Langmuir 2008, 24, 547-550
547
Preparation, Photocatalytic Activities, and Dye-Sensitized Solar-Cell Performance of Submicron-Scale TiO2 Hollow Spheres Yoshihiko Kondo,† Hirofumi Yoshikawa,† Kunio Awaga,*,† Masaki Murayama,‡ Tatsuo Mori,‡ Kayano Sunada,§ Shunji Bandow,⊥ and Sumio Iijima⊥ Research Center for Materials Science & Department of Chemistry, Nagoya UniVersity, Furo-cho, Chikusa, Nagoya 464-8602, Japan, Department of Electrical Engineering and Computer Science, Graduate School of Engineering, Nagoya UniVersity, Furo-cho, Chikusa, Nagoya 464-8603, Japan, Research Center for AdVanced Science and Technology, The UniVersity of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8904, Japan, and Department of Materials Science and Engineering, Faculty of Science and Technology, Meijo UniVersity, Nagoya 468-8502, Japan ReceiVed July 18, 2007. In Final Form: October 16, 2007 We prepared submicron-scale spherical hollow particles of anatase TiO2 by using a polystyrene-bead template. The obtained particles were very uniform in size, with a diameter of 490 nm and a shell thickness of 30 nm. The BrunauerEmmett-Teller surface area measurements revealed a large value of 70 m2/g. The photocatalytic property was investigated by the complete decomposition of gaseous isopropyl alcohol under UV irradiation. It was indicated that the activity of the hollow spheres was 1.8 times higher than that of the conventional P25 TiO2 nanoparticles with a diameter of 30 nm. Furthermore, we fabricated a dye-sensitized solar cell (DSC) using an electrode of the TiO2 hollow spheres, and examined the photovoltaic performance under simulated sunlight. Although the per-area efficiency was rather low (1.26%) because of a low area density of TiO2 on the electrode, the per-weight efficiency was 2.5 times higher than those of the conventional DSCs of TiO2.
Introduction Recently, nano- and submicron-scale inorganic materials with specific structural features have attracted much attention because of intriguing physical properties caused by their unusual sizes and shapes.1-3 Among them, the spherical hollow structures certainly provide advantageous features, such as a low density and a large surface area, so that they would have potential applications in magnetic devices, catalysis,4 photonic materials,5 drug delivery,6 and so forth. To prepare hollow spheres with a monodispersed size distribution, one of the most useful methods is soft template synthesis,7-9 in which template core particles, such as polystyrene (PS) beads, are coated with inorganic substances, and then removed by solvent liquation or calcination. In our previous works, we prepared hollow spheres of ccp- and hcp-Co, Co3O4, Fe, Fe3O4, and Fe2O3 with a diameter of ca. 600 nm and a shell thickness of ca. 40 nm using this method, and studied their peculiar magnetic properties.10-12 * To whom correspondence should be addressed. Tel: +81-52-7892487. Fax: +81-52-789-2484. E-mail:
[email protected]. † Research Center for Materials Science & Department of Chemistry, Nagoya University. ‡ Graduate School of Engineering, Nagoya University. § The University of Tokyo. ⊥ Meijo University. (1) An, K.; Lee, N.; Park, J.; Kim, C. S.; Hwang, Y.; Park, J.-G.; Kim, J.-Y.; Park, J.-H.; Han, J. M.; Yu, J.; Hyeon, T. J. Am. Chem. Soc. 2006, 128, 97539760. (2) Graf, C.; Dembski, S.; Hofmann, A.; Ru¨hl, E. Langmuir, 2006, 22, 56045610. (3) Kovalenko, V. M.; Bodnarchuk, I. M.; Lechner, T. R.; Hesser, G.; Scha¨ffler, F.; Heiss, W. J. Am. Chem. Soc. 2007, 129, 6352-6353. (4) Li, Y.; Zhou, P.; Dai, Z.; Hu, Z.; Sun, P.; Bao, J. New J. Chem. 2006, 30, 832-837. (5) Zhu, Y.-Z.; Chen, H.-B.; Wang, Y.-P.; Li, Z.-H.; Cao, Y.-L.; Chi, Y.-B. Chem. Lett. 2006, 35, 756-757. (6) Fujiwara, M.; Shiokawa, K.; Hayashi, K.; Morigaki, K.; Nakahara, Y. J. Biomed. Mater. Res. A 2007, 81, 103-112. (7) Shiho, H.; Kawahashi, N. J. Colloid Interface Sci. 2000, 226, 91-97. (8) Kawahashi, N.; Shiho, H. J. Mater. Chem. 2000, 10, 2294-2297. (9) Yoon, S. B.; Kim, J. Y.; Kim, J. H.; Park, S. G.; Kim, J. Y.; Lee, C. W.; Yu, J.-S. Curr. Appl. Phys. 2006, 6, 1059-1063.
