A detailed study of the optoelectronic processes occurring in ultrathin ${\mathrm{TiO}}_2$ nanocrystal-conjugated polymer, poly(p-phenylene vinylene) (PPV) composites is presented. Composites of ultrathin films (about 100 nm) are studied spectroscopically and as the active medium in photovoltaic devices of the structure (Al/composite/indium tin oxide). By varying the weight ratio of the nanocrystals and using results of photoluminescence efficiency, photocurrent, and photovoltaic measurements and time-resolved microwave conductivity, we are able to construct a well-defined picture of the relevant processes in the composite: including exciton dissociation, charge transport, and recombination. We combine the experimental results with a hopping model for charge transport in a nanocrystal lattice (random walk or biased random walk) to determine the probability of electron collection as a function of distance from a collecting contact in a nanocrystal lattice. The combined results indicate that most photogenerated excitons lead to charge separation at the interface between the polymer and the nanocrystals above 20-wt % ${\mathrm{TiO}}_2$ nanocrystals, but the electron collection efficiency in photovoltaic devices is limited by fast recombination. The transport model indicates that even for a relatively long recombination time and a well-ordered nanocrystal lattice, most of the collected charge will originate from the first several nanocrystal layers at the electrode rather than from sites throughout the film, due to the recombination process. We also argue that the existence of a photocurrent in these and related devices is not necessarily evidence of charge transport through a network of the nanocrystals (or other component), as the quantum yields can be accounted for by interfacial charge transfer at the contact alone. Quantum yields for collecting charge following direct band-gap excitation of the ${\mathrm{TiO}}_2$ are more than a factor of 10 larger than for excitation into the polymer, suggesting that either hole transfer to the polymer, or some preceding process, is rate limiting and much slower than the corresponding process following polymer excitation. We also examine the key differences between the mechanisms underlying conjugated polymer:nanocrystal devices and conventional, silicon pn devices.

• Received 5 August 1998

DOI:https://doi.org/10.1103/PhysRevB.59.10885