Ammonia, as a carbon-free and hydrogen-rich compound, with its high volumetric hydrogen density, mild liquefaction and storage transportation conditions, as well as a mature global supply chain, is gradually becoming an ideal hydrogen carrier. This article systematically reviews the research progress of the entire "ammonia-hydrogen" industrial chain. Firstly, this paper analyzes the high carbon emissions and low flexibility challenges faced by the traditional Haber-Bosch process in the energy transition, and discusses the design principles and performance bottlenecks of emerging green ammonia synthesis technologies such as the green Haber process, electrochemical nitrogen reduction, photocatalysis, and plasma catalysis. Secondly, the focus is placed on the diversified utilization of ammonia in the downstream conversion field. The technical paths, electrochemical performances and commercialization obstacles of direct ammonia solid oxide fuel cells, direct ammonia anion exchange membrane fuel cells and proton-conducting ceramic fuel cells are compared. Subsequently, based on this, an analysis from the perspectives of economy and environmental impact is conducted. Finally, the economic advantages of ammonia in the storage and transportation process, as well as its compatibility with existing infrastructure, give it the potential to replace traditional fuels in high-emission sectors such as shipping and power generation.

Against the background of global carbon peaking and carbon neutrality strategies, Compressed Air Energy Storage (CAES) has become a key technology for large-scale energy storage. This study takes the lining and seepage control materials of underground CAES caverns as the research core, which directly determine the operation safety and airtight performance of energy storage systems. Underground caverns, as mainstream gas storage structures, operate under harsh conditions including cyclic high pressure and drastic temperature changes, which put forward strict performance requirements for supporting and sealing materials. On the basis of sorting out domestic and foreign research results, this paper classifies common lining and seepage control materials, evaluates their engineering adaptability, and summarizes the latest progress in performance testing methods. Meanwhile, this work points out the technical defects and application bottlenecks of existing materials, and puts forward targeted research directions combined with engineering practice. The results prove that the innovation and optimization of key materials are the core driving force to promote the commercialization and large-scale application of CAES engineering.

Since the commercialization of lithiumion batteries in the 1990s, they have been extensively applied in portable electronics, electric vehicles, and energystorage systemsowing to their high energy density, long cycle life, low selfdischarge rate, and environmental benignity. As the core component determining the capacity and stability of lithiumion batteries, cathode materials play a decisive role in overall electrochemical performance. This paper briefly reviews the developmental history, fundamental structure, and operating mechanism of lithiumion batteries, then focuses on three major categories of commercial and researchoriented cathode materials: olivinetype lithium iron phosphate (LiFePO₄), spinelstructured lithium manganese oxide (LiMn₂O₄) and its highvoltage derivatives such as LiNi₀.₅Mn₁.₅O₄, as well as layered oxide systems including LiCoO₂, LiNiO₂, NiCoMn (NCM) ternary materials, and lithiumrich manganesebased layered oxides. It systematically compares their crystal structures, electrochemical characteristics, inherent drawbacks, and common modification strategies, with the purpose of offering theoretical support and technical guidance for the rational design and industrial application of advanced cathode materials.

Propane Direct Dehydrogenation (PDH) represents the dominating technique for propylene synthesis, a research hotspot in modern coal chemical and light hydrocarbon engineering. Pt-based catalysts manifest intrinsic superiority in C-H bond cleavage, while zeolites endow unique spatial confinement, tunable acid-base properties and strong metal-support interactions, which modulate Pt electronic states and particle dispersion, thus suppressing coke deposition and thermal sintering. This review systematically dissects the molecular-level dehydrogenation mechanism of Pt-zeolite catalysts, clarifying the intricate structure-activity relationships among active site micro-configuration, surface electronic effects, and reaction pathways. It further provides a comprehensive comparative evaluation of two mainstream synthetic strategies, post-synthetic modification and in-situ one-pot crystallization, focusing on their efficacy in the precise fabrication of subnanometric Pt active sites and the long-term catalytic durability of the resulting materials. Finally, we summarize the pivotal scientific insights and technical progresses achieved in this field, laying a solid theoretical foundation for the rational design and industrial upgrading of high-efficiency, stable PDH catalysts.

Ionogels have emerged as a class of advanced functional materials for flexible electronics owing to their unique integration of high ionic conductivity, wide electrochemical windows, and excellent mechanical flexibility. This review provides a systematic overview of recent progress in ionogels, covering material design principles, fabrication strategies, and performance regulation mechanisms. Particular emphasis is placed on the selection of ionic liquid/polymer systems, network structure engineering, and nanocomposite reinforcement approaches to achieve optimized conductivity, mechanical robustness, and environmental stability. The intrinsic relationships between microstructure and macroscopic properties are analyzed to elucidate ion transport mechanisms, mechanical enhancement strategies, and stability improvement methods. In addition, the diverse applications of ionogels in flexible sensors, energy storage devices, and biomedical systems are comprehensively discussed, highlighting their multifunctionality and adaptability. Despite these advances, several challenges remain, including high material cost, limited large-scale manufacturability, and long-term stability under harsh conditions. Future research directions are proposed, focusing on green synthesis, low-cost material design, and intelligent optimization assisted by data-driven approaches. This review aims to provide a comprehensive framework for the rational design and practical application of high-performance ionogels.

