Growth of Epitaxial Anatase Nano Islands on SrTiO3(001) by Dip

Publication Date (Web): February 28, 2013 ... The procedure developed here provides a very cheap and simple approach to preparation of oriented anatas...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/crystal

Growth of Epitaxial Anatase Nano Islands on SrTiO3(001) by Dip Coating Freddy E. Oropeza,†,∥ Kelvin H. L. Zhang,†,⊥ Anna Regoutz,† Vlado K. Lazarov,‡ Didier Wermeille,§ Christopher G. Poll,# and Russell G. Egdell*,† †

University of Oxford, Department of Chemistry, Inorganic Chemistry Laboratory, South Parks Road, Oxford OX1 3QR, U.K. Department of Physics, University of York, Heslington, York, YO10 5DD, U.K. § XMaS CRG Beamline, European Synchrotron Radiation Facility, 38043 Grenoble Cedex 9, France and Department of Physics, University of Liverpool, Liverpool, L69 7ZE, U.K. # Department of Materials, Imperial College London, Exhibition Road, London, SW7 2AZ, U.K. ‡

ABSTRACT: High temperature annealing of ultrathin anataseTiO2(001) thin films deposited on SrTiO3(001) substrates by a dip coating method leads to the self-assembly of an array of squareshaped epitaxial anatase islands with lateral dimensions of order 150 nm. The procedure developed here provides a very cheap and simple approach to preparation of oriented anatase nanocrystals with a distribution of surface terminations very different to that found in free-standing material. In particular, the {101} facets which are dominant for unsupported nanocrystalline anatase are completely absent, and the (001) islands are instead bounded by sloping {103} side facets.

■

INTRODUCTION Titanium dioxide (TiO2) has attracted widespread interest as a material able to mediate photocatalytic reactions such as the degradation of organic pollutants or the photo-induced splitting of water.1−6 The two most important polymorphs of TiO2 are anatase and rutile. The key materials properties of these two phases are summarized in Table 1. It can be seen that the band

Anatase is thermodynamically unstable with respect to rutile, albeit with a very small enthalpy difference of ∼0.7 kJ/mol between the two phases. However, anatase is usually obtained when amorphous TiO2 gels precipitated from solution are calcined at low temperatures. The nanocrystalline anatase formed in this way is metastable at room temperature but converts irreversibly into rutile at higher temperatures.13−17 The rate of transformation of high purity commercial anatase powder is very slow at 600 °C but rapid above 730 °C.13,15 In the present paper, we demonstrate that anatase is stabilized to a very much higher temperature when grown as an ultrathin epitaxial film on SrTiO3(001). The key to the epitaxial stabilization apparently lies in a tensile mismatch of only −3.1% between the a parameter for tetragonal anatase (a = 3.7845 Å) and the cubic lattice parameter of SrTiO3 (a = 3.905 Å), as compared to a mismatch of +17.6% for rutile (a = 4.5941 Å). However, the mismatch is still sufficiently large to promote a dewetting process that leads the anatase films to break up into an array of nanoscale square islands upon annealing at 1000 °C. Continuous epitaxial thin films of pure and transition metal doped anatase have been grown by a range of techniques including pulsed laser deposition,18−22 magnetron sputtering,23 and O-plasma assisted molecular beam epitaxy.24−28 Both SrTiO3(001) and LaAlO3(001) substrates have been employed, the cubic lattice parameter of the latter (a = 3.793 Å) being

Table 1. Some Important Materials Parameters of Rutile and Anatase Phases of TiO2 rutile space group lattice parametersa,b

volume per formula unit bandgapb,c a

anatase

P42/mnm a = 4.5941 Å c = 2.9589 Å Z=2 31.225 Å3

I4/amd a = 3.7845 Å c = 9.5143 Å Z=4 34.067 Å3

3.06 eV

3.46 eV (3.2 eV at room temperature)

Ref 10. bRef 11. cRef 12.

