Jianxin Xie

Reliable service of lightweight high-strength alloys in energy efficient structures is challenged by hydrogen embrittlement. Extensive research efforts have been devoted to uncovering the mechanism underlying the phenomenon for decades, primarily focusing on the interaction between hydrogen and dislocations. Nonetheless, the explicit role of hydrogen in the direct vicinity of a propagating crack tip, wherein the deformation carriers nucleate, remains obscure. Here, by using in-situ high-resolution environmental transmission electron microscopy, we directly capture hydrogen-enhanced diffusive deformation-mediated crack propagation in pure Al, manifesting vacancy formation and atomic diffusion on the crack tip surface. This is corroborated by density functional theory calculations revealing considerably weakened Al-Al bonds at the adsorption sites and lowered energy barriers for vacancy formation and adatom diffusion. This diffusive cracking pathway is applicable under high hydrogen fugacity and low strain rates, which imposes evident effects on bulk materials. The findings point to restraining atomic diffusion to alleviate the hydrogen-assisted cracking in advanced aluminum alloys.

Cracking, a major challenge for nickel-based superalloys processed via laser powder bed fusion (L-PBF), is highly sensitive to laser scanning parameters. Conventional parameter optimization approaches employing a trial-and-error method require a significant number of experimental trials. In this work, we adopted a machine learning (ML)-assisted strategy to optimize the L-PBF parameters of a typical crack-susceptible superalloy, CM247LC, with the aim of both inhibiting cracking and enhancing the strength-ductility trade-off. An L-PBF processing parameter-crack density dataset was generated through orthogonal experiments. Next, a machine learning-based crack prediction model was constructed, combined with data augmentation and Bayesian optimization to enhance predictive accuracy and global exploration capacity. Within two iterations, the crack density was significantly reduced by 99% compared to the best case in the orthogonal design, achieving nearly crack-free samples. Microstructure analyses revealed that the enhanced cracking resistance under ML-optimized parameters was primarily due to reduced grain boundary segregation of Hf and C, as well as decreased residual stresses resulting from lower thermal gradients. At 900 ℃, compared with cast CM247LC, the L-PBF-processed alloy in this study exhibited a 23% increase in ultimate tensile strength (937 ± 8 MPa), a 35% increase in yield strength (809 ± 5 MPa), and comparable elongation (10.8 ± 0.4%). The machine learning-based L-PBF process optimization model proposed in this work provides an efficient and economical approach for crack mitigation in crack-susceptible alloys processed via L-PBF.

Extracting data from scientific literature and constructing machine learning-ready (ML-ready) datasets is crucial for data-driven materials discovery. However, material properties under various compositions and processing conditions are often summarized in figures, making manual data extraction time-consuming and labor-intensive. In this study, we propose an automated figure data extraction tool to rapidly generate datasets from figures in materials science literature on Cu-Cr-X alloys used in high-end lead frames, aiming to support the development of next-generation high-strength, high-conductivity copper alloys. We collected 251 figures containing mechanical and electrical performance data from 146 papers and automatically extracted 3,018 high-quality Cu-Cr-X data records. Each record includes article metadata and composition/processing information manually extracted from text, enabling construction of a domain-specific dataset linking alloy composition, processing, and properties. The dataset summarizes 20 microalloying elements and mechanical and electrical performance under different processing conditions. This tool can significantly improve the efficiency of large-scale data accumulation in materials science.

Large lossless currents in high-temperature superconductors (HTS) critically rely on dense defects with suitable size and dimensionality to pin vortices, with dislocations being particularly effective due to their one-dimensional geometry to interact extensively with vortex lines. However, in non-metallic compounds such as HTS with rigid lattices, conventional deformation methods typically lead to catastrophic fracture rather than dislocation-mediated plasticity, making it a persistent challenge to introduce dislocations at high density. Here, we propose an asymmetric stress field strategy using extrusion to directly nucleate a high-density of dislocations in HTS by activating shear-driven lattice slip and twisting under superimposed hydrostatic compression. As demonstrated in iron-based superconductors (IBS), atomic displacements of nearly one angstrom trigger the formation of tilted dislocation lines with a density approaching that of metals. With further structural refinement, these dislocations serve as strong pinning centers that lead to a fivefold enhancement in the current-carrying capacity of IBS at 33 T, along with low anisotropy and a large irreversibility field. This work not only establishes a scalable route to engineer pinning landscapes in HTS, but also offers a generalizable framework for manipulating dislocation structures in rigid crystalline systems.