The synthesis of Metal-Organic Frameworks (MOFs) currently faces an environ-mental paradox. While these advanced materials are designed for energy and environmental remediation, their conventional fabrication routes are heavily reliant on toxic solvents and high-carbon-emission precursors. This Perspective proposes a sustainable, "waste-to-wealth"paradigm that upcycles post-consumer polyethylene terephthalate (PET) plastic into high-performance UiO-66 adsorbents for deep fuel desulfurization. By cross-validating recent lit-erature, we elucidate how PET-derived linkers not only facilitate green, aqueous synthesis pathways but also induce favorable "defect engineering." These ligand defects expand pore volumes while maintaining highly competitive surface areas (∼995 m2/g), thereby enhancing mass transfer dynamics for sterically hindered polycyclic sulfur compounds such as diben-zothiophene (DBT). By integrating quantitative Life Cycle Assessment (LCA) insights, we demonstrate that PET-upcycling effectively neutralizes major carbon hotspots. Our analysis reveals that this route potentially reduces the Global Warming Potential (GWP) by 40–60% compared to conventional petroleum-based synthesis ( 21–31.25 kg CO2-eq/kg MOF), effec-tively transforming fuel desulfurization into an ecologically and economically viable circular-economy solution.

With the rapid development of electronic information industries such as flexible displays, 5G/6G communications, and semiconductor packaging, colorless and transparent polyimide (CPI) has become an indispensable electronic substrate material due to its combination of high optical transmittance, excellent thermal stability, superior mechanical properties, and low dielectric characteristics. Traditional transparent polyimides typically incorporate fluorine atoms to inhibit the formation of intra- and intermolecular charge transfer complexes (CTC) to enhance transparency. However, fluorinated monomers face challenges including high costs, complex synthesis processes, and the demands of large-scale industrialization and green electronics development. Therefore, the development of fluorine-free transparent polyimide (FFPI) has emerged as a critical research direction in the field of electronic polymer materials. Centering on the raw material design and processing technologies of fluorine-free CPI, this paper reviews the regulation mechanisms of FFPI in terms of optical properties, thermal properties, dimensional stability, and dielectric properties. It also identifies current bottlenecks in FFPI regarding the balance of performances and processing stability. The paper serves as a reference for further research and process optimization of fluorine-free transparent polyimides in the electronics sector.

Although algorithm selection is still a practical difficulty when working with high-dimensional mixed variables, truss structure optimization is a basic topic in structural engineering. This work examines the development of truss optimization algorithms in a methodical manner, contrasting and evaluating metaheuristic algorithms, developing machine learning paradigms, and conventional mathematical programming. The study methodically assesses several algorithms from three perspectives: computing resource consumption, optimization mechanism, and adaptation to discrete-continuous mixed variables. Analysis reveals that while metaheuristic algorithms, like genetic algorithms and particle swarm optimization, can successfully handle mixed variable problems, their application in large trusses is frequently limited by the high computational cost of high-frequency finite element analysis. Traditional gradient algorithms are limited when dealing with discrete sections. Thus, this paper investigates the introduction of Multi-Agent Reinforcement Learning (MARL) and Deep Neural Network (DNN) surrogate models, pointing out that they offer an efficient means of reducing the previously mentioned computational bottleneck through graph structure collaborative search and quick mechanical response prediction. The objective of this study is to offer data-driven design technical references and methodical algorithm selection criteria for the intelligent optimization of intricate spatial trusses.

To address the high trial-and-error cost and computational workload in defect preparation of two-dimensional photonic crystals, this paper applies a BP neural network to model the relationship between energy band data and defect layer thickness. The dispersion data obtained by the MPB algorithm are used as input, while the defect layer thickness is taken as output. A single-hidden-layer BP neural network is constructed, completing dataset preparation, network design, and model training and testing. Model performance is evaluated using relative error and the coefficient of determination R². Results show that the model can rapidly and accurately predict defect layer thickness, with the test set R² exceeding 0.92. This significantly reduces computational cost and provides an efficient approach for optimizing photonic crystal microcavity design.

With the intensifying trend of global population aging, the incidence of neurological diseases such as stroke and spinal cord injury has risen significantly, leading to an increasingly prominent contradiction between the surge in rehabilitation medical demand and the shortage of professional treatment resources. Therefore, developing efficient and accessible rehabilitation assistive devices has become an urgent social need. This paper mainly investigates the evolution trajectory and other key technical mechanisms of rehabilitation robot technology from early passive motion assistance to neural integration models based on biosignals and motion prediction based on machine learning. The research aims to clarify the development history of rehabilitation robot technology, evaluate the clinical effectiveness and existing bottlenecks of current neural integration technology and artificial intelligence-assisted technology in promoting neuroplasticity, and provide a theoretical basis and development direction for building the next generation of rehabilitation systems with adaptive capabilities. This study adopts bibliometric analysis and technical comparison methods to systematically review core patents and academic achievements before 2025; combines typical clinical cases to compare the differences among three generations of technical architectures: passive assistance, active interaction, and neural integration; and comprehensively evaluates the logic and future trends of technical evolution through an interdisciplinary perspective (robotics, neuroscience, rehabilitation medicine).