Marine silicate alteration is tightly linked to carbon burial over geological time scales, especially in continental margins known as ‘hotspot’ due to active organic carbon-driven diagenetic processes. Volcanic materials are one of the most reactive silicate groups, which plays an important role in sedimentary Si processes in continental margins. Since the Last Glacial Maximum, continental margins have undergone strong sediment regime shift, including sediment source changes, due to sea level rise. However, how marine silicate alterations respond to such sediment regime shifts remain largely unexplored. This study uses a 155-meter-long sediment core (IODP Exp. 375 U1518) from the Hikurangi margin, eastern New Zealand, one of the most representative sites in the subduction zone that are enriched with volcanic materials and meanwhile suffers from drastic sea level rise over a hundred of meters in the past 20,000 years. A tailored Si isotope-based sequential chemical leaching protocol in combination with porewater chemistry and geochemical analysis enable us to interrogate the sediment core in great details and uncover sedimentary Si mass balance by constraining various Si endmembers and major Si geochemical processes. Our results show that the δ30Si values of Si endmembers can vary by 1‰ within 1.5–155.7 mbsf due to sediment source changes and early diagenetic modifications. This suggests that conventionally fixed δ30Si value of a single Si endmember through one sediment core may result in large uncertainty in constraining sedimentary Si processes. A potential organic-bounded Si phase with a δ30Si value of −0.19 ± 0.21‰ was also observed in this study. Simulations of a reactive-transport model further reveal enhanced clay mineral dissolution in the shallow Holocene sediments, leading to rapid porewater dissolved Si (DSi) increase and δ30SiDSi value decrease. Compared with the Pleistocene sediments, these changes directly elevate DSi diffusive flux by an order of magnitude with a lower δ30Si value toward seafloor. Meanwhile, slow incongruent volcanic material dissolution in the deep Pleistocene sediments drives the gentle increase of DSi concentrations and δ30SiDSi values, from 541 μmol/L and + 0.24 ± 0.16‰ at 37.10 mbsf to 686 μmol/L and +0.94 ± 0.23‰ at 155.70 mbsf, respectively. High vs. low silicate alteration rates between shallow and deep sediments generates a large DSi concentration gradient and downward diffusion. We propose that it is necessary to include Si phase-based isotope analysis for better constraining marine silicate alterations and its mass balance, particularly for those environments subjecting to sediment source changes. Our findings also highlight that simply attributing decrease of DSi concentration to authigenic clay precipitation (reverse weathering) may cause an overestimation of its role in regulating marine carbon cycle during glacial-interglacial transgressions.

To investigate the influence of carbon dioxide (CO2) on hydrogen (H-2) generation and reaction rates during serpentinization, we performed experiments at 300 degrees C and 2.0-3.4 kbar by reacting peridotite and olivine (25-50 mu m of grain sizes) individually with NaCl-KHCO3 solutions. These included 0.5 M NaCl combined with 0.1 M, 0.5 M, and 0.9 M KHCO3, corresponding to a range of CO2 concentrations. Hydrogen was quantified by gas chromatography, and serpentinization extent was determined using standard curves based on infrared spectroscopy. The results demonstrate that KHCO3 concentration strongly influences H-2 generation and serpentinization kinetics. In peridotite experiments, 0.1 M KHCO3 reduced H-2 production and slowed the serpentinization rate, whereas higher KHCO3 concentrations (0.5 and 0.9 M) enhanced serpentinization. Despite the accelerated reaction, H-2 yield decreased slightly at 0.5 M KHCO3 and dropped by an order of magnitude at 0.9 M KHCO3, with minor carbonates (<5 wt.%) observed after prolonged durations. In olivine experiments, CO2 consistently suppressed the reaction rates and H-2 formation, and olivine reacted more slowly and produced significantly less H-2 than peridotite in CO2-bearing fluids. This study indicates that low CO2 concentrations retard serpentinization in both peridotite and olivine, whereas higher CO2 levels promote the reaction specifically in peridotite while suppressing H-2 generation. These findings suggest that the CO2 content of geological fluids is a critical factor controlling H-2 generation that is a key energy source for deep subsurface life.

