Charge transport in porous nanocrystalline titanium dioxide

• Published in: Physica. E, Low-dimensional systems & nanostructures Low-dimensional systems & nanostructures, 2002-04, Vol.14 (1), p.197-202
• Main Authors: Eppler, Anuradha M., Ballard, Ian M., Nelson, Jenny
• Format: Article
• Language: English
• Subjects: Charge transport; Light; Photoconductivity; Recombination; Solar cells; TiO 2
• ISSN: 1386-9477; 1873-1759
• DOI: 10.1016/S1386-9477(02)00383-1

The dark conductivity and photoconductivity of porous, anatase titanium dioxide films have been studied in different ambient conditions. The films are nanocrystalline with a particle size of 5– 15 nm and porosity of around 50%. Films are resistive (10 4– 10 6 Ω m ) in the dark in ambient air, and exhibit space charge limited current–voltage behaviour, modified by the presence of traps. Vacuum reduces the dark conductivity by a factor of 10 2–10 3. This effect is tentatively attributed to the removal of water, which is known to adsorb dissociatively on TiO 2 surfaces and may dope the material by proton insertion and Ti 3+ formation. The photoconductivity in vacuum is 10 6 larger than that in air at maximum photocurrent and increases with decreasing pressure. In this case the effect is attributed to the loss of surface adsorbed oxygen, a known electron scavenger, in vacuum. Removal of oxygen extends the electron lifetime and results in a much larger saturation photocurrent. In vacuum, a point of inflexion is observed in the transient rise and the shapes of the curves are intensity dependent. Both these observations are consistent with the presence of traps. No correlation was observed between the photoconductivity decays and temperature, which suggests that the decay occurs by band-to-band recombination and not thermionic emission. On the basis of these observations, a model based on competition between photogeneration, trapping and scavenging has been developed. By varying the trapping and recombination rates we can simulate the effects of air and vacuum. The intensity dependent results can be simulated by changing the generation rate alone which allows us to estimate a trap density of less than 10 20 cm −3 . We propose that photoconductivity may be used as a direct probe of the electron lifetime and can serve to evaluate different chemical environments for dye sensitised solar cells, and to study photocatalytic function.