Electron injection from the transition metal complex Ru(dcbpy)(2)(NCS)(2) (dcbpy = 4,4'-dicarboxy-2,2'-bipyridine) into a titanium dioxide nanocrystalline film occurs on the femto- and picosecond time scales. Here we show that the dominating part of the electron transfer proceeds extremely rapidly from the initially populated, vibronically nonthermalized, singlet excited state, prior to electronic and nuclear relaxation of the molecule. The results are especially relevant to the understanding and design of molecular-based photovoltaic devices and artificial photosynthetic assemblies.

Time-resolved absorption spectroscopy was used to study the femtosecond and picosecond time scale electron injection from the excited singlet and triplet states of Ru(dcbpy) 2 (NCS) 2 (RuN3) into titanium dioxide (TiO 2 ) nanocrystalline particle film in acetonitrile. The fastest resolved time constant of ∼30 fs was shown to reflect a sum of two parallel ultrafast processes, nonergodic electron transfer (ET) from the initially excited singlet state of RuN3 to the conduction band of TiO 2 and intersystem crossing (ISC). The branching ratio of 1.5 between the two competing processes gives rate constants of 1/50 fs -1 for ET and 1/75 fs -1 for ISC. Following the ultrafast processes, a minor part of the electron injection (40%) occurs from the thermalized triplet state of RuN3 on the picosecond time scale. The kinetics of this slower phase of electron injection is nonexponential and can be fitted with time constants ranging from ∼1 to ∼60 ps.

A zinc phthalocyanine with tyrosine substituents (ZnPcTyr), modified for efficient far-red/near-IR performance in dye-sensitized nanostructured TiO(2) solar cells, and its reference, glycine-substituted zinc phthalocyanine (ZnPcGly), were synthesized and characterized. The compounds were studied spectroscopically, electrochemically, and photoelectrochemically. Incorporating tyrosine groups into phthalocyanine makes the dye ethanol-soluble and reduces surface aggregation as a result of steric effects. The performance of a solar cell based on ZnPcTyr is much better than that based on ZnPcGly. Addition of 3alpha,7alpha-dihydroxy-5beta-cholic acid (cheno) and 4-tert-butylpyridine (TBP) to the dye solution when preparing a dye-sensitized TiO(2) electrode diminishes significantly the surface aggregation and, therefore, improves the performance of solar cells based on these phthalocyanines. The highest monochromatic incident photo-to-current conversion efficiency (IPCE) of approximately 24% at 690 nm and an overall conversion efficiency (eta) of 0.54% were achieved for a cell based on a ZnPcTyr-sensitized TiO(2) electrode. Addition of TBP in the electrolyte decreases the IPCE and eta considerably, although it increases the open-circuit photovoltage. Time-resolved transient absorption measurements of interfacial electron-transfer kinetics in a ZnPcTyr-sensitized nanostructured TiO(2) thin film show that electron injection from the excited state of the dye into the conduction band of TiO(2) is completed in approximately 500 fs and that more than half of the injected electrons recombines with the oxidized dye molecules in approximately 300 ps. In addition to surface aggregation, the very fast electron recombination is most likely responsible for the low performance of the solar cell based on ZnPcTyr.

Electron injection from the transition metal complex Ru(dcbpy) 2 (NCS) 2 (dcbpy ) 2,2′-bipyridine-4,4′dicarboxylate) into a titanium dioxide nanoparticle film occurs along two pathways. The dominating part of the electron injection proceeds from the initially excited singlet state of the sensitizer into the conduction band of the semiconductor on the sub-hundred-femtosecond time scale. The slower part of the injection occurs from the thermalized triplet excited state on the picosecond time scale in a nonexponential fashion, as was shown in a previous study et al. J. Am. Chem. Soc. 2002, 124, 489). Here we show that the slower channel of injection is the result of the excited state being localized on a ligand of the sensitizer that is not attached to the semiconductor; hence, the electron cannot be injected directly from such an excited state into the semiconductor. Before being injected, it has to be transferred from the non-surface-attached ligand to the attached one. The results show that the interligand electron-transfer time is on the picosecond time scale, depends on the relative energies of the two ligands, and controls the electron injection from the excited triplet state of the sensitizer. The findings provide information relevant to the design of molecularbased assemblies and devices.

Photoinduced electron injection from the sensitizer Ru(dcbpy)2(NCS)2 (RuN3) into SnO2 and TiO2 nanocrystalline films occurs by two distinct channels on the femto- and picosecond time scales. The faster electron injection into the conduction band of the different semiconductors originates from the initially excited singlet state of RuN3, and occurs in competition with intersystem crossing. The rate of singlet electron injection is faster to TiO2 (1/55 fs-1) than to SnO2 (1/145 fs-1), in agreement with higher density of conduction band acceptor states in the former semiconductor. As a result of competition between the ultrafast processes, for TiO2 singlet, whereas for SnO2 triplet electron injection is dominant. Electron injection from the triplet state is nonexponential and can be fitted with time constants ranging from approximately 1 ps (2.5 ps for SnO2) to approximately 50 ps for both semiconductors. The major part of triplet injection is independent of the semiconductor and is most likely controlled by intramolecular dynamics in RuN3. The overall time scale and the yield of electron injection to the two semiconductors are very similar, suggesting that processes other than electron injection are responsible for the difference in efficiencies of solar cells made of these materials.

