This review revisits the state of the art of research efforts on the design of mechanical materials using machine learning.

Interface interaction can strongly modify contact angle, adsorption energy, interfacial tension, and composition of the contact area. In particular, the interfaces between gallium-based liquid metal (LM) and its intermetallic layer present many mysterious and peculiar wetting phenomena, which have not been fully realized up to now. Here in this study, we found that a gallium-based liquid metal droplet can quickly transform into a puddle on the CuGa surface through a spreading-wetting procedure. The mechanism lying behind this phenomenon can be ascribed to the formation of an intermetallic CuGa on Cu plate surface, which provides a stable metallic bond to induce the wetting behavior. For a quantitative evaluation of the interface force, a metallic bond-enabled wetting model is established on the basis of the density functional theory. The first-principles density functional calculations are then performed to examine the work function, density of states, and adsorption energy. The predicted results show that the work function of CuGa (010) is approximately 4.47 eV, which is very comparable with that of pure liquid Ga (4.32 eV). This indicates that the valence electrons between Ga and CuGa slab can exchange easily, which consequently leads to the strong valence electron hybridization and metallic bond. In addition, the adsorption energy of a single Ga atom on CuGa (010) slab has a larger value than In and Sn. The tested metallic bond wetting force at the interface is proportional to the average adsorption energy of the gallium-based LM adatom, and increases with the rising content of gallium. The simulation results demonstrate excellent consistency with the experimental data in this work.

Materials-by-design is a paradigm to develop previously unknown high-performance materials. However, finding materials with superior properties is often computationally or experimentally intractable because of the astronomical number of combinations in design space. Here we report an AI-based approach, implemented in a game theory–based conditional generative adversarial neural network (cGAN), to bridge the gap between a material’s microstructure—the design space—and physical performance. Our end-to-end deep learning model predicts physical fields like stress or strain directly from the material microstructure geometry, and reaches an astonishing accuracy not only for predicted field data but also for derivative material property predictions. Furthermore, the proposed approach offers extensibility by predicting complex materials behavior regardless of component shapes, boundary conditions, and geometrical hierarchy, providing perspectives of performing physical modeling and simulations. The method vastly improves the efficiency of evaluating physical properties of hierarchical materials directly from the geometry of its structural makeup.

No abstract

No abstract