The flash-photolysis time-resolved microwave conductivity technique (FP-TRMC) has been used to study photoinduced charge separation in bilayers consisting of a smooth, transparent, 80 nm thick layer of anatase TiO 2 onto which poly(3-hexylthiophene) (P3HT) sensitizer layers have been spin-coated. Interfacial charge separation, resulting from excitation of the polymer in the visible, is found to persist well into the millisecond time domain. Photoconductivity action spectra have been measured between 420 and 700 nm for P3HT layer thicknesses, L, from ∼2 to 200 nm. Using this electrodeless technique, the bilayers could be irradiated from either the polymer ("front") or semiconductor ("back") side. On front-side irradiation at 540 nm (close to the absorption maximum of the polymer), the efficiency of charge separation per incident photon (IPCSE) initially increased to a maximum value of 0.8% for L ≈ 10 nm. For thicker layers the IPCSE gradually decreased, eventually to 0.1% for L ≈ 170 nm. On back-side irradiation the IPCSE increased over the first 10 nm to a value close to the maximum found for front-side irradiation, and decreased only slightly for further increase in layer thickness. Analytical expressions for the thickness dependence based on exciton diffusion with a Lambert-Beer excitation profile have been used to fit the experimental data. Best fits were obtained for an exciton diffusion length, Λ () (Dτ) with D the diffusion coefficient and τ the natural lifetime), of 5.3 or 2.6 nm depending on whether excitons were taken to be reflected or quenched at the polymer/gas interface, respectively. The IPCSE decreased at high light intensities; an effect that is attributed to the occurrence of exciton-exciton annihilation within the polymer layer.

We have studied photoinduced charge separation in a bare, 3.4 microm thick layer of nanocrystalline ("nc") anatase TiO(2) and an nc-TiO(2) layer coated with free-base 5,10,15,20-tetrakis(4-carboxyphenyl) porphyrin (H(2)TPPC) using the electrodeless flash-photolysis time-resolved microwave-conductivity technique (FP-TRMC). Photoconductivity transients, resulting from the formation of mobile, conduction band electrons in the semiconductor have been measured on excitation with 3 ns pulses of UV (300 nm) and visible (410-700 nm) light. The product of the yield of formation of mobile charge carriers, phi, and the sum of their mobilities, Sigmamicro, has been determined from the maximum conductivity for light intensities varying from approximately 10(12) to approximately 10(16) photons/cm(2)/pulse. For the bare nc-TiO(2) layer at 300 nm and the coated layer at all wavelengths, phiSigmamicro initially increased with increasing intensity, reached a maximum, and eventually decreased at high intensities. The initial increase is attributed to the gradual filling of (surface) electron trapping sites. This effect was absent when the samples were continuously illuminated with background irradiation at 300 nm with an intensity of 6 x 10(13) photons/cm(2)/s (40 microW/cm(2)), thereby presaturating the trapping sites prior to the laser pulse. The trap-free mobility of electrons within these 9 nm nanoparticles is estimated to be 0.034 cm(2)/Vs at 9 GHz. The eventual decrease in phiSigmamicro at intensities corresponding to an electron occupancy of more than one electron per particle is unaffected by background illumination, and is attributed to a decrease in micro due to electron-electron interactions within the semiconductor particles. The photoconductivity action spectrum of the coated nc-TiO(2) layer closely followed the photon attenuation spectrum in the visible of the porphyrin, with a charge separation efficiency per absorbed photon of 18% at the Soret band maximum. The after-pulse decay of the photoconductivity showed a power law behavior over a time scale of nanoseconds to several hundreds of microseconds, which is attributed to multiple trapping and detrapping events at chemical or physical defects within the semiconductor matrix.

The optimization of interfacial charge transfer is crucial to the design of dye-sensitized solar cells. In this paper we address the dynamics of the charge separation and recombination in liquid-electrolyte and solid-state cells employing a series of amphiphilic ruthenium dyes with varying hydrocarbon chain lengths, acting as an insulating barrier for electron-hole recombination. Dynamics of electron injection, monitored by time-resolved emission spectroscopy, and of charge recombination and regeneration, monitored by transient optical absorption spectroscopy, are correlated with device performance. We find that increasing dye alkyl chain length results in slower charge recombination dynamics to both the dye cation and the redox electrolyte or solid-state hole conductor (spiro-OMeTAD). These slower recombination dynamics are however paralleled by reduced rates for both electron injection into the TiO2 electrode and dye regeneration by the I-/I3- redox couple or spiro-OMeTAD. Kinetic competition between electron recombination with dye cations and dye ground state regeneration by the iodide electrolyte is found to be a key factor for liquid electrolyte cells, with optimum device performance being obtained when the dye regeneration is just fast enough to compete with electron-hole recombination. These results are discussed in terms of the minimization of kinetic redundancy in solid-state and liquid-electrolyte dye-sensitized photovoltaic devices.

Charge transport rate at open-circuit potential (V(oc)) is proposed as a new characterization method for dye-sensitized (DS) and other nanostructured solar cells. At V(oc), charge density is flat and measurable, which simplifies quantitative comparison of transport and charge density. Transport measured at V(oc) also allows meaningful comparison of charge transport rates between different treatments, temperatures, and types of cells. However, in typical DS cells, charge transport rates at V(oc) often cannot be measured by photocurrent transients or modulation techniques due to RC limitations and/or recombination losses. To circumvent this limitation, we show that charge transport at V(oc) can be determined directly from the transient photovoltage rise time using a simple, zero-free-parameter model. This method is not sensitive to RC limitation or recombination losses. In trap limited devices, such as DS cells, the comparison of transport rates between different devices or conditions is only valid when the Fermi level in the limiting conductor is at the same distance from the band edge. We show how to perform such comparisons, correcting for conduction band shifts using the density of states (DOS) distribution determined from the same photovoltage transients. Last we show that the relationship between measured transport rate and measured charge density is consistent with the trap limited transport model.