Journal of Alloys and Compounds 397 (2005) L1L4

Letter

Solgel synthesis and characterization of nanocrystalline ZnO powders

Division of Materials Chemistry, Ruder Bokovi Institute, P.O. Box 180, Bijeni ka Cesta 54, HR-10002 Zagreb, Croatia s c c b Department of Physics, Faculty of Science, University of Zagreb, P.O. Box 331, HR-10002 Zagreb, Croatia Received 8 January 2005; accepted 28 January 2005 Available online 9 March 2005

Abstract Nanocrystalline zinc oxide (ZnO) powders were prepared by fast hydrolysis of zinc 2-ethylhexanoate dissolved in 2-propanol, adding a tetramethylammonium (TMAH) aqueous solution. XRD showed an average value of 2535 nm for the basal diameter of supposed cylinder (prism)-shaped crystallites, whereas the height of these crystallites was 3545 nm. TEM showed that the size of the majority of ZnO particles varied between 20 and 50 nm, thus indicating that particle and crystallite sizes in ZnO powders were approximately equal. The size of ZnO particles did not change signicantly for different amounts of zinc 2-ethylhexanoate in the precipitation systems investigated. Raman spectra of ZnO particles were interpreted taking into account the nanosize effect. 2005 Elsevier B.V. All rights reserved.
Keywords: Solgel; Nanocrystalline ZnO; XRD; FT-IR; Raman; TEM

1. Introduction Zinc oxide (ZnO) powders have found important applications in the production of paints, ceramics, catalysts, varistors, etc. The size of ZnO particles in these powders is a very important factor for the specic application. Chemical, microstructural and physical properties of ZnO powders are dependent on the synthesis procedure. Very different methods were used in the synthesis of ZnO powders. Musi et al. [1,2] investigated the inuence of the synthec sis procedure on the formation of ZnO powders. ZnO did not crystallize upon hydrothermal treatment of Zn(NO3 )2 aqueous solution containing urea, up to 160 C. Hydrozincite (Zn5 (CO3 )2 (OH)6 ) was precipitated instead. Upon heating at 260 C in air, Zn5 (CO3 )2 (OH)6 transformed to ZnO. The precipitates obtained by abrupt adding of a concentrated NH4 OH aqueous solution to the Zn(NO3 )2 solution yielded a complex compound, Zn5 (OH)8 (NO3 )2 (H2 O)2 x (NH3 )x , which upon autoclaving transformed to ZnO. Matijevi and coc workers [3,4] used urea and triethanolamine (TEA) process

Corresponding author. Tel.: +385 1 4680 107; fax: +385 1 4680 098. E-mail address: ristic@irb.hr (M. Risti ). c

ing to prepare uniform ZnO particles. Urea processing [5] was also used to prepare ZnO powders, which were further utilized in the preparation of ZnO varistors. Taubert and Wegner [6] investigated the formation of monodisperse ZnO particles by the hydrolysis of Zn2+ ions in a decomposing hexamethylenetetramine (HMTA) solution, into which a soluble starch was also added. In recent years researchers have focused more on the synthesis and properties of ZnO nanoparticles due to their application in advanced technologies. ZnO nanoparticles were prepared [711] by adding dry LiOH or its ethanolic solution to the ethanolic solution of Zn(ac)2 2H2 O. Inubushi et al. [12] prepared ZnO nanoparticles by the reaction of Zn(acac)2 with NaOH in ethanol. Mondelaers et al. [13] proposed the preparation of ZnO nanoparticles via an aqueous acetatecitrate gelation method. Microemulsion method was also used [14,15] to produce ZnO nanoparticles. In the present work, we report about a novel laboratory procedure for the preparation of nanocrystalline ZnO powders. This procedure is based on the solgel method. In the present case, a tetramethylammonium hydroxide (TMAH) solution was added to the alcoholic solution of zinc 2ethylhexanoate. TMAH is a strong organic alkali, which can

0925-8388/$ see front matter 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jallcom.2005.01.045

L2

M. Risti et al. / Journal of Alloys and Compounds 397 (2005) L1L4 c

be used to easily adjust pH values near 14. The advantage of TMAH in relation to a strong inorganic alkali (for example, NaOH) at very high pH is that it does not contaminate metal oxides with alkali metal cations, which in the next step would affect the ohmic conductivity of the oxide material. In specic cases, the tetramethylammonium cation may also serve as a templating agent.

laser with = 514.5 nm, as the excitation source. The scattered light was analyzed by a Dilor Z-24 Raman spectrometer. TEM observation of the samples was performed using a Philips transmission electron microscope (model Morgagni 268).

