Fig. 6: Virtual screening of the AI-predicted candidate 2D materials for the different renewable energy technologies.
From: An artificial intelligence-aided virtual screening recipe for two-dimensional materials discovery
The 2D materials that are suitable for efficient photovoltaic cells are shown in magenta. The candidate 2D materials for the photoelectrochemical splitting of water into oxygen and hydrogen are shown in blue. The candidate 2D materials that are predicted to be photocatalytically active for CO2 reduction and inactive for oxygen evolution from water are shown in red. The candidate 2D materials that are predicted to be photocatalytically active for N2 reduction and inactive for oxygen evolution from water are shown in both red and green. For the classification of the promising materials for different energy applications, we used the ML-predicted band gap, VBM, and CBM of the 2D materials and the alignment of the frontier energy bands with respect to the redox potentials of the chemical conversion reactions under study, typically at pH = 0 in aqueous solution vs. NHE, 25 ∘C, and 1-atmosphere gas pressure. The values of the properties are predicted by using the final ML models that have been trained on the entire training data.
