Interface Structure of Gold Particles on TiO 2 Anatase

Transcription

1 Materials Transactions, Vol. 52, No. 3 (2011) pp. 280 to 284 Special Issue on New Trends for Micro- and Nano Analyses by Transmission Electron Microscopy #2011 The Japan Institute of Metals Interface Structure of Gold Particles on TiO 2 Anatase Daiki Nagamatsu 1, Takashi Nemoto 1, Hiroki Kurata 1, Jinting Jiu 2, Motonari Adachi 3 and Seiji Isoda 4 1 Institute for Chemical Research, Kyoto University, Kyoto , Japan 2 ISIR, Osaka University, Ibaraki , Japan 3 Department of Chemical Engineering and Materials Science, Doshisha University, Kyotanabe , Japan 4 icems, Kyoto University, Kyoto , Japan Gold particles supported on titanium dioxide (TiO 2 ) are known to activate many catalytic reactions. In this study, interface structure of gold on anatase was examined in the case on an anatase {101} surface, since the surface is the most important for several photocatalytic reactions. Anatase TiO 2 nanorods having clear washboard-like {101} outer surfaces were synthesized by using a double surfactants system. From high resolution electron microscopy observation for gold particle on the {101} surface, the (113) plane of gold was found to be parallel to the surface of TiO 2. Such an epitaxy is corresponding to be the axial and planar orientations of [031] (200)Au == [311] (113)TiO 2, being originated from the smallest lattice mismatching, instead of the previous cases of [011] (111)Au == [021] (112)TiO 2 on the {110} and {112} surfaces of anatase TiO 2. [doi: /matertrans.mb201004] (Received September 13, 2010; Accepted November 2, 2010; Published December 29, 2010) Keywords: interface structure, gold particles, anatase titanium dioxide, high resolution electron microscopy 1. Introduction Semiconductor photocatalysis is an influential field that plays an important part in modern catalytic research. In particular, environmental applications of titanium dioxide (TiO 2 ) attract considerable attention in these decades as one of the most active photocatalysts. 1) Furthermore, by modification of TiO 2 with gold particles in a size range of a few nanometers, the modified TiO 2 were reported to be more active for many reactions at low temperature such as CO oxidation, propylene epoxidation and other reactions. 2,3) Furthermore, higher photocatalytic properties were also reported such as photo-degradation of molecules, 4 6) in which the reactions are especially influenced of the gold particles. 7 11) The activity of the gold particles should depend on their structures, morphologies and also on their interfacial relations with respect to substrates. For structural study of metal particles on TiO 2, highresolution transmission electron microscopy (HR-TEM) is a powerful tool to analyze the size, shape and surface structure of metal particles which could be responsible to the catalytic performance. 12,13) Many of the detailed structural experimental results found in the recent studies are on the rutile TiO 2 cases, 14,15) because this TiO 2 form is the more stable one and easy to prepare electron-transparent TEM samples by thinning down a commercially available rutile TiO 2 single crystals. However, the wide-bandgap semiconductor TiO 2 exists usually in two polymorphic phases; anatase and rutile. Both are organized in tetragonal lattices but in different unit cells exhibiting different physical and chemical behavior. It is the anatase that plays often an essential role in a number of charge-separating events, 16) such as dye-sensitized solar cells, so that it is quite indispensable to study interfacial structures of gold particles on the anatase surfaces. In this light, some important structural studies on the gold particles on anatase surface have been performed so far. 17,18) In those reports, gold particles were examined for their growth on the {110} and {112} surfaces of commercially available powdery anatase crystals. Lattice deformation of gold layers was concluded and an major epitaxial orientation, near (111) [011]Au == (112) [021]TiO 2, was proposed for gold particles on these anatase TiO 2 surfaces. However, actual catalytic particles are usually shaped with various crystallographic surfaces so that it is desirable to analyze on the diverse interface structures for understanding of the activity. For example, there has been no reports on gold particles growth on the {101} surface of anatase TiO 2, which is very important in dye-sensitized solar cells for accommodating efficiently a sensitizing dye through its carboxyl groups such as N719 dye of (Bu 4 N) 4 [Ru(dcbpy) 2 (NCS) 2 ]. 