- Email: [email protected]
Correlation between dispersion properties of TiO2 colloidal sols and photoelectric characteristics of TiO2 films
Journal of Colloid and Interface Science 279 (2004) 479–483 www.elsevier.com/locate/jcis
Correlation between dispersion properties of TiO2 colloidal sols and photoelectric characteristics of TiO2 films Hyun Suk Jung a , Sang-Wook Lee a , Jin Young Kim a , Kug Sun Hong a,∗ , Young Cheol Lee b , Kyung Hyun Ko b a School of Materials Science and Engineering, Seoul National University, 151-742 Seoul, South Korea b Department of Materials Science and Engineering, Ajou University, Suwon 442-749, South Korea
Received 1 December 2003; accepted 30 June 2004 Available online 9 September 2004
Abstract TiO2 film for use as dye-sensitized solar cell was prepared using the TiO2 colloidal sols (unpeptized sol and peptized sol). The optical properties and photocurrent–voltage characteristics of the resultant films were investigated. The optical transmittance of TiO2 thin film prepared from the peptized colloidal sol was over 90%, while that of TiO2 film from the unpeptized sol was under 80%. The TiO2 photoelectrode prepared from the peptized colloidal sol showed low photoelectric conversion efficiency (η), 1.30%, whereas the efficiency of photoelectrode from the unpeptized sol was 2.21%. The high optical transmittance and low conversion efficiency of TiO2 film from the peptized sol are discussed in terms of dense microstructure due to the drying nature of well-dispersed colloidal sol. 2004 Elsevier Inc. All rights reserved. Keywords: Dye-sensitized solar cell; TiO2 ; Peptization; Dispersion; Microstructure
1. Introduction Dye-sensitized solar cells are the most promising alternative to conventional solar cells conceived in recent years [1,2]. Nanosized TiO2 particles have received great attention for use as a photoelectrode in dye-sensitized solar cell systems. In general, the TiO2 photoelectrode consists of nanosized colloids that are sintered on a transparent conducting substrate, thus having a porous geometry [1,2]. Various preparation methods for nanosized TiO2 colloids have been suggested, such as soft chemistry, hydrothermal, and the sol–gel process [3–6]. In particular, the sol–gel process has the advantages of high purity, in comparison with the sulfate or chloride processes, and easy application to the fabrication of thin films [7,8]. Colloidal TiO2 sol can be synthesized via the hydrolysis of titanium alkoxide in excess water, followed by peptization with acid [9]. Peptization * Corresponding author. Fax: +82-2-886-4156.
E-mail address: [email protected] (K.S. Hong). 0021-9797/$ – see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2004.06.092
is the process of redispersing a colloid that has been coagulated. In our previous work, the well-crystallized anatase particles were successfully obtained by controlling the hydrolysis and peptization conditions [10]. In the present paper, we investigated the correlation between the peptization process of TiO2 colloidal sol with the photocurrent–voltage characteristics of the resultant TiO2 films, which were discussed in terms of intrinsic photocurrent (photocurrent of non-dye-adsorbed TiO2 films) and dye adsorption properties depending on morphologies of TiO2 films.
2. Experimental 2.1. Preparation of TiO2 colloidal sols and powders Titanium tetraisopropoxide (TTIP, Ti(OC3 H7 )4 , 97%, Aldrich Chemicals Co.) was hydrolyzed with excess water ([H2 O]/[Ti] = 50) at 80 ◦ C and stirred for 24 h at the same temperature. The resultant sol was named 80N sol. Nitric
480
H.S. Jung et al. / Journal of Colloid and Interface Science 279 (2004) 479–483
acid ([H+ ]/[Ti] = 0.5) was added to the 80N sol at an aging time of 24 h, after it was peptized at 80 ◦ C for 24 h. The peptized sol was designated as 80N80Y sol. The colloidal TiO2 sols were frozen instantaneously with liquid nitrogen. The frozen samples were dried in a freezedrying machine (−50 ◦ C and 10−3 Torr) for 24 h in order to prevent the TiO2 particles from agglomerating each other.
into the dye cell. The photocurrent–voltage characteristics of 80N and 80N80Y films were measured with a potentiostat (Model CHI 608A, CH Instruments) under an illumination of AM 1.5 (intensity 100 mW/cm2 ). In order to check the intrinsic photocurrent of bare TiO2 film, photocurrent of non-dye-adsorbed TiO2 electrodes was also measured under an illumination of UV light (306 nm, SANKYO). The electrolyte solution (SOLARONIX) was also used.