Electron injection and recombination dynamics of the dye Fluorescein 27 adsorbed to a nanocrystalline titanium dioxide (TiO2) thin film in acetonitrile (CH3CN) were studied by femtosecond pump−probe spectroscopy. After excitation of the dye at the absorption maximum, transient absorption spectra and kinetics were recorded in the spectral region between 400 and 2000 nm. It was found that most of the transient spectrum is dominated at early times by induced excited-state absorption (ESA). Even in the near-IR spectral region, where the ESA of the dye in aqueous solution is less than the noise level of our measurements, a pronounced ESA of the dye/TiO2 system has been observed. Photoproduct formation following electron injection from the dye molecule into the conduction band of the semiconductor can be resolved in spectral regions where ESA is either canceled by stimulated emission (SE) or is compensated by ground-state absorption bleach at early times. The remaining signals at these wavelengths reflect the dynamics of photoproducts, which are conduction-band electrons in TiO2 and oxidized Fluorescein 27 dye molecules. The kinetics of SE decay and the rise times of induced absorption of photoproducts are ultrafast and nonexponential, requiring at least three time constants ranging from <100 fs to ∼8 ps. The depopulation of the excited state monitored by SE decay corresponds very well to the formation of the photoproducts. Recombination is also nonexponential, part of it happens in tens of picoseconds, but the major part does not occur on the investigated time scale (until 500 ps). The possibility of resolving the dynamics of both precursor and product species in Fluorescein 27−TiO2 nanoparticle thin films makes it suitable for spectroscopicaly studying interfacial electron transfer as a function of system parameters.

The dynamics of photoinduced electron injection and recombination between all-trans-8'-apo-beta-caroten-8'-oic acid (ACOA) and a TiO(2) colloidal nanoparticle have been studied by means of transient absorption spectroscopy. We observed an ultrafast ( approximately 360 fs) electron injection from the initially excited S(2) state of ACOA into the TiO(2) conduction band with a quantum yield of approximately 40%. As a result, the ACOA(*)(+) radical cation was formed, as demonstrated by its intense absorption band centered at 840 nm. Because of the competing S(2)-S(1) internal conversion, approximately 60% of the S(2)-state population relaxes to the S(1) state. Although the S(1) state is thermodynamically favorable to donate electrons to the TiO(2), no evidence was found for electron injection from the ACOA S(1) state, most likely as a result of a complicated electronic nature of the S(1) state, which decays with a approximately 18 ps time constant to the ground state. The charge recombination between the injected electrons and the ACOA(*)(+) was found to be a highly nonexponential process extending from picoseconds to microseconds. Besides the usual pathway of charge recombination forming the ACOA ground state, about half of the ACOA(*)(+) recombines via the ACOA triplet state, which was monitored by its absorption band at 530 nm. This second channel of recombination proceeds on the nanosecond time scale, and the formed triplet state decays to the ground state with a lifetime of approximately 7.3 micros. By examination of the process of photoinduced electron transfer in a carotenoid-semiconductor system, the results provide an insight into the photophysical properties of carotenoids, as well as evidence that the interfacial electron injection occurs from the initially populated excited state prior to electronic and nuclear relaxation of the carotenoid molecule.

In most of the previous ultrafast electron injection studies of Ru(dcbpy)2(NCS)2-sensitized nanocrystalline TiO2 films, experimental conditions and sample preparation have been different from study to study and no studies of how the differences affect the observed dynamics have been reported. In the present paper, we have investigated the influence of such modifications. Pump photon density, environment of the sensitized film (solvent and air), and parameters of the film preparation (crystallinity and quality of the film) were varied in a systematic way and the obtained dynamics were compared to that of a well-defined reference sample: Ru(dcbpy)2(NCS)2-TiO2 in acetonitrile. In some cases, the induced changes in the dynamics were uncorrelated to the electron injection process. High pump photon density (not in the linear response region) and exposure of the sensitized film to air altered the picosecond-time-scale kinetics considerably, and the changes were attributed mostly to degradation of the dye. In other cases, changes in the measured kinetics were related to the electron injection processes: reducing the firing temperature of the nanocrystalline film or making the film via electron beam evaporation (EBE) resulted in a decrease of the overall crystallinity of the film, and the electron injection slowed. In the sensitized EBE films, in addition to an increased contribution of triplet excited-state electron injection, a new electron transfer (ET) process with a time constant of 200 fs was observed.