Ab-initio and ex-machina materials design for energy from computations of transport properties of electrons, phonons and ions.
Accurate atomistic computations of transport properties are a particularly important challenge and an opportunity for designing materials for energy conversion and storage. I will start by discussing thermoelectric materials, where quantum-level coupling of electronic and thermal carriers controls the transport properties and overall efficiency. We developed a new approach for quickly computing electronic transport properties of complex semiconductors and low-dimensional quantum materials, without empirically fitted parameters. The first-principles calculations of the electron-phonon coupling demonstrate that the energy dependence of the electron relaxation time varies significantly with chemical composition and carrier concentration, contrary to commonly made approximations. Using this method allowed us to computationally discover new thermoelectric alloy compositions that were confirmed to have leading performance and stability.
Understanding of ionic transport mechanisms is important for enabling high ionic conductivity in solid-state batteries. I will discuss how geometric frustration, disorder and collective motion influence ionic transport in solid polymer and ceramic electrolyte materials. We find that strong correlation in soft and liquid materials can lead to surprising transport phenomena, even reversing the effective charge of mobile cations. Computational tools for describing ionic transport in the strongly correlated regime enabled design of new solid polymer electrolyte compositions. Finally, I will also present recent progress and challenges in large-scale molecular dynamics simulations enabled by on-the-fly active machine learning, that are transforming the way we approach complex materials.