19) For an anode electrode of solar cells, anatase TiO 2 usually achieve a higher efficiency than rutile TiO 2. 20,21) Lu et al. reported that the incident photo-to-electron conversion efficiency values were much higher on the anatase {101} surface than on the other anatase surfaces. 22) Therefore it is a most important surface for dye-sensitized solar cell. Accordingly, in the present study, the interface structure of gold particles on the anatase {101} surface was examined. The aims of the present study are to know whether a regular interface is realized on the interface and what kind of crystallographic orientation relation is formed. This experiment became very easier owing to the success of a unique hydrothermal synthesis of anatase nanorods bounded with the {101} surfaces. 23) 2. Experimental Nanorod TiO 2 was prepared by a sol-gel method using a hydrothermal process as described in the previous reports. 24,25) The synthesized nanorods exhibit the anatase structure with nm length and nm width on an average, by using a double surfactants system of a blockcopolymer (F127) and cetyltrimethylammonium bromide in water. The nanorods have a tetragonal shape on the whole with their long axis of crystal along the h001i direction of anatase and the individual side surfaces of tetragonal rods are

2 Interface Structure of Gold Particles on TiO2 Anatase 281 TiO2 5 nm 1 µm 100 nm c-axis washboard surface Au particles Fig. 2 HR-TEM image of Au particles on washboard surfaces composed of {101} planes of anatase. Fig. 1 Low-magnified TEM image of Au particles on surfaces of anatase TiO2 nanorods. formed with clear washboard-like {101} surfaces,26) so that the surface can be covered effectively with Gra tzel dyes, exhibiting a high efficiency of light-to-electricity.23) The washboard composed of {101} surfaces provide a nice opportunity to examine crystallographic orientation of gold particles on the {101} of anatase. Gold particles of 2 5 nm in size were deposited on the surfaces of TiO2 nanorods by vacuum-evaporation under a vacuum of 10 4 Pa at room temperature. The particle shape of gold, their orientation on the substrate and the growth mechanism were examined by HR-TEM in addition to 3D TEM imaging and high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) in a plane-view and an edge-view imaging associated with a multislice image simulation using the Win HREM v. 3.0 software package.27) An electron microscope used in this study was a TEM/STEM JEM2200FS operated at 200 kv. HAADF-STEM images were collected over scattering angles from 50 mrad to 170 mrad. The 3D image was reconstructed by IMOD3D using tomography method by taking 120 TEM images tilted every 0.5 degrees. 3. Results and Discussion Figure 1 shows typical TiO2 nanorods observed by TEM, where the gold particles with diameter less than 5 nm are dispersed almost evenly on TiO2 nanorods. A typical washboard structure is shown in Fig. 2. The washboard structures are observed clearly on every surfaces by taking snapshots in 3D image. The terrace size of individual surface seems to be not so large there, and it is varied from a nanorod to a nanorod. The surfaces are considerably clean and sharp for a material synthesized through a liquid reaction. When an incident electron beam was set along the TiO2 [010] zone axis, parallel to edge surfaces of TiO2, the image indicates that TiO2 nanorod has a well-ordered flat surfaces which are composed almost of the and (1 