gap of anatase is slightly bigger than that of rutile and that the volume per formula unit is also bigger. Moreover, the bandgap of anatase is indirect, whereas rutile has a direct bandgap.7,8 In addition a higher electron mobility is found for anatase.9 Clearly, the energy threshold for photocatalytic activity mediated by anatase is bigger than that for rutile. However, there is a general consensus that anatase is a superior photocatalyst under ultraviolet (UV) irradiation. © 2013 American Chemical Society

Received: October 5, 2012 Revised: February 7, 2013 Published: February 28, 2013 1438

dx.doi.org/10.1021/cg301455q | Cryst. Growth Des. 2013, 13, 1438−1444

Crystal Growth & Design

■

better matched with anatase. Thus, anatase films could be grown on SrTiO3(001) or LaAlO3(001) by electron beam evaporation for substrate temperatures between 500 and 1000 °C, although growth on SrTiO3(001) at 1100 °C led to a mixture of anatase and rutile phases.29 Epitaxial films have also been grown on these substrates by low temperature solution based deposition techniques30,31 or by hydrothermal dissolution of SrO out of SrTiO3.32 In the present communication, a very simple and inexpensive wet chemical procedure is used to prepare the TiO2 films, which crystallize into small islands of the anatase phase upon thermal annealing at 1000 °C. The new route for preparation of supported anatase nanocrystals with a well-defined orientation and morphology opens the possibility of detailed study of photocatalysis mediated by nanocrystalline anatase with a very different distribution of surface terminations to that found for free-standing material.

■

Article

RESULTS AND DISCUSSION Panels a−d of Figure 1 show 5 μm × 5 μm AFM images of TiO2 samples obtained after two dip−hydrolysis cycles and

EXPERIMENTAL SECTION

Samples were prepared by a dip coating method. Titanium(IV) isopropoxide (Aldrich, 97%) was dissolved in isopropanol (Fischer Scientific, analytical grade) to give 25 cm3 of a solution with 0.33 M Ti concentration, which was acidified using HNO3 (HNO3/Ti molar ratio 0.15). The resulting solution was used to dip coat SrTiO3(001) substrates in a 25 cm3 wide neck flask. The substrates were cleaned by rinsing in acetone followed by sonication in isopropanol. After immersion, substrates were withdrawn from the alkoxide solution using a motor drive at a pulling rate of 3.5 mm/min. The wet substrates were immediately placed in flowing N2 gas saturated with H2O at room temperature, so that the alkoxide was hydrolyzed on the SrTiO3(001) surface. The amorphous film products of the hydrolysis were calcined for 4 h at temperatures ranging from 700 °C up to 1000 °C to form epitaxial films of anatase TiO2. By varying the numbers of dip-hydrolysis cycles, the coverage of the substrate can be controlled: three series of samples were prepared with two, four, and eight dips, although completely isolated nano islands only formed after calcinations of the two dip samples at 1000 °C. For larger coverages dewetting of the substrate was not complete and the crystalline islands retained some connectivity. Following alignment of the crystal specular θ-2θ X-ray diffraction profiles were measured using a PANalytical X’Pert Pro diffractometer incorporating a monochromated Cu Kα source (λ = 1.5406 Å). Further diffraction experiments were performed on three samples on the bending magnet beamline XMaS/BM2833 at the European Synchrotron Radiation Facility (ESRF), Grenoble France using a photon energy of 15 keV, corresponding to a wavelength λ = 0.8266 Å. The samples were mounted directly in air on a vertical scattering 4circle Huber diffractometer.33 The scattered photons were detected with a Pilatus 300 K two-dimensional (2D) detector, here effectively operated as a point detector by integrating counts in the central group of pixels. The SrTiO3 substrate was aligned using the (002) and (012) reflections, and it was verified that [010] direction of the anatase epilayer lay in the same plane as that for the substrate by scanning through the (013) epilayer reflection along the [100] direction in reciprocal space. Diffraction data from the ESRF are presented in terms of reciprocal lattice units of the substrate. Atomic force microscopy (AFM) images were recorded in a Digital Instruments Multimode Scanning Probe Microscopy instrument with a Nanoscope IIIa controller operating in a contact mode, as described in detail elsewhere. Specimens for cross-sectional transmission electron microscopy (TEM) were prepared by cutting and mechanical grinding down to 5−10 μm, followed by thinning to electron transparency by Ar ion beam milling using a Gatan 691 Precision Ion Polishing System (PIPS). Cross-sectional TEM images were collected using a JEOL3000F microscope operating at 300 keV. Digital Micrograph image software was used to process images. Bragg filtering was applied in order to remove some of the background noise, and contrast adjustment was applied uniformly across the images where necessary to enhance the display quality.

