NANO LETTERS
Use of Highly-Ordered TiO2 Nanotube Arrays in Dye-Sensitized Solar Cells
2006 Vol. 6, No. 2 215-218
Gopal K. Mor, Karthik Shankar, Maggie Paulose, Oomman K. Varghese, and Craig A. Grimes* Department of Electrical Engineering and Department of Material Science and Engineering, The PennsylVania State UniVersity, UniVersity Park, PennsylVania 16802 Received October 24, 2005
ABSTRACT We describe the use of highly ordered transparent TiO2 nanotube arrays in dye-sensitized solar cells (DSCs). Highly ordered nanotube arrays of 46-nm pore diameter, 17-nm wall thickness, and 360-nm length were grown perpendicular to a fluorine-doped tin oxide-coated glass substrate by anodic oxidation of a titanium thin film. After crystallization by an oxygen anneal, the nanotube arrays are treated with TiCl4 to enhance the photogenerated current and then integrated into the DSC structure using a commercially available ruthenium-based dye. Although the negative electrode is only 360-nm-thick, under AM 1.5 illumination the generated photocurrent is 7.87 mA/cm2, with a photocurrent efficiency of 2.9%. Voltage-decay measurements indicate that the highly ordered TiO2 nanotube arrays, in comparison to nanoparticulate systems, have superior electron lifetimes and provide excellent pathways for electron percolation. Our results indicate that remarkable photoconversion efficiencies may be obtained, possibly to the ideal limit of ∼31% for a single photosystem scheme, with an increase of the nanotube-array length to several micrometers.
Transparent porous semiconducting electrodes1 are of considerable interest for application in dye or solid-state heterojunction solar cells.2-8 Dye-sensitized solar cells (DSCs) are a relatively low cost solar cell technology that has achieved overall light-to-electricity conversion efficiencies of over 10.6%.6,7 The electron-collecting layer in a DSC is typically a 10-µm-thick nanoparticulate film, with a threedimensional network of interconnected 15-20-nm-sized nanoparticles.8 The large surface area of the nanoporous film enables efficient light harvesting, maximizing the amount of photogenerated charge. Electron transport is a limiting factor in the performance of these nanoporous nanocrystalline electrodes, hindering progress in achieving higher efficiencies. The structural disorder at the contact between two crystalline nanoparticles leads to enhanced scattering of free electrons, thus reducing electron mobility.9 An ordered and strongly interconnected nanoscale photoanode architecture offers the potential for improved electron transport leading to higher photoefficiencies. Several nanotubular architectures have been investigated for potential enhancement of electron percolation pathways and light conversion as well as improved ion diffusion at the semiconductor-electrolyte interface. Sol-gel,10 sonochemistry,11 and surfactant templating9-based approaches have been used to fabricate nanotubular structures. For example, Adachi et al.12 have fabricated DSCs’ with electrodes, 4-µm-thick, composed of disordered TiO2 single-crystalline nanotubes (10-nm diam* Corresponding author. E-mail:
[email protected]. 10.1021/nl052099j CCC: $33.50 Published on Web 12/31/2005
© 2006 American Chemical Society
eter, 30-300-nm length) using molecular assemblies to obtain efficiencies of 4.88%. We have reported previously on the use of highly ordered 6.2-µm-thick TiO2 nanotube arrays, made by anodization of a thick-film titanium foil,13 to generate hydrogen by water photoelectrolysis with a photoconversion efficiency of over 12.25% under 320-400-nm illumination.14 The highly ordered architecture allows for improved charge separation and charge transport, with a calculated quantum efficiency of over 80% for incident photons with energies larger than the titania band gap. However, in application to waterphotolysis the TiO2 nanotube arrays are anchored to an opaque Ti metal substrate, as the nanotube array is fabricated by anodization of a Ti foil approximately 250-µm-thick. Application of this material geometry to dye cells requires backside illumination, a less than optimal approach for lightto-electrical energy conversion because the platinized counter electrode partially reflects light, while iodine in the electrolyte absorbs photons in the near UV region. Recently, we reported the fabrication of transparent TiO2 nanotube arrays on conducting glass15 enabling front-side illumination; herein we report our initial results on application of these transparent, highly ordered nanotube arrays for use on the negative electrode in dye solar cells. A detailed methodology of fabricating optically transparent nanotube arrays has been published elsewhere;15 hence, only the key points of the fabrication process are summarized here. Titanium films, 500-nm-thick, are rf sputtered onto fluorinedoped tin oxide (FTO)-coated glass substrates held at 500
Figure 1. Key stages in the fabrication of a transparent TiO2 nanotube-array film: (a) sputter deposition of a high-quality Ti thin film; (b) anodization of resulting film, and (c) heat treatment to oxidize the remaining metallic islands.
