NANO LETTERS
Enhanced Photocleavage of Water Using Titania Nanotube Arrays
2005 Vol. 5, No. 1 191-195
Gopal K. Mor, Karthik Shankar, Maggie Paulose, Oomman K. Varghese, and Craig A. Grimes* Department of Electrical Engineering, and Department of Materials Science and Engineering, The PennsylVania State UniVersity, UniVersity Park, PennsylVania 16802 Received October 15, 2004; Revised Manuscript Received November 16, 2004
ABSTRACT In this study highly ordered titania nanotube arrays of variable wall thickness are used to photocleave water under ultraviolet irradiation. We demonstrate that the wall thickness and length of the nanotubes can be controlled via anodization bath temperature. We find that the nanotube wall thickness is a key parameter influencing the magnitude of the photoanodic response and the overall efficiency of the water-splitting reaction. For 22 nm inner pore diameter nanotube arrays, those fabricated in a 5 °C anodization bath, 224 nm length and 34 nm wall thickness produced a photoanodic response that was thrice that of a nanotube array fabricated in a 50 °C anodization bath, 120 nm length and 9 nm wall-thickness. At high anodic polarization, above 1 V, the quantum efficiency under 337 nm illumination was greater than 90%. For the 5 °C anodization bath samples (22 nm pore-diameter, 34 nm wall thickness), upon 320−400 nm illumination at an intensity of 100 mW/cm2, hydrogen gas was generated at the power−time normalized rate of 960 µmol/h W (24 mL/h W) at an overall conversion efficiency of 6.8%. To the best of our knowledge, this hydrogen generation rate is the highest reported for a titania-based photoelectrochemical cell.
The principal impetus toward fabricating nanodimensional materials lies in the promise of achieving unique properties and superior performance due to their inherent nanoarchitectures. Titanium dioxide is a particularly versatile material with technological application as a photocatalyst, photovoltaic material, gas sensor, optical coating, structural ceramic, electrical circuit varistor, biocompatible material for bone implants, spacer material for magnetic spin valve systems, and more. Therefore, the ability to control the architecture of titania down to nanoscale dimensions can be expected to positively impact a variety of economically important technologies. As such, there have been significant efforts to develop nanotubular titania, including sol-gel and templatebased approaches.1-4 Our interest has been in the fabrication of ordered titania nanotube arrays by anodization of a starting titanium thick or thin film.5-7 Titania nanotubes fabricated by anodization are highly ordered, high-aspect ratio structures with nanocrystalline walls oriented perpendicular to the substrate. The nanotubes have a well-defined and controllable pore size, wall thickness, and tube length. Applications of this unique nanoarchitectured material are only beginning. Previously, we have reported titania nanotube based resistive gas sensors that exhibit an amazing 1,000,000,000% change in electrical resistance upon exposure to 1000 ppm of hydrogen gas at room temperature.8 We have also reported on the photocatalytic self-cleaning ability of titania nanotubebased gas sensors9 and their photochemical properties.10 * Corresponding author. E-mail:
[email protected] 10.1021/nl048301k CCC: $30.25 Published on Web 12/08/2004
© 2005 American Chemical Society
In this study we examine the use of titania nanotube arrays for the photocleavage of water into hydrogen and oxygen gas under ultraviolet irradiation with particular emphasis on the rate of hydrogen generation and on the influence of the architecture of the titania nanotubes on the photochemical properties. Efficient light-induced photocleavage of water into hydrogen and oxygen gas is recognized as a “Holy Grail” of chemistry,11 by which we might obtain a renewable, nonpolluting and readily portable energy source. Titania nanotubes were grown from a starting titanium sheet (0.25 mm thick, 99.7% purity purchased from Aldrich Corporation) by potentiostatic anodization 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 as-anodized nanotubes were amorphous; to induce crystallinity, they were subsequently annealed at 500 °C in an oxygen ambient for 6 h with heating and cooling rates of 1 °C/min. The morphologies of the titania nanotubes were studied using a JEOL JSM-6300 field emission scanning electron microscope (FESEM). The electrochemical measurements were performed in a standard three-electrode configuration with a Pt mesh counter electrode and a saturated Ag/AgCl reference electrode. The photocurrent at the TiO2 electrode, i.e., the working electrode, was measured using a scanning potentiostat (CH Instruments, model CHI 600B). The potential was swept linearly at a scan rate of 20 mV/sec. The illuminated area of the working electrode is 1 cm2. An
