J. Am. Ceram. Soc., 92 [9] 1921–1925 (2009) DOI: 10.1111/j.1551-2916.2009.02970.x r 2009 The American Ceramic Society Journal Dye-Sensitized Solar Cells Based on TiO2 Coatings with Dual Size-Scale Porosity Lai Qi, Judith D. Sorge,w and Dunbar P. Birnie III Department of Materials Science and Engineering, Rutgers University, Piscataway, New Jersey Dye-sensitized solar cells (DSSC) with efficiencies greater than 4% were produced with templated ‘‘inverse opal’’ titania coatings. A novel one-step method produces uniform and crack-free coatings made using commercially available titania nanoparticles with high reproducibility and uniformity. In this research, a volatile solvent electrolyte was tested; however, it shows proofof-concept that larger pore volumes can be created for increased penetration of more viscous electrolytes that can be utilized in high-efficiency cells. This dual size-scale porosity film is a promising structure for DSSC applications, especially for those solidstate or quasi-solid-state cells that require polymer electrolytes. the short circuit current of the cell greatly. However, a high surface area also allows for a greater possibility of recombination between the electrolyte and the titania layer. If a titania particle is not covered in dye and an injected electron passes back to the electrolyte instead of through to the back electrode, current loss is encountered. A similar balance is observed for the pore volume as more pores allow the electrolyte to reach every dye molecule, which is necessary for sustained high current, but this can lead, again, to greater recombination between the electrolyte and the titania layer. It has been shown that the optimized TiO2 film geometric structure is that of mesoporous channels aligned in parallel to each other and perpendicular to the electrode substrate.10,11 A good model for this concept is organized TiO2 nanorod or nanowire arrays aligned perpendicular to the transparent conducting oxide (TCO) substrate, which can be produced by a sol– gel method with a template or in an oxidizing environment.12,13 However, aligning the nanorods/wires into an organized structure on top of a TCO substrate rather than on opaque metal sheets proves to be challenging.13,14 Another microstructure that fulfills Grätzel’s criteria is that represented by those found in ‘‘inverse opals.’’ The method for preparing these films is usually a two-step template-infiltration process, wherein the monosized template spheres are assembled first and then an organic titanium alkoxide precursor solution infiltrates by capillary forces.15–22 Inverse opal structures can be difficult and time consuming to produce due to shrinkage issues and have only produced relatively low-efficiency cells so far.16,18,21,23 In contrast, our research uses a production method very similar to that of conventional films, which produces a much larger and tunable pathway size for the electrolyte while maintaining the simple doctor blade deposition method. The use of crystalline nanoparticles of titania as a starting material also results in less shrinkage during firing and fewer defects. The end result is a microstructure that is a combination of a nanocrystalline network with an aligned macropore network of controllable pore size in the range of 0.2–2 mm. After testing in solar cells (as discussed below), we found that 1-mm polystyrene (PS) template particles produced the most efficient cell at a point when the overall porosity was 73%. About two-thirds of this porosity comes from the normal nanoscale interstitial spaces between the titania starting particles but the remainder of the space is derived from the interconnected network of micron-sized pores occupied by the template particles. I. Introduction Dye-sensitized solar cells (DSSC) are photoelectrochemical cells based on sensitized semiconductor–electrolyte interfaces,1,2 which were first introduced in 1991, and have demonstrated the