Figure 1. (a−d) 5 μm × 5 μm AFM images of TiO2 sample obtained after two dip−hydrolysis cycles calcined at the temperatures indicated between 700 and 1000 °C. (e) Expanded 2 μm × 2 μm AFM image of two-dip sample annealed at 1000 °C. (f) 2 μm × 2 μm AFM image of eight-dip sample annealed at 1000 °C.

calcination at temperatures between 700 and 1000 °C. After calcination at 700, 800, or 900 °C continuous but increasingly rough films were obtained, with obvious signs of development of square islands after annealing at 900 °C. The film morphology changed dramatically after annealing at 1000 °C with formation of a striking array of square islands with a fairly narrow size distribution. From the known bulk crystallographic directions of the substrate, it may be inferred that islands are all oriented with edges parallel to the [100] and [010] directions of the SrTiO3 substrate. Analysis of the AFM images indicates the square-shaped structures have an average height of approximately 15−20 nm and edge length of around 150 nm. These observations indicate that the originally flat film suffered a dewetting process upon high temperature annealing leading to the formation of a self-assembled array of square-shaped nanoislands. The sample initially crystallizes as a metastable continuous film that breaks down when kinetic barriers to dewetting can be overcome. Coll et al.34 reported a similar thermally induced dewetting process for (001) heteroepitaxial 1439

dx.doi.org/10.1021/cg301455q | Cryst. Growth Des. 2013, 13, 1438−1444

Crystal Growth & Design

Article

annealed at 700 °C and a slightly smaller value of 1.3° for the two- and four-dip samples annealed at 1000 °C. Expanded θ-2θ scans of the anatase (004) peak are shown in Figure 3. It can be seen that after calcination at 700 °C the

YBa2Cu3O7 films grown from trifluoroacetate precursors on SrTiO3(001) and LaAlO3(001). In that work it was possible to rationalize dewetting purely on the basis of surface and interface energies. Four- and eight-dip films underwent a similar process of roughening and island formation with increasing annealing temperatures, but as shown in Figure 1f for an eightdip film, dewetting was not complete for the thickest films, with no bare substrate visible even after annealing at 1000 °C. Small square islands form on the surface, but they remain connected and since the substrate remains covered, the z range in the AFM images is smaller than in the corresponding 2 μm × 2 μm image of a two-dip film. Figure 2 shows θ-2θ XRD profiles of two-dip films calcined at 700 and 1000 °C acquired in overnight scans on the

Figure 3. Expanded scan over anatase (004) diffraction peak in θ-2θ XRD pattern of a TiO2 film obtained after two dip−hydrolysis cycles and calcination at temperatures from 700 to 1000 °C showing shift to low angle with increasing calcinations temperature.

anatase (004) diffraction peak shows strong asymmetric broadening that gradually disappears to give a narrow and symmetric diffraction peak after the sample was annealed at 1000 °C. It can also be seen in Figure 3 that there is a shift toward lower angle upon increasing the annealing temperature. Similar results were found for four-dip and eight-dip samples, even though these films did not break down into isolated islands. The anatase c lattice parameter is plotted as a function of annealing temperature in Figure 4. There is a monotonic increase in c with increasing annealing temperature. This type of variation is expected for a thin epitaxial layer under tensile

Figure 2. θ-2θ XRD pattern of a TiO2 film obtained after two dip− hydrolysis cycles and calcination at temperatures of 700 and 1000 °C. Intensities in XRD patterns are presented on a logarithmic scale.

