High-Throughput Screening\nof II-Type/Z-Type Photocatalytic\ng‑GaN Heterojunctions Based on Machine Learning and Density\nFunctional Theory for Photocatalytic Splitting of Water

The direct Z-type van der Waals (vdW) heterojunction\nis perceived\nas a promising photocatalyst, offering more efficient carrier migration\nprocesses compared with conventional II-type vdW heterojunctions.\nNevertheless, the abundance of candidates and the expensive validation\nprocess pose hindrances to the further advancement of direct Z-type\nvdW heterojunctions. Leveraging a two-dimensional materials database,\nwe amassed a set of approximately 100 samples of II/Z-type g-GaN-based\nvdW heterojunctions and innovatively proposed a fusion descriptor.\nBased on this, we developed a binary machine learning model and applied\nthe first-principles nonadiabatic molecular dynamics for theoretical\nverification. The results indicate that the optimal model prediction\naccuracy of the fusion descriptor attains 95%, and the prediction\noutcomes are entirely in line with the theoretical calculation results.\nMeanwhile, based on the idea of high-throughput screening, we have\nconstructed a complete screening framework for g-GaN-based vdW heterojunctions.\nThis work provides assistance for the discovery and research of direct\nZ-type vdW heterojunctions with suitable band gaps for the photocatalytic\nsplitting of water.