Abstract
Two 8 μm thick TiO2 photoelectrodes have been sensitized separately by N719 dye molecules and CdS quantum dots for a comparison study. Photoinduced absorption (PIA) spectroscopy was employed to investigate the mechanistic properties of electrons under illumination conditions comparable to sunlight. The PIA spectrum of both electrodes (in the presence of electrolyte) is due to electrons in TiO2 and iodine radicals in the electrolyte. In the absence of redox electrolyte, both electrodes show long-lived photoinduced charge-separation with lifetime in a millisecond range (8.5 ms for Q-dot-sensitized TiO2 and 11.5 ms for dye-sensitized TiO2).
1. Introduction
Nanostructured solar cells sensitized by organic dyes (DSSCs) [1–6] or by inorganic short bandgap semiconductors (also called quantum dots, QDs) [7–10] have attracted a great deal of interest. They are capable to obtain efficient conversion of solar energy to electricity at a low cost comparative to conventional semiconductor photovoltaic devices [11, 12]. The approach of using semiconductor colloids for the design of optically transparent thin semiconductor films is considered as a unique and an alternative for the amorphous silicon solar cells. Using this approach, dye-sensitized solar cells based on bi- and polypyridyl ruthenium complexes have achieved solar-to-electrical energy conversion efficiencies of 10-11% under AM 1.5 irradiation [1–3]. On the other hand, wide bandgap semiconductors have also been sensitized by short bandgap quantum dots (CdSe/TiO2 [6], CdS/TiO2-SnO2 [10]) as alternative to dye sensitization. Vogel and coworkers [8] have investigated the sensitization of nanoporous TiO2, ZnO, and so forth by Q-sized CdS. Photocurrent quantum yields up to 80% and open-circuit voltages up to 1 V range were obtained. In contrast with the dye-sensitized solar cells, fundamental understanding of factors controlling the interfacial electron transfer reactions in QD sensitized solar cells is limited.
In this paper, we report photoinduced absorption spectroscopy of an organic (N719 dye)-sensitized and inorganic (Q-dot CdS) semiconductor-sensitized TiO2 (8 μm) photoelectrodes under illumination conditions comparable to sunlight in order to compare the mechanistic properties of electrons.
2. Experimental
2.1. Preparation of Nanostructured TiO2 Films
We take care more about similarity in TiO2 thicknesses in order to insure the same length of e-transfer from different sensitizer to external circuit. Before coating conducting glass ITO with TiO2 nanoparticles, first we coated a blocking layer by immersing ITO plates into 0.02 M TiCl4 solution at 70°C for 30 min. TiO2 layer was made with EPFL paste by the Doctor Blade technique. Working electrodes were then sintered at 450°C for 30 min using heat gun and cooled down to room temperature.
2.2. Surface Modification of TiO2 by Quantum Dots CdS
TiO2 metal oxide nanostructured electrodes were successively dipped into an aqueous solution of saturated Cd(ClO4)2 and 0.1 M Na2S for 1 and 2 min, respectively. After each CdS layer deposition, the electrodes were heated at 125°C for 5 min.
2.3. Surface Modification of TiO2 by N719 Dye
The relatively hot (~80°C) TiO2 nanostructured electrodes were immersed in a 0.5 mM ethanolic solution of N719 dye ((TBA)2-cis-Ru(Hdcbpy)2-(NCS)2, Solaronix, Switzerland) and left for about 2 hours only. (There was no need for overnight adsorption, just a reasonable adsorption to be able to investigate a good mechanistic study.)
2.4. Characterization Methods
UV-Vis spectra were recorded using a Hewlett-Packard 8453 diode array spectrometer. For PIA spectroscopy (Figure 1), excitation of the sample was provided by light from a blue LED (Luxeon Star 1 W, Royal Blue, 470 nm), which was square-wave modulated (on/off) by electronical means using an HP 33120 A waveform generator and a home-built LED driver system. The beam, with an intensity in the range of 0.5–30 mW/cm2, excited a sample area of about 1 cm2. White probe light was provided by a 20 W tungsten-halogen lamp. A cutoff filter (Schott RG715) was used to minimize excitation of the sample by the probe light where indicated. The transmitted probe light was focused onto a monochromator (Acton Research Corporation SP-150) and detected using a UV-enhanced Si photodiode, connected to a lock-in amplifier via a current amplifier (Stanford Research Systems models 830 and 570, resp.). For the time-resolved studies the output of the current amplifier was connected to a data acquisition board (National Instruments PCI-6052E). All PIA measurements were done at room temperature.
3. Results and Discussion
3.1. UV-Vis Absorption Spectra
Figure 2 shows UV-Vis absorption spectra of bare TiO2 film, CdS-sensitized, and N719-sensitized TiO2 films. It is clear that CdS and N719 dye both have extended further the absorption into the visible up to 600 and 750 nm, respectively.