PANalytical diffractometer. Aside from the SrTiO3(001) and SrTiO3(002) substrate Bragg peaks, the only reflection that can be seen in the patterns is the anatase (004) peak. This indicates crystal growth with the anatase c axis normal to the substrate surface, so that film and substrate have the epitaxial relationship: (001)TiO2 ||(001)STO

The mosaic spread for three of the samples was investigated by scanning through the (004) and (008) epilayer reflections along the [100] direction in reciprocal space. The full widths at half-maximum height ΔQ[100] in these transverse scans decreased slightly from 0.043 reciprocal lattice units for a four-dip 700 °C sample to 0.037 reciprocal lattice units for twoand four-dip samples annealed at 1000 °C. The full widths at half-maximum height were found to be twice as large in the scan through the (008) peak so that the ratio ΔQ[100]/Q[001] remains roughly constant, where Q[001] is the longitudinal wavevector transfer. This shows that the broad width of the peaks arises mainly through mosaic spread rather than domain size effects.35,36 Given that the wavevector transfer for the (004) peak is 1.64, these results suggest a mosaic spread of tan−1(ΔQ[100]/Q[001]), i.e., 1.5° for the four-dip sample

Figure 4. Experimentally determined c-parameter of anatase TiO2 film in the range of annealing temperature 700−1000 °C. Data for four-dip and eight-dip films are shown as well as for the two-dip films that form the main focus of the current paper. 1440

dx.doi.org/10.1021/cg301455q | Cryst. Growth Des. 2013, 13, 1438−1444

Crystal Growth & Design

Article

Figure 5. Left hand panels: reciprocal space scans through the anatase (013) reflection along the [010] direction for four-dip samples annealed at (a) 700 °C (b) 1000 °C. The dashed lines indicate the position of the peak for bulk anatase. Right hand panels: reciprocal space scans through the anatase (013) reflection along the [100] direction for four-dip samples annealed at (c) 700 °C and (d) 1000 °C. The dashed lines indicate the position of the peak expected when (010)TiO2∥(010)STO.

Figure 6. (a) Cross-sectional TEM image of an anatase island viewed down the [010] crystal direction of the substrate; (b) selected area electron diffraction pattern from the SrTiO3 substrate; (c) selected area electron diffraction pattern from the anatase TiO2 epilayer.

increases progressively above the bulk value, rather than converging toward it. This is a puzzling result in a system where it might be assumed that the epilayer should be placed under tensile stress by the substrate. One scenario is that the anatase films annealed at high temperatures may be placed under compressive stress as the film is cooled down from the annealing temperature. In principle this situation can arise because the coefficient of thermal expansion of anatase in the basal plane (α = 3.5 × 10−6 K−1 at room temperature, but increasing at higher temperature)10 is very much smaller than that of SrTiO3 (α = 29.8 × 10−6 K−1).38 However direct

stress from the substrate. After low temperature deposition, the epilayer is strained to match the substrate, leading to a contraction in the out-of-plane lattice spacing via the Poisson effect. Strain is expected to be larger close to the interface than toward the top of the film, leading to asymmetric broadening of out-of-plane Bragg peaks on the high angle side, as observed here. Higher temperatures allow strain relaxation and thus an increase in the out-of-plane lattice constant.37 However, a major problem with this simple argument is that for the sample calcined at 700 °C the c lattice constant is already close to the bulk value and as the calcination temperature increases c 1441

dx.doi.org/10.1021/cg301455q | Cryst. Growth Des. 2013, 13, 1438−1444

Crystal Growth & Design

Article

Table 2. Values of the Surface Energies of the Anatase Phase of TiO2 from References 39−42 and of SrTiO3 from Reference 43 surface energy J m−2

a

TiO2(100)

TiO2(001)

TiO2(101)

2.02a 0.71b 0.53c,d

2.79a 1.08b 0.90c,d

1.73a 0.61b 0.44c,d

TiO2(103) faceted 1.05b 0.83c,d

TiO2(103) smooth 1.14b 0.93c,d

SrTiO3(001) 1.36 (SrO)e 1.41 (TiO2)e

Ref 39. bRef 40. cRef 41. dRef 42 eRef 43 (HFLYP functional).

