Noble Metal (Cu, Ag, Au) Modified PdPS Monolayers for Adsorption of SF₆ Decomposition Gases (H₂S, SO₂, SOF₂, and SO₂F₂): A Density Functional Theory Investigation

Sulfur hexafluoride (SF 6 ) is extensively employed in gas-insulated switchgear (GIS) due to its exceptional insulating and arc-extinguishing properties. However, operational failures like arcing or overheating can lead to the decomposition of SF 6, releasing various gases such as H 2 S, SO 2, SOF 2, and SO 2 F 2, depending on the fault type. The urgent need for efficient real-time sensors to detect these decomposition gases is highlighted. This study examines the adsorption characteristics of sensing materials based on palladium phosphide (PdPS) and its composites with noble metals (Cu, Ag, and Au) utilizing first-principles density functional theory simulations. The analysis includes parameters such as adsorption distance, charge density, binding energy, adsorption energy, charge transfer, molecular frontier orbitals, and desorption kinetics. Optimal embedding sites for noble metals were identified at the T P positions on the PdPS surface, yielding binding energies of −1.529 eV for Cu, −0.987 eV for Ag, and −1.04 eV for Au. Notably, while pristine PdPS exhibited unfavorable adsorption profiles, the Cu-PdPS composite demonstrated significant adsorption capabilities with energies of −1.359 eV for H 2 S, −1.245 eV for SO 2, −0.902 eV for SOF 2, and −0.806 eV for SO 2 F 2 . Desorption kinetics reveal that the Cu-modified system offers rapid desorption times, notably 4.24 s for SO 2 F 2 at room temperature, positioning it as an excellent candidate for ambient detection. Additionally, the Cu-PdPS configuration shows enhanced versatility across varying temperatures, particularly excelling in high-temperature conditions. The findings underscore the potential of PdPS-based materials in developing advanced sensors, paving the way for integrating low-power, high-sensitivity monitoring systems for real-time SF 6 decomposition gas detection in GIS applications, with implications for future research toward AI-driven models for quantitative gas mixture analysis.