Abstract Surface meltwater that penetrates glacier interiors to the beds can warm the ice and enhance basal mass loss, exerting critical control over glacier dynamics. However, these processes remain poorly constrained in Tibetan Plateau glaciers, particularly those in the western sector, which have long been regarded as cold-based. Here, we used pollen radiocarbon ages from sediments in seven proglacial lakes, spanning catchments with 0.5–55.0% glacier cover, to distinguish old terminal and/or basal ice melt from young surface snow/ice melt of upstream glaciers previously assumed to be cold-based (four lakes) and polythermal (three lakes) over the past 150 years. We show that after the 1990 s, the melting of Tibetan glaciers shifted into a new regime, marked by accelerated mass loss of old basal ice under intensified surface melt conditions. These findings underscore the urgent need for integrated observations of both supra- and subglacial processes to resolve the evolving hydrothermal regimes in Tibetan glaciers under ongoing climate warming.
Abstract The delivery and evolution of carbonaceous materials to airless planetary bodies remain poorly constrained. This study characterized a porous, aggregated, carbon-rich dust particle (CE6-CDP) from Chang’e-6 lunar samples. CE6-CDP yielded a mean δ13CVPDB value of − 25.92 ± 3.62‰ and exhibited significant nitrogen enrichment. Nano-infrared spectroscopy further detected N–H, C = N, and C–H bonds, indicating that the carbonaceous material within CE6-CDP shows organic-like signatures. Graphitic carbon with tangled and concentric textures within CE6-CDP exhibited a comparable δ13CVPDB value (−27.31‰) but contained low nitrogen, suggesting possible graphitization of some organic precursors. The coexistence of organic-like carbonaceous material and graphitic carbon is consistent with dust ejection during an impact involving a larger carbon-rich projectile. This study indicates that extraterrestrial carbon may constitute an important lunar carbon reservoir and contribute to the dark surfaces of Mercury and Vesta. These findings provide new insights into carbon transport and evolution across airless planetary bodies.
Abstract Fluorescence imaging enables noninvasive visualization of mitochondrial dynamics, yet conventional fluorogenic probes are often compromised by high background fluorescence and limited target specificity. Herein, we report a molecular scaffold based on multichannel barrierless quenching (MCBQ), which suppresses background emission through barrierless transitions into dual twisted intramolecular charge transfer (TICT) states enabled by flexible pyridinium regioisomer design. Notably, the meta-position pyridinium configuration disrupts resonance delocalization, enhances charge separation, and drives ultrafast MCBQ-mediated quenching within 0.30 ps. Leveraging this MCBQ strategy, we developed a mitochondrial-targeted, ultralow-background fluorogenic probe (MBFP) platform that undergoes biochemical factor-triggered self-immolation to release a bright fluorophore, resulting in up to 2331-fold fluorescence enhancement. As a proof of concept, the H2O2-responsive MBFP enabled selective quantification of mitochondrial H2O2 variations in cellular and murine models, further supporting its translational potential in clinical serum samples. This work establishes MCBQ as a broadly applicable design strategy for ultralow-background mitochondrial fluorogenic probes and provides a versatile platform for real-time monitoring of mitochondrial dysfunction toward early disease diagnosis.
Abstract The impure phase in typical solution-processed two-dimensional/three-dimensional (2D/3D) perovskite heterostructures limit the performance and stability of perovskite solar cells. We overcome this limitation through a templated growth strategy utilizing excess lead iodide, which directs the formation of robust phase-pure 2D perovskite at the 3D perovskite surface. This strategy effectively suppresses phase segregation and ion interdiffusion at the 2D/3D perovskite interface, while promoting favorable energy alignment for charge extraction. The resulting 2D/3D devices achieved 26.8% efficiency (26.5% certified) and improved durability, retaining over 95% of their initial performance after 3000 h of continuous operation and maintaining nearly 100% of their initial performance throughout 90 days of outdoor testing, showcasing promising real-world stability.
