Regeneration and recombination kinetics was investigated for dye-sensitized solar cells (DSCs) using a series of different cobalt polypyridine redox couples, with redox potentials ranging between 0.34 and 1.20 V vs. NHE. Marcus theory was applied to explain the rate of electron transfer. The regeneration kinetics for a number of different dyes (L0, D35, Y123, Z907) by most of the cobalt redox shuttles investigated occurred in the Marcus normal region. The calculated reorganization energies for the regeneration reaction ranged between 0.59 and 0.70 eV for the different organic and organometallic dyes investigated. Under the experimental conditions employed, the regeneration efficiency decreased when cobalt complexes with a driving force for regeneration of 0.4 eV and less were employed. The regeneration efficiency was found to depend on the structure of the dye and the concentration of the redox couples. [Co(bpy-pz)2](2+), which has a driving force for regeneration of 0.25 eV for the triphenylamine based organic dye, D35, was found to regenerate 84% of the dye molecules, when a high concentration of the cobalt complex was used. Recombination kinetics between electrons in TiO2 and cobalt(iii) species in the electrolyte was also studied using steady state dark current measurements. For cobalt complexes with highly positive redox potentials (>0.55 V vs. NHE) dark current was found to decrease, consistent with electron transfer reactions occurring in the Marcus inverted region. However, for the cobalt complexes with the most positive redox potentials an increase in dark current was found, which can be attributed to recombination mediated by surface states.

The relaxation dynamics of the indoline dye D149, a well-known sensitizer for photoelectrochemical solar cells, have been extensively characterized in various organic solvents by combining results from ultrafast pump-supercontinuum probe (PSCP) spectroscopy, transient UV-pump VIS-probe spectroscopy, time-correlated single-photon counting (TCSPC) measurements as well as steady-state absorption and fluorescence. In the steady-state spectra, the position of the absorption maximum shows only a weak solvent dependence, whereas the fluorescence Stokes shift Dñ F correlates with solvent polarity. Photoexcitation at around 480 nm provides access to the S 1 state of D149 which exhibits solvation dynamics on characteristic timescales, as monitored by a red-shift of the stimulated emission and spectral development of the excited-state absorption in the transient PSCP spectra. In all cases, the spectral dynamics can be modeled by a global kinetic analysis using a time-dependent S 1 spectrum. The lifetime t 1 of the S 1 state roughly correlates with polarity [acetonitrile (280 ps) o acetone (540 ps) o THF (720 ps) o chloroform (800 ps)], yet in alcohols it is much shorter [methanol (99 ps) o ethanol (178 ps) o acetonitrile (280 ps)], suggesting an appreciable influence of hydrogen bonding on the dynamics. A minor component with a characteristic time constant in the range 19-30 ps, readily observed in the PSCP spectra of D149 in acetonitrile and THF, is likely due to removal of vibrational excess energy from the S 1 state by collisions with solvent molecules. Additional weak fluorescence in the range 390-500 nm is observed upon excitation in the S 0 -S 2 band, which contains short-lived S 2 -S 0 emission of D149. Transient absorption signals after excitation at 377.5 nm yield an additional time constant in the subpicosecond range, representing the lifetime of the S 2 state. S 2 excitation also produces photoproducts.

The ultrafast photoinjection and subsequent relaxation steps of the indoline dye D149 were investigated in detail for a mesoporous electrodeposited ZnO thin film and compared with experiments on sintered TiO 2 and ZrO 2 thin films, all in contact with air, using pump-supercontinuum probe (PSCP) transient absorption spectroscopy in the range 370-770 nm. D149 efficiently injects electrons into the ZnO surface with time constants from r70 fs (time-resolution-limited) up to 250 fs, without the presence of slower components. Subsequent spectral dynamics with a time constant of 20 ps and no accompanying change in the oscillator strength are assigned to a transient Stark shift of the electronic absorption spectrum of D149 molecules in the electronic ground state due to the local electric field exerted by the D149 + radical cations and conduction band electrons in ZnO. This interpretation is consistent with the shape of the relaxed PSCP spectrum at long times, which resembles the first derivative of the inverted steady-state absorption spectrum of D149. In addition, steady-state difference absorption spectra of D149 + in solution from spectroelectrochemistry display a bleach band with distinctly different position, because no first-order Stark effect is present in that case. Interference features in the PSCP spectra probably arise from a change of the refractive index of ZnO caused by the injected electrons. The 20 ps component in the PSCP spectra is likely a manifestation of the transition from an initially formed bound D149 + -electron complex to isolated D149 + and mobile electrons in the ZnO conduction band (which changes the external electric field experienced by D149) and possibly also reorientational motion of D149 molecules in response to the electric field. We identify additional spectral dynamics on a similar timescale, arising from vibrational relaxation of D149 + by interactions with ZnO. TiO 2 exhibits similar dynamics to ZnO. In the case of ZrO 2 , electron injection accesses trap states, which exhibit a substantial probability for charge recombination. No Stark shift is observed in this case. In addition, the spectroelectrochemical experiments for D149 + in dichloromethane and acetonitrile, which cover the spectral range up to 2000 nm, provide for the first time access to its complete D 0 -D 1 absorption band, with the peak located at 1250 and 1055 nm, respectively. Good agreement is obtained with results from DFT/TDDFT calculations of the D149 + spectrum employing the MPW1K functional.