Humidity-Sensitive Electrical Conductivity of a Langmuir−Blodgett

10066

Langmuir 2006, 22, 10066-10071

Humidity-Sensitive Electrical Conductivity of a Langmuir-Blodgett Film of Titania Nanosheets: Surface Modification as Induced by Light Irradiation under Humid Conditions Kazuko Saruwatari,*,† Hisako Sato,†,‡ Toshihiro Kogure,†,‡ Takayuki Wakayama,§ Masanori Iitake,§ Kosho Akatsuka,| Masaaki Haga,| Takayoshi Sasaki,‡,⊥ and Akihiko Yamagishi†,‡,# Graduated School of Science, The UniVersity of Tokyo, 7-3-1 Hongo, Bunkyo, Tokyo 113-0033, Tokyo, Japan, Core Research for EVolutional Science and Technology (CREST), Japan Science and Technology Agency, 4-1-8 Honcho, Kawaguchi, Saitama, 332-0012, Japan, National Institute of AdVanced Industrial Science and Technology, Tsukuba Central 1, Tsukuba, Ibaraki, 305-8561, Japan, Chuo UniVersity, 1-13-27 Kasuga, Bunkyo-ku, Tokyo, 112-8551, Japan, and AdVanced Materials Laboratory, National Institute for Materials Science, 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan ReceiVed July 23, 2006. In Final Form: September 16, 2006 Electrical conductivity of titania nanosheets was investigated for a single-layered Langmuir-Blodgett (LB) film deposited onto a comb-type electrode (5 or 10 µm (electrode spacing) × 8 mm (electrode width)). The photoresponsive electrical properties of the film were investigated by irradiating with a Xe lamp under various atmospheric conditions. The atmosphere was controlled by introducing either oxygen or nitrogen gases containing different amounts of water vapor. As a result, the LB film behaved as an insulator with little photoresponse under dry atmospheric conditions. It became conductive on illuminating with a Xe lamp under a wet oxygen atmosphere. Conductivity increased with the increase of irradiation time (0-30 min) to attain a stationary value in 1 h. The highest conductive state thus attained lasted for several hours in the dark. The impedance of the film was measured over the frequency range of 1 MHz to 50 Hz by varying the relative humidity of an atmosphere from 0 to 100%. The results were analyzed by assuming an equivalent circuit consisting of one resistance (R) with constant Warburg component (W) and one capacitance (C) in parallel. The R component depended remarkably on the relative humidity, while the C component stayed nearly at the constant value. The dependence of R on water vapor (PH2O) was expressed by R ) A[PH2O]n with A ) constant and n ) -2.9. The results were rationalized in terms of the surface modification of titania nanosheets to hydrophilic nature under the illumination of UV light.

I. Introduction Titanium dioxide has been one of the most extensively studied semiconductive materials due to its divergent functions under UV irradiation such as photocatalysis,1 photovoltaics,2 and superhydrophilicity.3 A variety of TiO2 morphologies have been developed by several synthetic methods.4 Among them, lepidocrocite-type titania nanosheets have been explored by being exfoliated into individual colloidal sheets through soft-chemical treatments.5,6 Nanosheets with a large aspect ratio (1 nm thickness and sub-micrometer width) were furthermore utilized to form a layer-by-layer film, employing the self-assembly or the Langmuir-Blodgett (LB) procedures.7-10 The multilayer ultrathin * To whom correspondence should be addressed. E-mail: [email protected]. † The University of Tokyo. ‡ CREST, Japan Science and Technology Agency. § National Institute of Advanced Industrial Science and Technology. | Chuo University. ⊥ National Institute for Materials Science. # Present address: Ochanomizu University, Graduate School of Science, 2-1-1 Otsuka, Bunkyo-ku, Tokyo 112-8610, Japan. E-mail: [email protected]. (1) Fujishima, A.; Honda, K. Nature (London) 1972, 238, 37. (2) Oregan, B.; Gratzel, M. Nature (London) 1991, 353, 737. (3) Wang, R.; Hashimoto, K.; Fujishima, A.; Chikuni, M.; Kojima, E.; Kitamura, A.; Shimohigashi, M.; Watanabe, T. Nature 1997, 388, 431. (4) Carp, O.; Huisman, C. L.; Reller, A. Prog. Solid State Chem. 2004, 32, 33. (5) Sasaki, T.; Watanabe, M.; Hasizume, H.; Yamada, H.; Nakazawa, H. J. Am. Chem. Soc. 1996, 118, 8329. (6) Sasaki, T.; Watanabe, M. J. Am. Chem. Soc. 1998, 120, 4682. (7) Sasaki, T.; Ebina, Y.; Fukuda, K.; Tanaka, T.; Harada, M.; Watanabe, M. Chem. Mater. 2002, 14, 3524.