3.2. PIA Spectroscopy of CdS/TiO2 System
Figure 3 shows typical PIA spectra of the CdS-nanostructured TiO2 system. In the absence of the redox electrolyte, the PIA spectrum clearly reflects the differential spectrum of CdS upon formation of the oxidized CdS (or hole formation), with a bleach of the main absorption band at 470 nm. The PIA spectrum after the onset of 550 nm should reveal the spectrum of injected electrons into TiO2 conduction band with a peak around 650 nm compared to CdS alone on ITO. In the later, large recombination occurs when it failed to be transported far away form the existing created holes. When the redox electrolyte is added, the PIA spectrum changes significantly. The oxidized CdS is rapidly reduced by the electrolyte, and thus little bleaching is observed under 530 nm. The brad absorption at wavelengths larger than 600 nm can be attributed to injected electrons in TiO2.
In Figure 4, PIA spectra of CdS-adsorbed TiO2 electrodes in the presence of electrolyte and under open circuit are shown. When the redox electrolyte, composed of 0.7 M LiI and 0.05 M I2 in 3MPN, is added, the PIA spectrum changes significantly. The absorption increases up to 600 nm and remains almost flat up to 800 nm. We attribute the PIA signals to electrons in nanostructured TiO2 and iodine radicals () [13]. Triiodide does not absorb at wavelengths larger than 600 nm, while exhibits a broad absorption peak at 750 nm with an extinction coefficient of 2200 M−1 [14].
Study of the kinetics in semiconductor sensitizing solar cells is not only feasible by laser flash photolysis but also possible using time-resolved PIA measurements. Figure 5 shows such a typical PIA transient decay recorded at 700 nm. Pseudo-first-order rate constant for e-injection from CdS conduction band to TiO2 conduction band was about 8.5 ms approximately which does not follow simple first-order kinetic but is characterized by a range of injection times. This relatively fast decay proves at least a well pore filling of TiO2 film by ultrafine CdS particles.
3.3. PIA Spectroscopy of N719/TiO2 System
Figure 6 shows typical PIA spectra of the dye-sensitized TiO2 system. In the absence of the redox electrolyte (Figure 6(a)), the PIA spectrum clearly reflects the differential spectrum of N719 upon formation of the oxidized dye, with a bleach of the ground state MLCT (metal-to-ligand charge transfer) main absorption band at 540 nm and an absorption peak at 780 nm (Figure 6(a)) which corresponds to MLCT transitions. Electrons in the TiO2 will also be present in the PIA. When the redox electrolyte composed of 0.7 M LiI and 0.05 M I2 in 3MPN is added (Figure 6(b)), the PIA spectrum changes significantly. The oxidized dye is rapidly reduced by the iodide in the electrolyte, presumably according to the following reactions [15–17].
The resulting PIA spectrum should reveal the spectrum of electrons in TiO2 electrons in TiO2 as well as triiodide in the electrolyte. The gradually increasing absorption with wavelengths larger than 600 nm can be attributed to electrons in TiO2 and iodine radicals [13]. Electrons in nanostructured TiO2 exhibit a very broad absorption and have an extinction coefficient of about 1000 M−1 cm−1 at 800 nm [18]. Triiodide does not absorb at wavelengths larger than 600 nm, while exhibits a broad absorption peak around 750 nm with an extinction coefficient of 2200 M−1 cm−1. Recently, it was shown in pulsed nanosecond laser studies that is an important long-lived intermediate in the dye-sensitized solar cell [17, 19]. The PIA spectrum of Figure 6(b) has contributions of both electrons and radicals.
Figure 7 shows typical PIA transient decay recorded at 750 nm. Pseudo-first-order rate constant for e-injection from dye excited state to TiO2 conduction band was about 12 ms. This does not follow simple first-order kinetic, but is characterized by a range of injection times. This has been explained by trapping of electrons within the TiO2 nanocrystals [16]. Similar pseudo-first-order rate constant of 11 ms has been measured using PIA transient decay for N719/TiO2 for electron recombination [20]. This relatively fast decay proves at least a well pore filling of TiO2 film by N719 dye molecules. The CdS/TiO2 system shows its first-order rate constant for e-injection of 8.5 ms which is relatively faster than that of the N719/TiO2 system (12 ms). This relatively faster kinetic process seen in CdS/TiO2 system may reflect its faster process of high kinetic processes in the range of picoseconds already observed with this system [21]. For solar cell performance the pseudo-first-order rate constant under steady-state conditions is a relevant parameter, as it can give direct information on possible recombination losses due to the reaction of electrons with oxidized dye molecules.
4. Conclusion
Photoinduced absorption spectroscopy where the excitation is provided by an on/off monochromatic light source can give direct information on electron-injection and hole-electron recombination rates using spectra of transient species, and their kinetics can be explored using time-resolved techniques. PIA can monitor slow processes and is cheaper compared to laser flash photolysis. The relatively faster kinetic process observed in CdS/TiO2 system compared to N719/TiO2 system could reflect its faster process of high kinetic processes in the range of picoseconds already observed.
Acknowledgments
The research project is funded by the National Plan for Science and Technology Program, Grant no. 09-NAN859-02, King Saud University, Riyadh, Saudi Arabia. Id. Bedja would like to thank King Abdullah Institute for Nanotechnology at King Saud University, Riyadh, Saudi Arabia, for a financial support. Thanks also go to Dr. Gerrit Boschloo for his helpful discussions.