Under global warming, an intensified ocean water cycle has led to pronounced changes in sea surface salinity (SSS). However, the response of South China Sea (SCS) SSS to climate warming prior to the satellite era before the 1980s remains poorly understood due to the scarcity of instrumental observation. Here, we reconstruct a 158-year (1851–2008) monthly resolved SSS record for the northern SCS using a Porites coral δ18O record from Yongxing Island. The reconstruction is validated through instrumental calibration and exhibits coherent variability with additional coral δ18O records from the northern SCS. The record indicates a long-term freshening trend of −0.16 ± 0.04 psu per century (1σ) since 1851 to 2008. Local freshwater forcing alone cannot adequately explain this freshening. Comparison with Luzon Strait transport suggests that reduced Kuroshio-derived salt input provides a physically consistent explanation for this reconstructed freshening tendency. On interannual to decadal timescales, northern SCS SSS variability is closely linked to the El Niño–Southern Oscillation (ENSO) and the Pacific Decadal Oscillation. However, over the length of our reconstruction, their coupling exhibits significant non-stationarity on multi-decadal to low-frequency timescales, with sign reversals in correlation occurring around 1880s and after 1960s. Given the 158-year span of the record, we frame these findings as indicative of low-frequency variability rather than a complete multi-centennial cycle. This non-stationarity cautions against assuming a time-invariant ENSO/PDO–SSS relationship over multi-decadal and longer periods. Our study confirms that coral δ18O can reliably extend SSS records into the pre-instrumental era and provides a critical baseline for distinguishing natural variability from anthropogenically forced salinity changes in the warming SCS.

Sedimentary biogenic silica (bSi) is a primary sink of oceanic silicon (Si), and its quantification is essential for constraining the marine Si budget. The traditional alkaline leaching method estimates bSi contents by extrapolating a regression of Si leachates at 2, 3, and 5 h, but this approach is biased by lithogenic Si (LSi) dissolution. We analyzed 59 marine sediment samples (0.2 to 57.0 wt% bSi) and used Si isotopes to re-assess this method. Samples with moderate bSi contents (5-20 wt%) show a marked decline in dissolution rates after 3-5 h of leaching. Si isotopes of the leachates reveal that the decline reflects either a shift from dominant bSi to LSi phase or changes in bSi species through time. Sediments with <5 wt% or > 20 wt% bSi display a stable Si dissolution rate throughout 8 h, likely due to unchanged dominant Si phases of lithogenic or biogenic Si. Additionally, we re-assessed the brucite co-precipitation method (MAGIC) for porewater dissolved Si (dSi) recovery. Mg/Si molar ratios <100 yield dSi recovery <90%, resulting in large isotope fractionation. Hence, an Mg/Si ratio of >= 300 is recommended to guarantee full recovery, and H2O2 pretreatment is necessary to eliminate matrix effects induced by high Fe, Mn and DOC concentrations. Our findings imply that the traditional leaching method likely underestimates current bSi contents, especially in clay or/and organic-rich sediments with moderate bSi contents.

Plate subduction transports crustal materials (e.g., oceanic crust) into the deep mantle, generating mantle compositional heterogeneities. However, the transformation and long-term evolution of these materials under deep mantle conditions remain poorly understood. In this study, we present molybdenum (Mo) isotope data of well-characterized alkaline basalts from Madeira Island in the eastern North Atlantic. Based on their stratigraphic relationships and eruption ages, the samples are divided into shield stage (5–0.7 Ma) and post-erosional stage (<0.7 Ma) groups. The Mo isotopic compositions (δ98/95Mo = -0.50‰ to -0.22‰, relative to NIST SRM3134) of the shield stage lavas are lighter than those of mid-ocean ridge basalts (MORB; δ98/95Mo = -0.19 ± 0.01‰), and the post-erosional stage lavas have δ98/95Mo values (-0.31‰ to -0.17‰) similar to those of MORB. The Mo isotopes in both lavas show good correlation with radiogenic Sr-Pb isotope ratios. Besides, the samples also exhibit higher Ce/Mo ratios (average ∼45.42) compared to MORB (average ∼37.18). Notably, δ98/95Mo values correlate with eruption ages: the shield stage lavas have low δ98/95Mo values, whereas the post-erosional stage lavas show relatively higher values. These geochemical variations are interpreted to reflect differing contributions from distinct portions of recycled oceanic lithosphere, with a minor sedimentary input. Specifically, the shield stage magmas were affected by a recycled hydrothermally altered upper basaltic section (eclogitized, with high 87Sr/86Sr and U/Pb values, and low δ98/95Mo values), whereas the post-erosional stage magmas reflect lower ultramafic components (gabbro) with mantle-like Sr-Mo isotopes and U/Pb values. The light Mo isotope signatures of the Madeira hotspot alkaline basalts thus confirm the importance of recycled oceanic crust in the genesis of alkaline ocean island basalts (OIBs), highlighting that Mo isotopes can be used to trace subduction processes.

