Tuning Mechanical Properties in Polycrystalline Solids Using a Deep Generative Framework

A framework for inverse design of tuning mechanical properties polycrystalline brittle materials is presented using artificial intelligence (AI). Crystalline solids, which often exhibit distinct at different orientations, can be used as building blocks composites. However, the space geometry and crystal misorientations typically intractable, all possible solutions cannot discovered experiment or numerical simulation. Herein, a deep learning (DL) alongside genetic algorithm (GA) adopted to generate composite materials, whereas raw material sensitive crystalline orientation, achieve in various combinatorial designs. The DL model, trained by full-atomistic simulations crystals with evolves autonomously yield desirable range target toughness values, exemplified maximizing minimizing toughness, are validated molecular dynamics (MD) simulations. It found that higher preferred high opposed lower overall low-toughness Notably, this shows mechanism extracted from AI algorithm. This materiomics method may ultimately change way nanomaterials designed, applied de novo biomaterial design, architected bioinspired structural materials.