examination of the in-plane lattice spacing by scanning through the (013) reflection along the [010] direction in reciprocal space (Figure 5a,b) shows that the in-plane lattice spacing for samples annealed at both 700 and 1000 °C is very close to the value expected from the bulk lattice parameter of anatase. We also note that scans through the (013) reflection along the orthogonal [100] direction (Figure 5c,d) show that the [010] direction of the epilayer coincides with that of the substrate. In summary, out-of-plane expansion does not appear to arise from in-plane compression. Moreover, similar variations in lattice parameter have been found for anatase deposited on LaAlO3(001) by a sol−gel route, even though there is very much better lattice matching in this system.29 These considerations suggest the lattice parameter variations are mainly due to variations associated with introduction of defects at higher temperatures. In a future publication we will present a detailed study of in-plane and out-of-plane lattice constants as a function of temperature for epitaxial anatase on SrTiO3 using synchrotron-based X-ray diffraction. Figure 6a shows cross-sectional HRTEM images of an anatase film (two dips) after annealing at 1000 °C. The images confirm that the sample has grown with the anatase (001) axis normal to the substrate surface. The image also reveals a reasonably sharp interface and lateral registry so that

where: A A Γ1 = (γ103 cosec θ103 − γ101 cosec θ101) i Γ2 = γ001 × (cot θ103 − cot θ101) A STO Γ3 = (γ001 + γ001 ) × (cot θ103 − cot θ101)

Here the γA and γSTO are surface energies for anatase TiO2 and SrTiO3 respectively, and γi001 is the interface energy. Using the surface energies for anatase calculated by Lazzeri et al.,41,42 along with surface energies for SrTiO3 calculated using the HFLYP functional Γ1 turns out to be positive with a numerical value of 0.82 J m−2. However, owing to the high surface energies for anatase TiO2(001) and SrTiO3(001),43 this term is outweighed by (1/2)Γ3, which has numerical values of 0.90 J m−2 and 0.94 J m−2 respectively for SrO and TiO2 terminated SrTiO3(001) surfaces. Thus provided the interface energy is less than about 0.20 J m−2 development of {103} side facets is favored on the SrO terminated surface (for the higher energy TiO2 terminated surface the corresponding value is 0.30 J m−2). Given the reasonable epitaxial match between anatase(001) and SrTiO3(001), a low value for the interface energy seems plausible. These considerations can be extended to analyze the appearance of {103} rather than {101} facets on the tops of the islands. Here the difference in energy for a facet of height ht and edge length lt is given by

(001)TiO2 ||(001)STO (100)TiO2 ||(100)STO

Et{103} − Et{101} = ltht[Γ1 − Γ4]

These epitaxial relationships are further confirmed by selected area electron diffraction patterns shown in Figure 6b,c. The top surface of the anatase island are parallel to the substrate and are of (001) orientation. However, there are well developed sloping side facets which make a contact angle of about 40° with the substrate. This establishes that these facets belong to the {103} family with a calculated contact angle θ103 = 39.94° rather than to {101}, where the contact angle is θ101 = 68.30°. This finding is somewhat surprising as the surface energy for anatase-TiO2(103) is calculated to be significantly higher than for anatase-TiO2(101)39−42 (see Table 2, which also includes calculated surface energies for SrTiO3(001)43). A Wulff construction suggests that {101} facets should predominate in free-standing anatase nanocrystals,39,41 and this prediction from theory is confirmed by experimental observations of the morphology of nanocrystals.44,45 However for an island of fixed volume, development of sloping side facets reduces the area of bare SrTiO3(001) substrate and the top anatase(001) surface and at the same time increases the interface area. Considering a square island of height he and edge length le, the difference in surface and interface energies which differentiates between development of {103} and {101} side-facets on the edges of the islands is given by

where Γ1 is as before and A Γ4 = γ001 × (cot θ103 − cot θ101)

