Research

Computational Catalysis and Electrochemistry

Computational catalysis leverages quantum chemical simulations—primarily DFT—to understand, predict, and design catalytic materials and their reactivity at the atomic level. This field enables the systematic exploration of reaction mechanisms, identification of active sites, and rational screening of promising catalysts, all while capturing the influence of electronic structure and local environment on performance. By simulating surface reconstruction, mapping multistep reaction networks, and tuning the interfacial environment under operating conditions, we aim to accelerate the discovery of efficient catalysts that support industrial decarbonization and sustainable energy conversion.

AI in Catalysis

AI is increasingly transforming the field of catalysis by offering powerful tools to address the vast complexity of catalytic materials and reaction networks. The chemical space of possible catalysts—ranging from metals and oxides to alloys and nanostructures—is enormous, and AI enables systematic exploration by learning from high-dimensional datasets and guiding simulations toward promising regions. In computational catalysis, AI methods such as supervised learning, active learning, and delta-learning can significantly reduce the cost of quantum chemical calculations and identify hidden patterns across different catalytic systems. These approaches help bridge the gap between idealized models and real-world systems, accelerating the discovery of catalysts that support clean energy conversion and sustainable chemical manufacturing.

AI for Science

In chemistry and materials science, AI for Science provides powerful tools to accelerate innovation across diverse domains. For instance, in retrosynthetic analysis, ML models can efficiently navigate vast chemical reaction spaces to propose viable synthetic pathways for complex molecules, drastically reducing trial-and-error in the laboratory. In 2D perovskites, AI accelerates the search for new materials with tailored ferroelectric properties by integrating high-throughput simulations, structure–property predictions, and stability assessments. Similarly, for seawater desalination, AI enables the design of advanced membranes and catalytic systems, optimizing performance while reducing energy consumption and material costs.