highest cell efficiency so far, 11%,3 attracting much attention worldwide.4–8 Further improvements in efficiency and stability are necessary for these cells to become a competitive renewable energy resource because, on average, the efficiencies found are much lower, in the range of 5%–7%. Cells built in general have a hard time besting the record set by Gratzel et al. and so much work is performed in a comparative manner so that a greater understanding of the properties of the cell and how the layers interact can be found. The conventional configuration of a DSSC is a layered structure with two electrodes, one an active layer of ruthenium dye chemisorbed to a nanocrystalline titanium dioxide film and the other a sputtered platinum layer, both coated onto transparent conductive oxide layers on glass substrates. Sandwiched between the two electrodes is an iodide/tri-iodide electrolyte, most commonly in a volatile liquid solution such that the cell must be sealed. As the light enters the cell, an electron in the ground state of the dye becomes energized to a higher state and then is injected into the titania layer. The titania layer acts as a conductor passing the electron to the outer electrode. As this occurs, the iodide donates an electron to the dye molecule as it oxidizes to form tri-iodide. A schematic of this type of cell is portrayed in Fig. 1. Much research has focused on increasing the device efficiency, which can be carried out by optimizing all the different layers. Research that has delved into the titania layer shows that it must maintain a high surface area for high dye adsorption while allowing maximum electrolyte penetration through its pore volume. It is also important to recognize that an improved microstructure could also allow for mass-production printing processes to be used and lower costs for these devices.9 The current of the cell is in direct proportion to the amount of the adsorbed dye, such that an increased surface area can improve II. Experimental Procedure The PS latex solutions were purchased from Duke Scientific Inc. (Duke Scientific, Palo Alto, CA) The TiO2 nanoparticles were from Evonik Degussa (P25, with an average diameter of 50 nm, Piscataway, NJ). An aqueous dispersion of P25 was obtained by ball milling P25 particles in ethanol for 48 h with 0.3 mm zirconia balls. The dispersion was separated with a syringe filter and then mixed with the PS latex solution. The volume ratio of P25 to PS varied from 15% to 45%. The final concentration of solids in the solution decides the film thickness. Small amounts S. Bose—contributing editor Manuscript No. 25115. Received August 18, 2008; approved December 30, 2008. w Author to whom correspondence should be addressed. e-mail: jdsorge@eden.rutgers. edu 1921 1922 Journal of the American Ceramic Society—Qi et al. Vol. 92, No. 9 Fig. 1. Schematic of a dye-sensitized solar cell illustrating the interaction of the various layers. of surfactant (Triton X-100; Sigma-Aldrich, St. Louis, MO), binders (polyethylene glycol (PEG), Mw 5 1600; Aldrich), and cosurfactant (ethanol; Aldrich) were added to adjust the surface tension, wettability, and coating uniformity. The coating process can be performed by doctor blading, screen printing, or simply by cast coating, depending on the viscosity of the solution. Fluorine-doped tin oxide (FTO)-coated glass was used as the TCO substrate (5–10 O/&; Flexitec, Curitiba, Brazil). Scotch tape (B50 mm, 3M) was used to define the coating area, location, and thickness. After coating, the films were dried in air for 1 h and then heated at 4501C for 30 min. This heat treatment provided the dual role of removing the PS templates and sintering the TiO2 particles to form the conductive network for electron flow. After heating, the TiO2 electrodes were soaked in an ethanol solution of ruthenium dye (3  10 3M, N535; Solaronix, Aubonne, Switzerland) for 24 h. The molecular structure of this dye is shown in