Γ4 has a numerical value of 0.72 J m−2 and so Γ1 > Γ4. The energetic difference is quite small, but on this basis {101} top facets should be favored. Because this situation does not pertain experimentally in our work, we must infer that the calculated (103) surface energy is a little too high relative to (101). In this context it should be noted that in a structural study of an anatase TiO2(001) surface prepared by chemical vapor deposition on SrTiO3(001), Herman et al.46 found that the (1 × 1) anatase TiO2(001) surface reconstructs into a twodomain (1 × 4) + (4 × 1) microfaceted surface upon sputtering and annealing. Analysis of angle-resolved mass spectroscopy of recoiled ions (AR-MSRI) favored a model involving (103) and (1̅03) microfacets. This further supports the hypothesis of a low (103) surface energy. However, we note that Marshall et al.47 recently found that anatase islands formed on SrTiO3(001) surfaces by prolonged annealing in ultrahigh vacuum developed {101} side facets, so the two energies appear to be finely balanced and may be influenced by the oxygen partial pressure.

■

CONCLUSIONS It is well established that mismatch between the lattice parameters for a substrate and an epitaxial film can promote

⎡ 1 1 ⎤ Ee{103} − Ee{101} = 4lehe⎢Γ1 + Γ2 − Γ3⎥ ⎣ 2 2 ⎦ 1442

dx.doi.org/10.1021/cg301455q | Cryst. Growth Des. 2013, 13, 1438−1444

Crystal Growth & Design

Article

(14) Yoganarasimhan, S. R.; Rao, C. N. R. Trans. Faraday Soc. 1962, 58, 1579−1589. (15) Czanderna, A. W.; Clifford, A. F.; Honig, J. M. J. Am. Ceram. Soc. 1957, 79, 5407−5409. (16) Iida, Y.; Ozaki, S. J. Am. Ceram. Soc. 1961, 44, 120−127. (17) Shannon, R. D.; Pask, J. A. J. Am. Ceram. Soc. 1965, 48, 391− 398. (18) Murakami, M.; Matsumoto, Y.; Nakajima, K.; Makino, T.; Segawa, Y.; Chikyow, T.; Ahmet, P.; Kawasaki, M.; Koinuma, H. Appl. Phys. Lett. 2001, 78, 2664−2666. (19) Park, B. H.; Huang, J. Y.; Li, L. S.; Jia, Q. X. Appl. Phys. Lett. 2002, 80, 1174−1176. (20) Matsumoto, Y.; Murakami, M.; Shono, T.; Hasegawa, T.; Fukumura, T.; Kawasaki, M.; Ahmet, P.; Chikyow, T.; Koshihara, S.; Koinuma, H. Science 2001, 291, 854−856. (21) Ong, C. K.; Wang, S. J. Appl. Surf. Sci. 2001, 185, 47−51. (22) Kennedy, R. J.; Stampe, P. A. J. Cryst. Growth 2003, 252, 333− 342. (23) Jellison, G. E.; Boatner, L. A.; Budai, J. D.; Jeong, B. S.; Norton, D. P. J. Appl. Phys. 2003, 93, 9537−9541. (24) Shao, R.; Wang, C. M.; McCready, D. E.; Droubay, T. C.; Chambers, S. A. Surf. Sci. 2007, 601, 1582−1589. (25) Weng, X.; Fisher, P.; Skowronski, M.; Salvador, P. A.; Maksimovc, O. J. Cryst. Growth 2008, 310, 545−550. (26) Xu, L. M.; Yang, M.; Li, X. Y.; Hu, P.; Li, S. W. Scr. Mater. 2010, 63, 113−116. (27) Chambers, S. A.; Thevuthasan, S.; Farrow, R. F. C.; Marks, R. F.; Thiele, J. U.; Folks, L.; Samant, M. G.; Kellock, A. J.; Ruzycki, N.; Ederer, D. L.; Diebold, U. Appl. Phys. Lett. 2001, 79, 3467−3469. (28) Chambers, S. A.; Wang, C. M.; Thevuthasan, S.; Droubay, T.; McCready, D. E.; Lea, A. S.; Shutthanandan, V.; Windisch, C. F. Thin Solid Films 2002, 418, 197−210. (29) Lotnyk, A.; Senz, S.; Hessa, D. Thin Solid Films 2007, 515, 3439−3447. (30) Jung, H. S.; Lee, J. K.; Lee, J.; Kang, B. S.; Jia, Q.; Nastasi, M. J. Phys. Chem. C 2008, 112, 4205−4208. (31) Chan, K. Y. S.; Goh, G. K. L. J. Electrochem. Soc. 2009, 156, D231−D235. (32) Zhang, Z. M. J. Mater. Res. 2005, 20, 292−294. (33) Brown, S. D.; Bouchenoire, L.; Bowyer, D.; Kervin, J.; Laundy, D.; Longfield, M. J.; Mannix, D.; Paul, D. F.; Stunault, A.; Thompson, P.; Cooper, M. J.; Lucas, C. A.; Stirling, W. G. J. Synchrotron Radiat. 2001, 8, 1172−1181. (34) Coll, M.; Gazquez, J.; Pomar, A.; Puig, T.; Sandiumenge, F.; Obradors, X. Phys. Rev. B 2006, 73, No.075420−1−8. (35) Babkevich, A. Y.; Cowley, R. A.; Mason, N. J.; Sandiford, S.; Stunault, A. J. Phys.: Condens Matter 2002, 14, 7101−7121. (36) Regoutz, A.; Zhang, K. H. L.; Egdell, R. G.; Wermeille, D.; Cowley, R. A. J. Mater. Res. 2012, 27, 2257−2264. (37) Zhang, K. H. L.; Lazarov, V. K.; Lai, H. H. C.; Egdell, R. G. J. Cryst. Growth 2011, 318, 345−350. (38) deLigny, D.; Richet, P. Phys. Rev. B 1996, 53, 3013−3022. (39) Olson, C. L.; Nelson, J.; Islam, M. S. J. Phys. Chem. B 2006, 110, 9995−10001. (40) Zhao, Z. Y.; Li, Z. S.; Zou, Z. G. J. Phys. - Condens. Matter 2010, 22, No. 175008−1−6. (41) Lazzeri, M.; Vittadini, A.; Selloni, A. Phys. Rev. B 2001, 63, No.155409−1−9. (42) Lazzeri, M.; Vittadini, A.; Selloni, A. Phys. Rev. B 2002, 65, No. 119901(E)−1. (43) Heifets, E.; Eglitis, R. I.; Kotomin, E. A.; Maier, J.; Borstel, G. Phys. Rev. B 2001, 64, 235417−1−5. (44) Yang, H. G.; Sun, C. H.; Qiao, S. Z.; Zou, J.; Liu, G.; Smith, S. C.; Cheng, H. M.; Lu, G. Q. Nature 2008, 453, 638−641. (45) Jun, Y.; Casula, M. F.; Sim, J.-H.; Kim, S. Y.; Cheon, J.; Alivisatos, A. P. J. Am. Chem. Soc. 2003, 125, 15981−15985. (46) Herman, G. S.; Sievers, M. R.; Gao, Y. Phys. Rev. Lett. 2000, 84, 3354−3357.