3. Results and discussion 2. Experimental Zinc 2-ethylhexanoate containing 1% of ethylene glycol monomethylether supplied by Alfa Aeser was used. TMAH [(CH3 )4 NOH, 25%, w/w aqueous solution, electronic grade, 99.99% (metal basis)] was also supplied by Alfa Aeser . Water free 2-propanol and absolute ethanol supplied by Kemika were used. Twice distilled water used for washing of the precipitates was prepared in our laboratory. Chemical conditions for the precipitation of samples S1S6 are given in Table 1. Samples S1S6 were precipitated at room temperature (RT) as follows: onto a proper weight of zinc 2-ethylhexanoate (starting chemical), 90 ml of 2propanol was added. After strong mixing a clear (transparent) solution was obtained. Then, 10 ml of 25% TMAH was added to this solution under strong mixing. The formed colloidal suspensions (S1S4) were aged for 30 min, washed three times with ethanol and two times with twice distilled water, then dried at 60 C. Washing of the precipitates was performed using a Sorvall RC2-B ultraspeed centrifuge (maximum operational range, 20,000 rpm). Samples S5 and S6 were precipitated in the same way as samples S1S4 with only difference in the ageing of the colloidal suspension before the separation of a solid phase from the liquid phase. The ageing time for samples S5 and S6 was 24 h. All samples were characterized by XRD, FT-IR and Raman spectroscopies and TEM. X-ray powder diffraction (Philips counter diffractometer, model MPD 1880) measurements were performed at RT. FT-IR spectra were recorded at RT using a Perkin-Elmer spectrometer (model 2000). The FT-IR spectrometer was coupled with a personal computer loaded with the IRDM (IR data manager) program to process the recorded spectra. The specimens were pressed into small discs using a spectroscopically pure KBr matrix. Raman spectra were recorded using a coherent Innova-100 argon
Table 1 Chemical conditions for the preparation of nanocrystalline ZnO powders Sample S1 S2 S3 S4 S5 S6 ZEH (g) 0.982 1.496 1.995 5.002 1.009 5.010 2-Propanol (ml) 90 90 90 90 90 90 TMAH (ml) 10 10 10 10 10 10

ZEH: zinc 2-ethylhexanoate containing 1% ethylene glycol monomethylether, TMAH: tetramethylammonium hydroxide, 25% (w/w) aqueous solution.

In the present work, we have found that very ne ZnO particles can be produced by a fast hydrolysis of zinc 2ethylhexanoate dissolved in 2-propanol. A stable colloidal (milky) dispersion was obtained by adding aqueous solution of TMAH to the alcoholic solution of zinc 2-ethylhexanoate. A gradual coagulation of colloidal particles was observed upon complete addition of the TMAH solution. For that reason the suspensions S1S4 were aged for 30 min, whereas the suspensions S5 and S6 were aged for 24 h. The aim of prolonged ageing was to check for a possible inuence of the coagulation process on the properties of ZnO particles. All samples were identied by XRD as a single phase containing ZnO (PDF card nos.: 89-1397; 89-0511; 36-1451) [16]. Diffraction lines of ZnO were broadened, and diffraction broadening was found dependent on Miller indices of the corresponding sets of crystal planes. For most samples the diffraction line 0 0 2 was narrower than the line 1 0 1, and that one was narrower than the line 1 0 0. This indicated an asymmetry in the crystallite shape. It was supposed that crystallites were in the form of cylinder (prism), having the height (direction of the crystal c-axis) bigger than the basal diameter (crystal axes, a1 and a2 ). The crystallite size was obtained by measurements of the broadening of diffraction lines and applying the Scherrer equation, after a correction for the instrumental broadening. An average value of the basal diameter of the cylinder-shaped crystallites was 2530 nm, whereas the height of the crystallites was 3545 nm. Fig. 1 shows a characteristic part of the XRD pattern of sample S1 (nanocrystalline ZnO). The XRD pattern of commercial ZnO powder, exibiting only instrumental broadening, is also given for comparison. FT-IR spectra of samples S1S6 are shown in Fig. 2. Samples S1S4, aged as suspension for 30 min, showed a very broad IR band with the peak centered at 430 cm1 and shoulders at 535 and 400 cm1 , whereas S5 and S6 (aged as suspension for 24 h) showed a very broad and strong band centered at 430 cm1 with a shoulder at 535 cm1 . The shape of the IR spectrum of ZnO particles is generally inuenced by particle size and morphology, the degree of particles aggregation, or the crystal structure of the ZnO polymorph. Hayashi et al. [17] compared the recorded and calculated spectra of ZnO. ZnO particles showed three distinct absorption peaks located between the bulk TO-phonon frequency (T ) and the LO-phonon frequency (L|| ). These absorption peaks shifted towards lower frequencies when the permitivity of the surrounding medium was increased. Serna and co-workers [18,19] considered the relationship between the