Titanium dioxide (TiO2) has been studied extensively because of its unique properties, such as photocatalytic reactions13 and high photovoltaic efficiencies.14 To improve these properties, the morphologies and macroscopic structures of TiO2 have been intensively studied.15-17 For example, a three-dimensional porous structure of TiO2 with a large surface area is known to exhibit an enhanced photocatalytic performance.18 The submicron-scale hollow spheres of TiO2 are promising because of their potential to provide a large surface/volume ratio. In addition, since this size is comparable to the wavelengths of the visible and UV lights, the diffractions on the hollow spheres and the reflections due to the shell structure would improve the functional properties of TiO219 (see Figure 1a). While preparations of the TiO2 hollow spheres have been reported by several groups,20-24 their physical properties have yet to be examined in detail. (10) Yoshikawa, H.; Hayashida, K.; Kozuka, Y.; Horiguchi, A.; Awaga, K. Appl. Phys. Lett. 2004, 85, 5287-5289. (11) Ohnishi, M.; Kozuka, Y.; Ye, Q.-L.; Yoshikawa, H.; Awaga, K.; Matsuno, R.; Kobayashi, M.; Takahara, A.; Yokoyama, T.; Bandow, S.; Iijima, S. J. Mater. Chem. 2006, 16, 3215-3220. (12) Ye, Q.-L.; Kozuka, Y.; Yoshikawa, H.; Awaga, K.; Bandow, S.; Iijima, S. Phys. ReV. B 2007, 75, 224404. (13) Fujishima, A.; Inoue, T.; Honda, K. J. Am. Chem. Soc. 1979, 101, 55825588. (14) O’ Regan, B.; Gra¨tzel, M. Nature 1991, 353, 737-740. (15) Kasuga, T.; Hiramatsu, M.; Hoson, A.; Sekino, T.; Niihara, K. Langmuir 1998, 14, 3160-3163. (16) Yu, Y.; Xu, D. Appl. Catal. B 2007, 73, 166-171. (17) Beyers, E.; Cool, P.; Vansant, E. F. Microporous Mesoporous Mater. 2007, 99, 112-117. (18) Li, Y.; Kunitake, T.; Fujikawa, S. J. Phys. Chem. B 2006, 110, 1300013004. (19) Li, H.; Bian, Z.; Zhu, J.; Zhang, D.; Li, G.; Huo, Y.; Li, H.; Lu, Y. J. Am. Chem. Soc. 2007, 129, 8406-8407. (20) Imhof, A. Langmuir 2001, 17, 3579-3585. (21) Eiden, S.; Maret, G. J. Colloid Interface Sci. 2002, 250, 281-284. (22) Kobayashi, Y.; Gu, S.; Kondo, T.; Mine, E.; Nagao, D.; Konno, M. J. Chem. Eng. Jpn. 2004, 37, 912-914. (23) Syoufian, A.; Inoue, Y.; Yada, M.; Nakashima, K. Mater. Lett. 2007, 61, 1572-1575. (24) Strohm, H.; Lo¨bmann, P. Chem. Mater. 2005, 17, 6772-6780.
10.1021/la702157r CCC: $40.75 © 2008 American Chemical Society Published on Web 12/19/2007
548 Langmuir, Vol. 24, No. 2, 2008
Figure 1. Multiple diffractions and reflections on the hollow spheres (a); SEM images of the PS particles (b), the titania-coated PS particles (c), and the TiO2 hollow spheres (d); TEM image of the TiO2 hollow spheres (e).
