13
2008, 112, 13-17 Published on Web 12/13/2007
Spectral Response of Opal-Based Dye-Sensitized Solar Cells A. Mihi,† M. E. Calvo,† J. A. Anta,‡ and H. Mı´guez*,† Instituto de Ciencia de Materiales de SeVilla, Consejo Superior de InVestigaciones Cientı´ficas (CSIC), Ame´ rico Vespucio 49, 41092 SeVilla, Spain, and Departamento de Sistemas Fı´sicos, Quı´micos y Naturales, UniVersidad Pablo de OlaVide, 41013 SeVilla, Spain ReceiVed: NoVember 2, 2007; In Final Form: NoVember 26, 2007
Herein we present an experimental study of the spectral dependence of the photogenerated current of opalbased solar cells. We analyze the incident photon-to-current conversion efficiency (IPCE) for dye-sensitized solar cells in which colloidal crystals are introduced in different configurations. We prove that a dye-sensitized nanocrystalline titanium oxide electrode moulded in the shape of an inverse opal shows a decrease of efficiency for the spectral region in which a photonic stop band opens up. Contrarily, when a standard thin film of disordered titania nanocrystallites is coupled to an inverse opal, the mirror effect of the photonic crystal at band gap frequencies increases the light harvesting efficiency of the cell and thus the IPCE. This effect is further demonstrated by coupling an inverse opal multilayer to a homogeneous electrode, with two welldefined spectral ranges of increased photogenerated current being detected.
Introduction Nanocrystalline dye-sensitized solar cells1 (DSSC) have been shown to be promising alternatives to the more efficient but expensive silicon solar cells. DSSC are composed of a working electrode, typically a 5-10 µm thick film of nanocrystalline titania, sensitized with a visible absorbing dye, a liquid-iodidebased electrolyte that fills the interstitial space, and a platinised counterelectrode. DSSCs have achieved record efficiencies of around 11%2 and have technical advantages or added value such as ease of fabrication, low cost, and transparency. Generally, the ways explored to improve the efficiency of dye sensitized solar cells have focused on upgrading the components of the cell to enhance the photovoltage (by changing the supporting oxide), the photocurrent (by using different dyes), or the stability (by changing the electrolyte or the sealer).3 An interesting and complementary approximation that also leads to higher efficiencies is to introduce optical elements in the cell that increase the optical path of light, thus increasing the probability for the photons to be absorbed.4 By this means, the output power of the cell can be enhanced in terms of a higher photogenerated current since it depends on the light harvesting efficiency (LHE) or absorptance of the cell.5 In this direction, a layer of packed, large (on the order of a micron) titania particles within6 or on top of7 the working electrode has become a common ingredient of optimized cells. Such elements scatter light diffusively and very efficiently, increasing the optical path length within the absorbing electrode and thus enhancing the performance of the cell. Unfortunately, these scattering layers also turn the cells opaque, leaving DSSC useless as window modules or any other application where transparency, one of the added values of these cells, was required from these devices. * To whom correspondence should be addressed. E-mail: hernan@ icmse.csic.es. † CSIC. ‡ Universidad Pablo de Olavide.
10.1021/jp7105633 CCC: $40.75
In 2003, Mallouk et al. proposed the use of novel optical elements to improve the efficiency of DSSC. Their proposal consisted of structuring the working electrode in the shape of an inverse titania opal,8 a very porous kind of three-dimensional photonic crystal.9 By doing this, they sought to achieve an enhancement of the photocurrent for wavelengths corresponding to the borders of the photonic bands, at which the group velocity is reduced and thus the probability of absorption increased. Surprisingly, the enhanced wavelength interval was much wider than what they expected based on their hypothesis (∆λ at which the enhancement was measured was 120 nm instead of the 15 nm expected). Another intriguing feature Mallouk et al. pointed out was that the enhancement of the photocurrent was observed only when their cell was illuminated from the counterelectrode side. In that orientation, light first impinged onto a thick nanocrystalline dyed titania layer that was deposited onto the photonic crystal as a consequence of the inverse opal fabrication procedure. Effects originated from the coupling of the electrode to the photonic crystal were pointed out as the origin of the enhancement, but no theoretical or experimental confirmation of this hypothesis was given. Those results boosted several groups to start research on the photoelectrochemical applications of photonic crystals, particularly opals, on the basis of the possibility they offered to enhance optical absorption.10-13 Recently, in order to achieve further understanding of the phenomena that caused such enhancement, a theoretical analysis was performed and a detailed explanation of the observations was provided.14 By using a scalar wave approximation,15 we were able to simulate the optical response of DSSCs structured in the form of or including 3D inverse photonic crystals. Several conclusions were extracted from this study. First, a dyesensitized electrode shaped as an inverse opal would not lead to higher efficiency overall, since the dielectric mirror effect that arises as a consequence of the photonic pseudogap would © 2008 American Chemical Society
14 J. Phys. Chem. C, Vol. 112, No. 1, 2008
Figure 1. Schemes of the dye sensitized solar cells (DSSCs) under study. Standard cell (a), inverse dyed-titania opal-based solar cell (b) and standard DSSC including one inverse titania opal (c) or two different lattice paremeter inverse opals (d).