2.2. Preparation of TiO2 films To obtain the TiO2 electrode, 80N and 80N80Y slurries were formed by ball milling 3 g of TiO2 powders (as-dried 80N and 80N80Y powders), 9.6 ml of deionized water, 10 ml of ethanol, and 0.4 ml of acetylacetone. Transparent conducting glass (indium tin oxide, Samsung SDI) was cut into 2 × 2 cm pieces. The three edges were covered with scotch tape (40 µm thick), and approximately 100 µl of the TiO2 slurry was spread over the surface with a glass rod sliding on the scotch tape spacer. Therefore, the real dimension of the TiO2 electrode was 0.5 × 0.5 cm. After drying at room temperature, the TiO2 electrode was heated for 1 h in an electrical furnace at 450 ◦ C. The thickness of resultant film for both 80N and 80N80Y measured with alpha step apparatus was approximately 8 µm. The prepared TiO2 slurry was diluted with 20 ml ethanol in order to control the viscosity appropriately to spinning. The spin-coated TiO2 thin films were prepared on the quartz substrate on the purpose of measuring the transmittance. The diluted 80N and 80N80Y slurries were spread on the quartz substrate, which was spun, after the application, for 2 min at 3000 rpm. The coated TiO2 thin films were heated to 450 ◦ C for 1 h. The thickness of both 80N and 80N80Y thin films was approximately 400 nm.
3. Results and discussion The photographs of 80N and 80N80Y sols are shown in Figs. 1a and 1b, respectively. The 80N sol was optically opaque whereas the 80N80Y sol was translucent. It indicates that the unpeptized 80N sol consisted of severely agglomerated secondary particles, which scattered with visible light and resulted in the opacity. In contrasts, the relatively high transparency of 80N80Y sol is indicative of well-dispersed state. The optical transmittance properties of both 80N and 80N80Y thin films deposited on quartz substrates are shown in Fig. 2. The 80N thin film showed transmittance value (under 80%) relatively lower than that of 80N80Y film (over 90%). It is clear that the peptization process improved the optical transparency of the resultant thin film and the optical properties of thin films are dependent on the degree of dispersion of colloidal sols. Considering the higher transmittance value of 80N80Y thin film, it would be expected to have a dense structure. XRD reflections of both the freeze-dried 80N and 80N80Y particles are shown in Fig. 3. Anatase (101) reflection
2.3. Characterization The crystal structure of the TiO2 films was investigated using X-ray diffraction (XRD). Data collection was performed in a 2θ range of 20◦ to 33◦ using CuKα radiation (Model M18XHF, Macscience Instruments, Japan). Room-temperature IR transmittance spectra in the range 450–4000 cm−1 were measured using a Fourier transform IR (FTIR) spectrometer (Model DA8-12, Bomen, Canada). The UV transmittance of the thin films was measured with a HP 8452A spectrometer. Absorption spectra of dye solution, desorbed from TiO2 films by soaking in alkaline alcoholic solution, were also obtained to investigate the degree of dye adsorption. Field-emitted scanning electron microscopy (FESEM, Model JEOL JSM 6330F) was used to investigate the morphologies of the TiO2 films. The TiO2 electrode was immersed overnight in a solution of ruthenium dye (ruthenium (2,2 -bipyridyl-4,4dicarboxilate)2(NCS)2 (SOLARONIX) that was dissolved in ethanol). The resulting dye-adsorbed films were assembled with a Pt-sputtered glass to form a sandwich-type dye cell. The electrolyte solution (SOLARONIX) was spread
Fig. 1. Photographs of (a) 80N and (b) 80N80Y colloidal sols.
H.S. Jung et al. / Journal of Colloid and Interface Science 279 (2004) 479–483
Fig. 2. UV–vis transmission spectra of 80N and 80N80Y thin films: (2) 80N and (!) 80N80Y.
481
Fig. 4. Photocurrent–voltage characteristics of dye-sensitized 80N and 80N80Y photoelectrodes: (") 80N and (!) 80N80Y. Table 1 Photocurrent–voltage characteristics of 80N and 80N80Y electrodes heattreated at 450 ◦ C Sample
Voc (V)
Jsc (mA/cm2 )
FF (%)
η (%)
80N 80N80Y
0.63 0.66
10.6 4.0
33 50.2
2.21 1.30
Fig. 3. XRD reflections of (a) 80N and (b) 80N80Y freeze-dried powders.
