Rational design of novel photo-catalysts for hydrogen generation and carbon dioxide reduction by molecular modeling combined with machine learning

  • Yu Li

Student thesis: Doctoral Thesis

Abstract

Greenhouse gas emissions are the primary driver of global warming and climate change, the problems derived from this issue make the shift of traditional fossil fuels towards cleaner, more sustainable energy sources an urgent matter. In this direction, converting solar energy through photocatalysis into valuable fuels such as carbon-based fuels and hydrogen has been established as an intriguing strategy. However, seeking robust photocatalysts to achieve higher visible-light conversion efficiency, quantum yields and selectivity remains a pressing challenge. Along that effort, theoretical calculations and high-throughput modeling approaches are becoming indispensable tools to provide physical insights into the key properties and mechanisms, complementing the experiments and guiding the effective design of photocatalysts.

Hence, the aim of this PhD thesis is to use quantum density functional theory (DFT) calculations and machine learning (ML) assisted modeling as an enabling tool for understanding and searching novel materials for clean energy applications, focusing on H2 evolution from H2S splitting and water splitting, as well as CO2 conversion. We first summarized the mechanism of semiconductor photocatalysis and different computational methods relevant to these applications in the literature review section to gain deep insight into the reaction processes. Furthermore, we have carried out systematic DFT calculations to investigate H2S and CO2 adsorption and dissociation on perfect and defective CdS and TiO2 surfaces, respectively. Then two top-down approaches were performed to screen possible transition metals (TMs) as co-catalysts doped on the defect CdS (110) and hydroxyapatite (HAP) surface for HER and CO2RR, respectively. Finally, we demonstrated a high-throughput method with the assistance of ML to rapidly and efficiently predicting structural stability and intermediate adsorption energies, and screening high-performance metal sulfide photocatalysts for H2 evolution and CO2 reduction.
Date of AwardAug 2023
Original languageAmerican English

Keywords

  • Hydrogen generation
  • Carbon dioxide reduction
  • Photocatalysis
  • Density Functional Theory
  • Machine learning

Cite this

'