics, we suggested that the Yangshan gold deposits have the characteristics between Carlin-type and orogenic gold deposits. They are the product of the western Qinling orogeny during the Late Triassic.

Hydrothermal ore deposits at convergent plate boundaries represent extraordinary metal enrichment in the continental crust. They are generally associated with felsic magmatism in extensional settings at high thermal gradients. Although their formation is common during accretionary orogeny, more and more ore deposits have been discovered recently in the collisional orogens of China. Because collisional orogeny was operated in a compressional regime at low thermal gradients, it is not favorable for mobilization of ore-forming elements and thus for the production of hydrothermal ore deposits. Nevertheless, continental collision is generally preceded by oceanic subduction, which enables the preliminary enrichment of ore-forming elements in the mantle wedge due to chemical metasomatism by subducting slab-derived fluids. This gave rise to metal pre-enriched domains in the overriding lithosphere, which may be reactivated by extensional tectonism for hydrothermal mineralization either immediately during accretionary orogeny or at a later time during and after collisional orogeny. It is these tectonic processes that have resulted in the progressive enrichment of ore-forming elements through the geochemical differentiation of the subducting oceanic crust, the metasomatic mantle domains and the mafic juvenile crust, respectively, at different depths. Finally, the reactivation of metal pre-enriched domains by continental rifting in the orogenic lithosphere is the key to the metallogenesis of collisional orogens.

The genesis of seamount chains along fossil spreading ridges reflects the mantle melting dynamics of dying spreading center. The South China Sea developed a seamount chain along its fossil ridge shortly after cessation of spreading. International Ocean Discovery Program Expedition 349 has recovered three volcanic layers, a shallow volcanic breccia layer and two deeper (upper and lower sections) basalt layers, at Site U1431 near the fossil ridge of the South China Sea. Despite the mid-ocean ridge basalt (MORB)-like geochemistry of the lower and upper section basalts, they show distinct geochemical differences that indicate a change of mantle source and a transition from spreading to intraplate volcanism. While the lower section basalts could reflect the last stage volcanism before the cessation of ridge spreading, the upper section basalts with whole-rock MgO up to 20wt.% were formed by intraplate volcanism with significant olivine accumulation once the spreading center had become inactive. The MORB type basalts at Site U1431 have Hawaii-like high olivine Ni and Fe/Mn and low olivine Ca and Mn and bulk-rock low CaO that are distinctly different from normal global MORBs, suggesting an anomalous mantle source in lithology. We suggest that an eclogite-/pyroxenite-rich component, possibly sourced from the Hainan hot spot, played a fundamental role in the transition from spreading to intraplate volcanism of the dying spreading ridge of the South China Sea.

The origin of A-type granites has been the subject of great debate, especially the enigmatic synchronous A1- and A2-type granites. Cretaceous (~125 Ma) A1- and A2-type granites are common throughout the Lower Yangtze River Belt (LYRB), eastern China. However, their genesis still remains unclear. In this study, in-situ zircon Oisotopic data and chemical compositions of the Huayuangong (HYG) A-type granites in Anhui province, provide new insights into the origin and evolution of A-type granites in the LYRB, as well as the genetic link for synchronous A1- and A2-type granites. The HYG granites include syenogranite (75.9 wt%–76.6 wt% SiO2) and quartz syenite (66.1 wt%–66.9 wt% SiO2). Both are metaluminous and belong to ferroan series. They are characterized by high alkalis (K2O + Na2O = 8.36 wt%–8.55 wt% and 11.7 wt%–11.9 wt%), high field strength elements (Zr + Nb + Ce + Y = 909 ppm–1269 ppm and 1092 ppm–1329 ppm) and high Ga/Al ratios (10,000 ∗ Ga/Al = 4.91–4.96 and 2.64–2.68). The zircon saturation thermometer results indicate high magmatic temperatures (896–964 °C and 860–882 °C). All those geochemical features show an A-type granite affinity. They can be further classified into A1-and A2-type granites, corresponding to reduced and oxidized A-type granites, respectively. Additionally, the in-situ zircon O-Hf isotope compositions are also distinctly different, with δ18O = 4.7‰–6.0‰ and eHf(t) = −1.5 to −3.6 for A1-type granites, and δ18O = 7.0‰–7.8‰ and eHf(t) = −3.3 to −6.9 for A2-type granites. The geochemical signatures and newly discovered δ18O and eHf(t)values of the two A-type granite subgroups, indicate that they were derived from different source components and under disparate physicochemical conditions (e.g.,temperature, redox state and water contents). Lithospheric mantle-like isotopic data from zircons of A1-type granites suggest fractional crystallization of reduced, anhydrous basaltic magmas resulting in the formation of A1-type granites. In contrast, A2-type granites with relatively high δ18O and negative eHf values were generated from partial melting of the lithospheric mantle which was metasomatized by slab-derived melts/fluids. The coexisting A1- and A2-type granites were formed under the extensional setting where lithospheric thinning and asthenosphere upwelling occurred.

