Physically-Based Descriptors of the Nitrogen Reduction Reaction on Transition Metal Nitrides Using Ab Initio DFT

Ammonia is one of the most widely produced chemicals in the world due to its use in fertilizers. Around 90% of industrial ammonia production is done through the Haber-Bosch process. Unfortunately, this process accounts for around 1% of total global carbon emissions and 2% of global energy consumption, annually. As a result, researchers have been intensely studying the electrochemical nitrogen reduction reaction (NRR) as a means to directly produce decarbonized ammonia using clean energy. <br/><br/>One exciting class of materials for the NRR is transition metal nitrides (TMNs), which use the Mars-van Krevelen (MvK) mechanism for the reaction. Crucially, the involvement of surface nitrogen vacancies in this mechanism facilitates N₂ activation and dissociation by lowering the associated activation energies. However, current state-of-the-art TMN catalysts still don’t achieve the DOE performance targets for Faradaic efficiency and production rates. Therefore, in order to advance beyond current TMN performance, we need a holistic fundamental understanding of what materials properties influence the reaction energetics most strongly. It has been shown that the NRR on TMNs is limited by weak first protonation of the surface nitrogen and weak surface nitrogen vacancy filling by N₂H intermediates. Consequently, our objective is to identify physically-based descriptors for TMNs that correlate with the reaction step energies to elucidate what material properties should be targeted to potentially improve NRR activity. We hypothesize that descriptors relating to the bonding character and strength between the lattice nitrogen and metal species will be most strongly correlated to the limiting step energies. Inspired by the predictive success of the metal d-band center and oxygen p-band center descriptors, this work uses density functional theory (DFT) to study binary rocksalt-structured TMNs and to calculate various physically-based descriptors: metal d- and nitrogen p-band centers, hybridization between nitrogen p and metal d orbitals, and average metal/nitrogen Bader charges. <br/><br/>We find that the energy of the first protonation of a surface nitrogen species is negatively correlated with the average nitrogen Bader charge. Due to the binary nature of these TMNs, this also means that the protonation energy is positively correlated with the average metal Bader charge. Furthermore, the surface nitrogen vacancy filling energy by an N₂H intermediate is found to be positively correlated with the difference between the nitrogen p and metal d band centers (directly relating to orbital hybridization). We also investigate how the material composition of each TMN (approximated by the number of d electrons each metal contains) relates to the descriptors. Our results show that the average metal Bader charge increases with decreasing d electron number, while orbital hybridization decreases. Combining the results from the energetics-descriptor and descriptor-composition correlations, we see that the energetics of the limiting MvK steps on rocksalt TMNs are dependent on the nature of the covalency between the lattice nitrogen-metal bonds. <br/><br/>Overall, our results suggest a promising, accelerated materials design pathway to enhance NRR activity of binary TMNs. By taking advantage of the two unique active sites available in the MvK mechanism (the surface nitrogen species and surface nitrogen vacancies), we can potentially tune the energetics of the two limiting reaction steps independently using different metal species at each site. This strategy can yield more active catalysts, providing a pathway towards viable industrial electrochemical ammonia synthesis.