Kontrola na poziomie atomowym heterointerfejsów AlGaN dla diod LED głębokiego UV

In conventional new materials development, crystal growth conditions are optimized by trial and error. In recent years, on the other hand, there is a need to innovate and apply *process informatics (PI)*, the development of new materials using artificial intelligence (AI). In this research project, a Digital Twin based on firstprinciples calculations and cellular automata will be developed to establish a framework for exploring materials processes through machine learning. This innovative PI technology will be used to develop deep-ultraviolet (DUV) LEDs that contribute to the destruction and inactivation of RNA and DNA of COVID-19 viruses and bacteria. Specifically, (1) the Polish team will analyze physical properties such as adsorption probabilities of atomic and molecular gases on the crystal growth surface, (2) the Bulgarian team will develop a digital twin that analyses the dynamic behavior of surface atomic steps using the obtained physical properties as parameters. (3) One of Japanese teams will use machine learning with the developed Digital Twin to predict the crystal growth conditions for obtaining atomically flat hetero-interfaces in the DUV-LED structure. Based on the above results, (4) another Japanese team will fabricate DUV-LEDs with world’s highest emission efficiency (> 10%) using the AlGaN metal-organic vapor phase epitaxy (MOVPE), contributing to the realization of a clean, safe, and secure society through the international joint research by four teams from three countries.