largely diminish the number of photons absorbed in the cell. Second, in bilayer structures formed by a thick dye-sensitized film coupled to an inverse titania opal, resonant modes confined within the overlayer as a consequence of the mirror effect of the photonic crystal cause an enhancement of the photocurrent. From our calculations, we could identify an increase in the calculated absorptance spectrum of the modeled bilayer and estimate a photocurrent enhancement factor very similar to that reported in ref 8. Thus, we proposed that this effect, the presence of resonant modes confined in the bilayer, was at the origin of the light harvesting efficiency enhancement of dye sensitized solar cells coupled to photonic crystals. Later, we proposed that the coupling of the dye-sensitized electrode to colloidal crystal multilayers could also lead to the observation of enhancement in wider spectral ranges.16 In this paper, we present experimental evidence that confirms the predictions reported in our previous theoretical work. We demonstrate that a dye sensitized solar cell in which the absorbing dye sensitized titania is moulded as an inverse opal is inefficient for their use in solar cells since poorer overall performance is attained when compared to a standard electrode. On the other side, if the inverse colloidal crystal is placed behind a nanocrystalline dye sensitized titania layer, enhancement of the light harvesting efficiency of the DSSC is found at wavelengths that correspond to the energy range of the photonic pseudo gap of the photonic crystal. Moreover, we have also been able to use such phenomenon to enhance photocurrent at two separate spectral regions of the absorption band of the dye by placing two different lattice parameter inverse opals behind the titanium oxide layer. Solar Cell Preparation The dye sensitized solar cells under study in this work are based in the models described in our previous works14,16 and are schematized in Figure 1, in which a standard cell (Figure 1a), a dye sensitized titania inverse opal-based cell (Figure 1b), a DSSC including an inverse titania opal coupled to the working electrode (Figure 1c), and a DSSC including two different lattice parameter inverse opals placed onto the titania electrode (Figure 1d) have been prepared. The optimum configurations of the colloidal-crystal-based cells present the inverse titania opal structure deposited after the nanocrystalline film. Such a configuration, to the best of our knowledge, is very difficult to achieve by the widely spread colloidal crystallization method of evaporation-induced self-assembly (EISA),17 due to the poor adhesion of the colloidal crystal grown in such a way onto the porous nanocrystalline substrate. Instead, we have employed a recently developed colloidal crystal deposition technique based on spin coating18 onto said substrates. For the sake of clarity, we structured this part in two sections: One devoted to
Letters describing the preparation of the different working electrodes and another to the assembly of the cell. The latter includes the adsorption of the dye onto the TiO2 electrode, the preparation of the transparent conducting substrates and the sealing of both to assemble the solar cell. Preparation of the Different Working Electrodes. A standard nanocrystalline titania working electrode was grown onto a transparent conductive oxide (TCO) (F:SnO2 covered glass provided by Hartford glass), in order to use it as a reference for all of the photoelectrical characterization measurements. A thick titania film is deposited by squeegeing a paste containing TiO2 nanocrystals onto the TCO. After heating at 100 °C for 30 min, the film is calcined to 450 °C in an oven for 30 min.19 An inverse titania opal working electrode is fabricated using the following procedure. A 3D photonic colloidal crystal made of 300 nm polystyrene spheres (Ikerlat) was grown onto a TCO glass by spin coating 200 µL of a concentrated dispersion of spheres with the following volume fractions of latex: ethanol (rectapur, purity 98%) and water (DI): (35%:43,3%:21,6%, respectively). The substrate rotation speed was 50 rps. The dispersion was dropped during the spinning of the substrate. After 1 s, the spin coater is stopped and the film is left to dry slowly within the chamber of the spin coater. After 5 min, the film is placed in a furnace at 80 °C to allow the necking of the spheres and the consolidation of the structure. The latex opal film is mechanically stabilized by depositing a thin silica layer on top of the spheres by chemical vapor deposition (CVD).20 This method will allow us to maintain the periodicity in the structure after the infiltration of the titanium dioxide precursor and removal of the organic matrix with temperature. Although this method would lead to poorer dye adsorption onto the TiO2 surface, we will be able to rule out that the main contribution to efficiency enhancement might come from diffuse light scattered from disordered