Hydrothermal synthesis and characterization of TiO 2 nanorod arrays on glass substrates Yuxiang Li a , Min Guo a , Mei Zhang a , Xidong Wang b, * a Department of Physical Chemistry, University of Science and Technology Beijing, Beijing 100083, PR China b College of Engineering, Peking University, No. 5 of Yiheyuan Street, Haidian District, Beijing 100871, PR China 1. Introduction TiO 2 -based one-dimensional structures, especially nanorod/ wire/tube arrays, have attracted much attention owing to their excellent properties and important applications. Compared with TiO 2 nanoparticles, TiO 2 nanorod/wire/tube arrays have lower recombination rate for excited electron–hole pair, unique optical and electric properties. Therefore, they can be widely applied in many fields, such as photocatalyst, photovoltaic devices, solar energy batteries and gas sensors [1–4]. In recent years, a number of approaches have been reported to fabricate TiO 2 nanorod/wire/tube arrays, including template- assisted method, electrochemical anodic oxidation method, chemical vapor deposition (CVD) and hydrothermal method [5– 8]. Among these methods, template-assisted technique combined with sol–gel or electrochemical deposit process is widely used. Lei et al. [9] reported highly ordered TiO 2 nanowire arrays which were prepared in anodic alumina membranes by a sol–gel method. Chu et al. [10] synthesized ordered TiO 2 –Ru and TiO 2 –RuO 2 nanorod arrays in porous alumina films through cathodic electrodeposition. However, when the AAO template is removed, the nanorod arrays tend to collapse due to the huge surface tension between each nanorod [11–13]. Electrochemical anodic oxidation of titanium can fabricate vertically oriented TiO 2 nanotube arrays. Paulose et al. [6] fabricated self-aligned TiO 2 nanotube arrays by potentiostatic anodization of Ti foil which has 134 m m in length, but it needs high-purity polished titanium foil, which results in a high production cost [14]. Moreover, the as-fabricated nanotubes are amorphous and need to be annealed, which results in forming polycrystalline TiO 2 structures and influencing their photoelectric properties. Wu and Yu [15] synthesized well-aligned rutile and anatase TiO 2 nanorods using CVD method. However, this method involves the use of metal catalyst, high-temperature and vacuum technique, which makes the preparation process complicated. Compared to the above methods, hydrothermal synthesis of TiO 2 nanorod arrays (TNAs) is a promising approach due to its simple process, fast reaction velocity and low cost. Therefore, it is fit for large-scale preparation of TNAs. Up to now, preparing TNAs by hydrothermal approach is rarely reported. Recently, Feng et al. [16] prepared TiO 2 nanorod films by a low-temperature hydrothermal approach on a glass wafer substrate. The as-prepared TiO 2 nanorod film looked like a pile of radial papillae. Such morphology limits its applications in many fields. For practical applications, especially used for solar cells and photocatalyst, it is better, for TNAs, to have a well-aligned orientation and uniform density distribution. Therefore, exploring a new route for fabricating well-ordered TNAs is very essential. As is well known, surface features of a substrate have strong impact on the morphology of one-dimensional material grown on the substrate. Therefore, pre-treatment of the substrate is an Materials Research Bulletin 44 (2009) 1232–1237 ARTICLE INFO Article history: Received 3 July 2008 Received in revised form 21 November 2008 Accepted 16 January 2009 Available online 22 January 2009 Keywords: A. Nanostructures B. Crystal growth D. Catalytic properties ABSTRACT Large-scale, well-aligned single crystalline TiO 2 nanorod arrays were prepared on the pre-treated glass substrate by a hydrother mal approach. The as-prepare d TiO 2 nanorod arrays were characterized by X- ray diffraction, scanning electron microscopy and