Materials informatics methods applied to spin splitting and band gap directness in semiconductors
As many solutions and new technologies are directly linked to the development of novel materials and the design of devices, the demand for more efficient and theory driven workflows is made evident. In this context, many advancements on computational methods are making possible the emergence of projects such as AFLOWlib, The Materials Project and OQMD, notable examples of the increasing effort of the Materials Science community towards a more data-driven research, a field currently named Materials Informatics. These databases are generated through high-throughput calculations based on Quantum Mechanics - specifically on the Density Functional Theory (DFT). The output of these calculations then can be used as a dataset for materials screening and as a starting point for higher complexity calculations and as input for Machine Learning methods. In this work we used this kind of methodology to study spin splittings in 2D and 3D materials, and also to learn physical properties (band gap directness in semiconductors) using machine learning algorithms.