01) planes. The similar image was observed when the incident beam was parallel to the TiO2 [100] by rotating with 90 around the long axis from the TiO2 [010]. The edge surfaces are the (011) and (01 1) planes equivalent to the and (1 01) 20 nm edge view plan view Fig. 3 HAADF-STEM image of Au particles on TiO2 nanorods, where the Au particles are observed with a brighter contrast. planes. Therefore, edge surfaces are composed of the {101} planes that realize higher conversion efficiency value for photocatalytic activity like dye-sensitized solar cells. Figure 3 shows the deposited gold particles with a brighter contrast by HAADF-STEM. Intensities in the HAADFSTEM image are directly related to the atomic number at the irradiated area of sample, Z-contrast, so that gold particles appear with higher intensity due to their higher atomic number than TiO2. As one can see here again, the edge surfaces of the TiO2 nanorods are quite sharp, since the gold particles deposited on the side edges seem to be adsorbed on flat planes. One can observe the deposited gold particles on TiO2 {101} surfaces from an edge-view and also from a plane-view. However, for further analysis, HR-TEM was adopted rather than HAADF-STEM, because of easiness of the former method in crystal orientation analysis. Firstly the plane-view projection in HR-TEM was carried out, but this method could not be suitable to analyze the interface of Au/TiO2. A typical image by plane-view is shown in Fig. 4(a), where the electron beam was transmitted nearly along the [111 ] of anatase TiO2 (Fig. 4(c)). If gold particles grow by following the growth relation of (111) [011 ]Au == (112) [021 ]TiO2 as reported previously, the

3 282 D. Nagamatsu et al. (b) (a) 011 TiO TiO TiO Au 2 2 (c) (112) (011) Fig. 4 HR-TEM plane-view projection of Au particles on anatase TiO 2 surface. (a) HR-TEM image, (b) fast Fourier transform of a circular region in (a), where the diffraction pattern can be indexed as a projection nearly along the [111] of TiO 2 as shown in (c). 111 reflection of gold should coincide to the 112 reflection of TiO 2 typically in the fast Fourier transform (FFT) of [021]TiO 2 projection image (Fig. 4(b)). However, such relationship was not observed in this case as well as the other gold particles, suggesting another interfacial relation on the TiO 2 {101} surfaces. Practically it was difficult to determine a relationship of structures between gold nanoparticle and TiO 2 nanorod from the plane-view imaging, since the gold particles are very small to adjust their orientation with respect to the TiO 2 surface. Therefore, the edge-view imaging was performed as shown in Fig. 5. These HR-TEM images were taken nearly along the [010] direction of TiO 2, parallel to the {101} surfaces. As can been seen for gold crystallites, the electron did not transmit along a specific crystallographic low index direction of gold, but deviated slightly from that. Although there observed a little distortion of gold lattice, probably due to a size effect in addition to an effect of electron beam irradiation, we do not go further on this issue of lattice distortion, since it is not so large amount in the present case. Because of the edge projection, some gold particles are overlapped in these TEM images, but a particular orientation relation was extracted from images; that is, the of TiO 2 makes an averaged angle of about 29 with respect to the (111) of gold as indicated in the images. This means that a certain crystallographic relation should be considered at the gold particles on the {101} surfaces of TiO 2. By noticing the fact that the (113) plane of gold makes angle of 29.4 with its (111), the (113) of gold is most probably considered to be parallel to the surface of TiO 2 ; a planar epitaxial relation of (311)Au == TiO 2. Finally an interface structure is discussed from a lattice matching for the case of (311)Au == TiO 2. In order to examine such an interfacial relation, both crystals of gold and TiO 2 were rotated to find any particular crystallographic