Dissociation, Dissolution and Diffusion of Nitrogen on VxFey and VxCry Alloy Membranes Studied by First Principles

N-2-selective metal membranes could be used to couple the extraction of N-2 from flue gas and NH3 synthesis within a single membrane reactor. The larger interstitial spaces of BCC metals make them attractive for fast permeation of the large N atom. The caveat is that N permeabilities reported thus far for pure BCC metals (e.g., V, Cr, Fe, Mo, and W) are insufficient for practical application. However, practical permeabilities could be achieved by alloying these metals. Here, we use density functional theory (DFT) to study how alloying V, a metal with high N solubility but low N diffusivity, with Fe and Cr impacts the dissociation, dissolution, and diffusion of nitrogen. Although N binding at vacancies was studied, binding at these defects was not particularly strong compared to regular interstitial sites, and thus, vacancies may not significantly affect solubility. The alloys present two types of interstitial octahedral (o) sites: V-rich (ol) and V-depleted (o2). Nitrogen binding strength correlates with the number of V nearest-neighbors; hence, binding in o1 sites is on average 1 eV stronger than in o2 sites. Ab initio thermodynamics suggests that this combination of strong and weak binding sites mitigates the reduction in solubility expected from the alloying (at least at relevant operating conditions). However, the heterogeneity of the interstitial sites generally leads to higher energy barriers for N hopping than those encountered in pure V (1.24 eV). The exception to this observation was the V0.25Fe0.75 alloy (1.01 eV). Based on our calculations, at 673 K and 5 bar (N-2 pressure), N solubility and diffusivity in V0.25Fe0.25 would be similar to 3 times smaller and similar to 53 times larger than pure V, respectively. According to the solution-diffusion model, these findings indicate that an similar to 18-fold higher permeability would be expected for V0.25Fe0.25 relative to V. Permeability is expected to be controlled by bulk diffusion rather than by surface processes, as in all alloys, we find the energy barrier for N-2 dissociation at the alloy surface to be lower than the barrier for bulk diffusion in the same alloy.