was observed in both 80N and 80N80Y particles. The intensities and full widths at half maximum (FWHM) of the anatase (101) reflections for both 80N and 80N80Y sols was almost same. It implies that both the 80N and 80N80Y sol had well-crystallized anatase structure and the peptization process did not influence the crystallinity of 80N sol [10]. Fig. 4 shows the photocurrent–voltage characteristics of dye-sensitized 80N and 80N80Y electrodes heat-treated at 450 ◦ C. It is apparent that the 80N electrode had a better solar cell property than the 80N80Y. The photocurrent–voltage characteristics of both the 80N and 80N80Y electrodes are summarized in Table 1. The short-circuit current density of the 80N electrode was approximately 10.6 mA/cm2 , higher than that of 80N80Y (4.0 mA/cm2). The overall conversion efficiency (η) of the dye-sensitized solar cell is determined by the short-circuit current density (Jsc ), the open-circuit voltage (Voc ), the fill factor of the cell (FF), and the inten-
Fig. 5. XRD reflections of (a) 80N and (b) 80N80Y electrodes heat-treated at 450 ◦ C.
sity of the incident light (W ) [1,11]: Jsc Voc FF . W The conversion efficiency of the 80N electrode showed a higher value, 2.21%, than that of 80N80Y, whose efficiency was 1.30%. It indicates that the peptization process of TiO2 colloidal sol greatly influenced the photocurrent–voltage property of the resultant TiO2 photoelectrode. The anatase (101) reflections for 80N and 80N80Y electrodes heat-treated at 450 ◦ C are presented in Fig. 5. Considering that the intensities and FWHM of anatase (101)
η=
482
H.S. Jung et al. / Journal of Colloid and Interface Science 279 (2004) 479–483
Fig. 6. FTIR transmission spectra of (a) 80N and (b) 80N80Y powders obtained from scratching each electrode.
reflections for both 80N and 80N80Y films were almost the same, the peptization process did not impact the crystallite size and crystallinity of the resultant films in this case. FTIR spectra of TiO2 particles obtained by scratching both the 80N and 80N80Y electrodes were measured in order to investigate the surface state, as shown in Fig. 6. Before the annealing, there were some impurities such as residual organics or nitrate group in 80N and 80N80Y powders. The absorption bands at 1234, 1350–1460, and 1697 cm−1 are related to residual organics in 80N and the N–O stretching mode at 1385 cm−1 originates from the addition of nitric acid during the peptization process [12–15]. However, the impurities were fully removed after the annealing process and only the hydroxyl group remained on the surface of TiO2 . The similar intensity of the absorption band of the hydroxyl group indicates that the numbers of hydroxyl group adsorbed on the 80N and 80N80Y powders is not so different. Thus, the surface states of 80N and 80N80Y electrodes are not considered an important factor influencing the photocurrent–voltage characteristics of the photoelectrodes. The SEM micrographs of both 80N and 80N80Y electrodes heat-treated at 450 ◦ C are shown in Figs. 7a and 7b, respectively. The 80N electrode had more porous structure than the 80N80Y. The morphology of 80N and 80N80Y film is considered to depend on the extent of aggregation of TiO2 colloidal sol. Relatively large peptized particles pack efficiently, to give a dense microstructure in the green state [16]. Considering the overall measured results including the crystal structures, surface states, and morphologies of the 80N and 80N80Y photoelectrodes, the difference in the photocurrent–voltage properties (Fig. 4) is possibly correlated with the intrinsic photocurrent characteristics of TiO2 films and the number of adsorbed dye molecules depend-
Fig. 7. SEM micrographs of (a) 80N and (b) 80N80Y electrodes heat-treated at 450 ◦ C.
Fig. 8. Photocurrent characteristics of non-dye-adsorbed 80N and 80N80Y photoelectrodes: (2) 80N and (!) 80N80Y.
ing on the morphologies of photoelectrodes, i.e., porous and dense structures. The photocurrent characteristics of non-dye-adsorbed 80N and 80N80Y photoelectrodes were measured to investigate the intrinsic photocurrent characteristics of bare TiO2 films (Fig. 8). Contrary to the photocurrent characteristics of dye-adsorbed photoelectrode, the porous TiO2 electrode (80N) exhibits a slightly poor intrinsic photocurrent characteristic. The photocurrent density of 80N photoelectrode is approximately 13 mA/cm2, smaller than that of 80N80Y (15.5 mA/cm2). This might be related to the contact of TiO2 with ionic species in electrolyte, which considered to result in the low fill factor value of the 80N film (Table 1) [17,18]. The good solar cell performance of the porous photoelec-
H.S. Jung et al. / Journal of Colloid and Interface Science 279 (2004) 479–483
483
electrode from the unpeptized sol, 2.21%. Considering the photocurrent of non-dye-adsorbed TiO2 film and UV–visible absorption spectra of desorbed dye, the low conversion efficiency of TiO2 film fabricated from the peptized sol was attributed to small amount of dye adsorption, which resulted from a dense structure of TiO2 film related to the drying nature of well-dispersed colloidal sol. In the present paper, we report that the peptization process of colloidal sol can affect the morphology of the resultant film, and thereby impact on the photocurrent–voltage characteristics.