The Oceanic Anoxic Event 1a (OAE 1a) represents one of the most significant environmental perturbations during the Cretaceous. During OAE 1a, changes in climate and ocean redox conditions exhibit noteworthy spatiotemporal heterogeneity. However, environmental changes in the Eastern Tethys during OAE 1a remain poorly constrained. In this study, high-resolution geochemical data of organic carbon isotope (δ13Corg), redox-sensitive trace elements (RSTEs), and total sulfur (TS) were used to correlate the OAE 1a interval of two sections in the Eastern Tethys: the Gucuo II section in Southern Tibet and the Ocean Drilling Program (ODP) Site 765C in the Argo Abyssal Plain. Paleoredox and paleoclimate proxies of these sections exhibit similar trends across OAE 1a, indicating synchronized environmental changes. RSTEs and TS contents increase before OAE 1a, remain elevated during OAE 1a, and decrease to near background values after OAE 1a, indicating preceding deoxygenation and prolonged reoxygenation in the Eastern Tethys. Weathering proxies, including the chemical index of alteration (CIA), chemical index of weathering (CIW), and Ti/Na ratios, exhibit a decreasing trend beginning before the onset of OAE 1a and continuing into early OAE 1a, followed by an increasing trend during late OAE 1a, indicating a delayed enhancement in weathering in the Eastern Tethys. Regionally, the delayed weathering enhancement might have resulted from dry climate and/or transient cooling episodes during early OAE 1a. This study elucidated the regional environmental response of the Eastern Tethys to OAE 1a, contributing to a more comprehensive understanding of the initiation and cessation mechanisms of OAE 1a.

Porphyry-type deposits are widely recognized as major contributors to global Cu resources and are traditionally linked to subduction-related magmatic arcs. Recent exploration has revealed that substantial porphyry mineralization can also occur in post-collisional tectonic environments, but the ore enrichment mechanism in such settings remain insufficiently understood. Deciphering the roles of magma source, fluid evolution, and metal transport remains critical for understanding their ore fertility. Here we present whole-rock Zn isotope data and Cl concentrations for barren and mineralized rocks from the post-collisional Qulong porphyry Cu deposit in the Gangdese belt, southern Tibet. The barren host rocks and associated mafic microgranular enclaves (MMEs) display systematically elevated delta 66Zn values relative to mantle-derived magmas, which cannot be explained by magmatic differentiation or crustal assimilation. Instead, these heavy Zn isotopic signatures are best interpreted to reflect incorporation of recycled carbonate-rich sediments into their magma sources, resulting in magma oxidation and volatile enrichment to enhance ore fertility in post-collisional settings. In contrast, mineralized porphyries exhibit significantly lower delta 66Zn values and variational Zn concentrations, coupled with elevated K2O, Rb/Ba, and Pb/Ce ratios, indicating extensive interaction with exsolved magmatic fluids. We propose that exsolved fluids preferentially extracted isotopically light Zn from the melts, generating metal-rich hydrothermal fluids with high mineralization potential. Subsequent fluid-rock interaction and sulfide precipitation induced kinetic Zn isotope fractionation, progressively removing light Zn into sulfides, driving remaining fluids toward heavier Zn isotopic compositions, and rapidly diminishing Cu-carrying capacity. Our results demonstrate that Zn isotopes sensitively record both magma source composition and fluid evolution, providing novel insights into metal transport mechanisms in post-collisional porphyry systems.

