Data-driven approaches are attracting wide attention in the field of materials science due to their capacity to unravel complex structure-activity relationships deriving from nonlinear interplay of materials properties across multiple scales. However, unlocking their potential in materials discovery and design requires addressing two main challenges: multi-disciplinary knowledge barriers across the entire materials data lifecycle (acquisition, processing, and analysis), and the absence of an infrastructure that can accommodate the continuous proliferation of data volume, algorithms, and models. Here, we propose a multirole collaborative and co-constructive materials design ecosystem that restructures both the productive forces and the relations of production in materials design. By establishing a structured division of labor and a customized materials design infrastructure with a workflow system that decouples control and data flows, our framework reduces inter-module dependencies and enables the flexible, scalable integration of heterogeneous resources. A case study on electrochemical storage materials design demonstrates that this approach can improve streamlined collaborative efficiency by at least 50%, highlighting its potential to accelerate materials design. This work establishes a new paradigm for building intelligent materials design platforms, characterized by dynamic composability instead of static integration, thereby fostering an open and sustainable ecosystem for future materials discovery.

Lithium metal anodes, with a theoretical capacity of up to 3860 mAh·g⁻¹, are regarded as the cornerstone for developing next-generation high-energy-density batteries. However, several key challenges hinder their practical applications, including dendrite formation, unstable solid electrolyte interphase (SEI), side reactions with electrolytes, and associated safety risks. This review systematically explores the mechanisms of lithium nucleation, growth, and stripping in both liquid and solid-state battery systems, analyzing critical theoretical concepts like heterogeneous nucleation thermodynamics, surface diffusion kinetics, space charge effects, and SEI-induced nucleation, which are crucial for understanding the genesis of dendrite growth. Additionally, the review discusses the electrochemical-mechanical coupling failures that lead to SEI degradation and the formation of dead lithium. For liquid systems, the review proposes strategies to mitigate dendrite formation and SEI instability, which include electrolyte optimization, artificial SEI design, and electrode framework design. In solid-state batteries, the review offers a granular analysis of the interface challenges associated with polymer, sulfide, and halide electrolytes and summarizes different solutions for different solid-state electrolytes. Meanwhile, the review emphasizes the importance of advanced characterization techniques and computational modeling in understanding and regulating the interface between lithium metal and electrolytes. Looking ahead, the review highlights future research directions that emphasize the integration of cross-disciplinary approaches to tackle these interconnected challenges. By addressing these issues, the path will be clear for the rapid commercialization and widespread application of lithium metal batteries, bringing us closer to realizing stable, high-energy-density batteries that can satisfy the escalating demands of modern energy storage applications across various industries.

Machine learning (ML) is widely applied to accelerate materials design and discovery due to its outperforming capability of data analysis and information extraction. However, experimental and computational errors typically lead to emerging data anomalies, harming the performance of ML models. Most currently used anomaly detection methods are purely data-driven, which has limited capability of learning complicated factors in materials data. Here, we propose a domain knowledge-assisted data anomaly detection (DKA-DAD) workflow, where materials domain knowledge is encoded as symbolic rules. Three detection models are designed for evaluating the correctness of individual descriptor value, correlation between descriptors, and similarity between samples, respectively, and one modification model is constructed for comprehensive governance. We construct 180 synthetic datasets by injecting noise into 60 structured materials datasets collected from materials ML studies, to validate its potential utility and applications. DKA-DAD achieves a 12% F1-score improvement in anomaly detection accuracy on synthetic datasets compared to purely data-driven approach and the ML models trained on materials datasets processed through DKA exhibit an average 9.6% improvement in R² for the property prediction. Our work provides a data anomaly detecting approach under the guidance of materials domain knowledge towards accelerating materials design and discovery based on ML.

The 0.93(Na_0.5Bi_0.5)_1-xSm_xTiO₃-0.07BaTiO₃ multifunctional ceramics were prepared by solid-phase reaction method. The phase structure, microstructure, electrical and photoluminescent properties were systematically studied. With increasing x, the ceramics undergoes the phase transition from rhombohedral to tetragonal with some rhombohedral distortion, along with a reduced grain size and increased relative density. On the other hand, the Sm³⁺ doping enhances the electric-field driven reversible phase transition and domain size, and reduces the domain walls, thereby contributing to improved piezoelectricity and decreased depolarization temperature (T_d) from 91 ℃ to 40 ℃. Excellent piezoelectric properties of d₃₃ = 213 pC/N and k_p = 29.9% are achieved in the x = 0.010 ceramic. Under excitation (407 nm), the Sm³⁺-doped ceramic exhibits bright reddish-orange fluorescence at 564, 599, 646 nm and 710 nm. A polarization-induced enhancement of photoluminescence is obtained in BNBT-xSm ceramics with an improved relative intensity of emission band at 646 nm. These results indicate that Sm³⁺-doped BNBT ceramics show great potential in electro-optic integration and coupling device applications.