In the present work, we prepared hollow spheres of TiO2 with a diameter of 490 nm and a shell-thickness of 30 nm using PS templates. We examined their photocatalytic activities, as well as the photovoltaic efficiencies of the dye-sensitized solar cells (DSCs) fabricated with these hollow spheres. Experimental Section Samples. All the chemicals were commercially purchased, and were used without purification. The water used in the following sample preparations was 18 MΩ Milli-Q filtered. Cationic PS beads of 460 nm were prepared by surfactant-free emulsion polymerization, using a cationic initiator, 2,2′-azobis(isomethyl-propionamidine)dihydrochloride (AMPAHCl): to 190 mL of water were added 10 mL of styrene and 0.144 g of AMPAHCl, and the solution was then heated at 70 °C for 24 h. The suspension was dialyzed against water in cellulose membrane tubing for 12 h. The solvent was replaced with ethanol by repeated centrifugation and redispersal. Then the PS cores were isolated and dried. The cationic PS spheres were coated with titania by the hydrolysis of titanium tetraisopropoxide (TTIP); 0.760 g of poly(vinylpyrrolidone) (Mw ) 40 000) and 2.5 mL of water were dissolved into 100 mL of ethanol. To this solution were rapidly added 10 mL of ethanol containing the cationic PS at 0.05 g/mL and 1 g of TTIP dissolved in 15 mL of absolute ethanol, and then the solution was stirred vigorously for 45 min. The coated particles were centrifuged and redispersed into ethanol several times. Then they were collected and dried. The PS cores were removed by calcination in a furnace at 480 °C in air, and the hollow spheres of TiO2 were obtained. Characterization. The morphologies of the samples were examined by a scanning electron microscope (SEM; Hitachi S-4300) and a transmission electron microscope (TEM; Hitachi H-800). The crystal structures were determined on an X-ray powder diffractometer
Kondo et al. (Rigaku MultiFlex, CuKR, λ) 0.15418 nm). The specific surface areas were estimated using the Brunauer-Emmett-Teller (BET) method (Quantachrome Autosurb-1). Photocatalytic Activity. Photocatalytic activities of the TiO2 hollow spheres were evaluated by the decomposition time of isopropyl alcohol (IPA). The sample (ca. 50 mg) was placed in an airtight glass chamber and illuminated with a UV light (λmax) 355 nm, National Panasonic Type 10 BL-B) with an intensity of 1 mW/cm2 for 12 h, and then the chamber gas was replaced with dry pure air. After this pretreatment, 100 vppm of IPA was added to the chamber, and then the chamber was left in darkness until adsorption equilibrium was reached. The chamber was then illuminated with the UV light, and the CO2 concentration in the chamber was monitored by gas chromatography (Shimadzu model GC-08). The photocatalytic activities of P25 (Aerosil) TiO2 nanoparticles with a diameter of ca. 30 nm were also studied by the same procedure. Fabrication of DSCs. The electrodes of the TiO2 hollow spheres were prepared on a fluorine tin oxide (FTO) glass plate by the deposition method. The TiO2 hollow spheres were dispersed in a 0.3 wt % sodium dodecyl sulfate (SDS) aqueous solution (18 mL), and then a surface-cleaned FTO glass plate was soaked in this solution for several days. After drying, the coated FTO was calcined in a furnace at 450 °C in air for 30 min. For comparison, electrodes using the P25 TiO2 nanoparticles were also prepared by the doctor blade technique. The electrodes were immersed for 12 h in a 0.5 mM indoline dye (referred to as D149)25 solution with a mixed solvent of acetonitrile and t-butyl alcohol (v/v, 1:1), which also contained 1.0 mM of chenodeoxy cholic acid to prevent the dye molecules from aggregating with each other. These dye-adsorbed electrodes (1 cm2 surface area) and Pt counter electrodes coated on an indium tin oxide (ITO) substrate were assembled into the open-cells. The electrolyte solution was composed of 0.60 M 1,2-dimethyl-3propylimidazolium iodide, 0.05 M I2, 0.10 M LiI, and 0.05 M 4-tertbutylpyridine in acetonitrile. Photocurrent versus voltage characteristics were measured on a homemade instrument, using simulated sunlight produced by a solar simulator (Seric XIL-03E). The light intensity, AM1.5 (100 mW/cm2), was calibrated with a pyranometer (Eko Instrument, MS62).
