Inorganic synthesis recommendation by machine learning materials similarity from scientific literature

Abstract

Synthesis prediction is a key accelerator for the rapid design of advanced materials. However, determining synthesis variables such as the choice of precursor materials, operations, and conditions is challenging for inorganic materials because the sequence of reactions during heating is not well understood. In this work, we use a knowledge base of 29,900 solid-state synthesis recipes, text-mined from the scientific literature, to automatically learn which precursors to recommend for the synthesis of a novel target material. The data-driven approach learns chemical similarity of materials and refers the synthesis of a new target to precedent synthesis procedures of similar materials, mimicking human synthesis design. When proposing five precursor sets for each of 2,654 unseen test target materials, the recommendation strategy achieves a success rate of at least 82%. Our approach captures decades of heuristic synthesis data in a mathematical form, making it accessible for use in recommendation engines and autonomous laboratories. Short title AI learning inorganic synthesis from literature TeaserDecades of heuristic data are automatically captured for guiding successful synthesis of novel inorganic materials.

In Pursuit of the Exceptional: Research Directions for Machine Learning in Chemical and Materials Science

“…35 The latter is partially addressed by ML approaches that use natural language processing on the literature to extract experiment plans (for training) and then generate plans based on that data. 161 (A parallel discussion of these ideas as they apply to organic chemistry can be found in Ref. 157 .)…”