In this talk, I will introduce two examples of materials design using database analysis: organic superconductors and magnetic high entropy alloys.

Over the past few decades, organic superconductivity has been reported in various carbon-based organic materials. In particular, polycyclic aromatic hydrocarbons (PAHs) have been intensively studied in recent years because of the highly conjugated characteristics of the delocalized electrons supplied by alkali metal doping. Here, I introduce a PAH superconductor, sodium-doped triphenylene (nominal composition of Na₃triphenylene), which was discovered by searching the database of PAH molecules with a small energy difference between the lowest unoccupied molecular orbital (LUMO) and LUMO + 1 states. As the first sodium-doped PAH superconductor, it shows the superconducting transition at T_C ≈ 15 K under ambient pressure.

 In the application of high entropy alloy (HEA), the crystal structure prediction should be preceded because their physical and mechanical properties are deeply related to the crystal structures. Recently, the learning based approach has been successfully applied to the prediction of structural phases. However, it is well known that the HEA requires vast cost for preparing dataset because multi-element alloy is involved in the training. When magnetic elements is included, the database generation and training becomes more complicated. Here, I will show an efficient approach to predict the structural phases of the magnetic high entropy alloy without preparing a large scale of training dataset. This accelerated prediction can be applied to the prediction of various structural properties of multi-elements alloys which do not exist in the current structural database.