doi:10.1016/j.ccr.2004.03.027 
Copyright © 2004 Elsevier B.V. All rights reserved.
Primary and final charge separation in the nano-structured dye-sensitized electrochemical solar cell
Klaus Schwarzburg, Ralph Ernstorfer, Silke Felber and Frank Willig
, 
Dynamics of Interfacial Reactions, Department SE4, Hahn-Meitner-Institut, Glienickerstrasse 100, 14109, Berlin, Germany
Received 31 December 2003;
accepted 31 December 2003.
Available online 22 July 2004.
References and further reading may be available for this article. To view references and further reading you must
purchase this article.
Abstract
A time constant of 13 fs was measured via transient absorption for electron transfer from the excited singlet state of the chromophore perylene into anatase TiO2, when the chromophore was bonded to the surface via a carboxyl group. This result suggests that a similarly short time constant should be valid also for primary charge separation from the bipyridyl ligand of the so-called N3 ruthenium dye into anatase TiO2. The electron transfer time of perylene became much longer, 3.8 ps, at a distance of about 1.3 nm. Laser pulse induced transients of the photocurrent were measured in the Grätzel-cell (G-cell) for illumination firstly through the SnO2/TiO2 interface (SE) and secondly through the opposite SnO2/electrolyte interface (EE). For EE illumination the transient showed two peaks, the first in the μs range and the second in the ms range. For SE illumination only the early peak was observed. The different shapes of EE compared to SE transients were easily modeled employing a time dependent diffusion equation and appropriate boundary conditions. The experimental rise to the early peak occurred faster than corresponds to diffusion controlled final charge separation at the SnO2/TiO2 interface if the same diffusion constant was assumed that was valid in the bulk of the nm-structured sponge-type TiO2/electrolyte layer. The significance of this experimental result for photovoltaic energy conversion of the G-cell is discussed.
Author Keywords: Solar cell; Photocurrent transients; TiO2; Modeling; UPS; Perylene; Femtosecond spectroscopy; Electron transfer; Transport
Fig. 1. Time resolved measurement of electron transfer from the excited singlet state of the chromophore perylene bonded via a carboxyl group with close to perpendicular orientation of its long axis on the surface of anatase TiO2.
Fig. 2. UPS difference spectrum showing the ground state energy, peak in the center, and the first excited singlet state energy, peak to the right, of the chromophore perylene relative to the Fermi level in sponge-like nm-structured anatase TiO2 where the chromophore perylene was attached via a carboxyl group.
Fig. 3. Rise of the perylene cation absorption after generating the first excited singlet state of perylene anchored with a long tripod-type anchor-cum-spacer group on anatase TiO2.
Fig. 4. Equivalent circuit representing the resistive and capacitive behavior of the G-cell with an external load Re. The current source js(t) is derived from a diffusion model with source term, lifetime, and appropriate boundary condition for charge separation.
Fig. 5. Upper panel: initial spatial distribution of photo-generated electrons in the G-cell for laser pulse illumination through the SnO2 electrode (SE illumination). Center panel: two typical measured photocurrrent transients for SE vs. EE illumination. Lower panel: initial spatial distribution of photo-generated electrons in the G-cell for laser pulse illumination through the electrolyte space (EE illumination). Two consecutive peaks can be seen in the photocurrent transient (center panel) for the EE case but not for the SE case.
Fig. 6. (a) First peak of the measured photocurrent transients for different values of the external resistor Re given as parameter value at the respecitve curve. (b) Plot of the magnitude of the measured first photocurrent peak (indicated at the abscissa as 1/RC-amplitude) for different values of the external resistor Re vs. the value of Re for the respective measured photocurrent peak value
Fig. 7. Simulated photocurrent transients for EE and SE incidence of the laser pulse. Curves are normalized at the maximum amplitude. Parameters: D = 2 × 10−6 cm2/s, L = 5.4 μm, α = 8×103 cm−1, τ = 1 s, kes= 100 cm/s, C = 500 nF, R = 50 Ω, Re = 0 Ω, width of Gauss PULSE = 500 ns FWHM.
Fig. 8. Simulated photocurrent transients for different values of the external resistor Re. Same parameter values as in Fig. 7, except for Re assuming values between 0 and 100 Ω indicated at the respective curve.
Fig. 9. Simulated photocurrent transients for different values of kes, otherwise same parameter values as for Fig. 7.
Fig. 10. Comparison of the experimental rise of the first peak of the photocurrent transient (solid squares) with a transient simulated with the diffusion model (dashed curve). Adding an instantaneous contribution to the photocurrent transient simulated with the diffusion model (solid curve) results in a perfect match with the experimental data.
Abstract
Purchase PDF (268 K)
E-mail Article
Add to my Quick Links
Cited By in Scopus (31)
Purchase PDF (124 K)
AgCl) via the selective reduction of H2O2 at the TiO2 surface. The analytical characteristics of different poly 