Table 1. Average Wall Thickness and Tube Length of 10 V Titania Nanotube Arrays Anodized at Different Bath Temperatures anodization temp
wall thickness (nm)
tube length (nm)
5 °C 25 °C 35 °C 50 °C
34 24 13.5 9
224 176 156 120
electrical contact was taken from the backside of Ti foil after removing the oxide layer by mechanical scribing. An electrically insulated copper wire was attached on the backside using conductive silver-epoxy adhesive. The backside of the sample was fixed to the glass substrate using a nonconductive epoxy, which also covered the edges of the foil. 1 M KOH purged with N2 was used as the electrolyte. A 50 W metal hydride lamp (EXFO Lite) was used as the UV light source. The light was passed through broadband optical filters, which allowed only wavelengths between 320 and 400 nm to be incident on the TiO2 photoanode at a measured intensity of 100 mW/cm2. Photocleavage of water was studied at a potential of -0.4 V vs Ag/AgCl. The gaseous photoproducts were identified and measured by a SRI 8610C gas chromatograph (GC) equipped with a thermal conductivity detector. The quantum efficiency at 337 nm wavelength was determined by using a 337 nm interference filter (Edmund Industrial Optics) that reduced the incident intensity to 2.7 mW/cm2. Titania nanotube arrays were anodized at 10 V at four different electrolyte bath temperatures: 5, 25, 35, and 50 °C. Figure 1 shows FESEM images of the morphology of titania nanotubes fabricated by anodization at 10 V at (a) 5 °C, (b) 25 °C, (c) 35 °C, and (d) 50 °C. Figure 2 shows an illustrative cross-sectional image of a 25 °C sample anodized at 10 V. The pore diameter is essentially the same (22 nm) for the 10 V anodized titania nanotube arrays fabricated at these different temperatures, whereas the wall thickness changes by a approximately a factor of 4 and the tube length changes by approximately a factor of 2. The wall thickness increases with decreasing anodization temperature from 9 nm at 50 °C to 34 nm at 5 °C. As the wall thickness increases with decreasing anodization temperature, the voids in the interpore areas fill; with the tubes becoming more interconnected, the discrete tube-like structure approaches a nanoporous structure in appearance. The length of the nanotubes (corresponding to the thickness of the nanotube layer) increases with decreasing anodization bath temperature from 120 nm at 50 °C to 224 nm at 5 °C. Table 1 shows the variation in 10 V nanotube array wall thickness and tube length as a function of anodization temperature. Figure 3displays FESEM images of titania nanotubes fabricated by anodization at 20 V at (a) 5 °C and (b) 25 °C, with resulting inner pore diameters of 76 nm. Here, the nanotube wall thickness is clearly seen to increase from 17 nm at 25 °C to 27 nm at 5 °C, confirming the trend of increasing nanotube wall thickness as a consequence of lower anodization temperature. 192
Figure 1. FESEM images, top view, of 10 V titania nanotube arrays anodized at (A) 5 °C, (B) 25 °C, (C) 35 °C, and (D) 50 °C. The pore size is nearly 22 nm for all samples. In (A) the average wall thickness is 34 nm, and in (D) is 9 nm.
Figure 2. Cross-sectional FESEM image of 10 V titania nanotube arrays anodized at 25 °C.
These changes can be explained within the context of the formation mechanism of titania nanotubes.6,12 The chemical dissolution of titania in the HF electrolyte plays a key role in nanotube formation, and is the factor ultimately limiting nanotube length. The formation of interpore voids is caused by oxidation of the metal in the interpore region followed Nano Lett., Vol. 5, No. 1, 2005
Figure 5. Squared photocurrent vs potential for 10 V samples as a function of bath temperature.
Figure 3. FESEM images, top view, of 20 V titania nanotube arrays anodized at (A) 5 °C, (B) 25 °C.
Figure 6. Photocurrent density vs applied potential (potential with respect to open circuit potential) for 10 V 5 °C sample illuminated by 337 nm radiation at 2.7 mW/cm2. Figure 4. Variation of photocurrent density versus measured potential [vs. Ag/AgCl] for 10 V samples anodized at four anodization bath temperatures, i.e., 5 °C, 25 °C, 35 °C, and 50 °C.
