Date of Award

Spring 5-15-2021

Author's School

Graduate School of Arts and Sciences

Author's Department


Degree Name

Doctor of Philosophy (PhD)

Degree Type



Metal oxide semiconductors have been used as promising photocatalysts to convert solar energy into useful fuels, such as water splitting, carbon dioxide reduction, and nitrogen fixation reactions. However, the typical metal oxide photocatalysts' efficiency often suffers from their weak visible light absorption and low charge carrier mobilities. One common strategy to mitigate these limitations is to introduce oxygen vacancies into the lattice of the semiconductor photocatalysts. This strategy has been applied to many metal oxide photocatalysts, such as titanium oxide, tungsten oxide, and zinc oxide, and oxygen vacancies indeed enhance their catalytic activities. Various roles of oxygen vacancies have all been proposed to affect the activity of defect-rich semiconductor photocatalysts. The dominant mechanism has not yet been elucidated. For example, oxygen vacancies create new electronic states between the band gap of the semiconductor. These defect states may serve as color centers and enhance the absorption in the visible-light range. Also, these defect states are reported to trap photogenerated charge carriers and enhance charge-carrier separation. Furthermore, calculations using density-functional theory showed that oxygen vacancies could act as preferential adsorption sites for substrate molecules and activate the adsorbents. The primary difficulty in resolving this problem is that different nanoparticles within a single batch can exhibit large variations in their oxygen vacancy concentration. This heterogeneity makes it difficult to connect specific surface properties with the ensemble average catalytic activity measured for a large group of particles. Therefore, characterization methods that correlate the spatial distribution of reaction activities with specific structural features on individual particles are desired. This dissertation aims to apply single-molecule fluorescence microscopy to overcome this obstacle and study the role of oxygen vacancies in different semiconductor photocatalysts. This technique monitors the photocatalytic reactions in-situ at single-molecule and single-particle scale. The spatial distribution of active regions can be obtained at a single-particle level, and the heterogeneity among different catalyst particles can be revealed. Chapter 3 studies the competing roles of oxygen vacancies in BiOBr nanoplates in photocatalytic reduction reactions. We demonstrate that the activity of individual nanoplates first increases and then decreases while oxygen vacancies are continuously created photochemically. Our observation suggests an activation-deactivation mechanism: oxygen vacancies improve the visible light absorption at low density while serve as recombination centers at high density. Chapter 4 correlates the active regions for photocatalytic oxidation and reduction on individual BiOBr nanoplates. The photocatalytic oxidation activity is higher at colocalized regions, and the reduction activity is not affected by the degree of colocalization. We show that photoexcited electrons and holes separate at defective regions, but they colocalize at defect-deficient regions. Chapter 5 demonstrates that clusters of oxygen vacancies serve as active sites of photocatalytic •OH generation on W18O49 nanowires. Chapter 6 studies the interaction between organic ligands and W18O49 nanowires surface. We reveal that reductive ligands generate new active sites during photocatalysis and enhance the activity of W18O49 nanowires. By studying two different semiconductor photocatalysts, tungsten oxide (W18O49) and bismuth oxybromide (BiOBr), we showed two different roles of oxygen vacancies in these catalysts. In BiOBr photocatalysts, oxygen vacancies improve the photocatalytic activity by expanding the absorption at visible light range, and visible-light activity was detected. In comparison, no visible-light activity was detected in defective W18O49 nanowires, and clusters of oxygen vacancies activate the adsorbed H2O molecules. Furthermore, the role of oxygen vacancies can vary in different photocatalytic reactions, even for the same photocatalyst. Therefore, to rational design high-performance photocatalysts, visible light response, surface adsorption of substrate molecules, the possible changes of oxygen vacancies, and targeted catalytic reactions should be taken into considerations. Moreover, only a small number of semiconductor photocatalysts have been studied using our technique. We expect this technique will reveal more information on the mechanisms of these catalysts, and we are looking forward to following up on new discoveries in this research field.


English (en)

Chair and Committee

Bryce F. Sadtler

Committee Members

Richard A. Loomis, Bryce F. Sadtler, William E. Buhro, Sophia E. Hayes,