South China is famous for its poly-metal deposits, with more than 50% of global W and Sb reserves, 20% of global Sn reserves, as well as abundant Cu-Mo-Pb-Zn-Au-Ag, Ta, Nb, REE and U deposits. This...

The Dunde iron–zinc deposit is a typical submarine volcanogenic iron oxide deposit in the eastern Awulale Metallogenic Belt of the western Tianshan. We report new petrography, rare earth elements (REE), and fluid inclusion microthermometry and C–O–Sr isotopes of the calcite from the skarn and the magnetite ores from the Dunde deposit. Calcite from both the skarn and magnetite ores displays high LREE/HREE (44–172), implying significant REE fractionation. We infer that the complexation of Cl (± F) during the fluid circulation and mineralogical control of calcite at the depositional sites were important for the LREE enrichment in the Dunde calcite. The restricted Y/Ho (26.7–32.1) suggests that calcite in the skarn and magnetite ores was co-genetic and the superchondritic Y/Ho may have resulted from the preferential incorporation of Y over Ho in calcite. Positive Eu anomaly (1.23–1.63) of the skarn calcite may have been inherited from the parental fluids, while the negative or slightly positive anomaly (0.65–1.04) of the magnetite ore calcite may have been resulted from the high Sr2 + content that suppressed Eu2 + in calcite. Fluid inclusion microthermometric results exhibit a wide range of homogenization temperatures (106 °C–364 °C) and salinities (0.2 wt.%–41.5 wt.% NaCl equiv.). Three kinds of ore-forming fluids can be distinguished, i.e., high-salinity (> 35 wt.% NaCl equiv.), medium-salinity (10 wt.%–35 wt.% NaCl equiv.) and low-salinity (< 10 wt.% NaCl equiv.) fluids. The high-salinity fluid was likely to be magma-derived and occurred in both the skarn- and magnetite ore calcite. The low-salinity fluid, possibly seawater-derived, was found only in the skarn calcite. The medium-salinity fluid, mainly of CaCl2–H2O aqueous- and brine-featured fluids, may have been the product of the fluid–rock interactions as the magmatic fluids and seawater circulated through the strata. The Dunde calcite has δ13C comparable to marine carbonates, but δ18O substantially lower, suggesting possibly a result of limestone dissolution. Calculated δ18O of the fluids in equilibrium with calcite vary from − 0.78‰ to 5.46‰ (− 0.78‰–1.85‰ and 2.15‰–5.46‰ for skarn calcite and magnetite ore calcite, respectively), implying magmatic fluids and seawater mixing. The Dunde calcite 87Sr/86Sr (0.7071 to 0.7078) correlate positively with 1/Sr, implying that two sources, i.e., basaltic/andesitic volcanic wall rocks (87Sr/86Sr: 0.7056–0.7066) and limestones/seawater (87Sr/86Sr: 0.7075–0.7089), may have been involved in the mineralization. Mass-balance calculation between fresh basaltic tuff and its neighboring skarn showed that extra Ca, Mg and Fe were added to the basaltic tuff during the hydrothermal alteration. The limestone dissolution during the fluid–rock interactions may have added Ca to the ore-forming fluids, while Fe and Mg may have been introduced from the deep-seated magmatic system via the heated Cl-rich fluid circulation.

