A materials informatics approach to the study of van der Waals materials

The explosion in the recent number of proposed van der Waals (vdW) materials, alongside the emergence of materials databases, has increased the promise of uncovering novel physics and developing new applications for 2D atomic crystals. Though this large number of newly identified vdW materials comes with the potential for discovery, immense challenges also emerge. That is, traditional experimental and theoretical methods for the study of a large database of materials are slow and expensive. A novel approach to materials investigation is desirable. With the advent of materials informatics (the combination of statistical tools, computational methods and materials science) one is able to efficiently screen a large database of materials and make predictions of desirable properties. We will discuss how materials informatics can be used for a high-throughput study of 2D atomic crystals. In particular, we show how machine learning can be used to efficiently predict the heat of formation (i.e. stability) of known 2D atomic crystals, as well as candidates for entirely new layered materials.