Fig. 9. Absorption spectra of desorbed dye solution from each TiO2 photoelectrode (solid line: 80N and dashed line: 80N80Y).
trode seems to be independent of the intrinsic photocurrent characteristic. The high short-circuit current density (Jsc ) of dyeadsorbed 80N electrode indicates that large numbers of dye molecules were adsorbed on the electrode, which ascribed to the porous structure. For the comparison of relative amounts of adsorbed dye, UV–visible spectra of dye which desorbed from each photoelectrode were obtained (Fig. 9). The large absorbance of dye, desorbed from the porous 80N film, clearly shows that the amount of adsorbed dye on the porous 80N film is larger than that on the dense 80N80Y film. To obtain high efficiency of the solar cell, appropriate porosity is required because the photocurrent is directly linked to the number of dye molecules adsorbed on the TiO2 electrode. The peptization process was originally believed to break up the large aggregates into smaller aggregates by the electrostatic repulsion of the charged particles. In the present study, it is clear that the peptization process results in TiO2 film having a dense structure, and consequently influences the photocurrent–voltage characteristics of the dyesensitized TiO2 films.
4. Conclusion In summary, the optical transmittance properties and solar cell properties of TiO2 films depending on the colloidal dispersion characteristics of nanosized TiO2 sol were investigated. The TiO2 thin film fabricated from the peptized colloidal sol showed higher optical transmittance (over 90%), whereas the optical transmittance of TiO2 film from the unpeptized sol was below 80%. The photoelectric conversion efficiency (η) of the TiO2 photoelectrode prepared from the peptized colloidal sol was 1.30% lower than that of the
Acknowledgments This work was supported by the Ministry of Science and Technology of Korea (“Support Project of National Research Program <2003>” supervised by KISTEP). The authors are also grateful to the Research Institute for Advanced Materials.
References [1] B. O’Regan, M. Grätzel, Nature 353 (1991) 737. [2] M.K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, E. Müller, P. Liska, N. Vlachopoulos, M. Grätzel, J. Am. Chem. Soc. 115 (1993) 6382. [3] H. Yin, Y. Wada, T. Kitamura, T. Sumida, Y. Hasegawa, S. Yanagida, J. Mater. Chem. 12 (2002) 378. [4] S.D. Park, Y.H. Cho, W.W. Kim, S.J. Kim, J. Solid State Chem. 146 (1999) 230. [5] C.-C. Wang, J.Y. Ying, Chem. Mater. 11 (1999) 3113. [6] M. Wu, J. Long, A. Huang, Y. Luo, S. Feng, R. Xu, Langmuir 15 (1999) 8822. [7] H.D. Nam, B.H. Lee, S.J. Kim, C.H. Jung, J.H. Lee, S. Park, Jpn. J. Appl. Phys. 37 (1998) 4603. [8] C.J. Brinker, G.W. Scherer, Sol–Gel Science, Academic Press, New York, 1990. [9] B.L. Bischoff, M.A. Anderson, Chem. Mater. 7 (1995) 1772. [10] H.S. Jung, J. Kim, D.S. Lee, K.S. Hong, H.J. Youn, submitted for publication. [11] M. Grätzel, Nature 414 (2001) 338. [12] H.C. Zeng, Mater. Res. Soc. Symp. Proc. 346 (1994) 715. [13] K. Terabe, K. Kato, H. Miyazaki, S. Yamaguchi, A. Imai, Y. Iguchi, J. Mater. Sci. 29 (1994) 1617. [14] K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination Compounds, Wiley, New York, 1997, p. 48. [15] P.S. Ha, H.J. Youn, H.S. Jung, K.S. Hong, Y.H. Park, K.H. Ko, J. Colloid Interface Sci. 223 (2000) 16. [16] R.G. Avery, J.D.F. Ramsay, in: M. Che, G.C. Bond (Eds.), Adsorption and Catalysis on Oxides, in: Studies in Surface Science and Catalysis, vol. 21, Elsevier, Amsterdam, 1985, p. 149. [17] S. Kambe, S. Nakade, T. Kitamura, Y. Wada, S. Yanagida, J. Phys. Chem. B 106 (2002) 2967. [18] C.J. Barbé, F. Arendse, P. Comte, M. Jirousek, F. Lenzmann, V. Shklover, M. Grätzel, J. Am. Ceram. Soc. 80 (1997) 3157.