Fig. 2. The TiO2 electrodes are then washed with ethanol to remove the physisorbed dye molecules. The counter electrode is a platinum layer, which had been deposited by radio frequency sputtering, with a thickness of 200 nm. The two electrodes were separated by a Teflon spacer with a thickness of 40 mm. While holding the electrodes in place with binder clips, a thin layer of epoxy sealant (Solaronix, Amosil 4) was applied to the edges with two holes left for electrolyte infiltration. The electrolyte (Solaronix: Iodolyte TG-50) filled the cell by capillary draw through the holes. The microstructure of the TiO2 films was examined by a fieldemission scanning electron microscope (DSM 982 Gemini; Carl Zeiss SMT, Cambridge, UK). The pore size distribution was measured by a mercury porosimetry (Autopore III 9400; Bu4N + – O Micromeritics, Norcross, GA). The powder density was measured by a pycnometer (Accupyc 1330; Micromeritics). The AM1.5 illumination was provided by a 300 W Xenon solar simulator (Oriel 91160A; Newport, Irvine, CA). The film thickness was measured by a profilometer (Alpha-step 200; Tencor, Milpitas, CA). III. Results and Discussion (1) Structural Characterization As shown in Fig. 3, the templated P25 films with PS particles 1 mm in diameter have a macroporous network with mesoporous titania found in the interstices. At high resolution (Fig. 3(a)), one can see that the P25 nanoparticles form a self-supporting network where the packing of the nanoparticles is similar to that found in common nanocrystalline films, which suggests that rapid electron conduction is possible in this arrangement. The high dispersion of the P25 particles with the template PS O OH O N N NCS Ru NCS N N O OH Bu4N +– O O Fig. 2. The molecular structure of the ruthenium 535 dye used to coat the titania layer (Solaronix). Fig. 3. Scanning electron microscopy micrographs of TiO2 films (P25; Degussa) templated by 1 mm polystyrene nanospheres. (a) High-resolution micrograph of TiO2 skeleton. (b) Interpenetrating network of TiO2 and macropores. (c) Low-resolution micrograph of film morphology with no obvious cracking observed. (d) Cross-sectional view of a 14 mm thick film on fluorine-doped tin oxide-coated glass. September 2009 1923 Dye-Sensitized Solar Cells Fig. 4. Scanning electron microscopy micrographs of templated TiO2 films (P25; Degussa) produced with varying polystyrene (PS) template sizes. (a and b): High- and low-resolution images of TiO2 templated with 160 nm PS nanospheres. (c and d): High- and low-resolution images of TiO2 templated with 600 nm PS nanospheres. particles after mixing is evidenced by Fig. 3(c), in which the spaces where the templates are missing are completely occupied by P25 particles. During the initial phase of drying, domains of approximately 10 mm of organized PS spheres assemble and are bound by areas of randomly packed template particles. At a later stage, P25 particles dry in the interstices of the packed PS particles forming a honeycomb-like structure. Figure 3(d) shows the cross-section of a P25 1-mm PS film with a thickness of 14 mm coated on FTO-coated glass. The macroporous arrangement in the cross-section view does not seem as ordered as that in the top views (Figs. 3(a–c)), which may be partially due to the rough transection of the surface. The coating quality of the templated films is surprisingly high in terms of film uniformity, adhesion, and cracking. The method here works with presynthesized anatase-phase TiO2 particles that suffer negligible shrinkage during annealing at 4501C compared with organic precursors. As a consequence, the prepared films are almost crack-free, as shown in Figs. 3(c), and 4(b) and (d), and the coating uniformity is high. Similar microstructures can be reproduced with templates varying in diameter from 160 nm to 2 mm. Figure 4 shows the scanning electron microscopy (SEM) micrographs of the 160 nm (Figs. 4(a and b)) and 600 nm (Figs. 4(c and