by field-assisted dissolution. Wet chemical etching is a temperature dependent process, with etch rates typically being exponential functions of the temperature. The solubility product of ions in any given electrolyte is also a strong function of temperature. It is therefore reasonable to assume that etching by fluoride ions as well as dissolution of the oxide will be weaker at lower temperatures, thus accounting for the larger tube length and wall thickness. The pore diameter is essentially constant as it is largely determined by the (fixed) anodization voltage. Figure 4 shows the I-V characteristics of the titania nanotube array electrodes, photocurrent density versus potential, measured in 1 M KOH electrolyte as a function of anodization bath temperature. At 1.5 V the photocurrent density of the 5 °C anodized sample is more than three times Nano Lett., Vol. 5, No. 1, 2005
the value for the sample anodized at 50 °C. The lower anodization temperature also increases the slope of the photocurrent-potential characteristic. For crystalline semiconductors, under certain limitations the squared photocurrent is a linear function of the applied bias.13 In Figure 5, squared photocurrent is plotted against electrode potential; the traces closely approximate a straight line at low values of applied bias indicative that photogenerated charges are being separated by the electric field of the depletion layer.14 At higher anodic polarization, the squared photocurrent-potential plot deviates from linearity for all the samples due to band bending.14 As can be seen in Figure 5, the sample anodized at 5 °C adheres to linearity over the largest bias range. Figure 6 shows the photoresponse of a 10 V/5 °C anodized sample to monochromatic 337 nm 2.7 mW/cm2 illumination. At high anodic polarization, greater than 1 V, the ratio of the photogenerated charges to incident photons, i.e., the quantum efficiency, is larger than 90%. 193
The titania nanotube array architecture results in a large effective surface area in close proximity with the electrolyte, thus enabling diffusive transport of photogenerated holes to oxidizable species in the electrolyte. Separation of photogenerated charges is assisted by action of the depletion region electric field.15,16 Minority carriers generated within a “retrieval” length from the material surface, that is a distance from the surface equal to the sum of the depletion layer width and the diffusion length, escape recombination and reach the electrolyte.17 The relevant structure sizes of the titania nanotube arrays, i.e., half the wall thickness, are all smaller than 20 nm, which is less than the retrieval length of crystalline titania,18 hence bulk recombination is greatly reduced and the quantum yield enhanced.18 J. van de Lagemaat and co-workers observed a substantial enhancement of the quantum yield in SiC made nanoporous by anodic etching in HF solution.17 Due to light scattering within a porous structure, incident photons are more effectively absorbed than on a flat electrode.19 However, while bulk recombination is reduced by the nanotube architecture, photogenerated minority carriers can be trapped by surface states. According to the model proposed by Lubberhuizen at al. to explain recombination of photogenerated charge carriers in nanoporous GaP, photogenerated charge carriers are instantaneously and effectively separated by the capture of holes in surface bonds.18 Hence recombination between separated electrons and holes is sensitive to the band-bending potential that the electrons have to overcome to reach the surface where (unwanted) recombination takes place. Therefore, the greater the band-bending potential, the smaller the charge carrier recombination. It is well-known from the study of particulate nanocrystalline photoelectrodes that a depletion layer cannot be formed in nanoparticles a few nanometers in size.20 In general, within the nanometer regime, as the size of the particle or grain decreases, so does its ability to sustain a significant amount of band bending. As a consequence, we can anticipate that the titania nanotube array with 34 nm wall thickness has greater band bending than a similar array with 9 nm walls. The enhanced band bending of the thicker walls decreases surface recombination rates, thus increasing the photocurrent. We make note that while the nanotubes fabricated at 5 °C are longer than those at 50 °C, from our previous studies we do not believe that the variation in length (120 nm to 224 nm) is enough to significantly influence the photoresponse of the nanotube array electrodes.10 The photoconversion efficiency η of light energy to chemical energy in the presence of an external applied potential is calculated:21 η (%) )
(total power output - electrical power output) × light power input 100
) jp
(E0rev - |Eapp|) × 100 I0
The photocurrent density jp is in mA/cm2, the total power 194
Figure 7. Photoconversion efficiency as a function of measured potential [vs. Ag/AgCl] for 10 V samples anodized at four temperatures [i.e., 5 °C, 25 °C, 35 °C, and 50 °C].
Figure 8. Hydrogen photoproduction in a photoelectrochemical cell consisting of a 5 °C titania nanotube array photoanode and Pt mesh cathode.