The Lower Yangtze River Belt in eastern China preserves evidence for widespread Mesozoic magmatism. Geochronological and geochemical studies have been conducted on the magmatic rocks and their associated mineral deposits; however, the relationships and melting conditions of spatially associated I- and A-type granites remain poorly constrained. The Qingyang–Jiuhuashan granitic complex in southern Anhui Province, eastern China, comprises primarily hornblende-bearing granodiorite, monzogranite, alkali-feldspar granite, and granitic porphyry, which yield zircon U Pb ages of 142.2 ± 1.3, 134.6 ± 1.2, 126.4 ± 1.9, and 125.9 ± 1.2 Ma, respectively. Zircon grains from an enclave in the monzogranite yield an age of 134.5 ± 1.1 Ma, similar to that of the host rock. The ages are consistent with the timing of regional magmatic events in southern Anhui Province. Whole-rock geochemical compositions indicate that the granodiorite and monzogranite represent I-type granites, whereas the alkali-feldspar granite and granitic porphyry display A-type signatures. The I- and A-type granites yield similar whole-rock Sr Nd isotopic compositions, but distinct zircon δ18O values. We infer that the I-type granites were produced through water-fluxed melting of Mesoproterozoic basement during subduction of the paleo-Pacific plate. The A-type granites were generated through partial melting of the lower crust, triggered by extensional upwelling of mantle-derived magmas that provided heat and alkali-rich fluids. We identify three stages of Cretaceous tectono-magmatic activity that occurred in response to ridge subduction and subsequent slab rollback beneath eastern China.

Rocks and ore deposits in subduction zones can provide clues to material cycling and related subduction processes. Here we demonstrate, based upon a new data compilation from more than 240 late-Mesozoic fluorite deposits in southeast China that most of them can be divided into three fluorite deposit belts (FDB) with distinct ages and orientations. The two older inland FDB with a NE-SW orientation were formed at 165–150 Ma and 150–130 Ma, respectively, whereas the younger coastal FDB, roughly perpendicular to the others, was generated at 110–70 Ma. The temporal-spatial distributions of the FDB are compatible with the subduction of the (paleo-) Pacific plate. We propose that the genesis of the earlier two inland FDB was due to northeastward slab rollback during low-angle subduction, while formation of the coastal FDB resulted mainly from high-angle northwestward hot/warm subduction. Large amounts of F, mainly derived from the decomposition of phengite (±apatite) in the subducted slabs, could be transported into the overlying crust via F-rich fluids and alkalic magmas. Subsequently, extensive crustal magmatism and fluid activities associated with subduction further concentrate F from the magmas and/or wall-rocks and facilitate fluorite mineralization. This study provides new insights into the interaction between the subducting (paleo-) Pacific plate and overlying eastern Eurasian continent in the late Mesozoic. It also highlights the potential genetic association of F transport/deposition with plate subduction that could be a common phenomenon in ancient and modern subduction zones.

In general, the arc magmas are oxidized, forming magnetite-series granitoids. However, there are also ilmenite-series granitoids at convergent margins. The mechanism that controls the formation of different types of granitoids remains obscure. Here we show that the temporal distributions of ilmenite-series granitoids in Japan have a good correlation with oceanic anoxic events (OAEs). The subduction of OAE sediments with high contents of organic matters releases reducing fluids/melts into the mantle wedge, leading to the formation of ilmenite-series granitoids. This is supported by sulfur isotope compositions and high concentrations of organic gas of the ilmenite-series granitoids in Japan. The temporal correlation between OAEs and Il-series granitoids suggests that the OAE sediments was predominantly distributed within short distances (

Much of the world’s Li deposits occurs as basinal brines in magmatic orogens, particularly in continental volcanic arcs. However, the exact origin of Li enrichment in arc magmatic systems is not clear. Here, we show that, globally, primitive arc magmas have Li contents and Li/Y ratios similar to mid-ocean ridge basalts, indicating that the subducting slab has limited contribution to Li enrichment in arc magmas. Instead, we find that Li enrichment is enhanced by lower degrees of sub-arc mantle melting and higher extents of intracrustal differentiation. These enrichment effects are favored in arcs with thick crust, which explains why magmatism and differentiation in continental arcs, like the Andes, reach greater Li contents than their island arc counterparts. Weathering of these enriched source rocks mobilizes and transports such Li into the hydrologic system, ultimately developing Li brines with the combination of arid climate and the presence of landlocked extensional basins in thickened orogenic settings., The exact origin of lithium enrichment in arc magmatic systems is unclear. Here the authors conduct a global systematics of lithium, explaining why volcanic arcs built on thickened crust are most lithium-enriched, which sheds light on the future exploration of lithium resources.