the development of self-assembled structures, usually on a nanometer length scale. Extensively studied examples include the quantum dots structures obtained in the hetero-epitaxial growth of III−V or elemental semiconductors by molecular beam epitaxy. Well known examples here are InAs quantum dots on GaAs, where there is a 7% lattice mismatch,48−50 and dome-shaped Ge islands grown on Si substrate.51,52 Island formation and dewetting processes have also been studied in a number of oxide on oxide heteroepitxial systems including BaTiO3 on anatase53 and PbTiO3,54 PbZr0.52Ti0.48O3,55 and SrZrO356 on SrTiO3. The present contribution extends this work to a phase of TiO2 thermodynamically unstable as freestanding bulk material. The morphology of the supported anatase nanocrystals prepared in the present work is very different to that of unsupported nanocrystalline material. In particular, {101} facets which are predominant for unsupported anatase nanocrystals are completely absent for the epitaxial islands prepared in the current work. At the same time, the epitaxial matching with the substrate extends the range of thermal stability of the anatase far beyond that found for unsupported material.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Present Addresses ∥

(F.E.O.) Department of Materials, Imperial College London, Exhibition Road, London, SW7 2AZ, UK. ⊥ (K.H.L.Z.) Pacific Northwestern National Laboratory, Chemical and Material Sciences Division, Material Sciences P.O. Box 999, K8−87, Richland, WA 99352, USA. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS F.E.O. was supported by Pembroke College Oxford and FUNDAYACUCHO (Caracas, Venezuela). K.H.L.Z. would like to thank the Oxford Clarendon Fund for financial support.