In recent years, data‐driven machine learning has significantly advanced the design of new materials and transformed the research and development landscape. However, its heavy reliance on data and the “black‐box” nature of its model‐mapping mechanisms have hindered its application in materials science research. Integrating material knowledge with machine learning to enhance model generalization and prediction accuracy remains an important objective. Such integration can deepen the understanding of material mechanisms by screening physical and chemical features to uncover explicit intrinsic relationships. Thus, it promotes the advancement of materials science, representing a promising avenue for artificial intelligence (AI) applications in this field. In this review, the algorithms, functionalities, and applications in materials underlying interpretable machine learning approaches are summarized and analyzed. The impact of composition and microstructure on material properties is explored and mathematical expressions for intrinsic relationships of materials are developed. In addition, recent advancements in data‐ and knowledge‐driven strategies for new material discovery, key property enhancement, multi‐objective design trade‐offs, and optimizing the entire preparation and processing workflow are reviewed. Finally, the future prospects and challenges associated with applying AI in materials science and its broader implications for the field are discussed.

A solid/liquid continuous composite casting technology was developed to produce brass-clad copper stranded wire billets efficiently with continuous casting speeds ranging from 200 mm/min to 1000 mm/min. As the casting speed increased, the microstructure of the brass cladding transformed at an angle to the radial direction. The wire billet prepared at a casting speed of 600 mm/min was then subjected to drawing. As the percentage reduction in area of the billet increased from 11.9 to 81.5% during the drawing process, the tensile strength improved from 336 MPa to 534 MPa, while the elongation after fracture decreased from 30.1 to 4.7%. Meanwhile, dislocation, dislocation cells, and microbands successively formed in the pure copper strand wires, while twins, shear bands, dislocation pile-ups, and secondary twins gradually formed in the brass cladding. During the drawing process, the interface between copper and brass remained metallurgically bonded, exhibiting coordinated deformation behavior. This paper clarified the evolution of microstructure and mechanical properties of brass-clad copper stranded wires in high-speed solid/liquid continuous composite casting and drawing, which could provide important reference for industrial production.

The extensive use of scarce, expensive, and toxic elements in high‐performance metal alloys restricts their sustainable development. Here we propose a novel alternative alloying‐element design strategy that combines physicochemical‐factor screening, a “black‐box” interpretative method based on SHapley Additive exPlanation analysis, and sensitivity analyses of elemental influence. A “white‐box” model of alloy compositions and properties is therefore established that enables the rational selection of abundant elements and the efficient designs of alloys with substitution for scarce alloying elements. The success of this design strategy is demonstrated by reducing the Co content in the C70350 alloy series (e.g., Cu‐1.3Ni‐1.4Co‐0.56Si‐0.03Mg (wt.%)). Indeed, Cu‐1.95Ni‐0.5Co‐0.6Si‐0.2Mg‐0.1Cr (wt.%) is obtained as an ultra‐low‐Co‐containing alloy by substituting only a trace amount of Cr for Co in the Cu‐Ni‐Co‐Si‐Mg system, followed by compositional optimization. Although the Co content is reduced by 64% (i.e., from 1.4 to 0.5 wt.%), the properties of the alloy (ultimate tensile strength and electrical conductivity of 850 MPa and 47.2%IACS, respectively) are comparable to those of the C70350 alloy. This study contributes to the sustainable and green development of metallic materials by providing a new avenue for substituting abundant elements for scarce and undesired elements in metal alloys.