Machine-learning-assisted thin-film growth: Bayesian optimization in molecular beam epitaxy of SrRuO3 thin films

摘要: Materials informatics exploiting machine learning techniques, e.g., Bayesian optimization (BO), have the potential to reduce number of thin-film growth runs for conditions through incremental updates models in accordance with newly measured data. Here, we demonstrated BO-based molecular beam epitaxy (MBE) SrRuO3, one most intensively studied materials research field oxide electronics, mainly owing its unique nature as a ferromagnetic metal. To simplify intricate search space entangled conditions, ran BO single condition while keeping other fixed. As result, high-crystalline-quality SrRuO3 film exhibiting high residual resistivity ratio over 50 well strong perpendicular magnetic anisotropy was developed only 24 MBE which Ru flux rate, temperature, and O3-nozzle-to-substrate distance were optimized. Our method provides an efficient experimental design that is not dependent on experience skills individual researchers, it reduces time cost, will accelerate research.Materials efficien...