Carbonatite-associated rare earth element (REE) deposits are currently the primary source of REE resources. Their formation requires REE enrichment during prolonged magma evolution, achieved by suppressing REE-rich mineral crystallization and promoting REE-enriched brine melt formation. Our experiments on the fractional crystallization of carbonatitic magmas indicate that pressure (emplacement depth) is the primary factor controlling REE enrichment. High-pressure ( >0.3 GPa) promotes early olivine crystallization, depleting silica and suppressing REE-rich apatite formation. Deep emplacement also delays aqueous fluid exsolution, thereby stabilizing brine melts that enhance phosphate dissolution and prevent REE dispersion into apatite. In contrast, low-pressure conditions ( <0.3 GPa) lead to exsolution of REE-poor hydrothermal fluids, dispersing REE into magmatic apatite and preventing the deposition of economically significant REE-carbonates in subsequent hydrothermal stages. Our pressure-dependent model highlights deep emplacement as crucial for passive REE enrichment in residual brine melts, driving large-scale mineralization through precipitation of burbankite and/or bastnäsite.

The Philippine Mobile Belt records a complex Cenozoic subduction history, yet the subduction initiation processes during the Cenozoic and the characteristics of subsequent magmatic evolution remain poorly constrained. Here, we present the first integrated petrological, geochronological, and geochemical study of Eocene-Oligocene intrusive rocks from Bohol Island, Central Philippines, including pyroxene diorite (44-41 Ma), medium-grained quartz diorite (39 Ma), and coarse-grained quartz diorite (35 Ma). The early pyroxene diorites are geochemically highly depleted with N-MORB-like REE distribution patterns but much lower REE contents, which resemble the geochemical characteristics of forearc basalts that are formed during the initial subduction stage. Furthermore, the enrichment of fluid-mobile elements such as high Pb, Sr, Rb, Ba, U, and K of these rocks suggests that more slab-derived fluids were added in the mantle source than those of forearc basalts. Combining the depleted Nd isotopes (8Nd(t) = 8.28-9.16) but relatively enriched Sr isotopes (87Sr/86Sri = 0.703366-0.704565), we suggest the mafic intrusive rocks were formed by partial melting of a depleted mantle with the contribution of fluids from the altered oceanic crust during the early-stage subduction after subduction initiation. Furthermore, whole-rock Pb (208Pb/204Pbi = 37.9770-38.2556; 207Pb/204Pbi = 15.4824-15.5659; 206Pb/204Pbi = 18.4860-18.5505) and Nd isotopes, and magmatic zircon Hf (8Hf(t) = 10.7-18.8) isotopic compositions of these intrusive rocks reveal the contribution from fluids released by the subducted Pacific-type oceanic crust and partial melting of an Indiantype MORB mantle source. In contrast, the ca. 39-35 Ma quartz diorite suites are more felsic, showing more evolved geochemical characteristics, e.g., fractionated REE patterns and elevated Th contents, suggesting that more slab melt is added into the mantle. This indicates that the tectonic regime evolved from an incipient subduction to a more mature stage during this period. This study identified the magmatic records from incipient subduction to mature subduction in the Philippine island arc during the Cenozoic, providing further petrological and geochemical constraints on the Cenozoic subduction initiation and subsequent tectonic evolution on the western margin of the Philippine Sea Plate.

