Fabrication of nanometer-sized anatase particles by a pulsed laser ablation method K. Kawasaki,a J. F. Despres,a S. Kamei,b M. Ishikawab and O. Odawara*a aDepartment of Electronic Chemistry, T okyo Institute of T echnology, 4259 Nagatsuta,Midori-ku, Yokohama 226, Japan bMitsubishi Research Institute, Inc., 2-3-6 Otemachi, Chiyoda-ku, T okyo 100, Japan TiO2 particles of<10 nm have been prepared by ablating a rutile target rod with a pulsed laser beam, which was assisted by the synchronized injection of an Ar–O2 mixture.By controlling the wavelength of laser beam, injecting pressure ( p) and O2 concentration in the gas mixture, ultrasmall anatase particles could be obtained. In the present work, l=532 nm, p=5×105 Pa and O2 concentration<5% resulted in anatase particles of <6 nm.TiO2 has been used in various optical materials including Experimental semiconductor photocatalysts since the pioneering work by Laser ablation experiment Honda and Fujishima.1 Nanometer-sized TiO2 particles show size dependence on optical properties (‘quantum size eVect’), Laser ablation was carried out on a TiO2 rod target prepared which corresponds to a shift of optical absorption spectrum.2–4 by pressing TiO2 powder and sintering at 1400 °C for 2 h.The The bandgap shift (DEg) relates to the radius (R) of nanometer- apparatus was a combination of a molecular beam generator17 sized particles as follows;3,5 and an ablation system as shown in Fig. 1. The vacuum system employed in the ablation experiments consisted of a stainlesssteel chamber pumped by three pumps; a 900 l s-1 18 in.DEg= h2 8R2C1 me + 1 mhD- 1.8e2 4peR (1) diVusion pump, a 3000 l s-1 10 in diVusion pump, and a 1500 l s-1 turbomolecular pump. The pressure in the chamber was <1.2×10-8 Pa. where h is Planck’s constant, me and mh are the eVective mass The TiO2 target was irradiated by a Q-switched Nd5YAG of the electron and hole, e is the electron charge, and e is the laser.The wavelength of irradiating light was varied from the relative permittivity of the semiconductor. Ultrasmall TiO2 fundamental to the second and fourth harmonic of the particles are useful materials for highly activated photocata- Nd5YAG laser. In this study, 266 nm (5 ns pulse width, 70 lysts. Anpo et al.2 reported that the photocatalytic activity of mJ pulse-1), 532 nm (5 ns, 260 mJ pulse-1) and 1064 nm (6 ns, TiO2 particles become stronger with decrease in particle size, 455 mJ pulse-1) were used.The repetition time was 10 Hz. and rutile particles below 12 nm and anatase particles below The target TiO2 rod was rotated (1 rpm) so that the laser 8.5 nm result in a large blue shift and large quantum yield. could irradiate the clear surface of the target constantly during There have been many studies concerning other ultrasmall the rotation. The target irradiated by the laser was excited and semiconductor particles. For example, CdS and ZnS crystalproduced a plasma plume on its surface.A high pressure line5–8 and colloidal9 nanoparticles of <10 nm have attracted Ar–O2 gas mixture (purity >99.9% for both gases) was particular interest owing to the large blue shift of their absorpinjected into this plume for ca. 200 ms to cool it, injection being tion spectra and photoluminescence. synchronized with the laser pulse. The synchronization was Nanometer-sized particles have been mainly prepared by very sensitive because of a time delay for the injected gas to sol–gel techniques, and TiO2 particles have been prepared by reach the plume on the surface of the target and the gas was hydrolysis of TiCl4.2–4 Other methods such as laser induced injected 1 ms earlier than the pulse.The injection pressure of reactions, pyrolysis of titanium isopropoxide10 and titanium the mixture gas was varied up to 5×105 Pa and six conditions alkoxide,11 and ignition of TiCl4–H2–O2,12 have also been were tested.The concentration of O2 was varied from 1 to 5 reported. In these studies, a low photon energy CO2 laser (l= to 10%. 10.6 mm) has been used to provide heat for reactant gases to produce a very high reaction temperature over a short time which prevents dissociation of gas molecules. To discuss the quantum size eVect more clearly, nanometersized particles require high purity since impurities lead to a serious influence on the bandgap of the semiconductor.Laser ablation is a novel technique for fabrication of homogeneous nanometer-sized particles incorporated with ultrahigh vacuum systems. The technique is to evaporate the ceramic target by irradiation of a high-energy pulsed laser assisted with injection of high-pressure noble or reactive gas synchronized with the laser pulse.For laser ablation, a high photon energy Nd5YAG laser (fundamental wavelength=1064 nm) and ArF (193 nm) or KrF (248 nm) excimer lasers have been used. There have been few studies about TiO2 fabrication except for thin films.13–16 In this study, ultrasmall TiO2 particles of <10 nm have been fabricated by pulsed Nd5YAG laser ablation and conditions of fabrication and