Inverse Design of Solid-State Materials via a Continuous Representation

The non-serendipitous discovery of materials with targeted properties is the ultimate goal of materials research, but to date, materials design lacks the incorporation of all available knowledge to plan the synthesis of the next material. This work presents a framework for learning a continuous representation of materials and building a model for new discovery using latent space representation. The ability of autoencoders to generate experimental materials is demonstrated with vanadium oxides via rediscovery of experimentally known structures when the model was trained without them. Approximately 20,000 hypothetical materials are generated, leading to several completely new metastable V_xO_y materials that may be synthesizable. Comparison with genetic algorithms suggests computational efficiency of generative models that can explore chemical compositional space effectively by learning the distributions of known materials for crystal structure prediction. These results are an important step toward machine-learned inverse design of inorganic functional materials using generative models. While a traditional strategy for materials design has been to use chemical intuition and empirical rules, combining it with data science and machine learning can significantly expand the search space and accelerate the new discovery. Machine-learning models in materials science have been most extensively developed to predict properties of candidate materials, which still requires the selection of candidates. Inverting the role of machine learning to generate a candidate material with selected properties requires development of generative models for materials, as demonstrated herein. The inverse design pipeline for inorganic solids presented here is based on an invertible image-based featurization and is applied to find new crystal polymorphs of vanadium oxides. This proof-of-concept demonstration opens a great possibility of inverse designing new inorganic solid-state functional materials with desired properties. The inverse design of new materials with desired properties is the ultimate goal of materials research, but demonstrating such a possibility for inorganic solid-state materials has been challenging, due partly to the invertibility of representation. Here, we demonstrate that the generative model using invertible image-based representation yields accurate reconstruction performance and can successfully rediscover experimentally known vanadium oxides. The model predicts several completely new compositions and polymorphs of vanadium oxides that are metastable and may be synthesizable.