Exploring Bayesian Optimization for Photocatalytic Reduction of CO2

Abstract

The optimization of photocatalysis is complex, as heterogenous catalysis makes its kinetic modeling or design of experiment (DOE) significantly more difficult than homogeneous reactions. On the other hand, Bayesian optimization (BO) has been found to be efficient in the optimization of many complex chemical problems but has rarely been studied in photocatalysis. In this paper, we developed a BO platform and applied it to the optimization of three photocatalytic CO2 reduction systems that have been kinetically modeled in previous studies. Three decision variables, namely, partial pressure of CO2, partial pressure of H2O, and reaction time, were used to optimize the reaction rate. We first compared BO with the traditional DOE methods in the Khalilzadeh and Tan systems and found that the optimized reaction rates predicted by BO were 0.7% and 11.0% higher, respectively, than the best results of optimization by DOE, and were significantly better than the original experimental data, which were 1.9% and 13.6% higher, respectively. In both systems, we also explored the best combination of the surrogate model and acquisition function for BO, and the results showed that the combination of Gaussian processes (GP) and upper confidence bound (UCB) had the most stable search performance. Furthermore, the Thompson system with time dependence was optimized with BO according to the selectivity of CH4. The results showed that the optimized reaction time of BO agreed with the actual experimental data with an error of less than 5%. These results suggest that BO is a more promising alternative to kinetic modeling or traditional DOE in the efficient optimization of photocatalytic reduction.

Lead-Free Halide Perovskites for Photocatalysis via High-Throughput Exploration

“…92,93 Alongside BO, Gaussian processes are leveraged for BO, and dimensionality reduction techniques like PCA play a critical role in visualizing and interpreting the complex data from HT experiments. 94 However, a traditional BO approach often faces limitations, notably, its inability to incorporate prior information. To address this, the integration of physics-based machine learning (ML) methods has become increasingly relevant.…”

“…This enables a deeper understanding of the physical and chemical properties of photocatalysts and facilitates the efficient optimization of material parameters to achieve desired functionalities such as enhanced CO 2 reduction and water splitting efficiency. Currently, the Bayesian optimization (BO) method is being employed to navigate low-dimensional parameter spaces, including aspects like composition, ligand selection for nanocrystal growth, and antisolvent choice. , Alongside BO, Gaussian processes are leveraged for BO, and dimensionality reduction techniques like PCA play a critical role in visualizing and interpreting the complex data from HT experiments . However, a traditional BO approach often faces limitations, notably, its inability to incorporate prior information.…”

From Characterization to Discovery: Artificial Intelligence, Machine Learning and High-Throughput Experiments for Heterogeneous Catalyst Design

“…A study by Zhang et al (2023) highlights the potential of BO compared to typical Design of Experiment (DoE) optimization. They focused on the photocatalytic reduction of CO 2 to CH 4 , using BO on three previously kinetically modeled photocatalytic systems.…”

Balancing computational chemistry's potential with its environmental impact

“…The potential of BO has been explored in climate-focused studies. Zhang et al 91 employed BO to optimize the partial pressures of CO 2 and H 2 O, as well as the reaction time, aiming to maximize the reaction rate of the photocatalytic reduction of CO 2 . Their approach reached optimal conditions faster compared to DOE and kinetic modeling.…”