Correlation between dispersion properties of TiO2 colloidal sols and photoelectric characteristics of TiO2 films Hyun Suk Jung a , Sang-Wook Lee a , Jin Young Kim a , Kug Sun Hong a,∗ , Young Cheol Lee b , Kyung Hyun Ko b a School of Materials Science and Engineering, Seoul National University, 151-742 Seoul, South Korea b Department of Materials Science and Engineering, Ajou University, Suwon 442-749, South Korea
Received 1 December 2003; accepted 30 June 2004 Available online 9 September 2004
Abstract TiO2 film for use as dye-sensitized solar cell was prepared using the TiO2 colloidal sols (unpeptized sol and peptized sol). The optical properties and photocurrent–voltage characteristics of the resultant films were investigated. The optical transmittance of TiO2 thin film prepared from the peptized colloidal sol was over 90%, while that of TiO2 film from the unpeptized sol was under 80%. The TiO2 photoelectrode prepared from the peptized colloidal sol showed low photoelectric conversion efficiency (η), 1.30%, whereas the efficiency of photoelectrode from the unpeptized sol was 2.21%. The high optical transmittance and low conversion efficiency of TiO2 film from the peptized sol are discussed in terms of dense microstructure due to the drying nature of well-dispersed colloidal sol. 2004 Elsevier Inc. All rights reserved. Keywords: Dye-sensitized solar cell; TiO2 ; Peptization; Dispersion; Microstructure
1. Introduction Dye-sensitized solar cells are the most promising alternative to conventional solar cells conceived in recent years [1,2]. Nanosized TiO2 particles have received great attention for use as a photoelectrode in dye-sensitized solar cell systems. In general, the TiO2 photoelectrode consists of nanosized colloids that are sintered on a transparent conducting substrate, thus having a porous geometry [1,2]. Various preparation methods for nanosized TiO2 colloids have been suggested, such as soft chemistry, hydrothermal, and the sol–gel process [3–6]. In particular, the sol–gel process has the advantages of high purity, in comparison with the sulfate or chloride processes, and easy application to the fabrication of thin films [7,8]. Colloidal TiO2 sol can be synthesized via the hydrolysis of titanium alkoxide in excess water, followed by peptization with acid [9]. Peptization * Corresponding author. Fax: +82-2-886-4156.
E-mail address: [email protected] (K.S. Hong). 0021-9797/$ – see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2004.06.092
is the process of redispersing a colloid that has been coagulated. In our previous work, the well-crystallized anatase particles were successfully obtained by controlling the hydrolysis and peptization conditions [10]. In the present paper, we investigated the correlation between the peptization process of TiO2 colloidal sol with the photocurrent–voltage characteristics of the resultant TiO2 films, which were discussed in terms of intrinsic photocurrent (photocurrent of non-dye-adsorbed TiO2 films) and dye adsorption properties depending on morphologies of TiO2 films.
2. Experimental 2.1. Preparation of TiO2 colloidal sols and powders Titanium tetraisopropoxide (TTIP, Ti(OC3 H7 )4 , 97%, Aldrich Chemicals Co.) was hydrolyzed with excess water ([H2 O]/[Ti] = 50) at 80 ◦ C and stirred for 24 h at the same temperature. The resultant sol was named 80N sol. Nitric
480
H.S. Jung et al. / Journal of Colloid and Interface Science 279 (2004) 479–483
acid ([H+ ]/[Ti] = 0.5) was added to the 80N sol at an aging time of 24 h, after it was peptized at 80 ◦ C for 24 h. The peptized sol was designated as 80N80Y sol. The colloidal TiO2 sols were frozen instantaneously with liquid nitrogen. The frozen samples were dried in a freezedrying machine (−50 ◦ C and 10−3 Torr) for 24 h in order to prevent the TiO2 particles from agglomerating each other.
into the dye cell. The photocurrent–voltage characteristics of 80N and 80N80Y films were measured with a potentiostat (Model CHI 608A, CH Instruments) under an illumination of AM 1.5 (intensity 100 mW/cm2 ). In order to check the intrinsic photocurrent of bare TiO2 film, photocurrent of non-dye-adsorbed TiO2 electrodes was also measured under an illumination of UV light (306 nm, SANKYO). The electrolyte solution (SOLARONIX) was also used.
