Metal halide ionic octahedra, serving as the fundamental optoelectronic unit in halide perovskites, enable near-infrared (NIR) luminescence via transition-metal ion incorporation. However, their intrinsically low radiative efficiency and inadequate operational stability have posed significant challenges for practical implementation. In this work, we report the first successful synthesis of a highly stable Mo-doped Sn-based perovskite NIR emitter via a one-step hydrothermal approach, which exhibits unprecedented dual broadband NIR emission (800-1630 nm). The oxidation of Sn2+ induces the formation of mixed Mo4+/3+ valence states, while the synergy of lattice distortion, spin-orbit coupling, and vibronic coupling activates multiple d-d transitions. Specifically, they promote the 1T2g/1Eg -> 3T1g transition of Mo4+ and the Gamma 8(2T1g) ->Gamma 8(4A2g) transition of Mo3+, achieving a high photoluminescence quantum yield (PLQY) of 68% at room temperature. Notably, this NIR-emitting halide maintains 88% of its room-temperature emission intensity at 423 K, demonstrating exceptionally low thermal quenching. Moreover, the precise control of Mo doping level and the introduction of Sn2+ enable the systematic tailoring of the NIR-I/II luminescence. This breakthrough not only provides fundamental design principles for developing next-generation broadband NIR-I and II emitting material but also establishes a new application platform in night-vision and vascular imaging for optoelectronic devices with superior performance.

The oxygen evolution reaction (OER) is crucial for renewable energy systems, such as water splitting and metal air batteries. However, the slow kinetics of OER significantly limits the overall energy conversion efficiency, necessitating effective catalysts. The multielement transition metal oxides offer a promising alternative compared to precious metal oxides, yet their composition optimization remains challenging due to the vastness of the combinatorial space. Traditional trial-and-error approaches are labor-intensive and inefficient. To address this challenge, we develop an innovative automated platform integrating machine learning (ML) with Bayesian optimization for rapid and cost-effective synthesis and evaluation of electrocatalysts. This platform allows the automation of the entire experimental process, from synthesis to evaluation, enabling real-time feedback and guiding subsequent experiments. In a continuous operation of 32 h, the platform conducted 96 experiments to optimize the composition of (Ni-Fe-Co-Mn-Mo)Ox , resulting in an electrocatalyst with an overpotential of 231 mV at 10 mA/cm 2. This automated approach significantly reduces manual intervention and enhances efficiency, proving to be a valuable tool for optimizing materials in complex, multidimensional spaces. (c) 2026 Published by Elsevier B.V. on behalf of Chinese Chemical Society and Institute of Materia Medica, Chinese Academy of Medical Sciences.

All‐solid‐state batteries (ASSBs) promise significant advances in energy density and safety, yet the rapid expansion of literature and heterogeneity of synthesis protocols impede efficient knowledge integration and reproducibility. We present Synthesis‐GPT, a retrieval‐augmented, multiagent large language model (LLM) system that operates over a curated text corpus to extract stoichiometrically accurate synthesis routes and key performance descriptors with full provenance and to organize them into a structured, extensible knowledge base. The system supports two expert workflows—domain question answering and synthesis route extraction—augmented by standardized visualizations (flowcharts and parameter tables) that enable rapid comparison and experiment planning. We demonstrate interactive use cases spanning material recommendation, property lookup, and recipe‐style guidance for high‐conductivity electrolytes. Beyond point queries, Synthesis‐GPT constructs a method‐structured database (solid‐state, liquid‐phase, ultrafast, and other) that supports cross‐route analytics and downstream modeling. By grounding generation in retrieval and coordinating specialized agents, the framework reduces hallucination, preserves domain fidelity, and improves auditability—providing practical infrastructure toward reproducible, closed‐loop, automated materials discovery.

