Virtual screening of inorganic materials synthesis parameters with deep learning. (December 2017)

• Record Type: Journal Article
• Authors: Kim, Edward; Huang, Kevin; Jegelka, Stefanie; Olivetti, Elsa
• Is Part Of: Npj computational materials. Volume 3:issue 1(2017)
• Page Start: 1
• Page End: 9
• Publication Date: 2017-12
• Subjects: Materials science -- Computer simulation -- Periodicals; Materials science -- Mathematical models -- Periodicals; Materials science -- Computer simulation; Electronic journals; Periodicals; 620.110285
• DOI: 10.1038/s41524-017-0055-6
• Languages: English
• ISSNs: 2057-3960

Abstract

Abstract Virtual materials screening approaches have proliferated in the past decade, driven by rapid advances in first-principles computational techniques, and machine-learning algorithms. By comparison, computationally driven materialssynthesis screening is still in its infancy, and is mired by the challenges of datasparsity and datascarcity : Synthesis routes exist in a sparse, high-dimensional parameter space that is difficult to optimize over directly, and, for some materials of interest, only scarce volumes of literature-reported syntheses are available. In this article, we present a framework for suggesting quantitative synthesis parameters and potential driving factors for synthesis outcomes. We use a variational autoencoder to compress sparse synthesis representations into a lower dimensional space, which is found to improve the performance of machine-learning tasks. To realize this screening framework even in cases where there are few literature data, we devise a novel data augmentation methodology that incorporates literature synthesis data from related materials systems. We apply this variational autoencoder framework to generate potential SrTiO3 synthesis parameter sets, propose driving factors for brookite TiO2 formation, and identify correlations between alkali-ion intercalation and MnO2 polymorph selection. Machine learning: Computer-generated recipes for materials synthesis Recipes for synthesizing inorganic materials can now be generated by machine learning. Elsa Olivetti and coworkers at the Massachusetts Institute of Technology have developed an algorithm capable of 'learning' new procedures for synthesizing compounds in the lab. 'Virtual screening' for synthetic procedures would be a useful tool to aid synthetic chemists, however the field is still nascent with progress hindered by the often complex nature of materials synthesis. Here, the authors simplified the problem by representing syntheses by some key parameters (e.g. reaction temperature), which were used to train the algorithm. As a proof-of-principle, Olivetti used the method to generate synthesis parameters for strontium titanate, as well as revealing insights into the formation of other inorganic oxides. The authors hope their method may eventually lead to a predictive 'virtual screening' approach to inorganic materials synthesis.