relation under a constraint that (311)Au == TiO 2. The most probably interface we found is shown in Fig. 6, where a lattice matching is excellent; the (200) planes of Au coincide nicely with the (113) planes of TiO 2 on their spacing and parallel alignment, resulting into a high interfacial stabilization in the lattice-lattice interaction. For the crystals of gold and TiO 2 in Fig. 6, the projection directions are the [031] and the [ 311], respectively. Figure 6 does not correspond directly to the HR-TEM images in Fig. 5, but the projection should be rotated about 35 around the vertical axis in Fig. 6 to coincide with the HR-TEM projections of TiO 2 in Fig. 5. The important point here is that the (200)Au == (113)TiO 2, indicating an epitaxial growth of gold on the {101} surface of TiO 2 under a lattice matching mechanism. On the other words, the lattice misfit is excellently small between these plans as listed in Table 1; that is, better than the already reported misfit of (111)Au == (112)TiO 2. The misfit value is known to be a dominant factor in epitaxy by point-on-line coincidence ) and by the equivalent coincidence in reciprocal lattice point. 31,32) These two coincidence models are equivalent in principle, but different presentation in the real space or the reciprocal one. Hoshino et al ) examined the firstly growing epitaxial monolayer with respect to a substrate surface by scanning tunneling microscopy and concluded that a depositing lattice is deformed slightly from the stable bulk one so as for the growing lattice to match its spacing to a surface spacing and to gain a stabilization energy

4 Interface Structure of Gold Particles on TiO2 Anatase nm TiO TiO TiO2 2 nm 5 nm Fig. 5 HR-TEM images of Au particles on {101} surfaces of anatase TiO2 nanorods. The angle between the of TiO2 and the (111) of Au is nearly the same. Table 1 Misfit values between Au and TiO2 crystals, where the misfit was calculated by misfit ¼ ðdau dtio2 Þ=dTiO2. (200) Au [0 3 1] projection for Au Au h, k, l (1 1 3) Au [3 1 1] projection for TiO2 (011) TiO2 (113) TiO2 Fig. 6 Schematic relation of Au and TiO2 interface, where the (200)Au with d ¼ 0:20393 nm == the (113)TiO2 with d ¼ 0:20400 nm. This is an example of smallest lattice misfit than the case of (110)Au == (102)TiO2. through the lattice-lattice interaction. The distortion could be minimized, when a lattice of a bulk crystal (non-deformed) is similar in spacing with that of the substrate. In the table, lattice misfits were calculated from an equation of misfit ¼ TIO2 d (nm) h, k, l d (nm) Misfit (%) 1, 1, , 1, :50 1, 1, , 1, 2 1, 1, : , 1, , 1, :40 ðd1 d2 Þ=d2 where d1 and d2 are the lattice spacings of deposited and substrate crystals, respectively. The table lists only the cases of lower misfit values, and the smallest misfit, the most stable interface in term of lattice-lattice interaction, is realized in the present lattice matching of (311)Au == TiO2 and the secondly small value is in the previously reported case of (111)Au == (112)TiO2 on the {110} and {112} surfaces of anatase TiO2. Then clean epitaxial interface can be concluded to be originated from the lattice matching mechanism, even though the size of surface area is rather limited in some terraces of {101}. As the cases of small crystals formed at liquid-solid interfaces,33,34) the lattice-lattice interaction could control almost the orientation of deposited crystals even at a limited area. A charge distribution between the TiO2 and metal nanoparticles may cause particular changes in electronic state at the surface, such as the Fermi level to shift.35) The present definite epitaxial growth of gold on the TiO2 anatase {101}