films thus prepared show interesting properties such as anisotropic UV absorption, photoinduced hydrophilicity,7 and electrochemical photoresponse.8 The first-principle calculation of such a titania nanosheet indicates the semiconductive band structure with 3.15 eV, suggesting a potential to develop electrical devices and sensors.11 It has been known that metal oxide semiconductors change their conductivity as well as work functions due to contacting with various gases since the pioneering work on zinc oxides as a gas detector.12 Titanium dioxides have been applied as sensors for several kinds of gases including water vapor.13-16 The humidity sensing mechanism of titanium dioxides is considered to be an ionic conduction enhanced by adsorbed water molecules.16 On increasing relative humidity, the adsorbed layer of water molecules changes from a monolayer to a multilayer. This results in the dependence of conductivity on relative humidity. (8) Sakai, N.; Ebina, Y.; Takada, K.; Sasaki, T. J. Am. Chem. Soc. 2004, 126, 5851. (9) Yamaki, T.; Asai, K. Langmuir 2001, 17, 2564. (10) Muramatsu, M.; Akatsuka, K.; Ebina, Y.; Wang, K.; Sasaki, T.; Ishida, T.; Miyake, K.; Haga, M. Langmuir 2005, 21, 6590. (11) Sato, H.; Ono, K.; Sasaki, T.; Yamagishi, A. J. Phys. Chem. B 2003, 107, 9824. (12) Seiyama, T.; Kato, A.; Fujiishi, K.; Nagatani, M. Anal. Chem. 1962, 34, 1502. (13) Tien, T. Y.; Stadler, H. L.; Gibson, E. F.; Zacmanidis, P. J. Am. Ceram. Soc. Bull. 1975, 54, 280. (14) Go¨pel, W.; Rocker, G.; Feierabend, R. Phys. ReV. B 1983, 28, 3427. (15) Schierbaum, K. D.; Kirner, U. K.; Geiger, J. F.; Go¨pel, W. Sens. Actuators B 1991, 4, 87. (16) Yeh, Y.-C.; Tseng, T.-Y.; Chang, D.-A. J. Am. Ceram. Soc. 1989, 72, 1472.

10.1021/la0621618 CCC: $33.50 © 2006 American Chemical Society Published on Web 10/28/2006

