Precursor recommendation for inorganic synthesis 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 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 2654 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.

Stability and Equilibrium Structures of Unknown Ternary Metal Oxides Explored by Machine-Learned Potentials

“…We also note that our definition of synthesizability does not take into account kinetic aspects and synthetic pathways. Although accounting for these factors would yield a more comprehensive analysis, it currently remains an ongoing challenge in the field of materials science. , It is found that 45 systems contain 48 stable compounds with negative E hull , where 20 compounds in 19 systems are highly stable ( E hull < −50 meV atom –1 ). Most of the newly discovered materials contain noble metals (35 systems), with Au being the most abundant (13 systems) (see the blue bars at the top of Figure ).…”

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

“…The former is partially addressed by the plethora of ML models for predicting ground state thermochemistry, along with a proper accounting for metastability . 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 . (A parallel discussion of these ideas as they apply to organic and medicinal chemistry can be found in refs and .…”

Soft Chemistry of Hard Materials: Low-Temperature Pathways to Bulk and Nanostructured Layered Metal Borides

“…As Tom Mallouk stated in an early 1990s review article, many solid-state materials are synthesized “the old-fashioned way (by accident).” Similarly, in 2015, C. N. R. Rao commented, “While one can evolve a rational approach to the synthesis of solid materials, there is always an element of serendipity, encountered not so uncommonly.” Every year the solid-state reaction toolbox grows. Thanks to thorough experimental investigations and developments in theory, − our understanding of the driving forces behind solid-state reactions is increasing. − When planning a synthesis, solid-state chemists routinely draw on their deep understanding of the reactivities of the elements and their compounds as well as predictions of how new compounds may form based on their targeted structures and compositions. Despite this, the statements by Mallouk and Rao remain relevant.…”