MXgap: A MXene Learning Tool for Bandgap Prediction

Dataset used to train the models developed in the manuscript titled "MXgap: A MXene Learning Tool for Bandgap Prediction". These trained models are available to use within the MXgap program, with more details provided in the corresponding MXgap GitHub repository.

In this dataset is available the full list of features and targets for 4356 different MXene structures, accounting for elemental, structural and PBE DOS features, along with the target PBE0 bandgap properties. The MXene (Mn+1XnT2) structures include combinations of M = Sc, Y, Ti, Zr, Hf, V, Nb, Ta, Cr, Mo, or W; X = C or N; T = F, Cl, Br, I, O, S, Se, Te, H, OH, or NH; with n = 1–3 and six geometries per MXene. The dataset also includes a dedicated sheet for La-based MXenes and an "info" sheet explaining all features.