(311d) Interfacial Electron Transfer in Dye Sensitized Solar Cells Measured by Time Resolved Terahertz Spectroscopy
AIChE Annual Meeting
2007
2007 Annual Meeting
Nanomaterials for Energy Applications
Nanomaterials for Photovoltaics
Tuesday, November 6, 2007 - 4:33pm to 4:54pm
Dye sensitized solar cells (DSSCs) rely on efficient interfacial electron transfer from photoexcited dye molecules into wide band gap semiconductor nanoparticles on which the dye is adsorbed. In this paper, we report the charge injection dynamics for two commonly used DSSC dyes into TiO2 and ZnO nanoparticles. Charge injection is monitored on picosecond to nanosecond times scales by time-resolved terahertz spectroscopy (TRTS), a non-contact optical probe of the semiconductor conductivity with sub-picosecond time resolution. Dye molecules inject electrons into the semiconductor after photoexcitation with a 400 nm or 800 nm pump pulse, causing an increase in the semiconductor's electron density which is detected by the THz probe pulse. The dynamics of the injection process are investigated by varying the delay time between optical pump and THz probe pulses.
Electron injection times from the ruthenium organometallic dyes N3, "red dye," and N749, "black dye," vary from time scales of hundreds of femtoseconds to hundreds of picoseconds depending on the combination of dye, semiconductor, and excitation wavelength. Injection dynamics and quantum efficiency are determined by the electronic coupling between available states in the semiconductor conduction band and the dye's excited states. We find that injection into TiO2 is generally ultrafast, approaching the ~500 fs instrument response time. Injection dynamics into TiO2 are independent of dye with 400 nm excitation. However, with 800 nm excitation, injection from black dye takes ~10's of ps while injection from red dye remains ultrafast. Injection into ZnO is much slower than into TiO2, taking ~100 ps. The dynamics are independent of dye and excitation wavelength. The exception that is that is essentially no injection from red dye into ZnO using 800 nm excitation since there is insufficient photon energy to excite dye states that are above the ZnO conduction band. Literature suggests that faster injection into TiO2 than ZnO is expected because electrons are injected into a high density of TiO2 3d orbitals while injection into ZnO is into a lower density of s and p orbitals. All of the measured time scales are much faster than the 59 nanosecond relaxation time of the excited dye molecule in solution and should result in efficient interfacial electron transfer in DSSCs.
We have also investigated injection into TiO2/ZnO, ZnO/TiO2, and ZnO/Al2O2 core/shell nanoparticles. The use of a shell layer with higher conduction band potential than the core can reduce recombination in DSSCs by confining electrons to the nanoparticle core. However, the presence of the shell can change the injection dynamics as well. Shell layers were grown on nanoparticle films in solution by successive ionic layer adsorption and reaction (SILAR). Increasingly thick TiO2 shells on ZnO nanoparticles show the development of an ultrafast injection component in addition to the slower injection found in ZnO. However, the injected electron density decreases compared to uncoated ZnO, indicating that electrons may become trapped at the interface or remain in TiO2instead of proceeding to the ZnO core. TiO2/ZnO and ZnO/Al2O2 core/shell nanoparticles show decreased electron intensity without any significant change in dynamics. The lower electron density in coated nanoparticles indicates that electrons can easily become trapped at the core-shell interface. Studies of the microstructure of these interfaces are required for further interpretation of the data.