transmission electron microscopy. X-ray diffraction results show that the main phase of TiO 2 is rutile. Scanning electron microscopy and transmission electron microscopy results demonstrate that the large-scale TiO 2 nanorod arrays grown on the pre- treated glass substrate are well-aligned single crystal and grow along [0 0 1] direction. The average diameter and length of the nanorods are approximately 21 and 400 nm, respectively. The photocatalytic activity of TiO 2 nanorod arrays was investigated by measuring the photodegradation rate of methyl blue aqueous solution under UV irradiation (254 nm). And the results indicate that TiO 2 nanorod arrays exhibit relatively higher photocatalytic activity. ß 2009 Elsevier Ltd. All rights reserved. * Corresponding author. Tel.: +86 10 8252 9083; fax: +86 10 6275 7532. E-mail address: xidong@coe.pku.edu.cn (X. Wang). Contents lists available at ScienceDirect Materials Research Bulletin journal homepage: www.elsevier.com/locate/matresbu 0025-5408/$ – see front matter ß 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.materresbull.2009.01.009 effective means to control the growth of TNAs. Guo [17] and Wang and co-workers [18] prepared well-aligned ZnO nanorod arrays by pre-treating the ITO glass substrate, and the results indicated that pre-treatment of the substrate played an important role for the growth morphology of ZnO nanorod arrays. In this study, the substrate was pre-treated by coating TiO 2 nanoparticles followed by heat treatment, and then the pre-treated substrate (PT-substrate) was put into hydrothermal solution to fabricate TNAs. The growth mechanism of TNAs and influence of PT-substrate on the morphology of the prepared TNAs have been discussed. Furthermore, the photocatalytic activity of the prepared TNAs was also investigated. 2. Experimental 2.1. Materials All chemicals were of analytical reagent grade and used without further purification. And all the aqueous solutions were prepared using de-ionized water. The microscope slides (1–1.2 mm thick), used as substrates, were firstly ultrasonic cleaned by acetone, ethanol and de-ionized water, respectively prior to use. 2.2. Preparation of the colloid solution for coating substrate Firstly, 17 ml tetrabutyl titanate was dissolved in the mixed solution of 12 ml ethanol and 5 ml diethanolamine, and then stirred for 1 h to obtain a precursor solution. Secondly, the mixed solution of 1.7 ml de-ionized water, 34 ml ethanol and 0.13 ml hydrochloric acid (37% HCl) were added to the above precursor solution and further violently stirred for 15 min. Finally, the as- prepared mixed solution was aged for 24 h at room temperature in order to form a homogeneous and stable colloid solution, which served as the coating solution. 2.3. Pre-treatment of the substrate The coating colloid solution was coated onto the cleaned glass substrate using spin coater (KW-4A, made by the Chinese Academy of Sciences) at the rate of 3000 rpm for 30 s. Subsequently, the substrate was dried in air and annealed at 700 8C for 30 min in muffle furnace. The process was repeated for two times to produce TiO 2 nanocrystal seeds layer with appropriate thickness, and thus the pre-treated substrate (PT-substrate) were prepared. 2.4. Hydrothermal growth of TNAs The hydrothermal precursor solution was prepared by mixing 0.05 M titanium trichloride (TiCl 3 ) aqueous solution saturated with sodium chloride. After the precursor solution was stirred for 5 min, it was transferred into a polytetrafluoroethylene vessel which was placed in a sealed kettle, and then PT-substrate was immersed into the precursor solution, and subsequently, the hydrothermal growth of TNAs was carried out at 160 8C for 3 h. 2.5. Characterization The morphology and size distribution of the as-prepared TNAs were characterized by the field-emission scanning electron microscopy (FE-SEM, ZEISS SUPRA TM 55 operated at 10 keV). Transmission electron microscope (TEM, PHILIPS-FEI TECNAI20) and high-resolution TEM (HRTEM, JEM-2010) were used to further elucidate the microstructure and phase characterization. X-ray diffraction (XRD) analysis was performed with a Rigaku D MAX -RB diffractometer using Cu K a radiation. 