Accurate detection of ascorbic acid (AA) levels is essential for diagnosing scurvy. However, traditional sweat-based AA sensors are hindered by complex, costly electrodes, and weakly hydrophobic substrates. In this work, we present a novel approach to sweat AA detection using simple, low-cost, and highly hydrophilic sensors fabricated from carbon electrodes printed on flexible, delignified wood substrates. The electrodes are fabricated by screen-printing carbon ink, resulting in low-cost, simple structures. The wood substrates are prepared through a delignification process, which enhances their flexibility and hydrophilicity, enabling direct sweat permeability due to the natural alignment of vessel channels in the wood structure. These fabrication methods and unique substrate structures give the sensors a high sensitivity (0.032 mu ALmu mol-1), excellent selectivity, a low detection limit (10 mu mol/L), and a wide linear range (10-1000 mu mol/L), along with a fast response time, and outstanding stability against mechanical bending, temperature fluctuations, and long-term storage. Selective detection of ascorbic acid is achieved even in the presence of common interfering substances such as glucose and dopamine. Moreover, the sensors have been successfully tested on real human skin, demonstrating their potential for reliable real-time monitoring of AA levels. These sensors could be applied in the prognosis, diagnosis, and management of scurvy.

In lithium-ion battery lifetime prediction research, the rapid growth in the number of related papers and highly dispersed experimental settings pose significant challenges for systematic literature surveys and method comparisons. To address these challenges, we propose a multi-agent framework for automated literature analysis and knowledge extraction. Through the collaborative work of Parsing, Retrieval, and Verifier agents, the framework automatically extracts, aligns, and integrates key experimental information from scientific publications. It structures data on essential aspects, including task types, dataset configurations, input windows, training scales, and prediction errors, and constructs a traceable experimental database. This database enables unified statistical analysis and cross-study comparisons, revealing population-level patterns such as convergence of input settings, dispersion in performance distributions, and the influence of data scale on lifetime prediction. Our work highlights the potential of multi-agent methods for literature analysis and research infrastructure development, offering a scalable framework for data-driven battery lifetime prediction.

The global challenge on energy crises has promoted the extensive research on energy storage, which positions the lithium-ion batteries (LIBs) as a core technology due to the exceptional performance. While conventional approaches for LIB research are limited by long development cycles, high resource demands, and heavy reliance on the expertise of researchers, data-driven methods powered by ML exhibit remarkable potentials on enabling the efficient and accurate predictions of key battery performance metrics, significantly reducing the time and economic costs on the R&D of batteries. Despite these advances, data scarcity remains a major obstacle to the widespread applications of these techniques due to the high costs and prolonged cycles associated with battery experiments. To investigate appropriate approaches for addressing these challenges, a comprehensive review on the data-driven methods for lithium battery research is provided, offering insights into mitigating the impact of data scarcity. To reach this goal, this review systematically examines the implementations of data-driven methods associated with the applications in the battery domain. Detailed studies have been conducted to investigate the root causes of data scarcity at various levels ranging from materials to devices and effective strategies to accommodate these issues. Finally, the review discusss future directions, emphasizing the need for collaborative data-sharing frameworks and adaptive ML models that balance domain expertise with computational innovation. By addressing data scarcity through these strategies, this work aims to accelerate the development of next-generation LIB while retaining the core insights of traditional methodologies.

Battery recycling via direct regeneration has emerged as a next-generation strategy to simultaneously address environmental pollution and resource waste issues caused by end-of-life lithium-ion batteries. While previous research on direct regeneration has mainly focused on materials or repair processes, it has overlooked the impact of lithium fluoride (LiF) impurities, which can significantly degrade the electrochemical performance of regenerated materials. Through combined experimental characterization and computational simulations, we have experimentally confirmed the ubiquitous presence of LiF on spent LiNi0.82Co0.12Mn0.06O2 cathodes and elucidated its formation mechanism via first-principles calculations. To address this challenge, we developed a direct recycling strategy combining an acid-wash pretreatment with a high-temperature regeneration process. The acid wash effectively eliminates residual lithium fluoride impurities, thereby enhancing the lithium-ion transport dynamics. The solid-sintering regeneration simultaneously achieves lithium replenishment along with structural restoration, including microcrack restoration and phase transformation from rock salt to a layered structure. The purified regenerated LiNi0.82Co0.12Mn0.06O2 cathodes produced through our strategy demonstrate an initial discharge capacity of 197.37 mAhg-1 with a 7.2% enhancement compared to their nonpurified counterparts. Remarkably, the purified regenerated LiNi0.82Co0.12Mn0.06O2 cathodes exhibit a superior rate capability and cycling stability. This work not only uncovers the formation mechanism of lithium fluoride and its critical role in degrading the performance of regenerated batteries but also offers an effective strategy for the practical direct recycling of spent Ni-rich cathodes.