parameters con- Fig. 1 Schematic diagram of the apparatus trolling particle size are discussed. J. Mater. Chem., 1997, 7(10), 2117–2120 2117Analysis of the products and Ti ion intensity is shown in Fig. 3. The TiO2 ion intensity increased with injection gas pressure, whereas the Ti ion The cluster ions of the plume were analyzed by a Q-mass intensity decreased.This result indicates that TiO2 species are analyzer. Results from the analyzer may be somewhat diVerent generated by collisions with high-injection pressure gas. For from the actual distributions because of ionization ratio eVects. the limitation of vacuum capability, the injection pressure was However, in this study, the ionization source was a highlimited to 5×105 Pa, and examination of laser wavelength and energy laser and laser excitation was not influenced by coexist- O2 concentration dependence was carried out on products at ent elements.Therefore, the ionization ratio of each ion was 5×105 Pa. estimated to be equal and the ion intensity analysis is thus believed to be accurate. Laser wavelength dependence Particles were captured on carbon coated grids (Fig. 2) and investigated by transmission electron microscopy (TEM). Fig. 4 shows the laser wavelength dependence of ion intensity of TiO2 and Ti in the plasma plume. There was a large amount Particles sizes were determined using the image mode and crystal structures have been identified using selected area of Ti ions and the TiO2 ion intensity using a wavelength of 266 nm was much lower than of Ti.On the other hand, TiO2 electron diVraction (SAED). Dependence of injection pressure, laser wavelength, and O2 ion intensity with 532 and 1064 nm wavelength light was four orders of magnitude higher than that at 266 nm and was concentration in the gas mixture were examined to determine the optimum fabrication conditions. Dependence of injection almost equal to the Ti ion intensity.The influence of laser power of each wavelength was not considered because the pressure was examined by ion intensity analysis using a Qmass analyzer. Laser wavelength dependence was related to behavior of Ti and TiO2 was obviously diVerent. The laser wavelength dependence might be related to the the ion intensity and the captured particle size and structure. The O2 concentration dependence was related to the size diVerence of the excitation mechanism of the target material.The diVerence between 266, 532 and 1064 nm light is due to of particles. the photon energy (4.66, 2.33 and 1.17 eV respectively). The crystal structure of the target was rutile-type, and its maximum Results and Discussion absorption wavelength is at 413 nm, so the target absorbs only 266 nm laser light.The target which absorbs laser light is Injection pressure dependence electronically highly excited and entirely decomposed into Ti TiO2 nanocrystals were obtained by injection gas cooling of and O ions. In this process, the probability of Ti and O ion the plasma plume. Thus pressure is one of the most important recombination is very low.Therefore, 266 nm wavelength laser parameters for the fabrication and it was directly related to the cooling rate. The injection pressure dependence of TiO2 Fig. 4 Laser wavelength dependence on TiO2 and Ti ion intensity (gas Fig. 2 Schematic diagram of the capture of particles mixture: Ar 90%–O2 10%, injection pressure: 5×105 Pa) Fig. 3 Injection gas pressure dependence on TiO2 and Ti ion intensity (gas mixture: Ar 90%–O2 10%) 2118 J.Mater. Chem., 1997, 7(10), 2117–2120Fig. 5 TEM and SAED images of spherical particles prepared by Ar 90%–O2 10% gas mixture injection. Wavelength (a) 532 nm, (b) 1064 nm. Fig. 6 TEM image of TiO2 particles prepared by Ar 95%–O2 5% gas mixture injection using a 532 nm wavelength laser was not appropriate for generating TiO2 species. 532 and 1064 nm wavelength laser breaks TiMO bonds by thermal excitation and in this case, many titanium oxide ions exist in the plasma (mainly TiO ions14,15). In all cases, ions in the plasma can react by injection of an Ar and O2 gas mixture. Oxygen atoms of oxygen gas attach to titanium ions or titanium oxide ions to form TiO2 species [the TiMO bond (6.92 eV) is stronger than OMO (5.12 eV)18,19].From these results, 532 or 1064 nm wavelength laser was required for formation of fine TiO2 particles. The most suitable excitation wavelength (532 or 1064 nm) could not be diVerentiated by ion intensity analysis. The particles prepared using 532 and 1064 nm light were captured and investigated by TEM and SAED as shown in Fig. 5. At 1064 nm, the average particle size was large with a wide distribution (particles>100 nm were found).The crystal structure of these particles corresponded to the rutile-type according to SAED. 1064 nm light breaks only a few TiMO bonds of the target because its photon energy is very low compared with the bond energy of Ti–O; the big particles are simply fragments of the target, and do not arise from the process of recombination of oxygen and Ti ions or