Abstract The evolution of the East Asian Summer Monsoon (EASM) has been linked to the development of the Tibetan Plateau, yet monsoon simulations vary considerably between different palaeogeographic models due to inherent uncertainties in reconstructing deep-time landscapes, both locally, regionally and globally. Using the fully coupled HadCM3BL–M2.1aD palaeoclimate model, we examine EASM evolution over the past 145 Myr across three independent palaeogeographies (Getech, Scotese, Robertsons). While all simulations capture broad hydrological trends, they diverge markedly in the timing and intensity of monsoon development. The Cenozoic strengthening of the EASM aligns with Tibetan and regional mountain orogeny, but Cretaceous behaviour reflects strong remote teleconnections driven by continental configuration. These results demonstrate that Asian monsoon systems are sensitive not only to local orography but also to global palaeogeographic context due to the impact of teleconnections on the EASM. Constraining such uncertainties is essential for resolving the monsoon’s long-term evolution and climatic influence.
Abstract Localized oscillatory patterns, confined to finite spatial domains or subsets of nodes in a network, have been observed in cortical neurons, metacommunity ecosystems, and chemical experiments. Previous studies have shown that wave bifurcations (oscillatory Turing instabilities) in undirected networked reaction–diffusion systems can induce such localized oscillations. Here, we propose an alternative mechanism for their emergence, namely Hopf-type instabilities acting on a subcritical Turing branch. Unlike wave bifurcations, this mechanism does not require interactions amongst at least three components. We further demonstrate its robustness in both the Brusselator and FitzHugh-Nagumo models in undirected Barabási–Albert, Erdős–Rényi, and Watts–Strogatz networks. This work extends the universality of such patterns in networked systems and provides a theoretical foundation for the control of collective dynamical phenomena in complex systems.
Abstract Expression quantitative trait loci (eQTLs) link genetic variation to gene regulation and complex traits, yet population-scale blood eQTL resources remain limited in East Asian populations. We mapped cis-eQTLs in 1,005 Chinese individuals and meta-analysed them with a Japanese cohort to generate an East Asian whole-blood resource (MAEEA; n = 2,024). MAEEA identified ∼2.8 million cis-eQTL associations involving 13,376 eGenes, with lead effects showing external concordance, including in an independent East Asian cohort. For East Asian complex traits, MAEEA showed greater utility than larger non–East Asian eQTL resources for heritability partitioning and candidate gene–trait prioritization; reciprocal analyses using phenotype-matched European GWAS were consistent with ancestry matching contributing to these differences. Highly differentiated eQTL SNPs showed evidence consistent with recent positive selection, and trait-associated eQTL SNPs were enriched for elevated population differentiation. These findings expand East Asian regulatory genetic resources and support ancestry-matched eQTL data for complex-trait interpretation.
Abstract The renewable electricity-driven conversion of biomass-derived compounds into dicarboxylic acids represents an alternative to fossil-based processes, but achieving high selectivity and activity under mild conditions remains a significant challenge. Herein, we report Mo-doped NiFe-layered double hydroxide (NiFeMox-LDH) catalysts for electrosynthesis of maleic acid (MA) from selective electrooxidation of biomass-derived furfurals. Specifically, NiFeMo0.3-LDH with Mo content of 1.75 wt% affords MA in the highest yield of 93.2% to date, in KHCO3 electrolyte under the voltage range of 1.8–2.2 V (versus reversible hydrogen electrode). The Mo-doping in NiFeMox-LDH enhances activity of lattice oxygen and alters adsorption modes of 5-hydroxymethylfurfural (HMF) and intermediate, thus steering the products from furandicarboxylic acid obtained over NiFe-LDH to MA. Additionally, succinic acid can be produced in 99.5% selectivity via MA electroreduction on a copper foam cathode paired with HMF electrooxidation over NiFeMo0.3-LDH. Membrane electrode assembly electrolyzer delivers a Faradaic efficiency of 65.2% for MA production from HMF electrooxidation with a production rate of 252 mmol g−1 h−1 and a partial current density of 326 mA cm−2.
By regulating Earth’s climate and buffering the impacts of anthropogenic pressures, the ocean plays a crucial stabilizing role in the Earth system – yet, in doing so, it is itself undergoing profound changes. This Perspective outlines the role of the ocean in the Earth system and discusses arguments for incorporating additional ocean processes into the Planetary Boundaries framework.
Interactions among soil viruses, microbes and soil fauna form a unified soil biological loop that shapes microbial necromass and controls soil carbon storage. This work extends the microbial carbon pump theory by introducing the virus-mediated microbial carbon pump, offering a new framework to explain soil carbon sequestration under global change.
