UC Berkeley

With their extremely high surface-to-volume ratios, nanocrystal surface morphologies, structures and compositions can have outsize effects on a nanoparticle's electronic, optical and catalytic properties when compared to their bulk system counterparts. Nanocrystal research has, in recent years, begun focusing on systematic characterization and manipulation of these surfaces for rational control of a nanocrystal's desired physical properties. The work presented in this dissertation provides further investigations of surface structure-function relationships with direct relationship to the catalytic and stability requirements of solar-to-fuel conversion systems.

In the first chapter, a brief and general review of quantum dot structure-function relationships in solar energy conversion schemes will be presented with an emphasis on photoelectrochemical devices. A discussion of general methods in nanoparticle synthesis and surface modification will be followed by a more in-depth analysis of the key physical principles of quantum dot (QD) photoelectrochemical and photocatalytic device architectures. Much of that discussion will concentrate on controlling the kinetics of a series of interfacial electron transfers. Finally, a review of methods in solar-to-fuel conversion chemistry will be presented with an emphasis on integrated water splitting devices, architectures employing an intimate semiconductor-catalyst-liquid or a semiconductor-metal oxide-liquid junction. This discussion will focus on the protection methods developed in the past four decades to combat destructive photocorrosion reactions.

The second chapter will present research directed at catalytic modifications to and structural characterizations of colloidal QDs. The goal of this project was to photocatalytically reduce protons from water using a nanocrystal light harvester and a surface bound, proton-reducing electrocatalyst. While we found that a covalently linked, homogeneous molybdenum-oxo electrocatalyst was photocatalytically inert, the decomposition product, identified as a structural relative of amorphous molybdenum trisulfide, was found to be highly active for photocatalytic proton reduction. X-ray absorption and photoemission structural characterizations of the amorphous catalyst before and after photocatalysis have been included. We found that the parent MoS3 structure identified before catalysis evolves into a relatively undercoordinated Mo-S bonding geometry: bridging disulfide linkages are converted into dative sulfides. This structure opens up the sulfide for ready protonation as a possible intermediate during catalysis. Such protonation is not available to the disulfide-containing derivative. The morphological conversion to an undercoordinated metal-ligand center is often invoked in catalyst activities, but rarely structurally identified.

The third chapter presents a study of ligand effects on charge transfer kinetics in a model system, W18O49 (WO2.72) nanoparticles. Tungsten oxide phases derived from WO3 are numerous due to the stability of the system even with high concentrations of oxygen vacancies. These vacancies result in significant electron density in the material's conduction band with the material class undergoing a metal-semiconductor transition at stoichiometries around WO2.8. These nanoparticles were synthesized with a moderately strongly bound ligand shell based on alkylamines. We found that when exposed to a sphectrophotometric redox indicator, namely an iron(III) tris-phenanthroline derivative, we could track the oxidation of electrons out of the nanoparticle conduction band, and into solution via the visible signal from the reduced iron complex. With that tag, we sought to investigate how the ligand affects the charge transfer rates. Hypothesizing that one of two mechanisms were in effect - outer sphere (tunneling) and inner sphere (dissociative) - we synthesized nanoparticles with varying ligand lengths in their shells and ran ligand concentration dependence studies. We found no correlation between ligand length and charge transfer rate, but a strong dependence of the rate on the concentration of free alkylamines in solution appeared. From this observation, we conclude that charge transfer occurs through uncoordinated surface sites whose concentration is dictated by parameters in surface binding isotherms, i.e. ligand binding coefficients, temperature and ligand-ligand interactions.

The fourth and final chapter will focus on photoelectrochemical water splitting employing a QD sensitized mesoporous titania thin film. To protect these light absorbers, a crosslinkable ligand was synthesized to passivate the vast majority of surface sites, thereby restricting the loci of charge transfer to accessible unbound sites. At these unbound sites, a water oxidation catalyst was deposited as a hole acceptor. Crosslinking was hypothesized to serve to reduce the native ligand's fluxionality on, off and over the surface of the QD by the chelate effect. In this hypothesis, ligand movement liberates new semiconductor surface sites to the corrosive aqueous environment. This can be tested by employing a ligand designed to react with nearest neighbors, suppressing ligand motion and desorption. Key characterization of the proposed architecture is presented via NMR, XPS and photoluminescent quenching studies. Photoelectrochemical testing indicates that the system does, in fact, produce oxygen, though at low current densities (~5 μA/cm2) and less than 100% Faradaic efficiency. While eventually unstable, we make the argument that many of this system's benefits warrant further investigations - namely the solution processability of their production and the rationality of their protection. Such prospects are discussed in a brief outlook section in the concluding section of this final chapter