d)) PS-templated structures. When templated with the 600 nm PS, the produced films have a similar multidomain appearance to that of 1 mm. Domain boundaries and point defects can be distinguished (Fig. 4(d)) but cracking is absent. However, in case of the 160 nm template, the resulting microstructure is composed of a randomly distributed pore network. The average particle size of P25 is around 50 nm, which is too large to fill in the gaps between the 160 nm PS assembly; hence, the titania particles cause interference with PS self-assembly. The macroscopic coating of the 160 nm films is, however, even more uniform than that by larger templates due to the disappearance of the domains (Fig. 4(b)). The dual porosity was measured by mercury porosimetry, which confirms what is seen in the SEM micrographs. Figures 5(a and b) show the curves of the cumulative and incremental pore volume versus pore diameter, respectively. The highest pressure used was 3  104 psi. The presence of double peaks in the incremental pore volume curves clearly indicates the coexistence of the dual porosity in the templated structures. Mesoporosity was consistently observed at 36 nm for all the samples, including the nontemplated P25 films. The average diameter of the macropores changes according to the size of the PS templates used. Table I lists the measured macropore and mesopore Fig. 5. Pore size distribution of polystyrene (PS)-templated P25 films measured by a mercury porosimeter. (a) Curves of cumulative pore volume vs. pore diameter. Solid symbols denote intrusion. The right arrows indicate the macroporosity while the down arrows indicate the mesoporosity. (b) Curves of incremental pore volume vs. pore diameter. The presence of two peaks proves the dual porosities. diameters, specific surface areas (SSAs), and porosities of the templated P25 films. The measured macropore diameters account for only 30%–50% of the corresponding template diameter, which may be because the size of the pore opening practically decided the intrusion pressure rather than the full template particle size. The SSA is largely decided by TiO2 particle size and mesoporosity after being normalized to weight. Because templates did not change either of them, SSA data for the templated structures show little change from that of the nontemplated data. The porosity increased after templating from 58% to above 70%. However, the porosity depends on the volume ratio of PS to TiO2, which varied from sample to sample. Table I. Measured Pore Diameters, Specific Surface Areas, and Porosities of Templated P25 Films Mercury porosimeter analysis Template size (nm) Macropore diameter (nm) Mesopore diameter (nm) Specific surface area (m2/g) Porosity (%) 500 210 85 NA 36 36 36 36 65.6 68.1 57.9 58.9 73.6 74.1 72.4 58.7 1000 600 160 No template NA, not applicable. 1924 Vol. 92, No. 9 Journal of the American Ceramic Society—Qi et al. The resulting microstructure of the present research is a combination of a nanocrystalline network with an aligned macropore network of controllable pore size in the range of 0.2–2 mm produced by a templated casting method.24 Both the PS latex templates and the TiO2 nanoparticles are commercial products, which render good data reproducibility. A high total porosity of 70%–80% has been achieved with the templated structures, which is beneficial for charge transport through the electrolyte, especially anticipating DSSC applications using polymer electrolytes.25,26 This structure is remarkable because the increased porosity is achieved by the creation of macroporous aligned channels rather than through the use of surfactants that evenly increase the average distance between the titania particles, degrading the electron conductivity.8 Often, polymer additives, i.e. PEG, cetyltrimethylammonium bromide, or a mixture of both, are added to the TiO2 during coating to guide the development of the microstructure.8,26–28 This produces structures with