regions in the photonic structure, as proposed elsewhere.21,22 For the sake of comparison, a reference cell built using standard nanocrystalline titania as the working electrode was exposed to the same silica CVD treatment. Eventually, the so stabilized colloidal film was infiltrated by soaking the opal structure in a solution containing titanium tetrachloride, ethanol, and water in a molar proportion of 1:40:10, respectively. Such a titanium dioxide precursor is widely used to improve performance in nanocrystalline TiO2based solar cells.23 Although it is an acidic compound, the silica coating prevents damage to the latex particles. Removal of the polystyrene matrix and crystallization of the oxide are carried out in an oven, heating at 450 °C for 30 min. Figure 2a is a FESEM (field emission scanning electron microscope) image of a cleaved edge of an inverse titania opal grown onto a glass slide. Finally, nanocrystalline working electrodes coupled to one or two inverse opals were built by depositing one or two 3D photonic colloidal crystal onto the nc-titania layer abovedescribed by spin coating colloidal dispersions of latex spheres.18 In the case of the electrode incorporating just one opal, we employed a precursor suspension of 240 nm diameter latex spheres. The original dispersion (latex:ethanol:water in volume fractions 35%:43.3%:21.7%) is employed to prepare a new dispersion with 5% vol of ethylene glycol. By the addition of this viscous alcohol, the fast drying process due to the high porosity of the substrate is slowed down. In the case of the cell integrating two opal films, we grow a second opal on top of the 240 nm diameter sphere opal using a dispersion of 300 nm diameter latex spheres (in volume fractions latex:ethanol:water
Letters
J. Phys. Chem. C, Vol. 112, No. 1, 2008 15
Figure 3. Pictures of (a) an inverse titania opal of 300 nm spherical voids that shows different colors due to Bragg diffraction whether it is dry (violet) or wet with ethanol (green). (b) DSSC in which the dyed nanocrystalline titania electrode is coupled to an inverse opal photonic crystal of 240 nm cavities filled with iodide-based electrolyte.
Figure 2. Scanning electron micrographs of a cleaved edge of an inverse titania opal (a), an inverse titania opal grown on top of a nanocrystalline titania layer (b) and two different lattice paramenter inverse titania opals grown on top of a nc-titania film.
of 35%:43.3%:21.7%) to prepare a new dispersion with 5% vol of ethylene glycol. A volume of 200 µL of such dispersion is dropped onto the surface of the already existing opal and extended by tilting and rotating the substrate. Once the dispersion is uniformly distributed, the substrate rotated at 25 rps for 1 s and then stopped. The new wet film is left to dry gently in the chamber of the spin coating. In all cases, as in the case of the electrode shaped as an inverse opal (Figure 2a), necking of the spheres was achieved by heating the ensemble in a furnace at 80 °C and further stabilized by depositing a layer of silica by CVD. The infiltration with TiCl4 in the opals was also carried out by soaking the films for 20 min in a solution containing titanium tetrachloride, ethanol, and water in a molar proportion
of 1:40:10, respectively. A TiO2 network is formed by the rapid evaporation of the ethanol and the hydrolysis of the chloride by the ambient moisture. Finally, the polystyrene opal film template is removed by thermal treatment at 450 °C. SEM images of the so fabricated electrodes are shown in Figure 2b,c. Dying of the Electrodes and Assembly of the Cell. Once the electrodes have been annealed at 450 °C and the temperature reaches 100 °C during the cooling process, in order to avoid water adsorption in the electrode that may spoil the cell, the structure is immersed in a 0.025 wt % solution of ruthenium bypiridile dye (rutenium 535-bis TBA, Solaronix) in ethanol (rectapur, purity of 98%) overnight in order to ensure a proper adsorption of the dye on the nc-TiO2 surface. After this, the electrode is put into electrical contact with a platinum-covered counterlectrode (Pt-catalyst T/SP, Solaronix) by infiltrating a liquid electrolyte (Iodolyte PN-50, Solaronix) in between them. The porous nature of the inverse opal allows the electrolyte to soak into the sensitized nc-TiO2 coating. Previously, a thin polymeric film window (Surlyn, jurasol DG, juraplast GmbH) that softens at 120 °C was used both as spacer and as sealant of the cell. These cells display a violet-colored reflection that turns into bright green when filled with electrolyte. Such behavior, evidenced in the photographs of the cells presented in Figure 3 for a working electrode (a) and an assembled cell (b), is due to the change in the dielectric contrast between the air or electrolyte filling the cavities and the titania of the host matrix. The change in the position of the Bragg peak (cause of the color observed) is an evidence of the order present in the structure.
