(216d) Hydrogen Solubility Predictions Via First-Principles Calculations: Investigations on Deviated Predictions from Experimental Solubility

Hydrogen solubility is an important parameter that affects hydrogen-selective membrane properties because the hydrogen permeability through a metallic membrane is half the product of the solubility and diffusivity of atomic hydrogen in the bulk metal. Accurate solubility predictions is crucial for the prediction of overall membrane properties. Our study may provide useful information in the search for new materials for hydrogen-selective membranes and hydrogen storage by identifying a proper alloy for the Pd-based and non-Pd-based membranes that have enhanced hydrogen solubility. In the last year, we have reported hydrogen solubility in ten transition metals (V, Nb, Ta, W, Ni, Pd, Pt, Cu, Ag, and Au) predicted by first principles based on density functional theory (DFT) combined with chemical potential equilibrium between hydrogen in the gas and solid-solution phases. This year, we have extended our work to further understand the deviations in the predictions of late transition metals. The solubility predictions in this study match experimental data within a factor of two in the cases of V, Nb, Ta and W, and within a factor of three in the cases of Ni, Cu, and Ag. Order of magnitude deviations are obtained in the cases of Pd, Pt, and Au. In the case of Pd, the deviation in solubility prediction is mainly attributed to the errors involved in the calculated vibrational frequencies of dissolved hydrogen. In Pt and Au, hydrogen in the octahedral interstitial site is less stable than in the tetrahedral site, contradicting the predictions based on the hard-sphere model. Potential energy surface analysis reveals a slightly downward concavity near the center of the octahedral sites in Pt and Au, which may explain the calculated imaginary vibrational frequencies in these sites and lead to unreliable solubility predictions.