Materials Informatics: Overview

Abstract

Recent years have seen the explosion of a field of research now commonly termed “materials informatics.” Though the roots of this research can be traced back to the early 1970s, and mostly driven by the chemistry community, the last 20 years have seen the pioneering and then systematic application of deterministic or stochastic methods based on large data collections to design or discover novel materials, where the materials data themselves are the outcome of calculations, or exploit these in synergy with experimental databases. This chapter aims to provide an overview of some of the most successful and exciting efforts worldwide in this area, with a focus on materials science and condensed-matter physics, but also including notable contributions in chemistry and molecular science. Contributions can be broadly assigned to two different areas: efforts dedicated to producing and/or storing curated or raw computational or experimental data and associated materials properties, and machine-learning efforts dedicated to leveraging those data to bypass the need of expensive, and typically first-principles, calculations. Both hold great promise in our quest to understand, predict, and design the properties of novel materials.