Understanding how terrestrial plant functional strategies (competitive, stress-tolerant, ruderal; CSR) respond to environmental conditions is crucial for predicting ecosystem dynamics under global climate change, yet remains unexplored at the global community level. Leveraging a machine learning approach, and utilizing multi-source satellite remote-sensing and field-collected sPlotOpen measurements data, we generated the first global community-level map of CSR functional strategy variations. Results show that S-selected strategies are globally dominant (C:S:R = 23.66:62.40:13.94%), with substantial spatial variations across biomes. This variability is strongly influenced by climatic variables (e.g. mean annual precipitation, diurnal temperature range) and soil properties (e.g. cation exchange capacity, total nitrogen). Future projections show that climate change favours S- (+0.33%) and R- (+0.31%) at the expense of C-selected strategy (-0.64%), alongside marked biome-specific shifts. Despite potential underestimation of localized climate uncertainties, these findings provide critical insights into global plant community dynamics under challenging abiotic conditions.
Abstract Asymmetric hydrofunctionalization of allenes has emerged as a valuable strategy for synthesizing privileged allyl skeletons. However, existing studies are largely restricted to special terminal allene substrates to avoid formidable deracemization issues. Herein, we present 1,3-diene principle as a new and broadly applicable mechanistic pathway to address this long-standing challenge and beyond. The 1,3-diene principle is demonstrated by the applications to over ten types of unexplored enantioselective transformations of racemic allenes. Stereodivergent synthesis, convergent preparation, and concise access to natural products further highlight the practical value and generality of this paradigm. Remarkably, the 1,3-diene manifold also enables the first allene migratory hydrofunctionalization, breaking the metal walking chemistry limited to alkene substrates. Both experimental results and theoretical calculations support the deracemization route involving 1,3-diene species, opening a new model for allene transformations.
Abstract Industrial robots underlie modern manufacturing automation, yet conventional deterministic control based on fixed trajectories and offline programming struggles under high-mix and flexible production. Embodied artificial intelligence (EAI) offers a promising alternative by coupling perception, reasoning, and action within closed-loop physical interaction, enabling industrial robots to adapt behaviors online rather than execute predefined tasks. Yet, general-purpose EAI remains difficult to deploy in industrial environments due to stringent requirements on precision, real-time performance, reliability, and safety. These challenges have motivated increasing interest in embodied artificial intelligence for industry (EAI4I). This paper presents a systematic survey of EAI4I from an industrial robotics perspective. Specifically, we first analyze the quantified requirements of industrial robots enabled by EAI4I. Afterwards, recent research progress is reviewed, covering core technologies for single- and multi-robot systems, dedicated hardware platforms, high-fidelity simulators, task-specific datasets, representative industrial application scenarios, and critical deployment challenges. Finally, promising directions toward EAI4I are discussed.
Abstract Major chronic diseases, including cancer, cardiovascular and neurodegenerative disorders, diabetes, and non-infectious inflammations, are fundamentally driven by dysregulated pathological chemical microenvironments characterized by oxidative stress, hypoxia, acidosis, protein misfolding, and abnormal carbohydrate and lipid metabolism. These aberrant chemical milieus not only drive pathogenesis but also create a pathological ‘soil’ that supports disease progression and confers therapeutic resistance. Traditional strategies predominantly target cellular components, overlooking this fundamental chemical basis. Nanocatalytic medicine has recently emerged as a novel paradigm that directly targets these pathological chemical microenvironments through controlled in-situ catalytic reactions. Rationally designed nanocatalysts enable precise regulation of disease-associated chemicals, including generating cytotoxic reactive oxygen species for tumor therapy, scavenging excessive oxidants to alleviate inflammatory damage, catalyzing oxygen production to relieve hypoxia, and treating biomolecular aggregates in neurodegenerative diseases. This Review delineates the chemical hallmarks of major diseases, elucidates nanocatalytic strategies for their regulation, and proposes design principles for optimizing nanocatalytic medicines. By positioning pathological chemical microenvironments as fundamental therapeutic targets, nanocatalytic medicine offers an innovative and powerful chemical toolbox for treating major chronic diseases.