a large surface area and improved porosity (B70%).26 However, the rapid electron transfer is compromised due to the increased space between the titania particles after polymer burnout. In contrast, while the templated structures presented in this work have similar high porosities, rapid electron transfer is maintained through the lack of a large amount of polymer additives so that the titania nanoparticles are in good contact after sintering. This also creates a very strong network such that these films adhere very well to the substrate, producing a robust film that, especially in the casing of the solar cell, is not in danger of degrading. The mesoporosity that still exists within these films, in conjunction with the macroporous network, is also highly beneficial for charge transport through the electrolyte. Papageorgiou29 reports a 30% loss of actual porosity in nanocrystalline films after dye adsorption and considerably lower mass transport in the electrolyte. The dual porosities in the structure of this work are believed to facilitate dye adsorption while maintaining large pathways for electrolyte diffusion during operation. Table II. IV Results of DSSC Made with TiO2 Films Templated with 1 mm PS Sample Cell Cell Cell Cell A B C D Thickness (mm) Voc (V) Jsc (mA/cm2) Fill factor Efficiency (%) 4.34 8.31 9.09 12.71 0.73 0.72 0.72 0.73 3.6 5.16 6.5 9.44 0.61 0.59 0.54 0.59 1.8 2.4 2.8 4.5 The incidence light intensity is 90 mW/cm2 for all the samples. IV, photovoltage–photocurrent; DSSC, Dye-sensitized solar cells; PS, polystyrene. by an irradiance meter (Daystar; Solaqua, El Paso, TX). The current is measured as different voltages are put on the system from 0 to 1 V and, from this data, the efficiency and other characteristics can be calculated. Both the efficiency and the shortcircuit photocurrent density (Jsc) increase as the TiO2 film thickness increases, with the highest efficiency of 4.5% coinciding with the thickest film of 12.7 mm. The open circuit voltage (Voc) and the fill factor (FF) do not show a clear dependence on the film thickness. The Voc was found to be almost constant at 0.73 V for all the cells. Table II lists the detailed IV characteristics of those cells with templated TiO2 structures. The optimal film thickness reflects the balance between changes in the roughness factor, the electrolyte diffusion distance, and the distance an electron must travel through the semiconductor to reach the TCO. The roughness factor is directly proportional to dye adsorption and increased current while increasing the electrolyte diffusion distance lowers the ionic conductivity and, in turn, the cell performance. For non- (2) Photovoltaic Characteristics The photovoltage–photocurrent (IV) characteristics of cells composed of templated TiO2 films with thickness variations from 4.3 to 13 mm are illustrated in Fig. 6. In order to vary the thickness of the cells, the number of scotch tape layers was varied in the doctor blading process. The illumination area of the cells varied from 0.25 to 0.5 cm2 and the PS template size was held constant at 1 mm. The cells were tested using solar simulation with an illumination intensity of 900 W/m2, as measured Fig. 6. The photovoltage–photocurrent curves of cells made with templated TiO2 films using 1 mm polystyrene templates. Film thickness varies in cells A–D from 4.3 to 13 mm. The corresponding dark curves for each cell are also shown. The incidence light intensity was 90 mW/cm2. Fig. 7. (a) Plot of cell efficiency and normalized current vs. film thickness, d, for templated cells. Linear increase illustrates possibility that optimum thickness has not yet been reached. (b) Efficiency vs. film thickness for cells produced without templates. September 2009 Dye-Sensitized Solar Cells templated TiO2 nanocrystalline films with randomly packed nanoparticles, the electrolyte diffusion length increases rapidly with film thickness due to the rise of the tortuous