The newly discovered giant Dahutang W–Cu–Mo deposit in the central part of the Jiangnan Orogen hosts an estimated resource of 1.1 million tons (Mt) of WO3, plus a proven resource of 0.65 Mt Cu and 0.08 Mt Mo. The Dahutang deposit is temporally and spatially associated with the Late Mesozoic S-type granites emplaced into the Neoproterozoic Jiuling granodiorite batholith and the Shuangqiaoshan Group. Based on petrographic observations, eight hydrothermal alteration/mineralization stages have been recognized, comprising pegmatitic (Stage I), potassic alteration (Stage II), albite alteration (Stage III), greisenization and main mineralization (Stage IV), polymetallic sulfide mineralization (Stage V), late scheelite mineralization (Stage VI), carbonate alteration (Stage VII) and supergene alteration (Stage VIII). Stage IV can be further divided into the early greisenization (Stage IV-A), the main wolframite mineralization (Stage IV-B) and the main scheelite mineralization (Stage IV-C). Fluid inclusion and H–O isotope analyses on nine Stage IV to VII ore/gangue minerals (wolframite, cassiterite, scheelite, sphalerite, fluorapatite, fluorite, quartz, calcite and chlorite) suggest the presence of an early W and a late Cu–Mo ore-forming fluid system. The early ore-forming fluids belong to a medium to high temperature (200°–420 °C), low salinity H2O–NaCl–CaCl2 system. This primary magmatic-hydrothermal fluid (δ18Ofluid = 5.4 to 8.8‰ and δD = −102 to −75‰) was likely exsolved from the Late Jurassic granites at Dahutang. Wolframite and scheelite coexist closely with fluorapatite and fluorite, respectively, indicating that the volatile-rich fluids at Dahutang were also rich in tungsten, fluorine and phosphorous. Tungsten was likely transported as tungstate species (e.g., WO42− and HWO4−), oxygen fluorine complexes (e.g., [WO2F4]2−, [WO3F2]2−) and phosphorous heteropolytungstate (e.g., [P(W12O40)]3−) in the fluids. Fluorapatite and wolframite may have precipitated first when the temperature dropped from 400° to 320 °C (along with pH increase) at an estimated depth of 7.8 km. This was likely followed by the extensive scheelite mineralization (with fluorite precipitation) that formed huge disseminated/veinlet-type W orebodies when the temperature further dropped to 200 °C (along with pH increase). After the W mineralization, extensive Cu–Mo polymetallic mineralization and the associated sericite-chlorite alterations may have formed by the granite porphyry and/or muscovite granite emplacement. The late ore-forming fluids belong to a medium to high temperature (200°–360 °C), low salinity H2O–NaCl–CaCl2 system. Molybdenum was likely transported mainly as Mo complexes (e.g., H2MoO4/MoO42−), and copper as Cu–Cl complexes. Molybdenite and chalcopyrite may have begun to precipitate when the fluids ascended and the temperature dropped to 330 °C. Meteoric water incursion may have then occurred (δ18Ofluid = 1.9–6.9‰ and δD = −99 to −68‰) and further cooled the fluid system to 250 °C, forming large Cu–Mo orebodies.