Abstract Minimizing power consumption and improving long-term stability of bioelectronic devices are crucial for ensuring prolonged operation and reducing the risk of tissue damage. Organic electrochemical transistors (OECTs) have gained widespread attention in biosensing applications due to their high transconductance and excellent biocompatibility. Although considerable efforts have been directed toward optimizing performance metrics (e.g. µC*) in OECTs, strategies to reduce power consumption have been largely overlooked, despite their critical importance for large-scale integration and in vivo sensing applications. Here, we present a unique self-doping strategy to achieve low power consumption in OECTs. Small molecules with certain end groups can undergo self-doping through ethylene glycol side chains, enabling n-type depletion-mode OECTs with near-zero threshold voltage (VTh) and outstanding long-term stability. Based on these, we developed biosensors with high signal-to-noise ratios while operating at an ultralow power consumption of 12 nW. Our findings highlight self-doping molecules for energy-efficient biosensing.
Abstract Embodied agents must continuously adapt to the physical world using interaction data collected across varying timescales, controllers, and environmental conditions. However, standard reinforcement learning assumes stationary dynamics and on-policy data, a premise often violated in reality where physical parameters drift and historical data becomes heterogeneous. The central challenge lies in the compound distribution shift: replayed transitions follow an occupancy distribution that diverges fundamentally from the current physical reality, leading to biased value estimation and catastrophic learning collapse. In this work, we propose Transition Occupancy Matching as a unifying principle to resolve policy and dynamics shifts within a single mathematical framework. We introduce Occupancy-Matching Policy Optimization (OMPO), a novel algorithm that optimizes a surrogate objective explicitly correcting for transition discrepancies. By leveraging a dual reformulation with a sign-free logarithmic link, OMPO transforms the intractable matching problem into a stable min-max optimization, amenable to arbitrary reward structures. Crucially, OMPO integrates a distributional critic and a multimodal encoder with a small-scale local buffer, allowing the agent to anchor massive historical data to the immediate physical context for rapid adaptation. Extensive evaluations across diverse benchmarks—including MuJoCo locomotion, DM-Control, Meta-World, and high-fidelity Panda robot manipulation—demonstrate that OMPO consistently outperforms specialized baselines in stationary, domain-shifting, and non-stationary settings. By unifying distribution correction across policy and dynamics shifts, OMPO addresses a fundamental bottleneck in transfer learning, providing a robust algorithmic framework for continual adaptation in changing physical conditions.
ABSTRACT Subsurface natural hydrogen is an emerging clean energy resource, but its origins, loss mechanisms and cycling timescales, particularly at greater depths, remain poorly constrained. Here, we report large disequilibrium in the clumped isotopes of H2 (2H2H) and hydrogen isotope fractionation between water and H2, measured in gas samples (containing 0.10%–1.59% H2) collected from deep (3000–7000 m) wells in the central Sichuan Basin. Because molecular hydrogen can equilibrate isotopes rapidly in the presence of water, such large disequilibrium indicates active kinetically controlled production and/or consumption. Analyses of geological and chemical evidence demonstrate that isotopic disequilibrium is driven by consumption by chemical reactions in host rocks coupled with partial hydrogen exchange with co-existing water. The results imply that measured compositions do not preserve the information on original sources and processes. More broadly, the dynamic cycling of hydrogen renders sediment-hosted deep gas reservoirs unfavorable for long-term or large-scale accumulation of natural hydrogen.
Abstract Intelligent mid-infrared (MIR) imaging integrates in-sensor preprocessing of thermal information, thereby reducing data redundancy and improving target-recognition robustness under low-visibility conditions. However, unlike in the visible regime, photoconductive programmability in the MIR typically relies on narrow-bandgap materials, which inherently limit stable and reconfigurable operation. Here, we report a programmable photothermoelectric (PTE) detector based on a suspended nanometer-thick SrTiO3 membrane. Dimensional scaling of the thermoelectric channel enhances sensitivity and enables a response time of ∼10 μs, over 104 times faster than bulk SrTiO3. Through dual-gate electrostatic modulation of the Seebeck profile, the device exhibits bidirectionally tunable photothermoelectric responsivity up to 50 V W−1 and supports 40 experimentally observed programmable response states. By leveraging a 3 × 3 programmable device array with device-to-device consistent gate-programmable photoresponse and reproducible polarity switching, a recognition accuracy of 83% is achieved under visually degraded conditions. Furthermore, the programmable array enables attention-guided thermal image weighting by enhancing target thermal signatures while suppressing background interference. These results demonstrate the dimensional scaling of perovskite oxides as an effective route toward adaptive, low-power machine vision and neuromorphic infrared electronics.