°C. The resulting films are then anodized at a constant potential of 12 V in an electrolyte of 0.5% HF + acetic acid mixed in a 7:1 ratio using a platinum counter electrode (99% purity Alfa-Aesar Corporation). The electrolyte temperature was kept constant at approximately 5 °C. The anodization process results in the formation of titania nanotube arrays 360-nm-thick. The samples are annealed at 450 °C in oxygen for 3 h with heating and cooling rates of 2.5 °C/min to induce crystallinity. The anodization process resulted in the formation of titania nanotube arrays with pore diameters of 46 nm (SD 8 nm), wall thicknesses of 17 nm (SD 2 nm), and lengths of 360 nm. Figure 1 shows visual images of (a) the deposited Ti film, (b) the film upon anodization, and (c) the film upon subsequent anneal. Illustrative field emission scanning electron microscope (FESEM) images of the transparent nanotube arrays are shown in Figure 2. We note that the length of the nanotube array, in this work 360 nm, is critically dependent upon the thickness of the starting Ti film. Our current efforts at making longer optically transparent nanotube arrays has been limited by our inability to fabricate high-quality Ti films greater than 500 nm in thickness. Prior to use, the annealed samples were placed in a 0.2 M TiCl4 solution for 60 min at room temperature within airtight bottles and then rinsed in ethanol and annealed in air at 450 °C for 30 min. We used a commercially available rutheniumbased dye RuL2(NCS)2:2 TBA (N-719, Solaronix Inc.) as the sensitizer; the nanotube samples were immersed overnight in 0.3 mM ethanolic solution of N-719. The electrolyte contained 0.5 M LiI, 0.05 M I2, 0.6 M N-methylbenzimidazole, 0.10 M guanidinium thiocyanate, and 0.5 M tertbutylpyridine in methoxypropionitrile (MPN). A conductive glass slide sputter-coated with 25 nm of Pt was used at the counter electrode in the fabricated DSCs. Electrode spacing was ensured by the use of Parafilms spacers, 120 µm in thickness. The electrolyte was introduced into the clamped electrodes by capillary action. Figure 3 illustrates the DSC geometry in cross section. 216
Figure 2. Top and lateral view FESEM images of titanium nanotubes grown from a 500-nm-thick Ti thin film (sputtered onto SnO2:F coated glass at 500 °C) anodized using a 0.5% HF electrolyte concentration at a potential of 12 V.
Figure 3. Integration of transparent nanotube array architecture into dye solar cell structure.
The photocurrent (I) and photovoltage (V) of the cell were measured with an active area of 0.25 cm2 using simulated sunlight at AM-1.5 produced by a 150-W Oriel Solar Simulator. The I-V characteristics of the devices are shown in Figure 4. The numbers are illustrative of approximately 50 nanotube array DSC samples that we have made. At 100% sun, TiCl4-treated 360-nm-thick nanotube array DSCs exhibit a Jsc of 7.87 mA/cm2, a Voc of 0.75 V, and a fill factor (ff) Nano Lett., Vol. 6, No. 2, 2006
Figure 4. Photocurrent-photovoltage characteristics of a transparent nanotube array DSC under 100% AM-1.5 illumination.