output is jp(E0rev), and the electrical power input is jp|Eapp|. The term E0rev is the standard reversible potential which is 1.23 V NHE-1. Emeas is the electrode potential (vs Ag/AgCl) of the working electrode at which the photocurrent was measured under illumination, and Eaoc is the electrode potential (vs Ag/AgCl) of the same working electrode at open circuit condition under same illumination and in the same electrolyte. Eapp ) Emeas - Eaoc, and I0 is the intensity of incident light in mW/cm2. Photoconversion efficiency as a function of potential for the different photoanodes is shown in Figure 7. A maximum conversion efficiency of 6.8% is obtained for nanotubes anodized at 5 °C. Figure 8 shows a quantitative description of the hydrogen evolution as a function of time for a photoelectrochemical cell with Pt mesh cathode and photoanode consisting of a 5 Nano Lett., Vol. 5, No. 1, 2005
°C anodized sample. Gas chromatographic analysis was used to verify that the volume ratio of the evolved hydrogen and oxygen was 2:1, which confirmed water splitting. During 1800 s of exposure to 320-400 nm 100 mW/cm2 illumination, 48 µmol of hydrogen gas was generated. Normalizing this rate to time and incident power, we find a hydrogen generation rate of 960 µmol/h W, or 24 mL/hW. To the best of our knowledge, this hydrogen generation rate is the highest for a titania based photocatalytic or photoelectrochemical cell. The photocurrent studies in this paper indicate that the architecture of the titania nanotubes, in this case wall thickness as controlled by anodization bath temperature, has a significant effect on the efficiency of water photocleavage achievable with these materials. The nanotubular architecture allows for more efficient absorption of incident photons as well as decreased bulk recombination, and the resulting quantum yield is over 90% at high anodic polarization. The nanotubes fabricated at lower anodization temperatures had greater wall thickness, which resulted in larger photocurrents. It is suggested that the architecture of nanotubes fabricated at lower anodization temperatures decreases the surface recombination. On a per watt basis, the hydrogen generation rate obtained for 24 mL/h W is among the highest reported for a titania-based photoelectrochemical cell, with a maximum photoconversion efficiency of 6.8%. References (1) Zhu, Y. C.; Li, H. L.; Koltypin, Y.; Hacohen, Y. R.; Gedanken, A. Chem. Commun. 2001, 2616-2617. (2) Peng, T. Y.; Hasegawa, A.; Qiu, J. R.; Hirao, K. Chem. Mater. 2003, 15, 2011-2016.
Nano Lett., Vol. 5, No. 1, 2005
(3) Zhang, M.; Bando, Y.; Wada, K. J. Mater. Sci. Lett. 2001, 20, 167170. (4) Wang, W. Z.; Varghese, O. K.; Paulose, M.; Grimes, C. A.; Wang, Q. L.; Dickey, E. C. J. Mater. Res. 2004, 19, 417-422. (5) Varghese, O. K.; Gong, D. W.; Paulose, M.; Grimes, C. A.; Dickey, E. C. J. Mater. Res. 2003, 18, 156-165. (6) Gong, D.; Grimes, C. A.; Varghese, O. K.; Hu, W. C.; Singh, R. S.; Chen, Z.; Dickey, E. C. J. Mater. Res. 2001, 16, 3331-3334. (7) Varghese, O. K.; Gong, D. W.; Paulose, M.; Ong, K. G.; Dickey, E. C.; Grimes, C. A. AdV. Mater. 2003, 15, 624-627. (8) Varghese, O. K.; Gong, D. W.; Paulose, M.; Ong, K. G.; Grimes, C. A. Sens. Actuators, B 2003, 93, 338-344. (9) Mor, G. K.; Carvalho, M. A.; Varghese, O. K.; Pishko, M. V.; Grimes, C. A. J. Mater. Res. 2004, 19, 628-634. (10) Mor, G. K.; Shankar, K.; Varghese, O. K.; Grimes, C. A. J. Mater. Res. 2004, 19, 2989-2996. (11) Foll, H.; Langa, S.; Carstensen, J.; Christophersen, M.; Tiginyanu, I. M. AdV. Mater. 2003, 15, 183. (12) Mor, G. K.; Varghese, O. K.; Paulose, M.; Mukherjee, N.; Grimes, C. A. J. Mater. Res. 2003, 18, 2588-2593. (13) Oliva, F. Y.; Avalle, L. B.; Santos, E.; Camara, O. R. J. Photochem. Photobiol., A 2002, 146, 175-188. (14) Pleskov, Y. V.; Krotova, M. D. Electrochim. Acta 1993, 38, 107109. (15) Sukamto, J. P. H.; Mcmillan, C. S.; Smyrl, W. Electrochim. Acta 1993, 38, 15-27. (16) Sukamto, J. P. H.; Smyrl, W. H.; Mcmillan, C. S.; Kozlowski, M. R. J. Electrochem. Soc. 1992, 139, 1033-1043. (17) Van de Lagemaat, J.; Plakman, M.; Vanmaekelbergh, D.; Kelly, J. J. Appl. Phys. Lett. 1996, 69, 2246-2248. (18) Lubberhuizen, W. H.; Vanmaekelbergh, D.; Van Faassen, E. J. Porous Mater. 2000, 7, 147-152. (19) Marin, F. I.; Hamstra, M. A.; Vanmaekelbergh, D. J. Electrochem. Soc. 1996, 143, 1137-1142. (20) Gratzel, M. Nature 2001, 414, 338-344. (21) Khan, S. U. M.; Al-Shahry, M.; Ingler, W. B. Science 2002, 297, 2243-2245.
NL048301K
195