Results and Discussion Preparation of TiO2 Hollow Spheres. Figure 1b shows an SEM image of the PS beads that were used as templates. Their sizes are monodispersed with a diameter of ca. 460 nm. Figure 1c shows an SEM image of the titania-coated PS beads with a narrow size dispersion from an average diameter of 590 nm and a shell thickness of 65 nm. The coating is very smooth and uniform. Figure 1d shows an SEM image of the TiO2 hollow spheres, obtained after removing the PS cores by calcination. The spherical shapes are maintained even after calcination. Figure 1e shows a TEM image of the TiO2 hollow spheres. The spherical dark contrast on the outer rim of each particle is consistent with the shell structure. The calcination also brings about shrinkages of the hollow spheres; the average diameter and shell thickness are 490 and 30 nm, respectively. This shrinking is caused by the evaporation of the PS beads and volatile elements in the inorganic layers during calcination. Figure 2 shows the X-ray powder diffraction (XRD) pattern of the hollow spheres, which is in good agreement with that of the anatase structure of TiO2. The half peak widths indicate a grain size of 16 nm. Figure 3 shows the nitrogen adsorptiondesorption isotherm of the as-prepared TiO2 hollow spheres, indicating a classic type-IV behavior for mesoporous structures or a type II behavior for macroporous structures. The inset of this figure depicts the pore size distribution, estimated from the adsorption branch of the isotherm using the Dollimore-Heal (25) Horiuchi, T.; Miura, H.; Sumioka, K.; Uchida, S. J. Am. Chem. Soc. 2004, 126, 12218-12219.
Submicron-Scale TiO2 Hollow Spheres
Langmuir, Vol. 24, No. 2, 2008 549
Figure 2. XRD patterns of the TiO2 hollow spheres.
Figure 3. Nitrogen adsorption-desorption isotherm of the TiO2 hollow spheres. Inset shows the pore size distribution obtained from the adsorption curve.
(DH) method. There is a wide size distribution, and the BET surface area and specific pore volume are calculated to be 70 m2/g and 0.23 cm3/g, respectively. These values are larger than the corresponding ones (50 m2/g and 0.20 cm3/g) of the P25 TiO2 nanoparticles with a particle size of 30 nm, suggesting a significant surface roughness for the TiO2 hollow spheres. Photocatalytic Activity. The photocatalytic activities of the TiO2 hollow spheres are examined together with the P25 nanoparticles by measuring the complete decomposition time of IPA by photocatalytic oxidation under UV irradiation. IPA (ca. 100 vppm) was added to a glass chamber that included the TiO2 hollow spheres or the P25 nanoparticles. After reaching adsorption equilibrium in darkness, the chamber was illuminated with the UV light. Figure 4a shows the IPA concentration cIPA before and during the irradiation. As soon as IPA was introduced into the reaction chamber, cIPA exhibited a quick decrease due to the adsorption of IPA onto the TiO2 particles. The values of cIPA become nearly zero within 50 min in the trial using the TiO2 hollow spheres. When the P25 TiO2 nanoparticles were used, in contrast, these particles did not completely adsorb IPA, and cIPA reached an equilibrium value of 30 vppm. This difference was probably due to the difference in surface area between the two particles. Figure 4b shows the time dependence of the yield of CO2, yCO2, which results from the photocatalytic decomposition of IPA. These yCO2 values were calculated from the theoretical value for the CO2 concentration after the complete decomposition of IPA. Under UV irradiation, yCO2 for the hollow spheres and nanoparticles exhibited quick increases within 50 min, but the difference in yCO2 values between the two particles became
Figure 4. Comparison between the photocatalytic reactions of the hollow spheres (O) and the P25 nanoparticles (b) of TiO2: the concentration of IPA (a) and the yield of CO2 (b). The photoirradiation starts at t ) 0.
significant after this initial spike. The full decomposition time for the hollow spheres is much shorter than that for P25; the photocatalytic activity of the hollow spheres is ca. 80% higher than that of P25. Note that the complete decomposition of IPA by P25 should include an adsorption process of the IPA gas that remained in the reaction cell before the irradiation. That is probably why the decomposition of IPA took a longer time for the P25 nanoparticles, in contrast to the initial rapid decomposition within 50 min. Since it is hard to clearly identify the rate-determining step, it would be worth noting other enhancement factors for the hollow spheres: the multiple diffractions and reflections19 due to the fact that the size of the hollow spheres is comparable to the wavelengths of the UV-vis light, and/or the very small grain size (16 nm), which could bring about a quick migration of the photoexcited electrons toward the particle surfaces. DSC. In this section, we study the application of an electrode composed of TiO2 hollow spheres to a DSC. Electrodes composed of these particles were prepared on an FTO glass plate by deposition in a 0.3 wt % SDS aqueous solution (18 mL), as described in the experimental section. By varying the amount (3, 4.5, 5, and 10 mg) of the TiO2 hollow spheres in the solution, we controlled the thickness of the TiO2 layer to optimize the DSC performance. The results showed that the DSC prepared with 4.5 mg of TiO2 exhibited the best performance among the four; the photocurrents of the latter two (5 and 10 mg) were significantly lower than those of the former two. Since the TiO2 layers in the latter two were found to include many cracks and much surface roughness, their poor performances were probably
550 Langmuir, Vol. 24, No. 2, 2008
Kondo et al. Table 1. DSC Performance of the TiO2 Electrodes Made from P25 Nanoparticles and Hollow Spheres
Figure 5. SEM image of the TiO2 hollow spheres on an FTO electrode.