■

REFERENCES

(1) Asahi, R.; Morikawa, T.; Ohwaki, T.; Aoki, K.; Taga, Y. Science 2001, 293, 269−271. (2) Di Valentin, C.; Finazzi, E.; Pacchioni, G.; Selloni, A.; Livraghi, S.; Paganini, M. C.; Giamello, E. Chem. Phys. 2007, 339, 44−56. (3) Qiu, X. F.; Burda, C. Chem. Phys. 2007, 339, 1−10. (4) Linsebigler, A. L.; Lu, G. Q.; Yates, J. T. Chem. Rev. 1995, 95, 735−758. (5) Wong, J. C. S.; Linsebigler, A.; Lu, G. Q.; Fan, J. F.; Yates, J. T. J. Phys. Chem. 1995, 99, 335−344. (6) Carp, O.; Huisman, C. L.; Reller, A. Prog. Solid State Chem. 2004, 32, 33−177. (7) Mo, S. D.; Ching, W. Y. Phys. Rev. B 1995, 51, 13023−13032. (8) Labat, F.; Baranek, P.; Domain, C.; Minot, C.; Adamo, C. J. Chem. Phys. 2007, 126, No.154703−1−12. (9) Tang, H.; Prasad, K.; Sanjines, R.; Scmidt, P. E.; Levy, F. J. Appl. Phys. 1994, 75, 2042−2047. (10) Rao, K. V. K.; Naidu, S. V. N.; Iyengar, L. J. Am. Ceram. Soc. 1970, 53, 124−126. (11) Semiconductors: Physics of Non-Tetrahedrally Bonded Binary Compounds II; Springer Verlag: Berlin, 1984; Vol. 17, Pt. f. (12) Tang, H.; Levy, F.; Berger, H.; Schmid, P. E. Phys. Rev. B 1995, 52, 7771−7774. (13) Czanderna, A. W.; Rao, C. N. R.; Honig, J. M. Trans. Faraday Soc. 1958, 54, 1069−1073. 1443

dx.doi.org/10.1021/cg301455q | Cryst. Growth Des. 2013, 13, 1438−1444

Crystal Growth & Design

Article

(47) Marshall, M. S. J.; Castell, M. R. Phys. Rev. Lett. 2009, 102, 146102−22−1−4. (48) Marzin, J. Y.; Gerard, J. M.; Izrael, A.; Barrier, D.; Bastard, G. Phys. Rev. Lett. 1994, 73, 716−719. (49) Schaffer, W. J.; Lind, M. D.; Kowalczyk, S. P.; Grant, R. W. J. Vac. Sci. Technol. B 1983, 1, 688−695. (50) Yoffe, A. D. Adv. Phys. 2001, 50, 1−208. (51) Medeiros-Ribeiro, G.; Bratkovski, A. M.; Kamins, T. I.; Ohlberg, D. A. A.; Williams, R. S. Science 1998, 279, 353−355. (52) Ross, F. M.; Tromp, R. M.; Reuter, M. C. Science 1999, 286, 1931−1934. (53) Zhao, L.; Chien, A. T.; Lange, F. F.; Speck, J. S. J. Mater. Res. 1996, 11, 1325−1328. (54) Seifert, A.; Vojta, A.; Speck, J. S.; Lange, F. F. J. Mater. Res. 1996, 11, 1470−1482. (55) Szafraniak, I.; Harnagea, C.; Scholz, R.; Bhattacharyya, S.; Hesse, D.; Alexe, M. F. Appl. Phys. Lett. 2003, 83, 2211−2213. (56) Langjahr, P. A.; Wagner, T.; Ruhle, M.; Lange, F. F. J. Mater. Res. 1999, 14, 2945−2951.

1444

dx.doi.org/10.1021/cg301455q | Cryst. Growth Des. 2013, 13, 1438−1444