Figure 5. Photovoltage-decay measurement of a TiCl4-treated transparent TiO2 nanotube array DSC at 100% sun.
of 0.49, with an overall conversion efficiency of 2.9%. We note that this efficiency is 5 times the efficiency of nonTiCl4-treated nanotubes; we hypothesize that the TiCl4 treatment facilitates improved bonding between the TiO2 and dye molecule, resulting in improved charge transfer. Open-circuit voltage-decay measurements were conducted to investigate the recombination kinetics of the nanotubular structure. Following the technique as reported by Zaban et al.,16 open-circuit voltage-decay measurements were performed by monitoring the Voc transient during relaxation from an illuminated quasiequilibrium state to the dark equilibrium, see Figure 5. When the AM-1.5 illumination on a DSC at open circuit is interrupted, the excess electrons are removed due to recombination, with the photovoltage decay rate directly related to the electron lifetime by the following expression: τn ) Nano Lett., Vol. 6, No. 2, 2006
[ ]
-kBT dVoc e dt
-1
(1)
Figure 6. Response time determined by open circuit photovoltage decay for a transparent TiO2 nanotube array DSC as well as response times for a TiO2 nanoparticle DSC replotted from ref 18.
The thermal energy is given by kBT, e is the positive elementary charge, and dVoc/dt is the derivative of the open circuit voltage transient. Appropriate use of eq 1 assumes that the recombination is linear with a first-order dependence on electron concentration and that electron recombination occurs only with the electrolyte.17 Figure 6 is the plot of the response time obtained by applying eq 1 to the data in Figure 5. For comparison, the response time data for TiO2 nanoparticles, replotted from ref 18, is also shown in Figure 6. In comparison to reported open circuit photovoltage decay measurements of nanoparticulate TiO2-based DSCs,16,18 the transparent nanotubes exhibit superior recombination characteristics, with the longer lifetimes indicating fewer recombination centers in the nanotubular films. Several aspects of our experimental results are worthy of consideration. The first is the photocurrent magnitude, 7.87 mA/cm2 under 1.5 AM illumination for a DSC having a 360nm-long nanotube array negative electrode. The relatively short nanotube array results in considerably less photoabsorption than, for example, the current DSC gold-standard of a 10-µm-thick layer of TiO2 nanoparticles. We note the fabrication of highly ordered nanotube arrays 6.2 µm in length from Ti foil,14 and suggest the possibility of fabricating highly efficient dye solar cells by increasing the length of the nanotube array on the negative electrode because the amount of the absorbed dye appears to be the limiting factor. Unfortunately, our work in this interesting field is currently limited by our inability to deposit high-quality Ti films suitable for anodization of greater thickness. A second factor for improvement is the fill factor, in our work 0.49, which is reduced with increasing series resistance. The series resistance will be increased, and hence fill factor reduced, with increasing barrier layer thickness and poor contact between the barrier layer and FTO substrate. The barrier-layer thickness can be reduced using a stepwise reduction in the anodization voltage, as described by Mor and co-workers,19 followed as needed by an acid rinse to further thin the barrier layer. Adhesion between the barrier layer and FTO substrate is a function of initial Ti film quality, which is in turn dependent upon deposition parameters. 217
Finally, we note that the resistance of the FTO substrate increases at least 1 order of magnitude with the oxygen annealing step needed to crystallize the nanotube array. Modification of this annealing step should facilitate retention of the FTO conducting properties thus increasing the fill factor. Reduction of the spacer thickness from 120 µm to the more conventional 25 µm thickness should further increase the fill factor by a significant amount. We believe that these processing issues are tractable and will result in a significant improvement in the photoconversion efficiency. In summary, we have fabricated transparent TiO2 nanotube arrays on conductive glass substrates and used them in frontside-illuminated dye-sensitized solar cells. Demonstrated conversion efficiencies of 2.9% from a 360-nm-thick nanotube array suggest transparent nanotubes to be a very promising electrode material. Comparison of the observed open circuit photovoltage decay in these devices to published reports suggest superior electron transport in nanotubular titania-based DSCs. Current challenges include reducing the barrier layer thickness during the anodization process, producing transparent nanotube arrays of micrometer length, and uniform deposition of the dye upon the nanotube arrays. Acknowledgment. Partial support of this work by the National Science Foundation, CTS-0518269, and a Seed Grant provided by the Penn State Hydrogen Center is gratefully acknowledged. C.A.G. thanks Robert Gray, Brian Hardin, and Greg Smestad for helpful discussions.