16 J. Phys. Chem. C, Vol. 112, No. 1, 2008
Letters
Figure 4. Optical and photoelectric spectral response of dye sensitized solar cells. (a) Specular reflectance spectrum (in black) of a dyed titanium oxide inverse opal (φ ) 300 nm) compared to the enhancement factor γ (in blue). (b) Specular reflectance spectrum of a DSSC including an inverse titania opal (φ ) 240 nm) after the working electrode (in black) compared with its corresponding enhancement factor. (c) Specular reflectance spectrum of a DSSC integrating two inverse titania opals of different lattice parameter (spheres sizes φ1 ) 240 nm and φ2 ) 300 nm) after the working electrode (in black) compared with its corresponding enhancement factor. In all cases, the enhancement factor γ is calculated from the ratio between the IPCE (incident photon to current conversion efficiency) of the considered cell (represented in d-f in a red dotted curve) and the IPCE of a standard cell that was subjected to similar silica deposition treatments for the sake of comparison (black dotted curves).
Photoelectric Characterization and Discussion of Results Dye sensitized solar cells including inverse TiO2 colloidal crystals in different arrangements have been characterized optically and photoelectrically, with the results being presented in Figure 4. Optical characterization was performed using a Fourier transform infrared spectrophotometer (BRUKER IFS66) attached to a microscope and operating in reflection mode. A X4 objective with a numerical aperture of 0.1 (light cone angle (5.7) was used to irradiate the solar cell and collect the reflected light at quasinormal incidence with respect to its surface. A spatial filter was used to selectively detect light from 0.6 mm2 circular regions of the sample. Incident photon to electric current conversion efficiencies (IPCE, a quotient between incoming photons and outgoing electrons under AM 1.5 solar spectrum irradiation) were measured in the spectral range comprised between 400 and 800 nm illuminating the front side of the cell with a plane parallel beam coming from a 450 W xenon lamp (Oriel) after being dispersed by a monochromator (Oriel) containing a 1200 lines/mm grating (Oriel). Slits were chosen to attain a 10 nm wavelength resolution. A silicon photodiode (Jaal) of known response was used as reference to extract the IPCE curves, and a Keithley multimeter was employed to record the photocurrent. IV curves were measured under white light illumination coming from the same light source plus UV and water IR filters. Currents were registered via a battery-operated potentiostat. The reflectance spectrum under normal incidence of a dye sensitized solar cell in which the absorbing material is structured in a dyed titania inverse opal structure and with the interstitial space filled with electrolyte is represented in Figure 4a (in black). Such a structure corresponds to the cell shown in Figure 1b and to the SEM picture in Figure 2a. In such a spectrum, a reflectance peak arises as a consequence of the photonic pseudogap that opens in the direction Γ-L of the opal film, which is perpendicular to the substrate surface. This pseudogap results from the periodic modulation of the dielectric constant
along the (111) direction of the opal film, as has been thoroughly documented. The dielectric contrast between the dyed titania of the walls (which is 35% porous, according to our SWA simulations, thus lower refractive index than in the case of dense titania is obtained nTiO2 ) 1.81 being found) of the inverse opal and the electrolyte embedding and filling the cavities is small (nelectrode TiO2+electrolyte ≈ 1.93 versus nelectrolyte ≈ 1.34), which explains the narrow width and low intensity of the Bragg peak. In Figure 4a is also plotted the enhancement factor γ, which is