2.2. Preparation of TiO2 films To obtain the TiO2 electrode, 80N and 80N80Y slurries were formed by ball milling 3 g of TiO2 powders (as-dried 80N and 80N80Y powders), 9.6 ml of deionized water, 10 ml of ethanol, and 0.4 ml of acetylacetone. Transparent conducting glass (indium tin oxide, Samsung SDI) was cut into 2 × 2 cm pieces. The three edges were covered with scotch tape (40 µm thick), and approximately 100 µl of the TiO2 slurry was spread over the surface with a glass rod sliding on the scotch tape spacer. Therefore, the real dimension of the TiO2 electrode was 0.5 × 0.5 cm. After drying at room temperature, the TiO2 electrode was heated for 1 h in an electrical furnace at 450 ◦ C. The thickness of resultant film for both 80N and 80N80Y measured with alpha step apparatus was approximately 8 µm. The prepared TiO2 slurry was diluted with 20 ml ethanol in order to control the viscosity appropriately to spinning. The spin-coated TiO2 thin films were prepared on the quartz substrate on the purpose of measuring the transmittance. The diluted 80N and 80N80Y slurries were spread on the quartz substrate, which was spun, after the application, for 2 min at 3000 rpm. The coated TiO2 thin films were heated to 450 ◦ C for 1 h. The thickness of both 80N and 80N80Y thin films was approximately 400 nm.
3. Results and discussion The photographs of 80N and 80N80Y sols are shown in Figs. 1a and 1b, respectively. The 80N sol was optically opaque whereas the 80N80Y sol was translucent. It indicates that the unpeptized 80N sol consisted of severely agglomerated secondary particles, which scattered with visible light and resulted in the opacity. In contrasts, the relatively high transparency of 80N80Y sol is indicative of well-dispersed state. The optical transmittance properties of both 80N and 80N80Y thin films deposited on quartz substrates are shown in Fig. 2. The 80N thin film showed transmittance value (under 80%) relatively lower than that of 80N80Y film (over 90%). It is clear that the peptization process improved the optical transparency of the resultant thin film and the optical properties of thin films are dependent on the degree of dispersion of colloidal sols. Considering the higher transmittance value of 80N80Y thin film, it would be expected to have a dense structure. XRD reflections of both the freeze-dried 80N and 80N80Y particles are shown in Fig. 3. Anatase (101) reflection
2.3. Characterization The crystal structure of the TiO2 films was investigated using X-ray diffraction (XRD). Data collection was performed in a 2θ range of 20◦ to 33◦ using CuKα radiation (Model M18XHF, Macscience Instruments, Japan). Room-temperature IR transmittance spectra in the range 450–4000 cm−1 were measured using a Fourier transform IR (FTIR) spectrometer (Model DA8-12, Bomen, Canada). The UV transmittance of the thin films was measured with a HP 8452A spectrometer. Absorption spectra of dye solution, desorbed from TiO2 films by soaking in alkaline alcoholic solution, were also obtained to investigate the degree of dye adsorption. Field-emitted scanning electron microscopy (FESEM, Model JEOL JSM 6330F) was used to investigate the morphologies of the TiO2 films. The TiO2 electrode was immersed overnight in a solution of ruthenium dye (ruthenium (2,2 -bipyridyl-4,4dicarboxilate)2(NCS)2 (SOLARONIX) that was dissolved in ethanol). The resulting dye-adsorbed films were assembled with a Pt-sputtered glass to form a sandwich-type dye cell. The electrolyte solution (SOLARONIX) was spread
Fig. 1. Photographs of (a) 80N and (b) 80N80Y colloidal sols.
H.S. Jung et al. / Journal of Colloid and Interface Science 279 (2004) 479–483
Fig. 2. UV–vis transmission spectra of 80N and 80N80Y thin films: (2) 80N and (!) 80N80Y.
481
Fig. 4. Photocurrent–voltage characteristics of dye-sensitized 80N and 80N80Y photoelectrodes: (") 80N and (!) 80N80Y. Table 1 Photocurrent–voltage characteristics of 80N and 80N80Y electrodes heattreated at 450 ◦ C Sample
Voc (V)
Jsc (mA/cm2 )
FF (%)
η (%)
80N 80N80Y
0.63 0.66
10.6 4.0
33 50.2
2.21 1.30
Fig. 3. XRD reflections of (a) 80N and (b) 80N80Y freeze-dried powders.