Anharmonic lattice dynamics (ALD) has proven to be a promising approach for the development of advanced superionic conductors for solid-state batteries.

Stable solid electrolytes are essential for advancing the safety and energy density of lithium batteries, especially in high-voltage applications. In this study, we designed an innovative high-entropy chloride solid electrolyte (HE-5, Li2.2In0.2Sc0.2Zr0.2Hf0.2Ta0.2Cl6), using multielement doping to optimize both ionic conductivity and high-voltage stability. The high-entropy disordered lattice structure facilitates lithium-ion mobility, achieving an ionic conductivity of 4.69 mS cm-1 at 30 degrees C and an activation energy of 0.300 eV. Integration of HE-5 into all-solid-state batteries (ASSBs) with NCM83 cathodes and a Li-In anode enables outstanding electrochemical performance, sustaining 70% capacity retention over 1600 cycles at a 4 C rate. Moreover, the high configurational entropy stabilizes the electrolyte's structure at elevated voltages, enabling stable operation at 5.0 V without significant degradation. Our work presents the dual advantages of high-entropy engineering in boosting high ionic conductivity and voltage stability, providing a broad roadmap for next-generation energy-dense ASSBs.

Machine learning has been massively utilized to construct data-driven solutions for predicting the lifetime of rechargeable batteries in recent years, which project the physical measurements obtained during the early charging/discharging cycles to the remaining useful lifetime. While most existing techniques train the prediction model through minimizing the prediction error only, the errors associated with the physical measurements can also induce negative impact to the prediction accuracy. Although total-least-squares(TLS) regression has been applied to address this issue, it relies on the unrealistic assumption that the distributions of measurement errors on all input variables are equivalent, and cannot appropriately capture the practical characteristics of battery degradation. In order to tackle this challenge, this work intends to model the variations along different input dimensions, thereby improving the accuracy and robustness of battery lifetime prediction. In specific, we propose an innovative EM-TLS framework that enhances the TLS-based prediction to accommodate dimension-variate errors, while simultaneously investigating the distributions of them using expectation-maximization(EM). Experiments have been conducted to validate the proposed method based on the data of commercial Lithium-Ion batteries, where it reduces the prediction error by up to 29.9 % compared with conventional TLS. This demonstrates the immense potential of the proposed method for advancing the R&D of rechargeable batteries.

In response to the rapid advancements and heightened competition within solid-state battery research, the sheer volume of publications presents a significant challenge for researchers seeking comprehensive insights. This paper introduces ChatSSB, an advanced research assistant designed to bolster scientific inquiry within this dynamic field. Leveraging the Retrieval-Augmented Generation (RAG) framework, ChatSSB excels in extracting precise information from the latest research publications through an intuitive Q&A interface. Beyond its foundational capabilities, ChatSSB boasts a customizable expert knowledge database, continuously updated through a dynamic feedback mechanism. This ensures researchers have access to cutting-edge and reliable information, overcoming the limitations of outdated or incomplete literature. Furthermore, the integration of multiagent collaboration and embedded tools within RAG facilitates robust quantitative analysis, enabling efficient data collection, visualization, and interpretation. Collectively, these features empower ChatSSB to deliver precise, actionable insights, significantly accelerating innovation in solid-state battery technology and propelling it toward the next frontier of materials science.

Solid-state lithium metal batteries hold great promise for next-generation energy storage, due to their safety and energy density. Entropy-engineered solid-state electrolytes (SSEs) offer enhanced ionic conductivity and interfacial stability but face phase separation issues during a conventional synthesis. Here, we report an ultrafast synthesis strategy for fabricating pure-phase medium-entropy Li(3x)Ln(2/3-x)TiO(3) (Ln = Nd, La, Gd, 0 < x < 0.16, ME-LLTO) SSEs. High-temperature in-situ and quasi-in-situ X-ray diffraction (XRD) analyses confirm that ultrafast synthesis employs ultrafast cooling to suppress undesired phase transitions, thereby stabilizing the cubic phase while significantly reducing the formation of intermediate phases. In contrast, conventional furnace synthesis leads to persistent intermediates and the coexistence of cubic and tetragonal LLTO phases. Microstructural and spectroscopic characterizations further reveal enhanced phase purity and elemental homogeneity in ultrafast-synthesized ME-LLTO. This work establishes ultrafast synthesis as an effective and scalable strategy for synthesizing high-performance entropy-engineered SSEs by suppressing phase separation.

