Adsorptive Removal of PH3, Cd, and Pb on Metals and Bimetallic Surfaces

DFT calculations and machine learning (ML) models are used to study the adsorption of toxic gas-phase contaminants PH₃, Cd, and Pb and predict novel adsorbent compositions. Ir, Pd, Pt, Rh, and Ru strongly bind these contaminants, making them suitable for simultaneous removal. PH₂ dissociation is rate-determining on Ir, Pd, Rh, and Ru, while dissociation of PH₃ to PH₂ and H is rate-determining on Cu and Au. The activation barriers for PH₃ dissociation are lower on Rh, Pt, Pd, and Ru as compared to those on Ag, Au, and Cu. Segregation and formation energy calculations support the formation of surface alloys after the PH₃ dissociation and Cd/Pb adsorption. Novel bimetallic adsorbents combining Ag/Au/Cu with Ir/Pd/Pt/Rh/Ru are identified for trace contaminant removal. ML models, including extra tree and random forest regression, predict contaminant adsorption energies on bimetallic surfaces using physical, electronic, and geometric properties, with key descriptors identified by extra tree regression.