Adakitic rocks with high MgO content and lower crust affinity at convergent margins provide important constraints on arc crust evolution and regional tectonics. An integrated study on the only adakitic pluton with high MgO content and its mafic microgranular enclaves (MMEs) in the Early Paleozoic North Qinling arc terrane was conducted in this paper. Zircons from 2 host rock samples yielded identical weighted mean ages of 428 Ma. The MMEs contain two types of zircons: type 1 zircons share similar morphology, chemical composition, and indistinguishable age with those from the host rock; type 2 zircons yielded an older age of 450 Ma. The similarity in zircon Hf isotopic composition for these zircons indicates a cognate origin of the MMEs. However, the lower Ce4+/Ce3+, higher crystallization temperature and the LREE overabundance of type 2 zircons indicate that mafic intrusions in the lower arc crust were subjected to a post-magmatic event at similar to 450 Ma. Samples from the Tangzang pluton are characterized by typical adakitic geochemical features, and high MgO (Mg#: 50-52), Cr, Ni, and Ba contents. Considering their similar isotopic compositions with the Neoproterozoic metabasites in the North Qinling terrane, we argue that the Tangzang pluton was generated by crustal anatexis of two sources with contributions from the metasomatized mantle wedge. The mantle connection was achieved by aqueous melt-fluxed melting of a "hot zone'' in the lower arc crust. This implies that andesitic continental crust with Mg# of similar to 0.5 may also be associated with fluxed-melting of the lower arc crust.

Numerous Late Cretaceous W–Sn deposits associated with granitic rocks were developed in the southern part of South China. However, the relationship among granitic magma, mineralization and tectonic setting remains disputed. The Yingwuling W–Sn deposit is spatially associated with K-feldspar granite and biotite granite in the southwestern part of Guangdong province, South China. LA-ICP-MS zircon U–Pb dating of K-feldspar granite, biotite granite and weathered ore-bearing granites yielded ages of ∼79 Ma, which agree well with the LA-ICP-MS cassiterite U–Pb age of 79.6 ± 0.9 Ma, indicating a close genetic link between the Yingwuling pluton and W–Sn mineralization. Geochemically, both K-feldspar granite and biotite granite show high SiO2, Na2O + K2O and low CaO, P2O5 contents, and they belong to shoshonitic series and high-K calc-alkaline series, respectively. High Rb/Sr ratios and low Nb/Ta and Zr/Hf ratios of the Yingwuling pluton indicate strong magmatic differentiation. Compared with K-feldspar granite, biotite granite experienced a higher degree of fractional crystallization as indicated by lower Fe2O3, TiO2, LREE, Sr, Eu, Ba and Zr contents and higher SiO2, HREE, Rb, Cs and Y contents. Their Sr and Nd isotope compositions (initial 87Sr/86Sr = 0.7027 and 0.7097; eNd(t) = −5.6 and −5.7) are less radiogenic than that of the South China basement. Thus, it is likely that some juvenile materials were involved in the formation of these granitic intrusions. This is also favored by the wide range of zircon Hf isotope values (eHf(t) = −6.6 to 0.8), which indicates mixing processes. A-type granite affinity of the Yingwuling pluton shows that it was possibly formed in an extensional setting. Late Cretaceous W–Sn deposits constitute an EW-trending belt from Guangdong to Yunnan provinces, which was compatible with the northward subduction of the Neo-Tethys plate rather than the northwestward subduction of the Pacific plate during Late Cretaceous. Combined with previous studies on magmatic rocks and tectonic activities during Late Cretaceous in South China, this study suggests that the subduction of the Neo-Tethys plate was most possibly responsible for the Late Cretaceous Yingwuling granitic intrusion and associated W–Sn mineralization.