was observed in both 80N and 80N80Y particles. The intensities and full widths at half maximum (FWHM) of the anatase (101) reflections for both 80N and 80N80Y sols was almost same. It implies that both the 80N and 80N80Y sol had well-crystallized anatase structure and the peptization process did not influence the crystallinity of 80N sol [10]. Fig. 4 shows the photocurrent–voltage characteristics of dye-sensitized 80N and 80N80Y electrodes heat-treated at 450 ◦ C. It is apparent that the 80N electrode had a better solar cell property than the 80N80Y. The photocurrent–voltage characteristics of both the 80N and 80N80Y electrodes are summarized in Table 1. The short-circuit current density of the 80N electrode was approximately 10.6 mA/cm2 , higher than that of 80N80Y (4.0 mA/cm2). The overall conversion efficiency (η) of the dye-sensitized solar cell is determined by the short-circuit current density (Jsc ), the open-circuit voltage (Voc ), the fill factor of the cell (FF), and the inten-
Fig. 5. XRD reflections of (a) 80N and (b) 80N80Y electrodes heat-treated at 450 ◦ C.
sity of the incident light (W ) [1,11]: Jsc Voc FF . W The conversion efficiency of the 80N electrode showed a higher value, 2.21%, than that of 80N80Y, whose efficiency was 1.30%. It indicates that the peptization process of TiO2 colloidal sol greatly influenced the photocurrent–voltage property of the resultant TiO2 photoelectrode. The anatase (101) reflections for 80N and 80N80Y electrodes heat-treated at 450 ◦ C are presented in Fig. 5. Considering that the intensities and FWHM of anatase (101)
η=
482
H.S. Jung et al. / Journal of Colloid and Interface Science 279 (2004) 479–483
Fig. 6. FTIR transmission spectra of (a) 80N and (b) 80N80Y powders obtained from scratching each electrode.
reflections for both 80N and 80N80Y films were almost the same, the peptization process did not impact the crystallite size and crystallinity of the resultant films in this case. FTIR spectra of TiO2 particles obtained by scratching both the 80N and 80N80Y electrodes were measured in order to investigate the surface state, as shown in Fig. 6. Before the annealing, there were some impurities such as residual organics or nitrate group in 80N and 80N80Y powders. The absorption bands at 1234, 1350–1460, and 1697 cm−1 are related to residual organics in 80N and the N–O stretching mode at 1385 cm−1 originates from the addition of nitric acid during the peptization process [12–15]. However, the impurities were fully removed after the annealing process and only the hydroxyl group remained on the surface of TiO2 . The similar intensity of the absorption band of the hydroxyl group indicates that the numbers of hydroxyl group adsorbed on the 80N and 80N80Y powders is not so different. Thus, the surface states of 80N and 80N80Y electrodes are not considered an important factor influencing the photocurrent–voltage characteristics of the photoelectrodes. The SEM micrographs of both 80N and 80N80Y electrodes heat-treated at 450 ◦ C are shown in Figs. 7a and 7b, respectively. The 80N electrode had more porous structure than the 80N80Y. The morphology of 80N and 80N80Y film is considered to depend on the extent of aggregation of TiO2 colloidal sol. Relatively large peptized particles pack efficiently, to give a dense microstructure in the green state [16]. Considering the overall measured results including the crystal structures, surface states, and morphologies of the 80N and 80N80Y photoelectrodes, the difference in the photocurrent–voltage properties (Fig. 4) is possibly correlated with the intrinsic photocurrent characteristics of TiO2 films and the number of adsorbed dye molecules depend-
Fig. 7. SEM micrographs of (a) 80N and (b) 80N80Y electrodes heat-treated at 450 ◦ C.
Fig. 8. Photocurrent characteristics of non-dye-adsorbed 80N and 80N80Y photoelectrodes: (2) 80N and (!) 80N80Y.
ing on the morphologies of photoelectrodes, i.e., porous and dense structures. The photocurrent characteristics of non-dye-adsorbed 80N and 80N80Y photoelectrodes were measured to investigate the intrinsic photocurrent characteristics of bare TiO2 films (Fig. 8). Contrary to the photocurrent characteristics of dye-adsorbed photoelectrode, the porous TiO2 electrode (80N) exhibits a slightly poor intrinsic photocurrent characteristic. The photocurrent density of 80N photoelectrode is approximately 13 mA/cm2, smaller than that of 80N80Y (15.5 mA/cm2). This might be related to the contact of TiO2 with ionic species in electrolyte, which considered to result in the low fill factor value of the 80N film (Table 1) [17,18]. The good solar cell performance of the porous photoelec-
H.S. Jung et al. / Journal of Colloid and Interface Science 279 (2004) 479–483
483
electrode from the unpeptized sol, 2.21%. Considering the photocurrent of non-dye-adsorbed TiO2 film and UV–visible absorption spectra of desorbed dye, the low conversion efficiency of TiO2 film fabricated from the peptized sol was attributed to small amount of dye adsorption, which resulted from a dense structure of TiO2 film related to the drying nature of well-dispersed colloidal sol. In the present paper, we report that the peptization process of colloidal sol can affect the morphology of the resultant film, and thereby impact on the photocurrent–voltage characteristics.