Electrical ConductiVity of Titania Nanosheets

Langmuir, Vol. 22, No. 24, 2006 10067

Based on this, the surface modification is thought to be an important factor for improving the sensitivity of humidity sensing. In this paper, we report the electrical properties of titania nanosheet films under the conditions of dry or wet (100% relative humidity) oxygen or nitrogen atmospheres. Attention was focused on the influence of UV light irradiation. The samples were prepared by the modified LB methods developed recently without amphiphilic templates.10 A LB film prepared with no amphiphile was thought to be desirable particularly as a conductive film because the amphiphile molecules could possibly act as a barrier against electrical conduction.17,18 The observed photoinduced electrical conductivity under humid conditions is discussed in relation to the surface hydrophilicity induced under UV irradiation. II. Experimental Section Preparation and Characterization of Films. H-type lepidocrocite titania (HxTi2-x/40x/4O4‚yH2O; x ∼ 0.7 with 0 ) vacancy) was synthesized via protonation of Cs-type titania (CsxTi2-x/40x/4O4) in an aqueous HCl solution as previously described.5 A suspension of titania nanosheets was prepared by exfoliating H-type titania with tetrabutylammonium hydroxide (TBAOH). The suspension was used as a subphase to form LB films without amphiphilic molecules as below.10 LB film preparations were carried out on a USI FSD-300 computer-controlled Langmuir trough with Teflon coating (effective area: 51.7 × 15 cm2). Surface pressure was measured with a Wilhelmy balance. The temperature of a trough was kept at 25 ( 0.5 °C by a thermostat. A subphase was prepared by diluting the stock nanosheet suspension to 0.008 g dm-3 with ultrapure water (g18 MΩ cm; Elgastat UHQ-III system, Kleiner, Switzerland). A substrate (a comb-type electrode) was dipped vertically in the subphase suspension. After 30 min, the surface was slowly compressed at a rate of 0.5 mm s-1 from zero surface pressure up to 15 mN m-1. On keeping of the surface pressure at 15 mN m-1 for 30 min, the dipped substrate was lifted up at the rate of 1 mm min-1. A singly deposited LB film thus prepared was characterized by atomic force microscopy (AFM; Nanoscope III scanning microscope, Digital Instruments) and scanning electron microscopy (SEM; S-4500 Hitachi, Japan). Electrodes and Conductivity Measurements. For a comb-type electrode, a bilayered metal electrode (100 nm Au on 5 nm Cr) was patterned lithographically on a thermally oxidized Si wafer covered with an oxide layer of ∼400 nm thickness. The spacing and total width of the electrode were 5 or 10 µm and 8 mm, respectively. The dc measurements were performed with an electrometer (Agilent Technology) by a two-probe method directly on the comb-type electrodes at room temperature under air condition. The ac impedance measurements were operated at the constant voltage of 1.0 V in the frequency range from 1 MHz down to 50 Hz with LC meter (Hioki 3532-50, Japan) on an electrode placed in a flow-through glass line under oxygen or nitrogen atmospheric conditions at different relative humidities (Figure 1). The present apparatus measured the impedance in the range of 0 < R < 107 Ω and 50 < f < 1 × 106 s-1. The above Si substrate was packaged with wire bonding as shown in Figure 1. For photoconductivity measurements, the film was irradiated with a 150 W Xe lamp (Hamamatsu Photonics, Japan) through a 10 cm quartz cell containing pure water. Relative humidity (0, 25, 50, 75, and 100% RH) was attained by mixing dry (0% RH) gas flow and water-saturated (100% RH) gas flow at appropriate proportions. The purity of oxygen and nitrogen gases was stated to be 99.9 and 99.99%, respectively. The analyses of the impedance data were performed using a software Z-View (Solarton). IR Measurements of a Cast Film. A film was prepared by casting the aqueous suspension of titania nanosheets onto a glass. The glass substrate was mounted in a glass cell with NaCl windows. The absorption FT-IR spectrum of the sample was measured between (17) Saruwatari, K.; Sato, H.; Kameda, J.; Yamagishi, A.; Domen, K. Chem. Commun. 2005, 15, 1999. (18) Saruwatari, K.; Sato, H.; Idei, T.; Kameda, J.; Yamagishi, A.; Takagaki, A.; Domen, K. J. Phys. Chem. B 2005, 109, 12410.

Figure 1. Schematic drawing of an experimental setup controlling the atmospheric conditions and the humidity.

Figure 2. (a) AFM image of a titania nanosheet LB film showing densely packing nanosheets and the I-V curve (c). (b) The AFM image of a titania nanosheet LB film with some cracks and the I-V curve (d). 3500 and 2500 cm-1 with a spectrometer FT-IR 460 (JASCO, Japan) under different relative humidities (0, 25, 50, 75, and 100% RH). As a baseline, the FT-IR spectrum of a bare glass plate was measured under dry (0% RH) conditions. No change was observed in the baseline spectrum when water vapor was introduced into a cell (100% RH). Before IR measurements, the sample was exposed under each humidity condition for 30 min.

III. Results and Discussion Characterization of a Titania Nanosheet LB Film. Figures 2a and 2b show the examples of the AFM images of a monolayer LB film together with the results of dc measurements. When particles were deposited so densely that they were in contact with each other at least partially (Figure 2a), a stable dc current was observed (Figure 2c). In contrast, when particles were deposited less densely with cracks (Figure 2b), no constant dc current was detected (Figure 2d). Figure 3 shows the SEM images of the same samples as in Figure 2. In the case of the sample in Figure 2a, the whole surface was mostly covered with nanosheets, leaving small cracks as indicated by the bright regions. When the same surface was imaged at higher magnification, the region between Au electrodes was uniformly covered with nanosheets (Figure 3c). In the case of the sample in Figure 3b, nanosheets were distributed over

10068 Langmuir, Vol. 22, No. 24, 2006

Saruwatari et al.

