Sub-nanometer metal clusters are active for a variety of industrially important reactions. However, their large phase space complexity makes it challengeing to characterize and study them computationally. Our group is developing methods that integrate electronic structure calculations, rare events methods, and machine learning to investigate their structure and reactivity.
With the advent of high throughput experiments we have access to large material property databases. However, many materials and properties remain unexplored due to limited data availability and high cost of experiments. Our group is developing machine learning methods to predict material properties in data-scarce regimes. We also work on uncertainty quantification, model interpretability, and feature engineering to ensure model reliability and transferability.
High-entropy alloys (HEAs) are alloys made of a large number of metals (typically more than 5 elements) mixed in relatively equal proportions. Recent studies have shown that they are active for several reactions including CO oxidation, hydrogen evolution, and ammonia decomposition. However, their configurational complexity leads to difficulty in characterization and has precluded precise understanding of their structure and mechanisms. Our group is developing computational methods to rigorously model HEAs.
Recent advances in virtual reality (VR) technology has opened avenues for immersive visualization of complex data. We are developing atomistic visualization tools in virtual reality (VR) and augmented reality (AR).