Spatial Selectivity of Photodeposition Reactions on Non-polar, Polycrystalline Oxide Photocatalysts
General Material Designation
[Thesis]
First Statement of Responsibility
Pisat, Ajay S.
Subsequent Statement of Responsibility
Rohrer, Gregory S
.PUBLICATION, DISTRIBUTION, ETC
Name of Publisher, Distributor, etc.
Carnegie Mellon University
Date of Publication, Distribution, etc.
2019
GENERAL NOTES
Text of Note
256 p.
DISSERTATION (THESIS) NOTE
Dissertation or thesis details and type of degree
Ph.D.
Body granting the degree
Carnegie Mellon University
Text preceding or following the note
2019
SUMMARY OR ABSTRACT
Text of Note
Spatial selectivity, that is the separation of reduction and oxidation reaction sites on the photocatalyst surface, is very important for efficient photocatalysis. It reduces charge recombination by driving photogenerated electrons and holes away from each other, reducing charge recombination, and thus loss of energy as heat. Additionally, it keeps the intermediates of reduction and oxidation reactions away from each other, reducing their reaction to form the original compound. Thus, studying spatial selectivity in photocatalysis is important to bring them a step closer to commercial and large-scale deployment. This is especially true for centrosymmetric non-polar photocatalysts, that cannot have built-in dipole moments to create spatial selectivity, unlike ferroelectrics. However, there are other features like the presence of different crystal facets and different surface chemistry that can create spatial selectivity. Ferroelastic domains can also create spatial selectivity, as has been observed on non-polar ferroelastic BiVO4. Non-polar materials being a large part of the available materials space, the focus of this work is thus on studying the three features listed above and how they create spatial selectivity. The central idea is to use photodeposition marker reactions, which are reduction or oxidation reactions that leave an insoluble product on the surface. These products can be imaged with an SEM or AFM to determine which areas are more reductive/oxidative. The first two features listed above, crystalline facets and chemical terminations, were investigated in the context of spatial selectivity on bulk, polycrystalline SrTiO3. Perovskites being highly tunable with respect to chemistry and band gap, showing spatial selectivity via these features on this crystal structure is important. Both of these features have been studied separately, but this is the first study that takes both into account simultaneously. After synthesizing pellets and polishing them, the arbitrarily oriented surfaces were seen to break into facets on annealing at a high temperature of 1250 °C. The orientations of these surfaces were identified to be {100} and {110} by using EBSD, AFM, and MATLAB. Photodeposition of Ag, MnOx, and PbOx was carried out to determine that {100} facets were reductive, and {110} facets were oxidative. Secondly, the {110} facets were changed to be partially reductive, when the samples were heated in a Sr-rich atmosphere. The charge separation from facets was obvious before effecting any thermochemical treatment, after which the chemical terminations dominated the charge separation on the {110} facets. Next, non-polar ferroelastic materials were investigated with regards to ferroelastic domain-driven spatial selectivity. A previous report showed that an effect similar to ferroelectrics was seen on BiVO4, a non-polar ferroelastic, where the photodeposition was specific to alternating ferroelastic domains, as if they were polarized positive-negative in an alternating manner. To ensure that this is not specific to BiVO4, another non-polar centrosymmetric ferroelastic with a different crystal structure - WO3, was investigated. The same effect was seen on the ferroelastic domains of WO3, with photodeposition being driven by alternating ferroelastic domains, along with a correlating piezoforce response. This supports that the ferroelectric domain-like effect may be general to the entire class of ferroelastics. Lastly, It is necessary to understand the origin of this effect to be able to tailor spatial selectivity on ferroelastic materials. Flexoelectricity had been proposed as a possible reason. It means that the breaking of symmetry due to a strain gradient can create dipoles even in centrosymmetric materials. There were no experiments carried out in literature to test this hypothesis. So, in the final part of this work, systematic experiments were carried out on different samples of BiVO4 to determine whether flexoelectricity could play a role. The stresses near the surface of a grain in a polycrystal were classified into two: Those arising due to the change in crystal system from the ferroelastic phase transition, and those due to the surrounding grains. The stresses from the surrounding grains were relaxed by milling out a trench around the region of interest. Further, these stresses were completely eliminated in the next set of experiments on hydrothermally-grown crystals. Finally, the bulk stresses from the lattice change were eliminated by heating the sample above the ferroelastic transition temperature. In all cases, spatial selectivity was observed. Most interestingly, it was present even in the absence of a ferroelastic domain structure, when photodeposition was carried out above the ferroelastic transition temperature (FTT). This indicates that a secondary mechanism contributes to domain selectivity, which persists for some time in the absence of domains; we discuss that segregation of charged defects to different domains as an example of such as secondary mechanism. It implies that there is a factor that is not directly driven by stress/strain (and hence, flexoelectricity) which creates spatial selectivity in BiVO4. Overall, the work presented herein shows that spatial selectivity can be tailored on non-polar perovskites using crystallo-chemical properties of the material, namely orientation (facets) and chemical termination of the exposed surfaces. Similarly, the origin of domain-specific spatial selectivity on non-polar ferroelastics appears to be a coupled crystallo-chemical property of the material, such as the distribution of charged ions between structural domains, though more work is needed to elucidate the details more clearly.