In Mg anode materials design, the second phase plays a significant role in influencing the alloy's properties. This paper proposes a fusion model based on graph neural networks (GNNs) aimed at optimizing alloy design by considering the impact of the second phase in the alloy. We calculate the types and mass percentages of the thermodynamically stable second phases in the alloy. A GNN-based fusion model is then used to capture the interactions between the second phases within the alloy to predict the voltage of the magnesium anode. Our model outperforms the traditional MLP model. The average error is reduced from 0.096 to 0.075 V, representing a 21.8% decrease in error. We use Bayesian optimization and Pareto front analysis for multi-objective optimization. The Mg-2Sr alloy, which has high voltage values at 5, 10 and 20 mA cm⁻², is selected for experimental validation. The GNN fusion model improves the accuracy of the model, providing a more scientific basis for the design and optimization of magnesium alloy anodes.

Hf₆Ta₂O₁₇ ceramic has outstanding thermal and mechanical properties, along with an extremely high phase transition temperature (2517 K), making it a promising candidate for next-generation thermal barrier coatings (TBCs). High-temperature corrosion of TBCs by calcium-magnesium-alumino-silicates (CMAS) is a major failure mode. To understand why Hf₆Ta₂O₁₇ ceramic resists CMAS corrosion well at high temperatures, the CMAS/Hf₆Ta₂O₁₇(010) system was analyzed via density functional theory to study CMAS – induced corrosion processes. Results show that above 1523 K, the diffusion coefficient of atoms in CMAS exceeds that in Hf₆Ta₂O₁₇(010). Also, there is significant mutual diffusion between (Ca, Si) and (Hf, Ta) with a low diffusion activation energy. The interaction between CMAS and Hf₆Ta₂O₁₇ has a low reaction energy, enabling the quick formation of dense corrosion products (CaTa₂O₆ and HfSiO₄) at the interface. These products have high phase stability and fast formation rates, remaining intact when in contact with residual CMAS. The interface-formed corrosion products greatly reduce the CMAS diffusion coefficient into Hf₆Ta₂O₁₇; for instance, calcium, the fastest-diffusing element, has its diffusion coefficient reduced by over five times. These mechanisms effectively limit CMAS corrosion of Hf₆Ta₂O₁₇, enhancing its potential as an ultrahigh temperature TBCs for the next generation.

In this study, the knowledge-constrained symbolic regression method was used to predict the creep life of Ni-based superalloys. Two forms of prediction formulas that can explain the creep mechanism of Ni-based superalloys were successfully constructed based on high-throughput data-driven approaches combined with machine learning algorithms. Through the selection and calculation of characteristic parameters, the integration factors VγʹTγʹ and 1/Γ were surprisingly found, which indicated the importance of γ′ phase strengthening and dislocation strengthening for the creep. Finally, the models were verified by experimental data, indicating that the prediction effect is excellent. It is notable that the models offer three key advantages: accurate creep life prediction, visual form, and interpretable mechanism.

The development of novel refractory high-entropy alloys (RHEAs) holds significant promise for advanced applications due to their exceptional properties. However, identifying optimal compositions of RHEAs within the vast alloy design space to meet specific property requirements remains a formidable challenge. In this study, we present an integrated machine learning (ML) framework to address this challenge, combining predictive models for material properties, a fingerprint map of composition distribution, a guided multiobjective search strategy, and a particle swarm optimizer to enable targeted exploration of promising RHEAs compositions. Using this approach, we successfully discovered several new RHEAs with outstanding mechanical performance, including Nb_0.189Ti_0.203V_0.203Mo_0.206Zr_0.197, Nb_0.204Ti₀₁₉V_0.207Mo_0.198Zr_0.198, Nb_0.174Ti_0.19V_0.251Mo_0.201Zr_0.181, Nb_0.242Ti_0.252To_0.001V_0.039Mo_0.209Zr_0.254, and Nb_0.164Ta_0.155Ti_0.186V_0.008W_0.153Mo_0.001Hf_0.168Zr_0.16. These alloys exhibit remarkable yield strengths ranging from 1580 to 1740 MPa and fracture strains between 23% and 27%. The integrated ML models make it possible to rapidly optimize multiple properties during other materials designing, thus overcoming the common problems of limited data and a vast composition space in complex materials systems, paving the way for efficient design of advanced materials tailored to diverse application requirements.

Ionic conductivity is a critical determinant of electrolyte performance in lithium-ion batteries, governing functionalities such as rate capability and low-temperature operability. Conventional optimizations, empirical or simulation-based, face significant limitations in either resource efficiency or predictive accuracy. To address these challenges, we developed an interpretable machine learning (ML) framework that combines least absolute shrinkage and selection operator (LASSO) regression with SHapley Additive exPlanations analysis to elucidate structure–property relationships in multicomponent electrolytes. This framework proposes a novel descriptor, model-input-weighted sum of LASSO features, which quantitatively captures the collective influence of molecular characteristics on ionic conductivity. Our approach achieves state-of-the-art predictive accuracy (RMSE = 1.33 mS cm⁻¹, R² = 0.88) while identifying two dominant molecular features: PEOE_VSA1, representing surface charge distribution, and NumAtomStereoCenters, reflecting stereochemical complexity. This led to the design of an optimized ternary electrolyte (1 mol L⁻¹ LiTFSI in MA:THF:DMF, 5:3:2 molar ratio) demonstrating unprecedented conductivity values: 15.74 mS cm⁻¹ at 25°C and 2.69 mS cm⁻¹ at −70°C. These results validate our framework's ability to guide the development of high-performance electrolytes for low-temperature applications. This study establishes a robust ML framework for accelerated electrolyte discovery, providing fundamental insights into molecular determinants of ionic conductivity.