and narrowed channels. However, in the templated TiO2 structures of this work, because the electrolyte ions can diffuse rapidly through the macroporous channels while the dye is adsorbed in the mesoporous structure, a better balance of film thickness can be achieved. As shown in Fig. 7(a), the cell efficiency and the current density both increase nearly linearly with film thickness for templated cells. Higher efficiencies are expected with even thicker layers of these templated materials up to a point when the system (titania and electrolyte) conductivities start to become a limiting factor. Figure 7(b) demonstrates thickness data in comparison with the efficiency for similar cells made without templates. The lower efficiencies in general as well as the earlier decrease of efficiency with thickness,demonstrate this ionic conduction bottleneck. Although in this work efficiency tests were not performed with varying PS particle sizes, some conclusions can be drawn from the microstructures shown earlier. The 160 nm PS gives microstructures similar to that of titania without templates because of the similar particle sizes, which cause interference in the packing structure. Whether the efficiency would increase with 600 nm PS particles instead of 1 mm would depend on the type of the electrolyte used. As the viscosity of the electrolyte increases with lower volatility, the size of the pores is theorized to greatly affect the efficiency of the cell. These structures are believed to be quite promising for solid-state or quasi-solid-state DSSC applications where significantly lower conductivity polymer electrolytes are required. IV. Conclusions Templated TiO2 structures with dual size-scale porosity were developed for DSSC applications. Our results demonstrate that the structure has excellent mass transport capability for regeneration of the dye by the electrolyte as well as excellent electron transport within the titania. The cell with the 12.7 mm-thick TiO2 film demonstrated 4.5% efficiency under simulated AM1.5 sunlight. This value is substantially higher than the baseline data for our process using the P25 coatings without templates. This dual porosity structure proved to be highly beneficial for dye adsorption and charge transport. The structure produced is believed to be even more promising for solid-state or quasi-solid-state DSSC applications where significantly lower conductivity polymer electrolytes are required. Acknowledgments We would like to thank the National Science Foundation, which funded this project at Rutgers through grant EEC-0509886. Support from the Malcolm G. McLaren Fellowship is also greatly appreciated. References 1 B. O’Regan and M. Grätzel, ‘‘A Low-Cost, High-Efficiency Solar Cell Based on Dye-Sensitized Colloidal TiO2 Films,’’ Nature, 353, 737–40 (1991). 2 M. Grätzel, ‘‘Photoelectrochemical Cells,’’ Nature, 414, 338–44 (2001). 3 M. K. Nazeeruddin, P. Pechy, T. Renouard, S. M. Zakeeruddin, R. HumphryBaker, P. Comte, P. Liska, L. Cevey, E. Costa, V. Shklover, L. Spiccia, G. B. Deacon, C. A. Bignozzi, and M. Grätzel, ‘‘Engineering of Efficient Panchromatic Sensitizers for Nanocrystalline TiO2-Based Solar Cells,’’ J. Am. Chem. Soc., 123, 1613–24 (2001). 4 1925 J. Bisquert, D. Cahen, G. Hodes, S. Ruhle, and A. Zaban, ‘‘Physical Chemical Principles of Photovoltaic Conversion with Nanoparticulate, Mesoporous DyeSensitized Solar Cells,’’ J. Phys. Chem. B, 108 [24] 8106–18 (2004). 5 M. Adachi, Y. Murata, J. Takao, J. Jiu, M. Sakamoto, and F. Wang, ‘‘Highly Efficient Dye-Sensitzied Solar Cells with a Titania Thin-Film Electrode Composed of a Network Structure of Single-Crystal-Like TiO2 Nanowires Made by the ‘‘Oriented Attachment’’ Mechanism,’’ J. Am. Chem. Soc., 126, 14943–9 (2004). 6 A. F. Nogueira, J. R. Durrant, and M. A. D. Paoli, ‘‘Dye-Sensitized Nanocrystalline Solar Cells Employing a Polymer Electrolyte,’’ Adv. Mater., 13 [11] 826–30 (2001). 