smact.dopant_prediction module#
The dopant prediction module facilitates high-throughput prediction of p-type and n-type dopants in multi-component solids. Candidates are filtered on accessible oxidation states, an n-type dopant being required to carry a higher charge than the host ion it replaces and a p-type dopant a lower one, with the sign of the site preserved. Survivors are scored using a data-mined substitution table or a species embedding. Scores for cation sites are combined by default with a selectivity term; anion substitutions take a fixed selectivity of 1.0 and so are ranked on that score alone. No ionic-radius criterion is applied; smact.dopant_prediction.doper explains why, and where to look if you do have a structure.
Doper, A class to identify possible dopants.get_dopants()returns a dictionary keyed by “n-type cation substitutions”, “p-type cation substitutions”, “n-type anion substitutions” and “p-type anion substitutions”.
See Using doper for a worked example, covering the default lambda table, a species embedding and the table and heatmap outputs.