(2gw) Catalyst Design for Water Treatment Using Ab Initio Simulation

Shortage of clean water due to population growth, industrialization, and climate change is a concerning problem worldwide. Physicochemical treatment processes were common approach for pollutant removal, which however will result in concentrated polluted solutions. Therefore, pollutant decomposition with heterogeneous catalysts has emerged as a promising strategy. In this research, we use ab initio simulation to unravel the fundamental photo-, electro-, and thermo-catalytic reaction mechanisms, which in turn will inform design strategies for better catalysts in perfluorocarboxylic acids (PFCA) or nitrate removal from groundwater.

Modeling photo-electrochemical (PEC) reactions occurring at catalyst/electrolyte interfaces is essential but challenging. Therefore, we first developed an electronic grand canonical density functional theory (GC-DFT) formalism to model PEC reactions on semiconductor surfaces in aqueous environment under constant potential. (Bhati and Chen et al. , J. Phys. Chem. C, 2020, 124, 49, 26625–26639). With the developed calculation method, we studied PFCA degradation. PFCA are widely used in everyday products, such as food packaging, clothing, and furniture, due to their resistance to heat, oil, and water. However, the outstanding stability prohibits PFCA from natural decomposition and thus PFCA accumulate in the environment, which leads to cancer and liver damage. Our collaborators in Prof. Michael S. Wong’s group reported that hexagonal boron nitride (hBN) outperforms state-of-the-art titanium dioxide (TiO₂) photocatalysts for degrading PFCA. This motivated our investigations on the photo-oxidative reaction mechanism governing PFCA activation on hBN using GC-DFT formalism: C_nF_2n+1COO^- + h⁺ → C_nF_2n+1• + CO₂. Also, we found that a nitrogen-boron substitutional defect (N_B), which generates a mid-gap state, can enhance UVC absorption and PFCA oxidation. This reveals introducing N_B defects is a promising engineering strategy to design better photocatalysts for PFCA degradation. (Chen et al. Environ. Sci. Technol., 2022, 56, 12, 8942–8952) Furthermore, another work by us showed that, on the hydrophobic basal plane of hBN surface, water molecules do not interfere with PFCA oxidation and thus PFCA is degraded more efficiently than the hydrophilic TiO₂ surface. This suggests that decreasing the hydrophilic edge defects on hBN can further improve its performance on PFCA decomposition. (Wang and Chen et al., submitted)

Besides PFCA, nitrate is another pervasive surface and groundwater contaminants, causing severe diseases such as blue baby syndrome. From chemical point of view, harmful nitrate can be transformed into valuable ammonia (NH₃, a fertilizer precursor) or harmless dinitrogen (N₂). We applied density functional theory (DFT) to investigate the reaction network of nitrate reduction and understand the NH₃/N₂ selectivity over Cu and Pd. We found that HNO* and NOH* formation are key steps to determine the NH₃/N₂ selectivity. Also, Pd surface is more selective to N₂ under high NO* coverage. Inspired by our discovery, our collaborators at Georgia Tech synthesized Pd nanocubes partially covered with Cu, where the nitrate is actively reduced on Cu domain to NO, which spillovers to Pd domain for further reduction to N₂ selectively. (Lim and Chen et al., ACS Catal. 2023, 13, 1, 87–98) Furthermore, we discerned how the electronic properties of the metal catalyst affects the nitrate reduction reaction mechanism, steering the products toward either N₂ or NH₃, which provides deep insight into catalyst screen and design for denitrification. (Chen et al. submitted)

Overall, my work provides insights into atomistic-level reaction mechanism and material properties and sets a basis for catalyst design for water treatment. I am currently a machine learning intern at AspenTech and working on building classification models on synthesized data for chemometric data drift detection. I am contributing to build a package to synthesize perturbed data from available calibration data automatically as well as create a classification model for drift detection and drift type identification. Also, I have experience in applying molecular dynamics (MD) simulation to investigate metal properties and innovate biological drug design.

Research Interests

For future post-doc positions, I am still interested in developing and applying computational modeling tools for investigating the reactions at interface at both the electronic and atomistic level for efficient energy conversion, storage, and utilization, as well as rational design of catalytic systems. Furthermore, combining the simulation methods and machine learning tools for catalyst screen and design would be a promising potential area for me to dig into.