7 N. Kopidakis, N. R. Neale, K. Zhu, J. V. D. Lagemaat, and A. J. Frank, ‘‘Spatial Location of Transport-Limiting Traps in TiO2 Nanoparticle Films in Dye-Sensitized Solar Cells,’’ Appl. Phys. Lett., 87, 202106 (2005). 8 K. D. Benkstein, N. Kopidakis, J. van de Lagemaat, and A. J. Frank, ‘‘Influence of the Percolation Network Geometry on Electron Transport in DyeSensitized Titanium Dioxide Solar Cells,’’ J. Phys. Chem. B, 107 [31] 7759–67 (2003). 9 K. Chittibabu and S. Hadjikyriacou, ‘‘Gel Electrolytes for Dye Sensitized Solar Cells’’; in Edited by U.S.P. Office Konarka Technologies, Inc., United States, (2005). 10 M. Grätzel, ‘‘Dye-Sensitized Solar Cells,’’ J. Photochem. Photobiol. C-Photochem. Rev., 4 [2] 145–53 (2003). 11 M. Gratzel, ‘‘Perspectives for Dye-Sensitized Nanocrystalline Solar Cells,’’ Prog. Photovoltaics, 8 [1] 171–85 (2000). 12 P. Hoyer, ‘‘Formation of a Titanium Dioxide Nanotube Array,’’ Langmuir, 12, 1411–3 (1996). 13 X. Peng and A. Chen, ‘‘Aligned TiO2 Nanorod Arrays Synthesized by Oxidizing Titanium with Acetone,’’ J. Mater. Chem., 14, 2542–8 (2004). 14 Y. Lin, G. S. Wu, X. Y. Yuan, T. Xie, and L. D. Zhang, ‘‘Fabrication and Optical Properties of TiO2 Nanowire Arrays Made by Sol–Gel Electrophoresis Deposition Onto Anodic Alumina Membranes,’’ J. Phys.: Condens. Matter, 15, 2917–22 (2003). 15 J. E. G. J. Wijnhoven and W. L. Vos, ‘‘Preparation of Photonic Crystals Made of Air Spheres in Titania,’’ Science, 281, 802–4 (1998). 16 P. Jiang, J. Cizeron, J. F. Bertone, and V. L. Colvin, ‘‘Preparation of Macroporous Metal Films from Colloidal Crystals,’’ J. Am. Chem. Soc., 121 [34] 7957– 8 (1999). 17 P. Ni, B. Cheng, and D. Zhang, ‘‘Inverse Opal with an Ultraviolet Photonic Gap,’’ Appl. Phys. Lett., 80 [11] 1879–81 (1998). 18 B. T. Holland, C. F. Blanford, and A. Stein, ‘‘Synthesis of Macroporous Minerals with Highly Ordered Three-Dimensional Arrays of Spheroidal Voids,’’ Science, 281, 538–40 (1998). 19 M. E. Abdelsalam, P. N. Bartlett, J. J. Baumberg, and S. Coyle, ‘‘Preparation of Arrays of Isolated Spherical Cavities by Self-Assembly of Polystyrene Spheres on Self-Assembled Pre-Patterned Macroporous Films,’’ Adv. Mater., 16, 90–3 (2004). 20 Z. Zhong, Y. Yin, B. Gates, and Y. Xia, ‘‘Preparation of Mesoscale Hollow Spheres of TiO2 and SnO2 by Templating Against Crystalline Arrays of Polystyrene Beads,’’ Adv. Mater., 12, 206–9 (2000). 21 A. Richel, N. P. Johnso, and D. W. McComb, ‘‘Observation of Bragg Reflection in Photonic Crystals Synthesized from Air Spheres in a Titania Matrix,’’ Appl. Phys. Lett., 76, 1816–8 (2000). 22 Z. Zhou and X. S. Zhao, ‘‘Opal and Inverse Opal Fabricated with a FlowControlled Vertical Deposition Method,’’ Langmuir, 21, 4717–23 (2005). 23 C. L. Huisman, J. Schoonman, and A. Goossens, ‘‘The Application of Inverse Titania Opals in Nanostructured Solar Cells,’’ Solar Energy Mater. Solar Cells, 85, 115–24 (2005). 24 L. Qi and D. P. Birnie III, ‘‘Templated Titania Films with Meso- and Macroporosities,’’ Mater. Lett., 61, 2191–4 (2007). 25 L. Schmidt-Mende and M. Gratzel, ‘‘TiO2 Pore-Filling and Its Effect on the Efficiency of Solid-State Dye-Sensitized Solar Cells,’’ Thin Solid Films, 500, 296– 301 (2006). 26 J. Jiu, F. Wang, M. Sakamoto, J. Takao, and M. Adachi, ‘‘Performance of Dye-Sensitized Solar Cell Based on Nanocrystals TiO2 Film Prepared with Mixed Template Method,’’ Solar Energy Mater. Solar Cells, 87, 77–86 (2005). 27 M. Zukalova, A. Zukal, L. Kavan, M. K. Nazeeruddin, P. Liska, and M. Grätzel, ‘‘Organized Mesoporous TiO2 Films Exhibiting Greatly Enhanced Performance in Dye-Sensitized Solar Cells,’’ Nano Lett., 5 [9] 1789–92 (2005). 28 J. Jiu, S. Isoda, F. Wang, and M. Adachi, ‘‘Dye-Sensitized Solar Cells Based on a Single Crystalline TiO2 Nanorod Film,’’ J. Phys. Chem. B, 110, 2087–92 (2006). 29 N. Papageorgiou, C. Barbe, and M. Grätzel, ‘‘Morphology and Adsorbate Dependence of Ionic Transport in Dye Sensitized Mesoporous TiO2 Films,’’ J. Phys. Chem. B, 102 [21] 4156–64 (1998). &