2.6. Photocatalytic activity measurement The photocatalytic activity of TNAs was investigated by measuring the photodegradation rate of methyl blue aqueous solution. Photocatalytic reaction was carried out in a 200 ml quartz glass vessel under which there was an 18 W UV lamp (254 nm wavelength). The 0.1 g/l methyl blue a queous solution (150 ml) and the as-prepared TNAs were placed into the reaction vessel, keeping the face of TNAs upward. At the given intervals of UV illumination, the specimen of methyl blue solution was collected and analyzed using an UV spectro- photometer (UNIC, UV-2100, absorption at l max =600nm for methyl blue). 3. Results and discussion 3.1. Morphology and crystal structure of TNAs Modification of the substrate has strong impact on the morphology and alignment ordering of TNAs. Fig. 1 shows the FE-SEM photographs of TNAs grown on the unmodified glass substrate. It can be seen from Fig. 1a that the unmodified substrate is clean and nothing is observed on its surface, and TiO 2 nanorods grow radially on the unmodified substrate and form many papillae in a random pattern, as illustrated in Fig. 1b and c. The FE-SEM photographs of TNAs grown on the pre-coated substrate without heat treatment are shown in Fig. 2. It can be seen that the pre- coated substrate (Fig. 2a) is different from the unmodified substrate (Fig. 1a), which covered with a layer of film composed of fine particles. Top view SEM images (Fig. 2b and c) show that TNAs grow more vertically on the substrate and have a better orientation after the substrate is coated with TiO 2 nanoparticles. This phenomenon can be explained that the glass substrate is amorphous and the distribution of nucleation sites is not uniform. After TiO 2 colloid solution is coated onto the glass substrate, the surface is covered with a layer of compact TiO 2 nanoparticles. Although TiO 2 nanoparticles are amorphous, they can provide uniform heterogeneous nucleation sites effectively. Therefore, the as-prepared TiO 2 nanorods have relatively better growth orienta- Fig. 1. FE-SEM images of TNAs grown on the unmodified glass substrates at a hydrothermal growth temperature of 160 8C and growth time of 3 h. (a) Unmodified glass substrate; (b and c) top view at different magnifications of TNAs. Y. Li et al. / Materials Research Bulletin 44 (2009) 1232–1237 1233 tion and density distribution compared with those grown on the unmodified substrate. Fig. 3 shows XRD patterns of the different substrates, which identify the phase structure of the substrates. From Fig. 3, it can be seen that a wide peak in the angle range of 2 u = 15–358 appears in all XRD patterns, which is due to the influence of amorphous phase of the glass substrate. Fig. 3a is XRD patterns of the unmodified glass substrate, and no other peaks appear. Fig. 3b shows XRD patterns of the substrate with pre-coated TiO 2 nanoparticles but not annealed. It can be seen that there is still no other peaks of crystalline TiO 2 , indicating that TiO 2 nanoparticles on the substrate are amorphous. When the substrate is coated with TiO 2 nanoparticles followed by heat treatment at 700 8C, it can be seen from Fig. 3c, although existing the noise signal from the glass substrate, the diffraction peaks of TiO 2 appear and can be indexed as the rutile phase (JCPDS No. 04-0551) and the anatase phase (JCPDS No. 01-0562) of TiO 2 , suggesting that most of TiO 2 nanoparticles transform from amorphous to rutile phase, but partial TiO 2 nanoparticles are anatase phase. When the substrate is coated with TiO 2 nanoparticles followed by heat treatment at 700 8C, it is clearly observed from Fig. 5a