Lithium dendrite growth due to uneven electrodeposition may penetrate the separator and solid electrolyte, causing inner short circuit and potential thermal runaway. Despite great electrochemical phase-field simulation efforts devoted to exploring the dendrite growth mechanism under the temperature field, no unified picture has emerged. For example, it remains open how to understand the promotion, inhibition, and dual effects of increased temperature on dendrite growth when using different electrolyte types. Here, by comprehensively considering the temperature-dependent Li⁺ diffusion coefficient, electrochemical reaction coefficient, and initial temperature distribution in phase-field model, we propose that the activation–energy ratio, defined as the ratio of electrochemical reaction activation energy to electrolyte Li⁺ diffusion activation energy, can be used to quantify the effect of temperature on dendrite morphology. Specifically, we establish a mechanism diagram correlating the activation–energy ratio, uniform initial temperature, and maximum dendrite height, which unifies the seemingly contradictory simulation results. Furthermore, results based on nonuniform initial temperature distribution indicate that a positive temperature gradient along the discharging current facilitates uniform Li⁺ deposition and local hotspot should be avoided. These findings provide valuable insights into the temperature-dependent Li dendrite growth and contribute to the practical application of Li metal batteries.

Propylene carbonate (PC)-based electrolytes have exhibited significant advantages in boosting the low-temperature discharging of graphite-based Li-ion batteries. However, it is still unclear whether they can improve the charging property and suppress lithium plating. Studying this topic is challenging due to the problem of electrochemical compatibility. To overcome this issue, we introduced graphite with phase defects. The results show that the pouch-type full batteries using PC-based electrolyte exhibit steady performance over 500 cycles and can be reversibly charged over 30 times at −20 °C with an average Coulombic efficiency of 99.95%, while the corresponding value for the conventional ethylene carbonate (EC)-based electrolyte sample is only 31.20%. This indicates that the use of PC-based electrolyte significantly suppresses lithium plating during low-temperature charging. We further demonstrate that the improved performance is mainly attributed to the unique solvation structure, where more ${\mathrm{P}\mathrm{F}}_6^{-}$ anions participate in solvation, leading to the formation of a stable F-rich solid state electrolyte interface on the graphite surface and a lower reduction tendency of Li⁺ ions. This work inspires new ideas for the design of PC-based electrolytes for low-temperature charging and fast-charging batteries.

Inorganic–polymer composite solid electrolytes (IPCSEs) obtained by filling the polymer matrix with inorganic materials usually have higher ionic conductivity compared with individual phases. This important increase in ionic conductivity is explained in terms of the new percolation paths formed by the highly conductive interface between inorganic filler and polymer. The conduction in such systems can be investigated using the effective medium theory (EMT) and random resistance model (RRM). EMT can be used to analyze the effect of filler size on the ionic conductivity of disordered IPCSEs, while RRM can describe the composites with inorganic fillers of various shapes (nano-particles, nano-wires, nano-sheets, and nano-networks) in ordered or disordered arrangement. Herein, we present software evaluating the ionic conductivity in IPCSEs by combining EMT and RRM. The approach is illustrated by considering the size, shapes, and arrangements of inorganic fillers. The ionic conductivities of different types of IPCSEs are predicted theoretically and found in good agreement with the experimental values. The software can be used as an auxiliary tool to design composite electrolytes.

Distortion manipulation emerges as an efficient approach to obtain desired perovskite phases for various applications. In part Ⅰ of this study, we propose a paradigm to quantify the structural distortion manipulation, which enables us to obtain desired perovskite phases by translating relevant materials research into a single mathematical question. As part Ⅱ of this continuous study, we construct normalized structures by introducing all possible couplings of dominant distortions into a cubic supercell and then compare them with variously shaped primitive/conventional cells known in the database. The structure comparison demonstrates that distortions are the only cause for phase and property variations. This confirms that our proposed distortion parameters can be directly used to construct phases, providing theoretical support for the paradigm in Part Ⅰ. Given the limited number of distortion types, we identify that the positional relations involved in distortion arrangements and couplings are the keys to describe numerous phases. Furtherly, a three-step workflow is proposed with core contents related to the positional relation, distortion hierarchy, and distortion-component-generation ordering in spatial dimension, respectively. The definition basis and value changes of distortion/model parameters in this workflow illustration provide guidelines about how to reveal the logic behind the perovskite phase evolution.