Laboratory for Research on the Structure of Matter

Title: “Molecular beam epitaxy of functional complex 4d & 5d transition metal oxides”

The discovery of high-temperature superconductivity in cuprates in 1986 ignited significant research into complex transition metal oxides. While 3d oxides have been extensively studied, 4d and 5d systems have been less explored due to challenges in achieving high-quality growth and controlling volatile oxide formation.

We have developed an electron-beam molecular beam epitaxy (EB-MBE) system equipped with electron impact emission spectrometry (EIES) for precise real-time control of elemental fluxes. This approach enables the synthesis of complex 4d and 5d transition metal oxides with tailored stoichiometries. To address the issue of volatile oxide formation, we employ a rate control system that compensates for element loss during oxidation.

To optimize the growth process, we have implemented a Bayesian optimization-based machine learning method. This approach allows for efficient exploration of the parameter space, enabling us to identify optimal growth conditions. By combining machine learning with EB-MBE, we can accelerate the discovery and synthesis of new materials with intriguing physical properties, such as palladates like Nd2−xCexPdO4.

A notable achievement is the synthesis of the first gold oxide, Sr5Au3O8, using molecular beam epitaxy. We will also discuss our recent progress in synthesizing complex osmates, a class of materials known for their high volatility and challenging growth conditions. Sr3OsO6, a double-perovskite related structure, exhibits the highest Curie temperature discovered so far among insulators. Our approach demonstrates the potential of EB-MBE and machine learning for exploring the frontiers of complex transition metal oxide materials science.