FIA-GNN

FIA-GNN is a graph neural network that predicts the fluoride ion affinity (FIA) of neutral p-block atom-based molecules in the gas and solution (dichloromethane) phase as defined by the fluorotrimethylsilane anchoring system. 13 different central atom types are included: B, Al, Ga, In, Si, Ge, Sn, Pb, P, As, Sb, Bi, and Te. The atoms of groups 14 and 15 are supported in their high (+IV, +V) and low (+II, +III) oxidation states.

For more details refer to our paper and the GitHub repository .