that there is a mass of uniform particles with rutile crystalline phase (Fig. 3c) on the substrate. Fig. 4 shows XRD patterns of standard rutile phase (JCPDS No. 04-0551), standard anatase phase (JCPDS No. 01-0562) of TiO 2 and TNAs grown on the different substrates. Fig. 4a shows XRD patterns of TNAs grown on the unmodified glass substrate, which shows that the main phase is rutile phase of TiO 2 . It can be seen from Fig. 4b that the intensity of 1 1 0 peak of rutile phase increases when TNAs grew on the substrate with pre-coated TiO 2 nanoparticles but not annealed. As shown in Fig. 4c, after the pre-coated substrate was annealed, the diffraction intensity of (1 1 0) planes increases even significantly. A number of works [19– 22] have reported that the prepared TiO 2 rods via hydrothermal approach in strongly acidic solution are usually pure rutile phase. This is in consistent with the results of XRD in this work. However, a small part of anatase phase of the as-prepared TNAs was detected by XRD, which might be due to the coated TiO 2 nanocrystal particles on the glass substrate. From Fig. 4, we also see that the diffraction intensity ratio of (1 1 0) to (1 0 1) becomes larger than that of the standard rutile phase of TiO 2 , indicating that the as- prepared TNAs prefer to grow along the (1 1 0) planes. As is shown in Fig. 5a, after annealed at 700 8C, there is a mass of uniform particles on the substrate serving as the seeds on which TiO 2 nanorods grow. Therefore, during hydrothermal reaction process, TiO 2 crystalline nucleus is apt to form on the coated TiO 2 particles and grow along the (1 1 0) planes. As a result, TNAs can grow on PT-substrate with well orientation and uniform density distribution, as can be seen from Fig. 5b. Fig. 5c is the side view SEM image of TNAs grown on PT- substrate. It can be seen that there is a layer of TiO 2 nanocrystal particles film of about 200 nm in thickness (marked by d1) between TiO 2 nanorod arrays and the glass substrate (marked by d2). The length of TiO 2 nanorods ranges from 300 to 500 nm. From the preliminary experiment, we found the following phenomena. When the substrate is coated for one time, the thickness of the seed layer is thinner (about 90 nm). However, the TiO 2 crystal seeds are not uniformly distributed on the substrate. As a result, the orientation of TiO 2 nanorod arrays grown on the substrate becomes poor. In addition, the diameter distribution of TiO 2 nanorods grown on PT-substrate is illustrated in Fig. 6. Statistics show that 70% of Fig. 2. FE-SEM images of TNAs grown on the TiO 2 -coated substrate without heat treatment at a hydrothermal growth temperature of 160 8C and growth time of 3 h. (a) PT- substrate without heat treatment; (b and c) top view at different magnifications of TNAs. Fig. 3. XRD patterns of the different substrates. (a) Unmodified glass substrate; (b) PT-substrate without heat treatment; (c) PT-substrate at an annealing temperature of 700 8C. Fig. 4. XRD patterns of standard rutile phase (JCPDS No. 04-0551), standard anatase phase (JCPDS No. 01-0562) of TiO 2 and TNAs grown on: (a) unmodified glass substrate; (b) PT-substrate without heat treatment; (c) PT-substrate at an annealing temperature of 700 8C. Y. Li et al. / Materials Research Bulletin 44 (2009) 1232–1237 1234 the nanorods have a diameter between 15 and 25 nm and the average diameter of TiO 2 nanorods is about 21 nm. Compared with the TNAs reported previously [11,16], the as-prepared TNAs in this work have been improved in the growth orientation, the density distribution and the aspect ratio. It can be seen from XRD patterns and SEM images, that the crystallinity of the prepared nanorods is low. In order to improve the quality of TiO 2 nanorods, TNAs were prepared under hydrothermal temperature of 190 8C. The SEM image (Fig. 7) shows obvious tetragonal structure of rutile, which exhibits a better crystallinity of TiO 