titanium oxide ions.By contrast, in the image of the product obtained with 532 nm wavelength, few large particles were observed, and almost all particles were small with comparatively uniform size. SAED reveals a ‘powder pattern’ whose peaks correspond to an anatase-type structure, which is strongly suggested to be stoichiometric.From these results, 532 nm wavelength laser light is appropriate for formation of homogeneous TiO2 particles. O2 concentration dependence Fig. 7 Size distributions of TiO2 particles prepared by 532 nm wave- The oxygen content in the gas mixture aVects the reaction in length laser. Gas mixture: (a) Ar 99%–O2 1%, (b) Ar 95%–O2 5%, (c) Ar 90%–O2 10%. the plasma plume.No reaction could occur in the case of Ar gas injection. Therefore, O2 content plays an important role in the product particle size and particles were prepared at low O2 concentration gas injection. The sizes of almost all particles all experiments. The oxygen concentration becomes lower when the collision frequency of oxygen decreases, which leads prepared using 5% O2 were <10 nm as shown in Fig. 6. The size distributions of the particles prepared by injection of 1, 5 to a slower reaction rate with titanium and titanium oxide ions, therefore, in the short reaction time, injection of a low and 10% O2 concentration in the gas mixture are shown in Fig. 7. The data at 1 and 5% diVered little but the distributions O2 concentration (<5% in the gas mixture) corresponds to a small amount of additional oxygen atoms leading to the were narrower and the average particle size smaller than that at 10% O2.These results indicate that small and uniform formation of ultrasmall TiO2 particles of <10 nm. particles are formed by injection of a low O2 concentration gas mixture. Addition of oxygen to titanium and titanium Conclusions oxide ions occurs only in the plasma over a very short time and limited area.In the plasma, not only titanium and titanium We have succeeded in the fabrication of ultrasmall anatase particles of <10 nm by a pulsed Nd5YAG laser ablation oxide ions but also a large amount of oxygen ions could be identified by Q-mass analysis. Oxygen ions are lighter than method. 532 nm wavelength laser light and high-pressure gas mixture injection were required for formation of small and titanium ions, and they primarily fly out from the reaction zone.The collision frequency does not depend on O2 concen- uniform TiO2 particles. The O2 concentration of the injected gas mixture was an important parameter in controlling particle tration in this work since the injection pressure was equal in J.Mater. Chem., 1997, 7(10), 2117–2120 21197 R. Rossetti, R. Hull, J. M. Gibson and L. E. Brus, J. Chem. Phys., size. The average particle size was ca. 6 nm at <5% O2 1985, 82, 552. concentration in the gas mixture. 8 A. P. Alivisatos, J. Phys. Chem., 1996, 100, 13 226. Laser ablation is one of the most useful techniques to 9 S. Baral, A. Fojtik, H. Weller and A.Henglain, J. Am. Chem. Soc., fabricate ultrasmall semiconductor particles of <10 nm. We 1986, 108, 375. are now studying the spectroscopic properties of plasma to 10 G. W. Rice, J. Am. Ceram. Soc., 1987, 70, C-117. 11 L. E. Depero, P. Bonzi, M. Zocchi, C. Cesale and D. D. Michele, understand the fabrication mechanism and are also investiga- J.Mater. Res., 1993, 8, 2709.ting the optical properties of TiO2 nanoparticles synthesized 12 T. Oyama, Y. Iimura, K. Takeuchi and T. Ishii, J.Mater. Sci. L ett., in this study for application as a highly activated photocatalyst. 1996, 15, 594. 13 H. Funakoshi, K. Fumoto, M. Okuyama and Y. Hamakawa, Jpn. J. Appl. Phys., 1994, 33, 5262. This research is conducted as a part of the industrial technology 14 H.A. Durand, J. H. Brimaud, O. Hellman, H. Shibata, S. Sakuragi, development promotion program of the Research Institute of Y. Makita, D. Gesbert and P. Meyrueis, Appl. Surf. Sci., 1995, Innovative Technology for the Earth. 86, 122. 15 N. Lobstein, E. Millon, A. Hachimi, J. F. Muller, M. Alnot and J. J. Ehrhardt, Appl. Surf. Sci., 1995, 89, 307. 16 C. Garapon, C. Champeaux, J. Mugnier, G. Panczer, P. Marchet, References A. Catheriot and B. Jacquier, Appl. Surf. Sci., 1996, 96–98, 836. 1 A. Fujishima and K. Honda, Nature (L ondon), 1972, 238, 37. 17 S. Kamei, M. Ishikawa, T. Hashimoto, S. Ishizuka and 2 M. Anpo, T. Shima, S. Kodama and Y. Kubokawa, J. Phys. Chem., I. Kusunoki, Bull. Res. Inst. Sci. Meas., T ohoku Univ., 1995, 44, 41. 1987, 91, 4305. 18 B. C. Guo, K. P. Kerns and A. W. Castleman, Jr., Int. J. Mass 3 E. Joselevich and I. Willner, J. Phys. Chem., 1994, 98, 7628. Spectrom. Ion Processes, 1992, 117, 129. 19 D. E. Clemmer, J. L. Elkind, N. Aristov and P. B. Armentrout, 4 N. Serpone, D. Lawless and R. Khairutdinov, J. Phys. Chem., 1995, J. Chem. Phys., 1991, 95, 3387. 99, 16 646. 5 L. E. Brus, J. Phys. Chem., 1986, 90, 2555. 6 L. E. Brus, J. Chem. Phys., 1984, 80, 4403. Paper 7/03816J; Received 2nd June, 1997 2120 J. Mater. Chem., 1997, 7(10), 2117–2120