The widespread usage of rechargeable batteries in portable devices, electric vehicles, and energy storage systems has underscored the importance for accurately predicting their lifetimes. However, data scarcity often limits the accuracy of prediction models, which is escalated by the incompletion of data induced by the issues such as sensor failures. To address these challenges, we propose a novel approach to accommodate data insufficiency through achieving external information from incomplete data samples, which are usually discarded in existing studies. In order to fully unleash the prediction power of incomplete data, we have investigated the Multiple Imputation by Chained Equations (MICE) method that diversifies the training data through exploring the potential data patterns. The experimental results demonstrate that the proposed method significantly outperforms the baselines in the most considered scenarios while reducing the prediction root mean square error (RMSE) by up to 18.9%. Furthermore, we have also observed that the penetration of incomplete data benefits the explainability of the prediction model through facilitating the feature selection.

Anharmonic lattice dynamics (ALD) has proven to be a promising approach for the development of advanced superionic conductors for solid-state batteries. However, the relationship between ALD and ion diffusion remains poorly understood due to the coupling between lattice dynamics and the potential energy surface. In this study, we demonstrate that in beta-Li3VO4, the enhanced ALD of OI atoms is coupled with the motion of LiII ions, resulting in increased activation and diffusion of LiII ions as temperature increases. Rietveld refinement analysis of the high-temperature X-ray diffraction (HTXRD) patterns indicates that ALD primarily involves LiII and OI atoms, with thermal vibration factors increasing significantly with temperature. In situ Raman spectroscopy combined with first-principles calculations reveals that three phonon modes associated with LiII-OI vibrations exhibit strong anharmonicity. Among these, one mode is linked to the activation of Li ions, while the other two are associated with the diffusion process. Based on these observations, we propose an atomic-scale mechanism to describe the ALD process. Our findings provide deeper insights into how ALD enhances ion diffusion and support the idea of precisely controlling ion mobility in superionic conductors through phonon engineering.

Zinc-bromine flow batteries (ZBFBs) are promising for sustainable energy storage due to their high energy density and cost-effectiveness. However, the sluggish kinetics of the Br2/Br- redox reaction at the cathode limits their performance. Here, we developed Fe/N co-doped micro-mesoporous carbon nanofibers (Fe-N-CNFs) as a high-performance cathode catalyst. Synthesized via electrospinning Fe/Zn-ZIFs with PAN/PVP, followed by carbonization, the Fe-N-CNFs exhibited a hierarchical pore structure with a specific surface area of 1057 m2 g- 1 and an average pore size of 2.5 nm. The optimized catalyst, doped with 4 wt% Fe, achieved an energy efficiency of 81 % at 80 mA cm- 2 and maintained a Coulombic efficiency of 98.4 % over 200 cycles. This work demonstrates the potential of integrating electrospinning with MOF-derived catalysts to enhance ZBFB performance, offering a scalable solution for high-efficiency energy storage systems.

While lifetime prediction of rechargeable batteries is crucial for ensuring the reliability and sustainability of electric devices, the accuracy and robustness of prediction models are often impacted by practical non-idealities in operational scenarios. In order to ensure the reliability of battery lifetime prediction, this work is dedicated to addressing a specific challenge posed by missing information in training data, which can be induced by multiple practical factors. To address this issue, this paper investigates multiple modeling strategies for handling missing data challenges, among which a novel multi-view imputation strategy is proposed that explores the diversity of underlying data patterns, thereby substantially improving the prediction accuracy. Experiments have been conducted to quantitatively evaluate the efficacy of the modelling techniques, where the proposed method is highlighted with substantial improvements in prediction accuracy and robustness, such that the root mean square error (RMSE) was reduced by up to 35.7 % under intensive missing data conditions compared to conventional approaches. Through offering an innovative solution for accommodating missing data in predictive modeling, this study has advanced the development of efficient and reliable battery management systems.