2 nanorod. However, it can be seen from Fig. 7 that the growth orientation becomes poor, and the diameter distribution of TiO 2 nanorods becomes wide. The microstructure of TiO 2 nanorods was further analyzed using TEM and HRTEM. Two rods with a diameter of about 4 nm were selected for study. Fig. 8a shows the TEM image and the associated selected area electron diffraction (SAED) pattern of TiO 2 nanorods. The SAED inserted in the upper left corner of Fig. 8a demonstrates that TiO 2 nanorod is a single crystal. High-resolution TEM image of one single TiO 2 nanorod is shown in Fig. 8b. The lattice fringes indicate that TiO 2 nanorod is well crystallized. The spacing between two fringes is 0.325 nm, corresponding to the interplanar distance of the (1 1 0) index planes of the rutile TiO 2 , indicating that the as-prepared TiO 2 nanorods via hydrothermal approach are the rutile structure. Furthermore, it is observed that the (1 1 0) crystal planes are perpendicular to the c-axis, implying Fig. 5. FE-SEM images of TNAs grown on PT-substrate at an annealing temperature of 700 8C and at a hydrothermal growth temperature of 160 8C and growth time of 3 h. (a) PT-substrate; (b) top view of TNAs; (c) side view of TNAs. Fig. 6. Diameter distribution of TiO 2 nanorods grown on PT-substrate at an annealing temperature of 700 8C and at a hydrothermal growth temperature of 160 8C and with growth time of 3 h. Fig. 7. SEM image of TiO 2 nanorod arrays grown on the substrate at an annealing temperature of 700 8C at a hydrothermal growth temperature of 190 8C and growth time of 3 h. Y. Li et al. / Materials Research Bulletin 44 (2009) 1232–1237 1235 that TiO 2 nanorods grow along the (1 1 0) crystal planes, and this is consistent with the results of XRD. Considering the effect of annealing temperature, the pre-coated substrates were annealed at 450 and 600 8C, respectively. However, after hydrothermal reaction, TiO 2 nanorods have not grown on the substrate with annealed at 450 8C, TiO 2 nanorods begin to grow but the morphology is very poor when the annealing temperature increases to 600 8C. Therefore, 700 8C was chosen as the annealing temperature of the PT-substrates in the present experiment. 3.2. Growth mechanism of TNAs Growth of TiO 2 nanorods on the substrate is strongly related to rutile’s inherent growth habit and hydrothermal preparation conditions. The structure of rutile consists of chains of TiO 6 octahedra along the c-axis [23]. The TiO 6 octahedra in the chain connect through one shared edge, while the TiO 6 octahedra between two chains connect through points, which implies that the chemical band in the chain is stronger than that between the chains. According to PBC theory [24–26], periodic bond chain constructs the crystal, and the direction of the strongest chemical bond is usually the direction in which the crystal has the fastest growth velocity. Therefore, for rutile TiO 2 crystal, the growth velocity along the [0 0 1] direction is faster than that of the [1 1 0] direction. However, as for the structure of anatase TiO 2 , all the TiO 6 octahedra connect through shared edges, which results in the same growth velocity in all directions. It is the reason that the as- prepared TiO 2 nanorods are rutile phase instead of anatase phase. The formation of rutile crystal nucleus in TiCl 3 strongly acidic precursor solution can be described by the following process [27]. First, single polymer [TiO(OH 2 ) 5 ] 2+ forms by the following reactive equations: Ti 3þ þ 6H 2 O !½TiðOH 2 Þ 6  3þ (1) ½TiðOH 2 Þ 6  3þ !½TiðOHÞðOH 2 Þ 5  2þ þ H þ !½TiOðOH 2 Þ 5  þ þ 2H þ (2) 4½TiOðOH 2 Þ 5  þ þ O 2 þ 4H þ ! 