During the early Cenozoic greenhouse period, counterintuitive contractions in tropical Pacific oxygen-deficient zones have been linked to enhanced deep-ocean ventilation, yet direct geological evidence remains limited. Here we present molybdenum (Mo) isotopic records from International Ocean Discovery Program Site U1509. Raman spectroscopy shows that Mn and Fe occur mainly as MnO2 and Fe-Mn oxides. The delta 98/95Mo values show a sustained negative shift from similar to 1 parts per thousand to -2 parts per thousand between similar to 59.5 and 50.5 Ma, and this trend is best explained by strengthened Mo adsorption onto Mn and Fe bearing oxides at the sediment-water interface. This mechanism provides a direct basis for interpreting the Mo isotope depletion, indicating prolonged bottom-water oxygenation within early Cenozoic greenhouse conditions. At similar to 50.5 Ma, a sharp decline in silicate epsilon Nd values and coeval rises in Ti, Zr, and Sc enrichment factors indicate intensified continental weathering that supplied isotopically light Mo. These results provide new constraints on early Cenozoic redox evolution.

Hydration of the lithospheric mantle is a critical part of deep water cycling and is essential in shaping Earth's habitable environments. At low water/rock ratios, water is stored in nominally anhydrous minerals or hydrous minerals; however, at high water/rock ratios, the leaching effect of water on mantle minerals remains poorly understood. Here, we present hydrothermal experiments of the 'phlogopite + H2O' system at 500 degrees C-950 degrees C and 0.9-1.8 GPa, showing that incongruent dissolution of phlogopite in water produces corundum (Al2O3) as an insoluble residual phase at pressures above similar to 1.5 GPa. Water leaches soluble components (K, Mg and Si) from phlogopite, leaving corundum as the residue under lithospheric mantle P-T conditions. This provides a mechanistic explanation for sapphire formation in the lithospheric mantle, consistent with natural sapphire gemstones occurring as mantle xenoliths and xenocrysts within alkaline basalts. Sapphires form through hydrous leaching of phlogopite at depths greater than similar to 50 km and are subsequently transported to the surface by alkaline basaltic magmas originating from at least similar to 70 km depth. The widespread sapphire deposits associated with alkaline basalts in East China indicate extensive hydrous leaching of the lithospheric mantle, with water ultimately sourced from the subducted slab, mantle transition zone or mantle plumes.

Hydration and complexation are crucial processes for dissolving metal elements and transporting metal complexes in hydrothermal fluids. However, the impact of hydration and complexation on the transport and enrichment of metal elements, such as those involving rhodium and chloride, has not been thoroughly evaluated. Here, this study employed a hydrolysis experiment of K3RhCl6 at 200-600 degrees C and 100 MPa to determine the controlling factors and thermodynamics of Rh-Cl complexes in chloride-rich fluids. The results show that the dominant Rh-Cl complex is RhCl3-6at 200-400 degrees C, gradually converting into Rh(III)-OH-Cl complexes over 400 degrees C. The hydrolysis equilibrium constant (LnK) of RhCl3-6 at 200-400 degrees C affected by temperature (T (K)) is calculated as: lnK 49 06 7 01 48802 3896 0 T Accordingly, the DrHmH and DrSHm of the hydrolysis reaction were obtained to be 405.8 +/- 32.39 kJ mol-1 and 407.9 +/- 58.30 J mol-1 K-1, respectively. Thermodynamic parameters reveal the dependence of the stability of Rh-Cl complexes on temperature in chloride-rich fluids. For instance, the formation constants (lnb) of RhCl3-6 vary from 0.0184 +/- 0.0022 to-0.0079 +/- 0.0012 as the temperature rises from 150 to 400 degrees C. Geochemical modeling illustrates that low-temperature and acidic fluids can enhance the stability of Rh-Cl complexes, which can be dominated by Cl concentration (over 0.5 wt.%). Hydrothermal fluids with low pH and high Cl content, typically occurring in the mid-ocean ridge, promote Rh transport and subsequent enrichment in encrustations and minerals by substituting Mn and Fe for isomorphism, or in the form of alloys, forming a substantial Rh reservoir in the ocean. (c) 2026 China University of Geosciences (Beijing) and Peking University. Published by Elsevier B.V. on behalf of China University of Geosciences (Beijing). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

