Computer simulations are applied to understand the influence of network geometry on the electron transport dynamics in random nanoparticle networks, and the predicted results are compared with those measured in one class of random nanoparticle networks: dye-sensitized nanocrystalline TiO2 solar cells. The model is applicable to all classes of random nanoparticle networks, such as highly disordered quantum dot arrays. The random nanoparticle networks are simulated by the step-wise condensation of a diffusion-limited aggregate. The fractal dimension of the nanoparticle films was estimated from the simulations to be 2.28, which is in quantitative agreement with gas-sorption measurements of TiO2 nanoparticle films. Electron transport on the computer-generated networks is simulated by random walk. The experimental measurements and random-walk simulations are found to be in quantitative agreement. For both a power-law dependence of the electron diffusion coefficient D on the film porosity P is found as described by the relation: D ∝ | P -Pc |μ. This power-law relation can also be derived from percolation theory, although only qualitatively. The critical porosity Pc (percolation threshold) and the conductivity exponent μ are found to be 0.76 ± 0.01 and 0.82 ± 0.05, respectively. It is estimated that during their respective transit through 50 and 75% porous 10-μm thick films as employed in the dye-cell, the average number of particles visited by electrons increases by 10-fold, from 106 to 107.

The steady state photoconductivity up of n-type, p-type and intrinsic a-Si:H has been studied up to photocarrier generation rates of G=5×1027cm-3s-1. In the 20ppm B2H6/SiH4 p-type sample photoconduction switches from holes at low G to electrons at high G. The electron photoconduction at high G is increased by n-type and decreased by p-type doping. This is explained by the charge state of the dominant electron recombination centers. The σp(G) curves of doped and intrinsic a-Si:H merge at the highest G used.

We have studied the dependence of the photoconductivity σp on photocarrier generation rate G in intrinsic a-Si:H at 300K between G=1012cm−3s−1 and 1028cm−3s−1. Below a certain value Go, we find σo =AGγ with γ=0.9±0.05 and the values of A vary considerably with defect concentration Nd which signifies monomolecular recombination through defects. Above Go the recombination is bimolecular, γ=0.5±0.02 and A=(6±3)×10−15 Ω−1cm1/2s1/2 is indpendent of Nd. The transition value Go is about 3×1020cm−3s−1 for high quality annealed a-Si:H and increases with Nd. A simulation of σp(G) assuming conduction in and recombination from extended states fits our experiments within a capture coefficient Ct=(6±2)×10−9cm3s−1 of carriers to their opposite tail states. Our Ct is close to the value (5±2)×10−9cm3s−1 obtained from optical measurements but higher than (0.5±0.1)×10−9cm3s−1 determined from photoelectric studies. Below T=150K our model calculations overestimate σp because the tunneling transitions, becoming important for recombination and conduction, are neglected.

An undoped and a compensated a-Si:H sample have been degraded by 17–34 ns laser pulses and by steady light (CW) at 78K and 300K. The light-induced defect concentration N is monitored by the increase in subgap absorption a. For the same change in a pulses degrade the photoconductivity σp more than CW light and more strongly for exposures at 78K than at 300K. The lack of correlation between σp and N suggests that light soaking causes additional damage besides an increase in N. This additional effect is more pronounced for pulse and low temperature exposures.