The development of wearable energy sto rage and harvesting devices is pivotal for advancing next-generation healthcare technologies, facilitating continuous and real-time health monitoring. Traditional wearable devices have been constricted by bulky and rigid batteries, limiting their practicality and comfort. However, recent advancements in materials science have enabled the creation of flexible, stretchable, and lightweight energy storage and harvesting solutions. The integration of energy storage and harvesting technologies is essential for developing self-sustaining systems that minimize reliance on external power sources and enhance device longevity. These integrated systems ensure the continuous operation of sensors and processors vital for real-time health monitoring. This review examines recent significant progress in wearable energy storage and harvesting, focusing on the latest advancements in wearable devices, solar cells, biofuel cells, triboelectric nanogenerators, magnetoelastic gene rators, supercapacitors, lithium-ion batteries, and zinc-ion batteries. It also discusses key parameters crucial for their wearable applications, such as energy density, power density, and durability. Finally, the review addresses future challenges and prospects in this rapidly evolving field, underscoring the potential for developing innovative, self-powered wearable systems for healthcare applications. • The latest advancements in energy storage and harvesting systems for wearable healthcare devices are discussed. • Flexible supercapacitors, lithium-ion batteries, solar cells, TENGs and other devices are systematically introduced. • Factors influencing wearable energy devices including energy density, power density, and durability are analyzed. • Future perspectives in wearable energy systems are explored, particularly emphasis on the role of AI and LLMs. • Flexible and wearable energy storage and harvesting systems offer a promising path for healthcare applications.

The precise synthesis of high-purity materials is crucial in accelerating materials discovery. However, the lack of theoretical understanding and practical guidance poses challenges, particularly for materials with compositional and structural complexity. Here, we propose a feasible principle toward synthesizing complex inorganic solids. This principle involves the introduction of an inducer that induces crucial intermediates, which in turn guide the synthesis pathway toward the target materials through structural templating, named inducer-facilitated assembly through structural templating (i-FAST). We validate this principle with three distinct oxides: garnet Li6.5La3Zr1.5Ta0.5O12, perovskite BaCo0.8Sn0.2O3, and pyrochlore Gd1.5La0.5Zr2O7. This structural templating approach enables synthesis along predesigned pathways, forming intermediates that are thermodynamically favored for prior formation and kinetically preferred for the final product, resulting in precisely synthesizing high-purity target materials. This study not only represents a substantial advancement in comprehending the interplay between thermodynamics/kinetics and phase evolution in complex solid synthesis but also provides an effective strategy for guiding exploratory solid-state synthesis.

All‐solid‐state lithium metal batteries offer enhanced safety and energy density by replacing flammable liquid electrolytes with solid‐state electrolytes (SSEs). High‐entropy (HE) SSEs, leveraging multi‐principal‐element compositions, present a vast design space to achieve exceptional ionic conductivity and electrochemical stability. However, the chemical complexity of HE SSEs introduces challenges in interfacial instability with lithium metal anodes due to the unavoidable inclusion of reactive elements. While conventional garnet‐type SSEs are considered stable, it is revealed that five HE garnets (HE‐LLZOs) undergo corrosion and partial dissolution upon lithium contact. Here, a rational design strategy is introduced to stabilize HE‐LLZO by combining thermodynamic assessments of interfacial reactivity with targeted compositional engineering. Through systematic exploration of element‐specific degradation mechanisms, selection criteria for lithium‐compatible principal elements are established. Guided by computational screening, unstable dopants are excluded (e.g., Nb, Mo, W, Cr, Bi) that drive interfacial degradation and synthesize a novel HE‐LLZO (Li 6.6 La 3 Zr 0.4 Sn 0.4 Hf 0.4 Sc 0.2 Ta 0.6 O 12 ) that exhibits high ionic conductivity (3.69 × 10 −4 S cm −1 ) and stable cycling over 2,500 h. X‐ray photoelectron spectroscopy confirms the interfacial stability of Zr, Sn, and Ta while identifying Nb as a destabilizing element. This work provides an integrated computational‐experimental framework for understanding element‐property relationships in HE oxides, advancing durable SSEs design.