TiO2
Voc (V)
Jsc (mA)
FF
efficiency (%)
P25 hollow spheres
0.73 0.64
8.12 4.32
0.446 0.455
2.64 1.26
The two kinds of electrodessi.e., those based on the hollow spheres and those based on the P25 nanoparticlesswere immersed in a solution of D149 for 12 h to adsorb the dyes. These electrodes were colored in red purple (see the inset of Figure 6).26 The color of the hollow-sphere electrodes were lighter than those of the P25 electrodes, reflecting the difference between their TiO2 area densities. Using these electrodes (1 cm2 surface area), we fabricated DSCs with an electrolyte solution, as described in the experimental section. We examined the photocurrent (I)-voltage (V) with an active area 1 cm2 using simulated sunlight. The results are shown in Figure 6, and the photovoltaic performance parameters are summarized in Table 1. The closed circles in Figure 6 show the results for the P25 DSC. This exhibited a performance with an overall conversion efficiency of ca. 2.6%, which is typical for the TiO2 DSCs. The open circles in Figure 6 show the results for the hollow-sphere DSC. This exhibited a short-circuit current density (Jsc) of 4.32 mA/cm2, an open circuit voltage (Voc) of 0.64 V, and a fill factor (FF) of 0.455, with which the overall conversion efficiency was calculated to be 1.26%. This value is much lower than those for the P25 DSC (2.6%), although the difference is understandable, since the TiO2 area density of the hollow-sphere DSC is one-fifth lower than that of the P25 DSC. If these coefficients are transformed into values per a unit of weight, the per-weight efficiency of the hollow-sphere DSC is 2.5 times higher than those of the P25 DSC. This could be attributable to the specific diffraction and/or reflection of the illuminated lights on the TiO2 hollow spheres of submicron size. Another possibility is a difference between the amounts of the adsorbed D149 on the two electrodes, although we could not measure their values.26 However, it is not likely, because the photocurrent of the P25 DSC is only twice as large as that of the hollow-sphere DSC, although the unit-area weight of TiO2 on the former is 5 times larger than that on the latter.
Conclusions
Figure 6. I-V curves of the DSCs made from the TiO2 electrodes; the hollow spheres (O) and the P25 nanoparticles (b). Inset shows the photographs of the TiO2 electrodes based on the hollow spheres and P25 nanoparticles.
due to the difficulty of forming thick layers of the submicron particles without defects. Hereafter we describe the performance of an electrode prepared by deposition of 4.5 mg of TiO2. Figure 5 shows an SEM image for this electrode; it can be seen that the spherical hollow structures of the TiO2 are maintained. The area density of TiO2 is calculated to be 0.26 mg/cm2 from the weight of TiO2. The XRD pattern is nearly the same as that in Figure 2 (data not shown). We also prepared electrodes with the P25 TiO2 nanoparticles (30 nm) by the doctor blade technique. To obtain the highest coefficient, we optimized the area density to be 1.31 mg/cm2 for this electrode. This density was much larger than that of the hollow-sphere electrode.
TiO2 hollow spheres were prepared by hydrolysis of a titanium alkoxide in the presence of the cationic PS templates, followed by removal of the PS cores under calcination. The hollow spheres were found to possess a large BET surface area of 70 m2/g. Photocatalytic activities of the hollow spheres were confirmed under UV irradiation. The DSC of the hollow-sphere electrode exhibited a Jsc of 4.32 mA/cm2, a Voc of 0.64 V, and an FF of 0.455 with an overall conversion efficiency of 1.26%. The perweight efficiency calculated from this value was 2.5 times higher than those of the conventional DSCs of TiO2. It was demonstrated that the photofunctions of TiO2 can be enhanced in the submicronscale hollow spheres. Acknowledgment. The authors are indebted to Prof. Kazuhito Hashimoto for his support in the photocatalytic activity measurement of Figure 3 and helpful discussion. This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology (MEXT). LA702157R (26) We tried to estimate the amounts of the adsorbed dyes, after separating them from TiO2 with a treatment of NaOH(aq) solution, but reliable values could not be obtained because of incomplete separations of the dyes.