Figure 4. Impedance plots under dry oxygen conditions and UV irradiations.

Figure 3. SEM images of titania nanosheet LB films on comb-type electrodes. (a) and (c) are photos of the densely packed LB film whose AFM images are Figures 2a and 2c. (b) and (d) are photos of the cracked LB film whose AFM images are Figures 2b and 2d.

both Au electrodes and silicon substrate with many cracks. These were in accord with the AFM observations. Electrical Conductivity under a Dry Atmosphere and the Effect of UV Irradiation. Figure 4 shows the results when dry oxygen gas was introduced to the cell containing a LB film in the dark. The impedance plot was close to a linear curve, indicating that the film was an insulator in the dark. On irradiating the sample, the impedance plots gradually became a part of a larger

semicircle, corresponding to the appearance of conductivity. The change was regarded as positive photoresponse. The decay of photoinduced conductivity was so slow that the plot returned to a linear curve in days after irradiation was turned off. The switching on and off of a Xe lamp were repeated to confirm the reproducibility of the above change. In the second and third irradiation experiments, the effect of positive photoresponse was less eminent with the faster return to the initial plots in the dark as shown in Figure 4. No increase of conductivity was detected in the forth and further irradiation experiments. These results suggest that the irradiation of UV light caused the increase of a number of carriers in the initial few irradiations. The generation of the carriers was, however, suppressed after the repeated UV irradiation. It might be due to the irreversible structural change of the titania nanosheet surface. Electrical Conductivity under a Wet Atmosphere and the Effect of UV Irradiation. An oxygen gas saturated with water vapor (RH ) 100%) was introduced to the cell mounting another film in the dark. The impedance was gradually lowered until it was stable after several hours (solid circles in Figure 5a). More specifically, the impedance plots became semiarcs with two segments: a conductive higher frequency part (1 × 106 s-1 > f > 100 s-1) and a resistive lower frequency part (100 s-1 > f > 50 s-1). When the film was subsequently irradiated by a Xe lamp, the radius of the semiarc became smaller in 30 min as shown by the plot in Figure 5a. No further change was observed when the irradiation time increased from 30 to 60 min. The impedance plot with the smallest radius as obtained in this way lasted for several hours after the Xe lamp was turned off. For a longer period, the radius of the impedance plot increased and exceeded that of the initial plot until it was stable in days. These results indicated that the present titania film is conductive even in the dark when it was in contact with a water vapor. Furthermore, the conductivity increased on the irradiation of UV light and the better conductivity lasted in the dark for several hours. These facts suggested that the observed conductivity was not simple photoconductivity generating carriers reversibly but it might be accompanied by the photomodification of a surface. It should be noted here that such a light-induced modification took place exclusively under wet oxygen conditions. The same sample was irradiated again to see reproducibility after the conductivity was lowered to a stationary value in the dark. The radius of the impedance plot became smaller on increasing the irradiation time. The small radius of the impedance plots was maintained in the dark for several hours. Thereafter, the radius increased in days as seen in the first experiment. The impedance data were analyzed in terms of an equivalent electrical circuit as given in Figure 5b. The circuit consists of

Electrical ConductiVity of Titania Nanosheets

Langmuir, Vol. 22, No. 24, 2006 10069

Figure 6. R variation during and after UV irradiation under wet nitrogen conditions calculated from the equivalent circuit in Figure 5b.