Ti-Al-based intermetallic compounds are promising candidates for high-temperature structural applications owing to their outstanding mechanical properties. Ti₂AlNb alloys, characterized by complex multiphase microstructures, present significant challenges for hot deformation modeling because of their atypical flow behavior and sensitivity to processing conditions. In this study, we systematically investigated the hot deformation behavior of Ti₂AlNb through experiments and compared conventional constitutive models with advanced machine learning approaches. The conventional strain-compensated Sellars (SCS) model showed limited accuracy for Ti₂AlNb, especially across complex microstructural transitions, while performing well for simpler alloy systems like Ti4822. To address these limitations, we developed a dynamic physics-guided neural network (DPGNN) that integrates physical constraints with data-driven learning via an adaptive gating mechanism. The DPGNN model significantly outperformed the SCS model and three purely data-driven baselines, achieving high accuracy (test R² > 0.98) and robust generalization across both Ti₂AlNb and Ti4822 alloys. These findings highlight the value of embedding physical principles within machine learning frameworks, providing a robust and generalizable tool for predicting hot deformation behavior in advanced alloys.

Precise tuning of dielectric constants (ε_r) in oxide glasses is critical for high-frequency devices in 5G/6G systems, where ε_r directly governs signal propagation efficiency. A machine learning framework combining data augmentation and physicochemical descriptor integration is developed to address data scarcity. Validated pseudo-labels are generated via ensemble learning, expanding the dataset from 1503 to 11,029 compositions without distributional shift. The XGBoost model trained on the augmented dataset achieved superior accuracy, with an R² of 0.96 and an MSE of 0.14. For prediction tasks on unseen data, it reduced the error rate by 48% compared to the non-augmented model and improved generalization performance by 43% over GlassNet. B₂O₃ and SiO₂ are identified as ε_r suppressors and BaO and TiO₂ as enhancers through SHAP analysis, aligning with network former/modifier roles. Cation-specific polarizabilities are derived via Clausius–Mossotti regression (R² = 0.909). Integration of physicochemical descriptors (coordination number and bond strength) enables transferable predictions for Y₂O₃ and La₂O₃ containing glasses, with mean deviation 2.46%–4.76%. Crucially, structural descriptors dominate polarizability with 69.9% feature importance, establishing network engineering as the optimal design paradigm. A data-driven pathway for rational dielectric glass development is thus established.

Traditional methods for material discovery and optimization are time-consuming and resource-consuming. Recent advancements in artificial intelligence (AI), particularly machine learning, offer a revolutionary opportunity for accelerating novel material discovery. This review overviews AI enhancement on high-throughput synthesis and screening methods for faster and more efficient material discovery, focusing on electrocatalysis and energy storage materials. The integration of AI with autonomous laboratories allows real-time data analysis and closed-loop optimization, accelerating material characterization and analysis. Despite challenges in data quality and model transparency, integration of AI with experimental workflows significantly advances materials science.

In this work, seven RE–Cu (RE = Ce, Pr, Sm, Eu, Tb, Er and Lu) binary systems were optimized using the CALPHAD method based on the reported experimental data. The liquid phase and terminal solid solution phases were modeled using the substitutional solution model, while the intermetallic compounds were treated as stoichiometric compounds. Self-consistent thermodynamic parameters of seven RE–Cu binary systems were obtained, which can be used to reproduce the experimental results including phase diagram and thermodynamic properties. Furthermore, in combination with the reported assessments of six RE–Cu (RE = La, Nd, Gd, Dy, Ho, and Yb) binary systems, phase equilibria, and thermodynamic properties of 13 RE–Cu (RE = La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Yb, and Lu) binary systems were examined systematically. Generally, it was observed that as the RE atomic number increases, the formation temperatures of the RE–Cu intermetallic compounds increase gradually, and the enthalpy of mixing of liquid RE–Cu alloys and the enthalpy of formation of the RE–Cu intermetallic compounds become increasingly negative. These results provide a comprehensive set of thermodynamic parameters for the RE–Cu binary systems, which will serve as a crucial foundation for developing a thermodynamic database of RE–Cu-based alloys.

Bipolar electrochemistry is a high-throughput corrosion testing method capable of applying a quasi-linear potential gradient across test specimens. This study employs—bipolar electrochemistry corrosion testing to investigate the influence of gravity on pitting corrosion of type 304L and 420 stainless steel across a broad range of applied potentials. Gravity modifies the distribution of current density on the bipolar electrode without altering the potential distribution. The impact of gravity on pitting corrosion is achieved through its effects on the dilution of the electrolyte and the removal of the salt film within the pits. Pits oriented in a face up position demonstrate smoother morphologies, larger cross-sectional areas and pit volumes. In contrast, pits oriented in perpendicular and facedown positions exhibit pit shape. Under conditions governed by diffusion and activation control, pits can up to over 100 μm. Additionally, crystallographic pits are observed to form in areas subjected to high applied potentials.

High-performance light alloys, including aluminum, titanium, magnesium alloys, etc., are utilized in aerospace, aviation, transportation and medical applications. A key challenge for these alloys is achieving both improved strength and stress corrosion cracking (SCC) resistance by optimizing the relationships between composition, processing, microstructure, and macroscopic properties. Artificial intelligence (AI)-driven multi-modal machine learning offers opportunities for materials design and prediction. Proposed strategies include applying machine learning-based approaches for concurrent improvement of alloy strength and SCC resistance, conducting in situ high-throughput experiments to investigate SCC microcrack initiation mechanisms under combined mechanical, microstructural, and corrosion conditions to support database development and developing correlative AI models for alloy microstructure evolution and macroscopic SCC failure behavior in service environments.