4½TiOðOH 2 Þ 5  2þ þ 2H 2 O (3) Then the single polymers [TiO(OH 2 ) 5 ] 2+ combine through dehydrating each other to form straight chain polymer by the edge connection. Finally, the straight chain polymers connect through points to form rutile crystal nucleus. As is well known, the heterogeneous nucleation of crystalline phase in solution is easier than homogeneous nucleation [28]. Therefore, TiO 2 nanocrystal particles coated onto the glass substrate followed by heat treatment can be served as the seeds of heterogeneous nucleation. Moreover, the structure of crystalline seeds matches that of the prepared single crystal nanorods, which make well-aligned nanorod arrays be prepared. In addition, Cl À plays a significant role when TiO 2 growth units deposit on TiO 2 crystal seeds, and it can promote TiO 2 crystal to grow into rods instead of particles. As is shown in Fig. 9, when NaCl is not added to the hydrothermal precursor solution, TiO 2 arrays grown on PT-substrate almost do not grow into rods. It can be explained that the (1 1 0) plane of rutile is positive polar-face and Cl À will be adsorbed preferentially on its surface, which prevents the contact of the TiO 2 growth units on the (1 1 0) surface and thus greatly limits the crystal growth along the (0 0 1) planes. There- fore, Cl À may be the key factor for TiO 2 crystal growing along [0 0 1] direction to form rod-like structure. 3.3. Photocatalytic property of TNAs Fig. 10 shows photodegradation rate curves of TNAs grown on different substrates and TiO 2 nanoparticles for methyl blue Fig. 8. TEM images of TiO 2 nanorods grown on PT-substrate at an annealing temperature of 700 8C and at a hydrothermal growth temperature of 160 8C and with growth time of 3 h. (a) TEM image and the associated selected area electron diffraction of TiO 2 nanorods; (b) lattice fringes image of high-resolution TEM of one single TiO 2 nanorod at the side surface. Fig. 9. FE-SEM image of the as-prepared TiO 2 arrays grown on PT-substrate when NaCl is not added to the hydrothermal precursor solution at a hydrothermal growth temperature of 160 8C and with growth time of 3 h. Fig. 10. Photodegradation rate curves of TiO 2 nanoparticles (a) and TNAs grown on different substrates for methyl blue solution: (b) PT-substrate without heat treatment; (c) PT-substrate; (d) unmodified substrate. Hydrothermal growth temperature: 160 8C and growth time: 3 h. Y. Li et al. / Materials Research Bulletin 44 (2009) 1232–1237 1236 solution. The curves of C/C 0 versus irradiation time reveal the concentration change of methyl blue solution as the photocatalytic reaction proceeds. The used TiO 2 nanoparticles are prepared by sol– gel method and annealed at 450 8C, which is pure anatase phase. The TiO 2 nanoparticles have a diameter of approximately 11 nm, and the specific surface area is about 40.74 m 2 /g. In the contrast experi- ments, the mass of TiO 2 nanoparticles and three kinds of TNAs adding to methyl blue solution is 0.02 and 0.006 g (excluding the pre-coated particle on the substrate), respectively. From Fig. 10,it can be seen that although the addition quantity of TNAs is much less than that of TiO 2 nanoparticles, it has relatively higher photode- gradation rate. In addition, three kinds of TNAs grown on different substrates have similar photocatalytic activity. The higher photo- catalytic activity of TNAs may result from its crystal structure and distribution of the arrays. TiO 2 nanoparticles have much more grain boundary than that of single crystalline TNAs, which will make it easier for the recombination of electron and hole pair. On the other hand, TiO 2 nanoparticles may be agglomerated together and reduce the surface area, while the distribution of the TNAs will protect it from agglomeration. As a result, the photocatalytic activity greatly improves with the prepared TNAs. 4. Conclusions Large-scale, well-aligned TNAs were fabricated on the modified glass substrate using hydrothermal approach. It has been demonstrated that TNAs grown on PT-substrate have a better growth orientation and uniform density compared with those grown on the unmodified substrates. TiO 2 nanorod is a single crystal rutile structure with