(17b) Solid-State Synthesis Planning for Inorganic Materials

High-throughput simulations have become valuable tools for materials discovery, but the process of realizing hypothetical materials from these simulations in either the laboratory or in manufacturing remains slow and unaided by theory. We present a data-driven theoretical framework and corresponding software implementation to assist solid-state synthesis. This software, termed "piro", identifies optimal reaction routes to a given inorganic solid by comparing tradeoffs between estimated catalytic nucleation routes and potential parasitic reactions. This enables both explicit predictions of which synthesis are most likely to be successful and detailed information on either the relative rate or selectivity bounds of any reaction in a given chemistry. We benchmark our tool using reported syntheses from the literature for popular functional materials and demonstrate that piro recovers most known syntheses and rationalizes the potential failure of those that have not been reported. Lastly, we discuss pathways to enable full retrosynthetic analysis of inorganic compounds or arbitrary complexity.