Yevgeny Rakita

Sustainability of energy conversion materials depends on their functional resilience. Such can be achieved by self-healing, SH, which requires mass transfer inside the material. Unfortunately, we lack guidelines as to which physical and chemical material properties can help predict if a material can self-heal and how. We report a data-driven approach to find such properties, using experimental and computational data from materials databases. We focus on four properties: optical deformation potential (ODP), relative structural polarization (RSP), ionicity, and hardness to distinguish SH materials from others. ODP with RSP predict the two main SH mechanisms in semiconductors: mass transfer of kinetically stabilized (class I) or thermodynamically stabilized (class II) defects. In crystals with covalent bonding, the correlation of hardness with ionicity & RSP emphasizes effects of the nature of the chemical bonding on the SH mechanism. Our approach and results present significant steps to find SH materials for various applications, especially for energy conversion.