Fig. 9. Absorption spectra of desorbed dye solution from each TiO2 photoelectrode (solid line: 80N and dashed line: 80N80Y).
trode seems to be independent of the intrinsic photocurrent characteristic. The high short-circuit current density (Jsc ) of dyeadsorbed 80N electrode indicates that large numbers of dye molecules were adsorbed on the electrode, which ascribed to the porous structure. For the comparison of relative amounts of adsorbed dye, UV–visible spectra of dye which desorbed from each photoelectrode were obtained (Fig. 9). The large absorbance of dye, desorbed from the porous 80N film, clearly shows that the amount of adsorbed dye on the porous 80N film is larger than that on the dense 80N80Y film. To obtain high efficiency of the solar cell, appropriate porosity is required because the photocurrent is directly linked to the number of dye molecules adsorbed on the TiO2 electrode. The peptization process was originally believed to break up the large aggregates into smaller aggregates by the electrostatic repulsion of the charged particles. In the present study, it is clear that the peptization process results in TiO2 film having a dense structure, and consequently influences the photocurrent–voltage characteristics of the dyesensitized TiO2 films.
4. Conclusion In summary, the optical transmittance properties and solar cell properties of TiO2 films depending on the colloidal dispersion characteristics of nanosized TiO2 sol were investigated. The TiO2 thin film fabricated from the peptized colloidal sol showed higher optical transmittance (over 90%), whereas the optical transmittance of TiO2 film from the unpeptized sol was below 80%. The photoelectric conversion efficiency (η) of the TiO2 photoelectrode prepared from the peptized colloidal sol was 1.30% lower than that of the
Acknowledgments This work was supported by the Ministry of Science and Technology of Korea (“Support Project of National Research Program <2003>” supervised by KISTEP). The authors are also grateful to the Research Institute for Advanced Materials.
References [1] B. O’Regan, M. Grätzel, Nature 353 (1991) 737. [2] M.K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, E. Müller, P. Liska, N. Vlachopoulos, M. Grätzel, J. Am. Chem. Soc. 115 (1993) 6382. [3] H. Yin, Y. Wada, T. Kitamura, T. Sumida, Y. Hasegawa, S. Yanagida, J. Mater. Chem. 12 (2002) 378. [4] S.D. Park, Y.H. Cho, W.W. Kim, S.J. Kim, J. Solid State Chem. 146 (1999) 230. [5] C.-C. Wang, J.Y. Ying, Chem. Mater. 11 (1999) 3113. [6] M. Wu, J. Long, A. Huang, Y. Luo, S. Feng, R. Xu, Langmuir 15 (1999) 8822. [7] H.D. Nam, B.H. Lee, S.J. Kim, C.H. Jung, J.H. Lee, S. Park, Jpn. J. Appl. Phys. 37 (1998) 4603. [8] C.J. Brinker, G.W. Scherer, Sol–Gel Science, Academic Press, New York, 1990. [9] B.L. Bischoff, M.A. Anderson, Chem. Mater. 7 (1995) 1772. [10] H.S. Jung, J. Kim, D.S. Lee, K.S. Hong, H.J. Youn, submitted for publication. [11] M. Grätzel, Nature 414 (2001) 338. [12] H.C. Zeng, Mater. Res. Soc. Symp. Proc. 346 (1994) 715. [13] K. Terabe, K. Kato, H. Miyazaki, S. Yamaguchi, A. Imai, Y. Iguchi, J. Mater. Sci. 29 (1994) 1617. [14] K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination Compounds, Wiley, New York, 1997, p. 48. [15] P.S. Ha, H.J. Youn, H.S. Jung, K.S. Hong, Y.H. Park, K.H. Ko, J. Colloid Interface Sci. 223 (2000) 16. [16] R.G. Avery, J.D.F. Ramsay, in: M. Che, G.C. Bond (Eds.), Adsorption and Catalysis on Oxides, in: Studies in Surface Science and Catalysis, vol. 21, Elsevier, Amsterdam, 1985, p. 149. [17] S. Kambe, S. Nakade, T. Kitamura, Y. Wada, S. Yanagida, J. Phys. Chem. B 106 (2002) 2967. [18] C.J. Barbé, F. Arendse, P. Comte, M. Jirousek, F. Lenzmann, V. Shklover, M. Grätzel, J. Am. Ceram. Soc. 80 (1997) 3157.