Apatite is ubiquitous in A-type granites, and can be used to elucidate the volatile contents of the silicate melt, which reflect its source characteristics. A-type granites have been recognized as a distinct group of granites. A1- and A2-type subgroups are produced under different extensional settings. However, the details of the mechanisms behind the distinctive geochemical characteristics of A1- and A2-type granites remain obscure. Belts of Cretaceous A1- and A2-type granites occur along the Lower Yangtze River Belt in eastern China. Here we investigated the major and trace element compositions of apatites from contemporary A1- and A2-type granites at different localities along the Lower Yangtze River Belt, in order to decipher their discrepant source processes. Apatites from A1- and A2-type granites show similar major and trace elements, but differ in their F and Cl concentrations. Apatites from A1-type granites in the eastern part of the Lower Yangtze River Belt have much lower F and higher Cl concentrations compared to A2-type granites in the western part. Moreover, from the east to the west, the F concentrations of apatites from A1-type granites increase, while the Cl concentrations decline. In a subducted plate, F is retained by amphibole, chlorite, serpentine and mica minerals through the amphibolite stage, and finally by phengite and lawsonite during the eclogite stage, whereas, Cl is controlled by amphibole, chlorite and serpentine. The high and varied Cl concentrations in A1 subgroup apatites, therefore, may be attributed to the breakdown of amphibole, chlorite and/or serpentine decomposition during partial melting of subducted oceanic crust releasing a large amount of Cl at shallower depth. In contrast, F is transported to deeper depths in the subducted oceanic crust, and released through breakdown of phengite and lawsonite, making an important contribution to the formation of A2-type granites. Apatites from A1- and A2-type granite samples show regular changes in LREE/HREE, LREE/MREE and MREE/HREE ratios with increasing distance from the location of the subduction zone, probably as the result of nonsynchronous dehydration of the serpentine and phengite at different stages during subduction. We propose that A1- and A2-type granites on the Lower Yangtze River Belt were derived from sources metasomatised by fluids originating from the breakdown of amphibole, chlorite and/or serpentine with higher Cl, lower F, and from phengite and/or lawsonite with relatively higher F but lower Cl, respectively.

The connection between prolonged granitic magmatism and the formation of giant tungsten (W) polymetallic deposits has long been disputed. In this study, we present 6 mica Ar-Ar plateau ages and 22 molybdenite Re-Os model ages data on the newly discovered giant Dahutang W-Cu-Mo deposit in South China, which is one of the largest W deposits in the world. New and published zircon U-Pb, mica Ar-Ar, and molybdenite Re-Os age data reveal that the Mesozoic Dahutang magmatism and mineralization occurred in two major periods: (1) the Late Jurassic (ca. 153-147 Ma), forming the hydrothermal breccia, large wolframite-bearing quartz vein, and scheelite-dominated disseminated/veinlet type orebodies, which is mainly associated with the emplacement of porphyritic biotite granite; (2) the Late Jurassic to Early Cretaceous (ca. 146-130 Ma), forming the Cu-Mo-W +/- Sn mineralization overprinting the Late Jurassic W-Mo +/- Cu orebodies, which is mainly related to the successively emplacement of the Early Cretaceous granites. We suggest that continuous accumulation of mineralization for a long period of time (151-130 Ma) have contributed to the formation of the giant Dahutang deposit.

The Dachang tin-dominant polymetallic district in western South China, hosts a total reserve of 110 Mt ore @ 1% Sn, 3–5% by-products of Zn, Pb, Sb, In and Cu, and 100 to 300 g/t Ag. Economic tin orebodies in the district mainly developed in Devonian limestone, black shale and siliceous rocks, as massive/stratiform and vein/stockwork mineralization. LA-ICP-MS U–Pb dating on seven cassiterite samples from the Tongkeng-Changpo and Gaofeng Sn–Zn–Pb deposits, and the Dafulou, Huile and Kangma Sn–Zn deposits, yields weighted mean 206Pb/238U ages ranging from 90.3 ± 1.8 Ma to 95.4 ± 4.9 Ma. These dates are significantly younger than the Devonian ore-hosted strata, but agree well with the previously published zircon U–Pb ages of 96.6–93.9 Ma for the Longxianggai biotite granite within analytical uncertainties, showing a close temporal link between the granitic magmatism and tin mineralization. Cassiterite grains from various ore types in the Dachang district show clear growth zonations as reflected from their cathodoluminescence images. LA-ICP-MS trace element analyses of cassiterite grains indicate that they have relatively high Ti, Fe and W but low Nb and Ta contents, which corresponds to the clear luminescencing. Chemical variations of cassiterite from different ore types, i.e. cassiterite-sulfide ores and cassiterite-bearing quartz veins, are indistinguishable, except for more enriched Fe in cassiterite from cassiterite-quartz veins. Combined with Cretaceous U–Pb ages, elevated Fe and W contents of cassiterite suggest tin-dominant polymetallic deposits in the Dachang district were formed in a granite-related magmatic-hydrothermal system controlled by the Longxianggai biotite granite, precluding their syngenetic origin.

