How Concerted Are Ionic Hops in Inorganic Solid-State Electrolytes?

Despite being fundamental to the understanding of solid-state electrolytes (SSEs), little is known on the degree of coordination between mobile ions in diffusive events, thus hindering a detailed comprehension and possible rational design of SSEs. Here, we introduce an unsupervised k-means clustering approach that is able to identify ion-hopping events and correlations between many mobile ions and apply it to a comprehensive ab initio MD database comprising several families of inorganic SSEs and millions of ionic configurations. It is found that despite two-body interactions between mobile ions being the largest, higher-order n-ion (2 < n) correlations are most frequent. Specifically, we prove a general exponential decaying law for the probability density function governing the number of concerted mobile ions. For the particular case of Li-based SSEs, it is shown that the average number of correlated mobile ions amounts to 10 ± 5 and that this result is practically independent of the temperature. Interestingly, our data-driven analysis reveals that fast-ionic diffusion strongly and positively correlates with ample hopping lengths and long hopping spans but not with high hopping frequencies and short interstitial residence times. Finally, it is shown that neglection of many-ion correlations generally leads to a modest overestimation of the hopping frequency that roughly is proportional to the average number of correlated mobile ions.

Universal ion-transport descriptors and classes of inorganic solid-state electrolytes.

Solid-state electrolytes (SSEs) with high ion conductivity are pivotal for the development and large-scale adoption of green-energy conversion and storage technologies such as fuel cells, electrocatalysts and solid-state batteries. Yet, SSEs are extremely complex materials for which general rational design principles remain indeterminate. Here, we combine first-principles materials modelling, computational power and modern data analysis techniques to advance towards the solution of such a fundamental and technologically pressing problem. Our data-driven survey reveals that the correlations between ion diffusivity and other materials descriptors in general are monotonic, although not necessarily linear, and largest when the latter are of vibrational nature and explicitly incorporate anharmonic effects. Surprisingly, principal component and k-means clustering analyses show that elastic and vibrational descriptors, rather than the usual ones related to chemical composition and ion mobility, are best suited for reducing the high complexity of SSEs and classifying them into universal classes. Our findings highlight the need for considering databases that incorporate temperature effects to improve our understanding of SSEs and point towards a generalized approach to the design of energy materials.