Halide Double Perovskites (HDPs) are quarternary materials with chemical formula A2BB'X6, which have recently been regaining scientific attention because they show beneficial properties for photovoltaic, X-ray detection, sensing, photocatalysis, and spintronic applications.  However, with over 40.000 potential HDP compositions, much of the available landscape remains underexplored. For this, we have generated a database of spin-polarized, hybrid functional (HSE06) DFT electronic structure data, combined with a chemical bonding analysis using LOBSTER (www.cohp.de). This was done for all HDPs with Cesium on the A-site and that are predicted to be stable by Bartel's tolerance factor[1], resulting in a database of quantum-chemical data on >2700 HDP compositions. 

I will highlight some of the interesting findings in this database, e.g., 134 predicted half-metals and some interesting correlations between the chemical bonding analysis and electronic features of interest. Furthermore, I will show how we are using interpretable machine learning techniques to expand the compositional space into alloyed HDPs and dimensionality reduction (UMAP) to visualize the chemical landscape.  [1]: Bartel et al., Sci. Adv. 2019,5,2, DOI: 10.1126/sciadv.aav0693