Examining the molecular origins of anomalously high H 2 O generation at oxide-passivated metal surfaces for plasma applications

• Published in: Journal of physics. Condensed matter 2023-11, Vol.35 (45), p.455001
• Main Authors: Leung, Kevin, Goeke, Ronald S, Ilgen, Anastasia G, Wilson, Alexander, Cuneo, Michael E, Bennett, Nichelle
• Format: Article
• Language: English
• ISSN: 0953-8984; 1361-648X
• DOI: 10.1088/1361-648X/ace9aa

Elucidating the mechanisms responsible for sub-microsecond desorption of water and other impurities from electrode surfaces at high heating rates is crucial for understanding pulsed-power behavior and optimizing its efficiency. Ionization of desorbed impurities in the vacuum regions may create parallel loads and current loss. Devising methods to limit desorption during the short time duration of pulsed-power will significantly improve the power output. This problem also presents an exciting challenge to and paradigm for molecular length-scale modeling and theories. Previous molecular modeling studies have strongly suggested that, under high vacuum conditions, the amount of water impurity adsorbed on oxide surfaces on metal electrodes is at a sub-monolayer level, which appears insufficient to explain the observed pulsed-power losses at high current densities. Based on density functional theory (DFT) calculations, we propose that hydrogen trapped inside iron metal can diffuse into iron (III) oxide on the metal surface in sub-microsecond time scales, explaining the extra desorbed inventory. These hydrogen atoms react with the oxide to form Fe(II) and desorbed H O at elevated temperatures. Cr O is found to react more slowly to form Cr(II). H evolution is also predicted to require higher activation energies, so H may be evolved at later times than H O. A one-dimensional diffusion model, based on DFT results, is devised to estimate the water outgassing rate under different conditions. This model explains outgassing above 1 ML for surface temperatures of 1 eV often assumed in pulsed-power systems. Finally, we apply a suite of characterization techniques to demonstrate that when iron metal is heated to 650  C, the dominant surface oxide component becomes -Fe O . We propose such specially-prepared samples will lead to convergence between atomic modeling and measurements like temperature-programmed desorption.