Abstract

Contemporary battery anodes rely on intercalation to store the active ion, typically lithium between the layers in graphite. Moving beyond lithium to sodium or potassium batteries could provide the reduction in cost required to disrupt industries that are reluctant or unable to electrify, for example grid-scale storage banks at renewable energy sources. Intercalation of these larger ions is much less effective, instead we turn our attention to alloying anodes that exhibit complex reaction pathways upon cycling; the typical examples of silicides, phosphides and stannides can provide theoretical gravimetric capacities an order of magnitude greater than Li in graphite. To predict and understand the performance of a candidate anode with a particular active ion, a crucial first step is to figure out which phases are thermodynamically stable. In this talk, I shall introduce how this can be tackled computationally with crystal structure prediction techniques and first-principles modelling. Some of our recent successes in applying ab initio random structure searching (AIRSS), data mining and evolutionary methods to binary electrode systems will be discussed, and additionally I will present some work in progress on the K-Sn-P ternary system and discuss the challenges of scaling our approaches to exponentially larger design spaces.