5 284 D. Nagamatsu et al. might contribute more for a formation of a characteristic electronic state at the interface. It would be an attractive next issue to understand effects of modification with gold particles on the {101} surfaces as well as on other surfaces. 4. Conclusions Gold nano-particles supported on TiO 2 are known to activate many catalytic reactions at low temperatures. In this study, interface structures of gold particles were examined in the case on anatase {101} surface. Nanorod TiO 2 particles synthesized in the present study exhibit the anatase structure. The nanorods have a tetragonal shape on the whole with their long axis along the h001i direction and an individual side surface is formed with clear washboard-like {101} surfaces. From high resolution electron microscopy observation on the interface between gold particles and the TiO 2, the (113) plane of gold was found to be parallel to the surface of TiO 2. Such an epitaxy is considered to be originated from the axial and planar orientations of [031] (200)Au == [311] (113)TiO 2, being relating to the smallest lattice misfit, instead of the previous cases of [011] (111)Au == [021] (112)TiO 2 on the {110} and {112} surfaces of anatase TiO 2. Acknowledgement This work was supported partly by a Grant-in-Aid for Scientific Research (B) No and (C) No Figure 6 was drawn with VENUS developed by R. A. Dilanian and F. Izumi. REFERENCES 1) M. R. Hoffmann, S. T. Martin, W. Y. Choi and D. W. Bahnemann: Chem. Rev. 95 (1995) ) M. Haruta: CATTECH 6 (2002) ) M. Kotobuki, R. Leppelt, D. A. Hansgen, D. Widmann and R. J. Behm: J. Catalysis 264 (2009) ) A. Orlov, D. A. Jefferson, M. Tikhov and R. M. Lambert: Catalysis Commun. 8 (2007) ) G. L. Chiarello, L. Forni and E. Selli: Catalysis Today 144 (2009) ) C.-J. Liu, T.-Y. Yang, C.-H. Wang, C.-C. Chien, S.-T. Chen, C.-L. Wang, W.-H. Leng, Y. Hwu, H.-M. Lin, Y.-C. Leed, C.-L. Cheng, J. H. Je and G. Margaritondo: Mater. Chem. Phys. 117 (2009) ) M. Haruta, S. Tsubota, T. Kobayashi, H. Kageyama, M. Genet and B. Delmon: J. Catal. 144 (1993) ) M. Valden, X. Lai and D. W. Goodman: Science 281 (1998) ) T. Hayashi, K. Tanaka and M. Haruta: J. Catal. 178 (1998) ) F. Boccuzzi, S. Coluccia, G. Marta and N. Ravasio: J. Catal. 188 (1999) ) M. Haruta: Catal. Today 36 (1997) ) E. J. Braunschweig, A. D. Logan, A. K. Datye and D. J. Smith: J. Catal. 118 (1989) ) M. H. Yao, D. J. Smith and A. K. Datye: Ultramicroscopy 52 (1993) ) T. Akita, K. Tanaka, M. Kohyama and M. Haruta: Surf. Interface Anal. 40 (2008) ) N. Shibata, A. Goto, K. Matsunag, T. Mizoguchi, S. D. Findlay, T. Yamamoto and Y. Ikuhara: Phys. Rev. Lett. 102 (2009) ) R. Hengerer, B. Bolliger, M. Erbudak and M. Grätzel: Surf. Sci. 460 (2000) ) S. Giorgio, C. R. Henry, B. Pauwels and G. Van Tendeloo: Mater. Sci. Eng. A 297 (2001) ) T. Akita, K. Tanaka, S. Tsubota and M. Haruta: J. Electron Microsc. 49 (2000) ) M. K. Nazeeruddin, S. M. Zakeeruddin, R. Humphry-Baker, M. Jirousek, P. Liska, N. Vlachopoulos, V. Shklover, C. H. Fischer and M. Grätzel: Inorg. Chem. 38 (1999) ) C. J. Barbe, F. Arendse, P. Comte, M. Jirousek, F. Lenzmann, V. Shklover and M. Grätzel: J. Am. Ceram. Soc. 90 (1997) ) N. G. Park, J. van de Lagemaat and A. J. Frank: J. Phys. Chem. B 104 (2000) ) Y. F. Lu, D. J. Choi, J. Nelson, O.-B. Yang and B. A. Parkinson: J. Electrochem. Soc. 153 (2006) E131 E ) J. Jiu, S. Isoda, F. Wang and M. Adachi: J. Phys. Chem. B 110 (2006) ) J. Jiu, F. Wang, S. Isoda and M. Adachi: Chem. Lett. 34 (2005) ) T. Kurata, Y. Mori, S. Isoda, J. Jiu, K. Tsuchiya, F. Uchida, F. Wang and M. Adachi: Current Nanosc. 6 (2010) ) K. Yoshida, J. Jiu, D. Nagamatsu, T. Nemoto, H. Kurata, M. Adachi, S. Isoda and T. Komatsu: Mol. Cryst. Liq. Cryst. 491 (2008) ) K. Ishizuka: Ultramicroscopy 90 (2002) ) A. Hoshino, S. Isoda, H. Kurata and T. Kobayashi: J. Appl. Phys. 76 (1994) ) A. Hoshino, S. Isoda, H. Kurata and T. Kobayashi: J. Cryst. Growth 146 (1995) ) A. Hoshino, S. Isoda, H. Kurata, T. Kobayashi and Y. Yamashita: Jpn. J. Appl. Phys. 34 (1995) ) S. Stemmer, P. Pirouz, Y. Ikuhara and R. F. Davis: Phys. Rev. Lett. 77 (1996) ) Y. Ikuhara and P. Pirouz: Mater. Sci. Forum (1996) ) S. Isoda, T. Nemoto, E. Fujiwara, Y. Adachi and T. Kobayashi: J. Cryst. Growth 229 (2001) ) T. Nemoto, Y. Adachi, E. Fujiwara and S. Isoda: Mol. Cryst. Liq. Cryst. 389 (2002) ) M. Jakob, H. Levanon and P. V. Kamat: Nano Lett. 3 (2003)