defined as the ratio between the IPCE of the opal shaped electrode with respect to that of the standard electrode, which can be compared to Figure 4d. The enhancement factor in Figure 4a is below the unity for wavelengths lying within the photonic pseudogap, where the reflectance spectrum (Figure 4a in black) shows a clear maximum. These results show that, as predicted from our calculations,14 a dye-sensitized electrode shaped as an inverse opal is not a suitable structure for solar devices by itself, since the dielectric mirror effect would prevent photons from being absorbed by the ruthenium dye at those wavelengths where the structure exhibits its forbidden energy interval. On the contrary, when the inverse structure is located behind the absorbing layer (Figure 2b), the IPCE enhancement factor increases for wavelengths corresponding to the reflectance maximum, as seen in Figure 4b. The reflectance spectrum (black line) corresponds to an electrolyte filled DSSC in which an inverse titania opal of 240 nm cavities is coupled to the nanocrystalline titanium oxide electrode. That increment in the IPCE measured for the cell, found when compared with the standard cell, coincides with the spectral window at which the pseudogap of the inverse opal appears, as can be seen in Figure 4e. This result is also in good agreement with the theoretical predictions reported in our previous work.14 It should be noticed that the reflectance peak of the inverse opal grown onto the thick nanocrystalline titania layer is less intense than that of the inverse colloidal crystal grown onto a glass substrate (Figure 4a). This is due to the optical coupling of both structures and
Letters the absorption of the thick dyed film. The origin of the enhancement of IPCE found in this case comes from resonant modes partially confined within the absorbing layer as a consequence of the dielectric mirror effect of the inverse opal in a certain wavelength interval. The higher the quality of the structure exhibiting the photonic band gap, the greater the enhancement would be. Regarding the angular response of opalbased solar cells, the spectral range at which photocurrent enhancement is observed is expected to shift to shorter wavelengths as we increase the incident angle, just like the photonic pseudogap does. Enhancement of the photocurrent can be attained in wider spectral ranges by using photonic crystals with wider photonic band gaps. Unfortunately, in the case of three-dimensional colloidal crystals, this can only be achieved by substituting the materials employed for others that present a higher dielectric contrast, thus altering the electron transport dynamics in the cell. However, this can also be achieved by coupling a tandem structure composed of two inverse opals with different lattice parameters (240 and 300 nm diameter) to the thick nc-titania layer.16 For this cell, two reflectance maxima can be distinguished in the reflectance spectrum represented in Figure 4c, corresponding to each of the photonic crystals present in the structure. In this case, γ shows an increase in a much wider spectral range. These results provide a proof of the theoretical concepts presented in refs 14 and 16. However, the power conversion efficiencies (η) of the DSSCs herein described and calculated from IV measurements are much lower than current standards as expected since all of the samples (including the reference cell) were subjected to a silica CVD treatment, which reduces the surface available for the dye adsorption. This treatment permits us to build a robust ordered structure in which the effect of imperfections is minimized. Efficiencies are η ) 0.501% for the reference, η ) 0.588% for the one photonic crystal-based cell, and η ) 0.625% for the two photonic crystalbased cell. Therefore, the