Jie Xiong

Overcoming the strength-ductility trade-off remains a central challenge in structural materials. Here we report an optimized thermomechanical-processing state of AlCoCrFeNi _2.5 high-entropy alloy, identified with the assistance of an alloy trajectory optimization network. The optimized alloy exhibits an exceptional combination of a yield strength of 1.6 GPa, an ultimate tensile strength of 1.8 GPa, and a large fracture elongation of 20% at room temperature. At 650 ^o C, it retains a yield strength of 1.2 GPa, an ultimate tensile strength of 1.4 GPa, and a fracture elongation of 14%. The unprecedented mechanical properties originate from a unique heterogeneous microstructure consisting of spindle-shaped FCC domains, fine lamellar regions, and recrystallized dual-phase regions. This architecture enables cooperative deformation among ductile domains, strengthening lamellae and recrystallized dual-phase regions, thereby promoting appropriate strain partitioning, sustained work hardening and delayed damage localization. These results demonstrate that thermomechanical-processing optimization can transform the complex eutectic architecture into a mechanically cooperative microstructural state with superior strength, ductility, and intermediate-temperature resistance.

Eutectic high‐entropy alloys (EHEAs) exhibit asynchronous phase evolution under identical thermal conditions, yet the underlying mechanisms remain unclear. Here, we investigated the isothermal annealing of AlCoCrCuFeNi alloy using phase‐resolved experiments and atomistic simulations. Electron microscopy revealed sharply contrasting structural pathways: the initially disordered BCC‐FeCr phase progressively developed long‐range order, whereas the B2‐NiAl phase retained its macroscopically ordered framework while undergoing local structural relaxation. Molecular dynamics and Monte Carlo simulations traced this discrepancy to phase‐dependent defect energetics. In B2‐NiAl, low vacancy formation and migration energies enable efficient vacancy‐mediated transport, promoting local structural refinement without disrupting the ordered lattice. In BCC‐FeCr, constrained vacancy transport together with accessible low‐energy interstitial configurations gives rise to a cooperative vacancy‐interstitial transport mechanism required for the disorder‐to‐order transition. These heterogeneous defect transport mechanisms govern asynchronous ordering in EHEAs and provide a phase‐resolved basis for designing thermally stable complex alloys.

Refractory high-entropy alloys (RHEAs) are promising for extreme environments, yet the vast non-equiatomic design space hinders discovery of compositions with tailored strength-ductility balance. Here we develop a neuroevolution potential (NEP)-centered, multi-fidelity design framework for Ta–Nb–Mo–V–W RHEAs. The NEP achieves DFT-level accuracy for energies and forces, and its rapidly fitted Young’s modulus agrees with experimental benchmarks. SVR surrogates trained on NEP-labeled elasticity data enable rapid screening of more than 50 thousands compositions within seconds, NEP provides medium-fidelity validation within minutes, and tensile MD offers high-fidelity verification from 300 K–1200 K. Guided by different performance demands, the workflow identifies Ta _0.20 Nb _0.05 Mo _0.35 V _0.05 W _0.35 with superior strength and Ta _0.20 Nb _0.06 Mo _0.34 V _0.34 W _0.06 that sustains a combined strength-toughness advantage over equiatomic TaNbMoVW across 300 K–1200 K. The approach accelerates discovery by coupling models and simulations with complementary speed and fidelity, enabling scalable, data-efficient screening, reliable validation, and temperature-aware optimization across complex compositional spaces.

Efficient exploration of vast compositional and processing spaces is essential for accelerated materials discovery. Bayesian optimization (BO) provides a principled strategy for identifying optimal materials with minimal experiments, yet its adoption in materials science is hindered by implementation complexity and limited domain-specific tools. Here, we present Bgolearn, a comprehensive Python framework that makes BO accessible and practical for materials research through an intuitive interface, robust algorithms, and materials-oriented workflows. Bgolearn supports both single-objective and multi-objective Bayesian optimization with multiple acquisition functions (e.g., expected improvement, upper confidence bound, probability of improvement, and expected hypervolume improvement etc.), diverse surrogate models (including Gaussian processes, random forests, and gradient boosting etc.), and bootstrap-based uncertainty quantification. Benchmark studies show that Bgolearn reduces the number of required experiments by 40-60% compared with random search, grid search, and genetic algorithms, while maintaining comparable or superior solution quality. Its effectiveness is demonstrated not only through the studies presented in this paper, such as the identification of maximum-elastic-modulus triply periodic minimal surface structures, ultra-high-hardness high-entropy alloys, and high-strength, high-ductility medium-Mn steels, but also by numerous publications that have proven its impact in material discovery. With a modular architecture that integrates seamlessly into existing materials workflows and a graphical user interface (BgoFace) that removes programming barriers, Bgolearn establishes a practical and reliable platform for Bayesian optimization in materials science, and is openly available at https://github.com/Bin-Cao/Bgolearn.

Accurate determination of crystal structures is central to materials science, underpinning the understanding of composition–structure–property relationships and the discovery of new materials. Powder X-ray diffraction (XRD) is a key technique in this pursuit due to its versatility and reliability, yet current analysis pipelines still rely heavily on expert knowledge and slow iterative fitting, limiting their scalability in high-throughput and autonomous settings. Here we introduce a physics-guided contrastive learning framework, XRD–Crystal Contrastive Pretraining (XCCP), which aligns powder diffraction patterns with candidate crystal structures in a shared embedding space to enable efficient structure retrieval and symmetry recognition. The XRD encoder employs a dual-expert design with a Kolmogorov–Arnold Network projection head: one branch emphasizes low-angle reflections reflecting long-range order, while the other captures dense high-angle peaks shaped by symmetry. Coupled with a crystal graph encoder, contrastive pretraining yields physically grounded representations. XCCP demonstrates strong performance across tasks, with structure retrieval reaching 88.98% and space group identification attains 93.39% accuracy. The framework further generalizes to compositionally similar multi-principal element alloys and demonstrates zero-shot transfer to experimental patterns. Together, these results establish XCCP as a robust, interpretable, and scalable approach that offers a new paradigm for PXRD analysis, facilitating high-throughput screening, rapid structural validation, and integration into autonomous laboratories.

Accurate determination of crystal structures is central to materials science, underpinning the understanding of composition-structure-property relationships and the discovery of new materials. Powder X-ray diffraction is a key technique in this pursuit due to its versatility and reliability. However, current analysis pipelines still rely heavily on expert knowledge and slow iterative fitting, limiting their scalability in high-throughput and autonomous settings. Here, we introduce a physics-guided contrastive learning framework termed as XCCP. It aligns powder diffraction patterns with candidate crystal structures in a shared embedding space to enable efficient structure retrieval and symmetry recognition. The XRD encoder employs a dual-expert design with a Kolmogorov-Arnold Network projection head, one branch emphasizes low angle reflections reflecting long-range order, while the other captures dense high angle peaks shaped by symmetry. Coupled with a crystal graph encoder, contrastive pretraining yields physically grounded representations. XCCP demonstrates strong performance across tasks, with structure retrieval reaching 0.89 and space group identification attains 0.93 accuracy. The framework further generalizes to compositionally similar multi principal element alloys and demonstrates zero-shot transfer to experimental patterns. These results establish XCCP as a robust, interpretable, and scalable approach that offers a new paradigm for X-ray diffraction analysis. XCCP facilitates high-throughput screening, rapid structural validation and integration into autonomous laboratories.