In this NEDO project, the National Institute of Advanced Industrial Science and Technology (NAIST) has developed a method to predict the yield of catalytic reactions by computer simulation based on catalyst structure and actual yields. In a halogen-free epoxidation reaction, 14 types of phosphonic acid molecules were characterised by atomic charge, infrared absorption and 30 more parameters. These data were correlated with experimental yields of the actual epoxidation reaction in which phosphonic acid was used as a catalyst, and machine learning was applied to correlate computer-based catalyst data and actual yields.. With this forecasting technology, parameters were obtained that contributed to predicting yields.

NEDO news release, January 31, 2018