growth direction along the (1 1 0) crystal planes. About 67% of TiO 2 nanorods have a diameter between 15 and 25 nm and the average diameter is about 21 nm, and the length of TiO 2 nanorods ranges from 300 to 500 nm. In addition, it indicates that TiO 2 nanoparticles coated on glass substrate are not completely transformed into rutile phase after heat treatment at 700 8C, and the (1 1 0) planes are the preferred orientation. The photodegradation experiment implies that single crystal TNAs have a relatively higher photocatalytic activity compared with polycrystal TiO 2 nanoparticles. This work demon- strates that the substrate pre-treatment has a strong impact on the growth morphology of TNAs and provides a very promising approach for the nanorod arrays preparation. Acknowledgements The authors would like to thank the National Science Foundation of China for the financial support (Nos. 50425415 and 50772004) and National Basic Research Program of China (973 Program: 2007CB613608). References [1] X.H. Xia, Y.S. Luo, Z. Wang, Y. Liang, J. Fan, Z.J. Jia, Z.H. Chen, Mater. Lett. 61 (2007) 2571. [2] Q. Wei, K. Hirota, K. Tajima, K. Hashimoto, Chem. Mater. 18 (2006) 5080. [3] G.K. Mor, K. Shankar, M. Paulose, O.K. Varghese, C.A. Grimes, Nano Lett. 6 (2006) 215. [4] I.D. Kim, A. Rothschild, B.H. Lee, D.Y. Kim, S.M. Jo, H.L. Tuller, Nano Lett. 6 (2006) 2009. [5] W. Ho, J.C. Yu, J.G. Yu, Langmuir 21 (2005) 3486. [6] M. Paulose, K. Shankar, S. Yoriya, H.E. Prakasam, O.K. Varghese, G.K. Mor, T.A. Latempa, A. Fitzgerald, C.A. Grimes, J. Phys. Chem. B 110 (2006) 16179. [7] S.K. Pradhan, P.J. Reucroft, F. Yang, Dozier, J. Cryst. Growth 256 (2003) 83. [8] Z.R. Tian, J.A. Voigt, J. Liu, B. Mckenzie, H. Xu, J. Am. Chem. Soc. 125 (2003) 12384. [9] Y. Lei, L.D. Zhang, G.W. Meng, G.H. Li, X.Y. Zhang, C.H. Liang, W. Chen, Appl. Phys. Lett. 78 (2001) 1125. [10] S.Z. Chu, K. Wada, S. Inoue, S.I. Hishita, K. Kurashima, J. Phys. Chem. B 107 (2003) 10180. [11] J. Zuo, C.G. Nie, X. Gu, Y.K. Lai, Y. Zong, L. Sun, C.J. Lin, Mater. Lett. 61 (2007) 2632. [12] M. Barbic, J.J. Mock, D.R. Smith, S. Schultz, J. Appl. Phys. 91 (2002) 9341. [13] D. Xu, Y. Xu, D. Chen, G. Guo, L. Gui, Y. Tang, Chem. Phys. Lett. 325 (2000) 340. [14] O.K. Varghese, D. Gong, M. Paulose, K.G. Ong, E.C. Dickey, C.A. Grimes, Adv. Mater. 15 (2003) 624. [15] J.J. Wu, C.C. Yu, J. Phys. Chem. B 108 (2004) 3377. [16] X.J. Feng, J. Zhai, L. Jiang, Angew. Chem. Int. Ed. 44 (2005) 5115. [17] M. Guo, P. Diao, S.M. Cai, J. Solid State Chem. 178 (2005) 1864. [18] T. Ma, M. Guo, M. Zhang, Y.J. Zhang, X.D. Wang, Nanotechnology 18 (2007) 035605. [19] Y.H. Jiang, H.B. Yin, Y.M. Sun, H. Liu, L.X. Lei, K.M. Chen, Y.J. Wada, Appl. Surf. Sci. 253 (2007) 9277. [20] M.N. Tahir, P. Theato, P. Oberle, G. Melnyk, S. Faiss, U. Kolb, A. Janshoff, M. Stepputat, W. Tremel, Langmuir 22 (2006) 5209. [21] K. Kakiuchi, E. Hosono, H. Imai, T. Kimura, S. Fujihara, J. Cryst. Growth 293 (2006) 541. [22] J. Yang, S. Mei, J.M.F. Ferreira, P. Norby, S. Quaresma ˆ , J. Colloid Interf. Sci. 283 (2005) 102. [23] X. Bokhimi, A. Morales, M. Aguilar, J.A. Toledo-Antonio, F. Pedraza, Int. J. Hydrogen Energy 26 (2001) 1279. [24] P. Hartman, W.G. Perdok, Acta Cryst. 8 (1955) 49. [25] P. Hartman, W.G. Perdok, Acta Cryst. 8 (1955) 521. [26] P. Hartman, W.G. Perdok, Acta Cryst. 8 (1955) 525. [27] E.W. Shi, Z.Z. Chen, R.L. Yuan, Y.Q. Zheng, Hydrothermal Crystallography, Science Press, Beijing, 2004, p. 150. [28] L. Vayssieres, K. Keis, S.E. Lindquist, A. Hagfeldt, J. Phys. Chem. B 105 (2001) 3350. Y. Li et al. / Materials Research Bulletin 44 (2009) 1232–1237 1237 . Hydrothermal synthesis and characterization of TiO 2 nanorod arrays on glass substrates Yuxiang Li a , Min Guo a , Mei Zhang a , Xidong Wang b, * a Department. the orientation of TiO 2 nanorod arrays grown on the substrate becomes poor. In addition, the diameter distribution of TiO 2 nanorods grown on PT-substrate