The Synthesis Genome: Data Mining for Synthesis of New Materials

Interpretable  machine-learning  (ML)  models  were  developed  to predict  two  key  solid-state synthesis  conditions that  must  be specified for any reaction: heating temperature and heating time.

 

Our ML results achieve a mean absolute error (MAE) of  140 °C ∼for heating temperature prediction. The predicted synthesis time is typically within a factor of 2 of the reported value.

 

Heating  temperature  prediction  is  dominated  by  precursor properties, which we hypothesize to be linked to reaction kinetics. Heating  time  prediction  is  dominated  by  experimental  operations, which may be indicative of human selection bias.

 

Moreover, our model formally proves that Tamman’s rule, which was firstly applied to alloys, can be extended to ionic compounds as well.

Designing Materials to Revolutionize and Engineer our Future (DMREF)