The redistribution and oxidation of iron during serpentinization are critical for quantifying hydrogen production and redox fluxes in subduction zones. This study investigates serpentinites from the Heimulin area in the Qinling Orogen, central China, to constrain their evolution and identify the factors controlling Fe3+ incorporation into serpentine minerals and associated H2 generation. Whole-rock geochemical compositions and the chemistry of relict primitive Cr-spinel domains indicate that the protoliths were highly refractory forearc mantle peridotites. The enrichment patterns of fluid-mobile elements closely resemble those of forearc serpentinites, such as those from the Mariana forearc, suggesting the protolith was hydrated by slab-derived fluids in a shallow forearc setting. Mineralogical analyses identify four distinct types of lizardite, reflecting the sequence of serpentinization. Initial low-temperature serpentinization produced some Type-1 lizardite grains, in which Fe3+ was incorporated into the crystal lattice through dioctahedral substitution. When silica activity became sufficiently low, hydration of olivine produced Fe-rich brucite and additional Type-1 lizardite and some Type-2 lizardite grains that lack detectable Fe3+ substitution. As serpentinization progressed, local increases in silica activity, driven by mineralogical heterogeneity (e.g., orthopyroxene dissolution) or influx of SiO2-rich external fluids, induced the breakdown of Fe-rich brucite and promoted the formation of fine-grained magnetite together with additional Type-2 lizardite grains, in which Fe3+ was incorporated into the crystal lattice through cronstedtite substitution. Since the major episode of Fe oxidation occurs during the interaction between brucite and SiO2-rich fluids, and given the widespread occurrence of cronstedtite-type substitution in natural lizardite, the formation of Fe3+-serpentine with tetrahedral Fe3+ may therefore mark the principal stage of hydrogen generation. During subsequent recrystallization, reduced Fe3+ availability favored the transformation of lizardite to chrysotile. Simultaneously, silica-poor conditions promoted the preservation of brucite, which aggregated and recrystallized to form economically significant fibrous brucite deposits. The chemical compositions of lizardite in recrystallized serpentinites, together with variations in bulk-rock Fe3+/ΣFe ratios, further indicate that Fe3+ was redistributed during recrystallization. Overall, low-temperature serpentinization of refractory peridotites under SiO2-poor conditions, followed by recrystallization, creates favorable conditions for the formation of economically valuable chrysotile asbestos and fibrous brucite deposits.

To reveal the fluid variations in the mantle sources of arc lavas from a fluid-dominated subduction zone and to further constrain its implications for global potassium (K) recycling, we present the K isotope compositions of thirty-two arc lavas from a 220 km arc-front to back arc transect across the Kamchatka Peninsula. Our results show that arc lavas from Eastern Volcanic Front (EVF, -0.48 f 0.02 %o to -0.19 f 0.03 %o, average at -0.37 f 0.08 %o, n = 15) and Central Kamchatka Depression (CKD, -0.39 f 0.03 %o to -0.25 f 0.05 %o, average at -0.32 f 0.05 %o, n = 9) display slightly higher delta 41K than those from Sredinny Range (SR, -0.44 f 0.03 %o to -0.37 f 0.02 %o, average at -0.41 f 0.03 %o, n = 5) and the mantle value (-0.42 f 0.08 %o), which indicate the incorporation of slab-derived fluids within their mantle sources. Notably, the delta 41K values of CKD arc lavas show positive correlations with delta 18O, and fluid indices (e.g., K/Nb, Sb/Ce, Pb/Ce and As/Ce ratios), suggesting the contributions of fluids derived from the subducted Hawaii-Emperor Seamount Chain. In addition, this study reveals that arc lavas from North Central Kamchatka Depression (NCKD) have higher delta 41K values (-0.32 f 0.03 %o to -0.31 f 0.02 %o, average at -0.32 f 0.01 %o, n = 3) than the mantle value. Their high delta 41K values were accompanied with high Sr/Y ratios, indicating that the K isotope data of NCKD arc lavas may reflect the influence of slab-derived melts. This study demonstrates that recycling of a slab-derived fluid components can modify the K isotope compositions in the mantle source of the arc magmas. Consequently, when combined with other indicators (e.g., Sb/Ce, As/Ce ratios and Pb-O isotopes), the K isotope system helps to constrain recycling of fluidmobile elements in subduction settings and to reveal how crustal materials recycle back via dehydration fluids from the subducted slab to the Earth's surface.