Models for electron transport and back reaction in dye-sensitized nanocrystalline solar cells were investigated by developing novel measurement techniques and the results were used to test two complementary models; diffusive electron transport within the TiO2 medium and the quasi-static approximation to deal with non steady-state conditions where trapping plays a role. These will be shown to be partly correct and the shortfalls highlighted and discussed. In the end it was found that more knowledge of the parameters governing the behaviour of electrons is required to further test and develop the models. The incorporation of a secondary sensing electrode allowed the internal quasi-Fermi level (QFL) within the TiO2 to be probed. The behaviour of the voltage measured by the secondary sensing electrode was in accordance with diffusive electron transport in the TiO2. This was confirmed by measuring the QFL along the current-voltage curve of the cell, and by the temperature dependence of the measured QFL. Discrepancies concerning the behaviour of the ideality of the open-circuit voltage (and hence the electron lifetime) between experiment and modelling are highlighted and discussed throughout. Assuming an Arrhenius relationship simple expressions for the temperature dependence of the open-circuit voltage were derived and experimentally tested. The trapped electron density was measured along the current-voltage curve. With the inclusion of the secondary sensing electrode and measuring the trap distribution, the way the trapped charge varied could be modelled and compared to experiment. This provided an important link between the free and trapped electron density profiles but again highlighted shortcomings of the applied models. The quasi-static approximation was tested against a full numerical solution (continuum model) to determine the phase space in which it is applicable. Knowing this, an almost ideally behaving cell was used to test the quasi-static approximation. Having shown that it was valid for the given cell, the quasi-static approximation was used to determine how the conduction band electron lifetime varied with temperature, resulting in an Arrhenius dependence of the back reaction rate of electrons. A strong temperature dependence of the electron lifetime, and hence a strong temperature dependence of the electron diffusion length was demonstrated.

Fresh produce is increasingly associated with foodborne outbreaks. In order to develop effective intervention and measures to reduce microbial risks, it is essential to attain a better understand the mechanisms of attachment of foodborne pathogens to fruits and vegetables. Using lettuce as a model, the attachment of Escherichia coli O157:H7 to produce surfaces was studied. Strains expressing various extracellular proteins (curli, O157-antigen, and intimin) known to influence attachment of E. coli to intestinal cells were evaluated for their physicochemical properties and ability to adhere to cut edge and whole leaf lettuce. Escherichia coli O157:H7 strains included: 0018, 43894 and 43895 (curli producing and non-producing); 86-24 (WT), F-12 (O157-antigen negative), pRFBE (O-antigen replaced on plasmid); and 86-24, 86-24Ã eae10 (intimin negative). The eleven strains were surveyed for their hydrophobicity and cell charge using hydrophobic interaction chromatography (HIC) and electrostatic interaction chromatography (ESIC) techniques. Iceberg lettuce squares (2 x 2 cm) were inoculated with E. coli O157:H7 strains separately (7.0 log CFU/square) and dried in a laminar flow hood. Lettuce was sampled before (unrinsed) and after being rinsed twice with sterile de-ionized water (rinsed). Strips (2 mm wide) of each cut edge of the lettuce were aseptically removed. Cut-edge and whole-leaf samples were homogenized and spiral plated onto Luria-Bertani agar, supplemented with nalidixic acid (50ppm), to assess levels of bacteria remaining on the lettuce leaf after rinsing. The rinse steps were not effective in significantly removing bacteria from lettuce (p>0.05). Curli-producing and non-producing strains preferentially attached to cut edge versus the whole leaf portions of lettuce (p<0.05); however the 86-24 strains showed no preference for attachment. With the exception of 0018 curli-producing and non-producing strains, presence/absence of extracellular proteins surveyed did not influence attachment of E. coli O157:H7 to either cut edge or whole leaf lettuce. There was significantly greater attachment of the curli-producing 0018 strain over the curli non-producing 0018 strain to cut and whole lettuce surfaces (p<0.05). Production of curli and O-polysaccharide significantly increased (p<0.05) the cell's overall hydrophobicity of the cell; however this did not affect attachment (p<0.05). The overall cell charge of all strains was negative; however, charge did not affect attachment of E. coli O157:H7 to lettuce. The presence of extracellular appendages (curli, O157-antigen, intimin) as well as hydrophobicity and cell charge properties had no affect on attachment of the cell to lettuce. / Ph. D.