Figure 5. (a) Impedance plots under wet oxygen conditions at the first irradiation experiment. (b) Equivalent electrical circuit most fitted into the experimental impedance plots and an example of the experimental result (solid line) and the fitting line (dashed one). (c) and (d) Variation of R and C components, which were estimated by the equivalent circuit (b), under wet oxygen conditions and UV irradiation.

one resistance (R) with constant Warburg component (W) and the capacitance (C) in parallel. Figure 5b also shows the typical results of reproducing the impedance plot in terms of the equivalent circuit. Figures 5c and 5d show the variations of R and C components in the dark and under the Xe irradiation, respectively. In contrast to the remarkable change of R during or after the light irradiation, C stayed constant under the investigated conditions. Since the resistance component decreased under light irradiation, the photoresponse under an oxygen atmosphere was regarded as positive photoresponse. Thus, it was concluded that the light-induced surface modification reversibly took place under the illumination of light and recovered to the initial state in the dark very slowly. The role of oxygen in the above effect was examined by replacing it with nitrogen. When the atmosphere of the sample cell at the lowest conductive state was changed from a wet oxygen gas (RH ) 100%) to a wet nitrogen gas (RH ) 100%), little change was observed in the impedance plot. No further change was observed when the film was left for more than 5 days in the dark. Thereafter, the effect of light was examined under wet nitrogen conditions. The R component was obtained by analyzing the impedance plots in terms of the same equivalent circuit as in Figure 5b (Figure 6). The R value increased during the irradiation of the initial 15 min and then decreased to the original value on continuing further irradiation. The initial increase of R on irradiation was regarded as negative photoresponse in contrast to positive photoresponse. The low value of R maintained for several hours after the irradiation stopped. R increased very slowly

in the dark until it reached a stationary value in 2 days. When the second irradiation experiment was performed, the same negative photoresponse was observed except for the shorter duration compared with the first negative photoresponse. When the third and forth irradiation experiments were performed on the same sample after reaching the stationary value, the negative photoresponsive behaviors were again observed but both lasted for much shorter duration in a similar manner. It was noted that the above negative photoresponse was particular to a wet nitrogen atmosphere. Moreover, the phenomenon was certainly reversible, suggesting that titania nanosheets underwent a photomodification in a way different from the oxygen atmosphere. The detailed change of a surface was not clear at this stage. Variation of ac Impedance under Different Relative Humidities. The above experiments show that a titania LB film becomes more conductive by light irradiation under a wet oxygen atmosphere. The appearance of such a state suggested a possibility of applying a titania nanosheet for humidity sensing. For pursuing the possibility, the ac impedance of a titania LB film was measured at various humidity conditions. Figure 7 shows the results when the relative humidity (RH) was varied from 0 to 100% in an oxygen atmosphere. The measurements were carried out in the dark after the film exhibited the highest conductance under UV irradiation. The atmosphere was replaced from 100% RH oxygen to oxygen with lower RH successively. The measurements under a new atmosphere were done after the stable impedance appeared (after 3-5 h). Over the investigated RH range, the cole-cole plot changed from the composite of two segments with a small semicircle at higher frequency and a part of a larger semicircle at lower frequency (RH ) 100 to 50%) to a part of a large circle (RH ) 25 and 0%). The decrease of RH resulted in the enlargement of the radius of the small semicircle in the higher frequency region. For RH < 50%, the plot showed two fragments no more, giving a part of one large semicircle. The R value was calculated on the basis of the equivalent circuit as shown in Figure 4b. Figure 7c shows the dependence of the calculated R value on RH. The results show that the R component varied over 2 orders for the change of RH from 0 to 100%. The C value, which was calculated on the same circuit, was constant around ∼4.3 × 10-11 F for the same change of RH. The dependence of electrical conductivity on humidity was expressed approximately by the following equation:

R ) C[PH2O]n with n ) -2.9

(1)

in which PH2O denotes vapor pressure in the atmosphere. Similar electrical properties were reported for the mesoporous TiO2 membrane19 and the porous TiO2 thick films under humidity conditions.20 (19) Vichi, F. M.; Tejedor-Tejedor, M. I.; Anderson, M. A. Chem. Mater. 2000, 12, 1762.

10070 Langmuir, Vol. 22, No. 24, 2006

Saruwatari et al.

Scheme 1. Schematic Figure Showing the Surface Modification of a Titania Nanosheet Surface under UV Irradiation at Wet Oxygen Conditions

FT-IR Spectra of a Cast Film of Titania Nanosheets. Figure 8 shows the FT-IR spectra of a cast film of titania nanosheets under various vapor pressures. All samples showed the peaks at 2975, 2923, and 2853 cm-1, which corresponded to the C-H stretching vibration in -CH3 and symmetric and asymmetric -CH2- stretching vibration bands, respectively. These peaks

Figure 7. (a) Impedance plots under different relative humidity conditions. (b) Enlarged impedance plot of (a). (c) Variation of R components toward different relative humidities. R values were calculated by the equivalent circuit in Figure 5b.

