Virtues and Vices of an Organic Dye and Ti-Doped MCM-41 Based Dye-Sensitized Solar Cells

Abstract

We report on femtosecond to millisecond (fs-ms) studies of the interactions of an efficient triphenylamine organic dye (TPC1) for photovoltaics with titanium-doped high porous MCM-41 (TiMCM-41) silicate materials in half and complete dye-sensitized solar cells (DSSCs). Stationary UV–visible absorption results indicate a higher dye loading (3–5 times) per Ti atom in the TiMCM-41-based solar cells in comparison with those based on typical titania nanoparticles (NPs). However, the dye loading per Ti atom decreases with increasing the Ti doping, and the total dye content is still smaller in the TiMCM-41 solar cells than in the NP ones. Time-resolved emission studies showed that the average electron injection times (from TPC1 to the titania conduction band) are about 2.5 times longer for the TiMCM-41 solar cells (12 ps) than using the classical titania NP ones (4.5 ps). However, taking into account the slow internal deactivation of the dye in both materials, the yield of electron injection is >90%. Nanosecond to millisecond (ns-ms) flash photolysis studies of films show that the (back) electron recombination from titania to the dye cation is slower using TiMCM-41 than that in the titania NPs (19 μs vs 7 μs). Similar experiments for complete solar cells indicate faster dye regeneration due to the electrolyte in the TiMCM-41 than in the titania NP devices (3 μs vs 8 μs). Therefore, from the point of view of interfacial charge separation, the results indicate comparable or even better performance of the TiMCM-41 than the typical titania nanoparticle. However, the solar cell performance and the total efficiency are much lower in the TiMCM-41 than in the NP devices. Additionally, the values of short-circuit current per Ti atom are higher for the NP devices, despite the higher dye content per Ti atom in TiMCM-41 samples. This suggests a strong limitation in the electron transport process along the TiMCM-41 channels. Hence, despite the promising interfacial properties observed in fast and ultrafast time scales, this kind of material needs further modification in terms of improving the total dye loading and the conductivity, in order to be suitable for DSSCs.