Paleoproterozoic A-type granites are widely outcropped in the North China Craton (NCC), particularly in the Trans-North China Orogen. However, their genesis and tectonic significance remain obscure. Here we report systematic studies on geochronology and geochemical characteristics of A-type granite in Huayangchuan, south margin of the NCC. The samples are enriched in total alkali (K2O + Na2O > 8.97 wt%), and depleted in MgO (0.84–0.93 wt%), CaO (1.28–1.90 wt%) and P2O5 (0.18–0.20 wt%), with high FeOT/MgO (5.69–6.67). They are characterized by high Zr + Y + Nb + Ce values (1293–1392 ppm) and 10,000 × Ga/Al ratios (3.14–3.35), which are typical characteristics of A-type granite. The Huayangchuan A-type granite can be further classified as A1-type subgroup based on particular geochemical features, e.g., low Y/Nb (0.87–1.00) and Yb/Ta (0.88–1.10). High precision zircon U–Pb dating of the A-type granite by secondary ion mass spectrometry (SIMS) yields Paleoproterozoic 207Pb/206Pb ages of 1829.5 ± 2.5 Ma. The low zircon ɛHf(t) values (−6.97 to −10.45), along with zircon Hf model age of 2.7–2.9 Ga, indicate that the Huayangchuan A-type granite was derived from partial melting of the ancient continental crust with contribution of enriched mantle components. The low zircon δ18O composition (4.00 to 6.78‰) indicates that the zircons were crystallized from low δ18O magmas, which derived from the crust metasomatized by low δ18O mantle fluids or melts. The E-W trend A1-type granitic plutons in the NCC are generally outcropped in a rift tectonic regime, which is consistent with the development of the mantle plume in the Xiong'er district. The large volume of basaltic magmas, generated by mantle plume head, underplated the lower continental crust and formed the Huayangchuan A-type granite.

Serpentinites are important volatile and fluid mobile element repositories in oceanic lithosphere and subduction zones, and thus provide significant constraints on global geochemical cycles and tectonic evolution at convergent margins. In this contribution, two types of serpentinites from the Mianlue suture zone in the Qinling orogenic belt, central China, are identified on the basis of detailed mineralogical and geochemical study. Serpentinites from the Jianchaling region (Group 1) are composed of lizardite/chrysotile + magnesite + magnetite. Most of these serpentinites (Group 1a), consist of pseudomorphic orthopyroxene and olivine, and are characterized by low Al2O3/SiO2, high MgO/SiO2 and Ir-type PGEs to Pt ratios, suggesting a residual mantle origin. Meanwhile, the U-shape REE pattern and positive Eu, Sr and Ba anomalies of these serpentinites indicate that serpentinization fluids have interacted with gabbroic cumulates at moderately high temperatures or associate with the chlorinity and redox conditions of the fluid. Considering the limited mobility of U in the hydrating fluids for the Group 1a serpentinites, hydrating fluids for these serpentinites are most likely derived from the dehydrated slab, and have been in equilibrium with subducting sediments. There are also some serpentinites with low-grade metamorphic recrystallization from the Jianchaling region (Group 1b), represented by recrystallized serpentine minerals (antigorite). The trace element compositions of these Group 1b serpentinites suggest that partial dehydration of serpentinites associated with the transformation from lizardite to antigorite in subduction zone is also likely to affect the geochemistry of serpentinites. Serpentinites from the Liangyazi region (Group 2) are composed of antigorite + dolomite + spinel + magnetite. The high Cr number (0.65–0.80) and low Ti concentrations of spinels in Group 2 serpentinites indicate a refractory mantle wedge origin. Fertile major element compositions (e.g., high Al2O3 content and Al2O3/SiO2) and conjoint enrichment in light rare earth elements and high field strength elements, however, suggest melt-rock interactions before serpentinization. Combined with their geochemical affinity to “subducted serpentinites”, we conclude that their protoliths (refractory mantle wedge peridotite) experienced melt-rock interactions and then were incorporated into the subduction channel before serpentinization. Studies on these two types of serpentinites indicate that serpentinites from the orogenic belt are most likely characterized by multi-source, multi-stage and multi-genesis, further providing important constraints on subduction channel processes.