Figure 8. FT-IR spectra of titania cast film under different relative humidity conditions (RH ) 0, 25, 50, 75, and 100%).

originated from TBA+ ions that were electrostatically adsorbed on titania nanosheets. When the relative humidity was increased, the broad peak around 3400 cm-1 gradually increased. The peak was assigned to liquidlike water.21,22 Thus, the amount of water molecules adsorbed on titania nanosheets increased with the increase of relative humidity to form a liquid like water. The peak due to solid water, which is reported to be at 3280 cm-1, was not apparent probably because of the peak overlapped by the large band due to liquid water in the same wavenumber region. No evidence was obtained for the hydroxyl groups of titania nanosheets around 3700 cm-1. Possible Mechanisms of Electrical Conductivity. The present film exhibited little photoresponse under repeated UV irradiation even with the slight decrease of conductivity under the dry atmosphere, while it showed a remarkable positive photoresponse under humid aerobic conditions. Moreover, the attained conductive state lasted for several hours, even after turning off an irradiating light. These aspects suggested that the phenomena were not directly related to the photogeneration of electrons or holes in a bulk. Instead, the increase and decrease of impedance were most probably related to the ionic conductivity caused by the surface modification. Such conductivity might result from the increase of water adsorption as observed in the FT-IR measurements. The recent quartz crystal microbalance study reported the increase of water adsorption on a TiO2 film after UV irradiation and discussed the mechanism related to the photoinduced hydrophilicity on anatase surface.23 Two possible mechanisms have been argued since the discovery of photoinduced hydrophilicity: one is caused by water molecules physisorbed on the surface, which is cleaned by catalytic decomposition of contaminants,24,25 and the other is caused by chemisorbed water molecules resulting in the formation of surface hydroxyl groups.3,26,27 Although either case would accelerate the adsorption of water molecules, proton carriers could be produced only through the formation of surface hydroxyls.26,27 The conductivity thus appeared may last several hours until the mobile protons decayed accompanied by the recovery of the hydrophobic surface state as reported for anatase. These possible mechanisms of conductivity are shown by Scheme 1. The present finding on the humidity-sensitive conductivity of titania nanosheets might open a possibility of developing a novel type of humidity sensor based on layered oxides in combination with light irradiation. In fact, it has been confirmed that the conductivity changed critically as a function of water vapor (eq 1). (20) Faia, P. M.; Furtado, C. S.; Ferreira, A. J. Sens. Actuators B 2004, 101, 183. (21) Ewing, G. E. J. Phys. Chem. B 2004, 108, 15953. (22) Asay, D. B.; Kim, S. H. J. Phys. Chem. B 2005, 109, 16760. (23) Uosaki, K.; Yano, T.; Nihonyanagi, S. J. Phys. Chem. B 2004, 108, 19086. (24) White, J. M.; Szanyi, J.; Henderson, M. A. J. Phys. Chem. B 2003, 107, 9029. (25) Zabkov, T.; Stahl, D.; Thompson, T. L.; Panayotov, D.; Diwald, O.; Yates, J. T., Jr. J. Phys. Chem. B 2005, 109, 15454. (26) Sakai, N.; Fujishima, A.; Watanabe, T.; Hashimoto, K. J. Phys. Chem. B 2001, 105, 3023. (27) Sakai, N.; Fujishima, A.; Watanabe, T.; Hashimoto, K. J. Phys. Chem. B 2003, 107, 1028.

Electrical ConductiVity of Titania Nanosheets

Conclusion A single-layered titania nanosheet LB film showed humiditysensitive electrical conductivity. The photoresponse depends on the atmospheric conditions and the humidity contents. The irradiation under the different atmospheric conditions modified the titania nanosheet surface, leading to the conductivity change.

Langmuir, Vol. 22, No. 24, 2006 10071

Acknowledgment. This work was supported by a Grantin-Aid on Priority Areas (417) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japanese Government. LA0621618