enhancements found for the DSSC including one or two inverse opal are ∆η1 PhC ) 8.7% and ∆η2 PhC ) 12.4%, respectively. Conclusions We have clarified how photonic crystals can be used to improve the performance of photovoltaic devices. Dye sensitized solar cells where the working electrode was moulded in an inverse opal shape exhibited less IPCE than an equivalent reference cell with no structure at all. However, when the same inverse titania colloidal crystal was deposited onto a common working electrode of dyed nanocrystalline titanium oxide, an enhancement in the measured photocurrent was found for frequencies comprised within the photonic pseudogap of the inverse opal structure. The dependence of the increased incident
J. Phys. Chem. C, Vol. 112, No. 1, 2008 17 photon to current conversion efficiency of the cell with the position of the forbidden interval range was herein employed to enhance absorption at two different spectral ranges of the ruthenium dye by coupling to the sensitized titania working electrode two inverse titania photonic crystals with different lattice parameter. Acknowledgment. This research has been funded by the Spanish Ministry of Science and Education under Grant MAT2005-03028 and the Ramo´n Areces Foundation. M.E.C. thanks CSIC for an I3P contract. J.A.A. thanks Spanish Ministry of Science and Education under Grant ENE2004-01657/ALT and Project HOPE CSD2007-00007 (Consolider-Ingenio 2010) and Junta de Andalucı´a under Project P06-FQM-01869. References and Notes (1) Oregan, B.; Gratzel, M. Nature 1991, 353, 737. (2) Gratzel, M. Nature 2001, 414, 338. (3) Kroon, J. M.; Bakker, N. J.; Smit, H. J. P.; Liska, P.; Thampi, K. R.; Wang, P.; Zakeeruddin, S. M.; Gratzel, M.; Hinsch, A.; Hore, S.; Wurfel, U.; Sastrawan, R.; Durrant, J. R.; Palomares, E.; Pettersson, H.; Gruszecki, T.; Walter, J.; Skupien, K.; Tulloch, G. E. Prog. PhotoVoltaics 2007, 15, 1. (4) Usami, A. Chem. Phys. Lett. 1997, 277, 105. (5) Tachibana, Y.; Hara, K.; Sayama, K.; Arakawa, H. Chem. Mater. 2002, 14, 2527. (6) Ferber, J.; Luther, J. Sol. Energy Mater. Sol. Cells 1998, 54, 265. (7) Hore, S.; Vetter, C.; Kern, R.; Smit, H.; Hinsch, A. Sol. Energy Mater. Sol. Cells 2006, 90, 1176. (8) Nishimura, S.; Abrams, N.; Lewis, B. A.; Halaoui, L. I.; Mallouk, T. E.; Benkstein, K. D.; van de Lagemaat, J.; Frank, A. J. J. Am. Chem. Soc. 2003, 125, 6306. (9) Yablonovitch, E. Phys. ReV. Lett. 1987, 58, 2059. (10) Huisman, C. L.; Schoonman, J.; Goossens, A. Sol. Energy Mater. Sol. Cells 2005, 85, 115. (11) Kavan, L.; Zukalova, M. T.; Kalbac, M.; Graetzel, M. J. Electrochem. Soc. 2004, 151, A1301. (12) Chen, J. I. L.; von Freymann, G.; Choi, S. Y.; Kitaev, V.; Ozin, G. A. AdV. Mater. 2006, 18, 1915. (13) Rodriguez, I.; Atienzar, P.; Ramiro-Manzano, F.; Meseguer, F.; Corma, A.; Garcia, H. Photonics and Nanostruct.-Fundam. Appl. 2005, 3, 148. (14) Mihi, A.; Miguez, H. J. Phys. Chem. B 2005, 109, 15968. (15) Mittleman, D. M.; Bertone, J. F.; Jiang, P.; Hwang, K. S.; Colvin, V. L. J. Chem. Phys. 1999, 111, 345. (16) Mihi, A.; Lopez-Alcaraz, F. J.; Miguez, H. Appl. Phys. Lett. 2006, 88. (17) Jiang, P.; Bertone, J. F.; Hwang, K. S.; Colvin, V. L. Chem. Mater. 1999, 11, 2132. (18) Mihi, A.; Ocana, M.; Miguez, H. AdV. Mater. 2006, 18, 2244. (19) Barbe, C. J.; Arendse, F.; Comte, P.; Jirousek, M.; Lenzmann, F.; Shklover, V.; Gratzel, M. J. Am. Ceram. Soc. 1997, 80, 3157. (20) Miguez, H.; Tetreault, N.; Hatton, B.; Yang, S. M.; Perovic, D.; Ozin, G. A. Chem. Commun. 2002, 2736. (21) Hore, S.; Nitz, P.; Vetter, C.; Prahl, C.; Niggemann, M.; Kern, R. Chem. Commun. 2005, 2011. (22) Halaoui, L. I.; Abrams, N. M.; Mallouk, T. E. J. Phys. Chem. B 2005, 109, 6334. (23) Huang, S. Y.; Schlichthorl, G.; Nozik, A. J.; Gratzel, M.; Frank, A. J. J. Phys. Chem. B 1997, 101, 2576.