218
References (1) Ito, S.; Yamada, Y.; Kuze, M.; Tabata, K.; Yashima, T. J. Mater. Sci. 2004, 39, 5853-5856. (2) Gratzel, M. Inorg. Chem. 2005, 44, 6841-6851. (3) Bach, U.; Lupo, D.; Comte, P.; Moser, J. E.; Weissortel, F.; Salbeck, J.; Spreitzer, H.; Gratzel, M. Nature 1998, 395, 583-585. (4) (a) Gratzel, M. Nature 2001, 414, 338-344. (b) Tunesi, S.; Anderson, M. A. Langmuir 1992, 8, 487-495. (5) Anderson, M. A.; Gieselmann, M. J.; Xu, Q. Y. J. Membr. Sci. 1988, 39, 243-258. (6) O’Regan, B.; Moser, J.; Anderson, M. A.; Gratzel, M. J. Phys. Chem. 1990, 94, 8720-8726. (7) O’Regan, B.; Gratzel, M. Nature 1991, 353, 737-740. (8) Gratzel, M. J. Photochem. Photobiol., A 2004, 168, 235-235. (9) Peng, T. Y.; Hasegawa, A.; Qiu, J. R.; Hirao, K. Chem. Mater. 2003, 15, 2011-2016. (10) Zhang, M.; Bando, Y.; Wada, K. J. Mater. Sci. Lett. 2001, 20, 167170. (11) Zhu, Y. C.; Li, H. L.; Koltypin, Y.; Hacohen, Y. R.; Gedanken, A. Chem. Commun. 2001, 24, 2616-2617. (12) Adachi, M.; Murata, Y.; Okada, I.; Yoshikawa, S. J. Electrochem. Soc. 2003, 150, G488-G493. (13) Cai, Q.; Paulose, M.; Varghese, O. K.; Grimes, C. A. J. Mater. Res. 2005, 20, 230-236. (14) Varghese, O. K.; Paulose, M.; Shankar, K.; Mor, G. K.; Grimes, C. A. J. Nanosci. Nanotechnol. 2005, 5, 1158-1165. (15) Mor, G. K.; Varghese, O. K.; Paulose, M.; Grimes, C. A. AdV. Funct. Mater. 2005, 15, 1291-1296. (16) Zaban, A.; Greenshtein, M.; Bisquert, J. ChemPhysChem 2003, 4, 859-864. (17) Bisquert, J.; Zaban, A.; Greenshtein, M.; Mora-Sero, I. J. Am. Chem. Soc. 2004, 126, 13550-13559. (18) Fabregat-Santiago, F.; Garcia-Canadas, J.; Palomares, E.; Clifford, J. N.; Haque, S. A.; Durrant, J. R., Garcia-Belmonte, G.; Bisquert, J. J. Appl. Phys. 2004, 96, 6903-6907. (19) Mor, G. K.; Varghese, O. K.; Paulose, M.; Mukherjee, N.; Grimes, C. A. J. Mater. Res. 2003, 11, 2588-2594.
NL052099J
Nano Lett., Vol. 6, No. 2, 2006