M. Risti et al. / Journal of Alloys and Compounds 397 (2005) L1L4 c

L3

Fig. 1. Characteristic parts of X-ray powder diffraction patterns: (a) sample S1 (nanocrystalline ZnO) and (b) commercial ZnO powder. Measurements taken at room temperature.

Fig. 3. Raman spectra of samples S1S6. The spectra recorded at room temperature.

Fig. 2. FT-IR spectra of samples S1S6. The spectra recorded at room temperature.

shape of the IR spectrum on one side, and the physical shape and aggregation of ZnO particles on the other. Tanigaki et al. [20] prepared ZnO particles by the high-temperature oxidation of zinc powder. They observed different shapes in the spectra of ZnO particles as a result of the used preparation route. Samples S1S6 were also investigated by Raman spectroscopy. The corresponding spectra are given in Fig. 3. These spectra are characterized by a very strong band at 438 cm1 . The band at 332 cm1 is well pronounced; the bands at 392 and 418 cm1 , however, are visible as shoulders. A broad band at 578580 cm1 with a shoulder at 536542 cm1 and a very weak band at 652656 cm1 are also observed in these spectra. Damen et al. [21] investigated the Raman effect on ZnO crystals of the size of several millimeter, and measured: two E2 vibrations at 101 and 437 cm1 , one transverse A1 at 381 cm1 and one transverse E1 at 407 cm1 , one longitudinal A1 at 574 cm1 and one longitudinal E1 at 583 cm1 . Calleja and Cardona [22] investigated the resonance effect of Raman scattering on a ZnO single crystal by E2 , A1T , E1L and E1T phonons including several second-order features with photon energies between 1.6 and 3 eV. Exarhos and Sharma [23] applied Raman spectroscopy in the investigation of ZnO lms (wurtzite phase) and concluded that these lms exhibited a certain degree of residual stress inferred from the E2 Raman shift relative to the single crystal position of this mode. The shape of the Raman spectrum of ZnO was inuenced by doping of ZnO particles with antimony ions [24]. With an

L4

M. Risti et al. / Journal of Alloys and Compounds 397 (2005) L1L4 c

Acknowledgement The authors wish to thank Professor Nikola Ljubei for sc his assistance with electron microscopy.

References
[1] S. Musi , S. Popovi , M. Maljkovi , . Drag evi , J. Alloys Compd. c c c c c 347 (2002) 324. [2] S. Musi , . Drag evi , M. Maljkovi , S. Popovi , Mater. Chem. c c c c c Phys. 77 (2002) 521. [3] M. Castellano, E. Matijevi , Chem. Mater. 1 (1989) 78. c [4] Q. Zhong, E. Matijevi , J. Mater. Chem. 6 (1996) 443. c [5] E. Sonder, T.C. Quinby, D.L. Kinser, Am. Ceram. Soc. Bull. 65 (1986) 665. [6] A. Taubert, G. Wegner, J. Mater. Chem. 12 (2002) 805. [7] L. Spanhel, M.A. Anderson, J. Am. Chem. Soc. 113 (1991) 2826. [8] P.V. Kamat, B. Patrick, J. Phys. Chem. 96 (1992) 6829. [9] E.A. Meulenkamp, J. Phys. Chem. B 102 (1998) 5566. [10] V. Noack, A. Eychm ller, Chem. Mater. 14 (2002) 1411. u [11] M.S. Tokumoto, S.H. Pulcinelli, C.V. Santilli, V. Briois, J. Phys. Chem. B 107 (2003) 568. [12] Y. Inubushi, R. Takami, M. Iwasaki, H. Tada, S. Ito, J. Colloid Interface Sci. 200 (1998) 220. [13] D. Mondelaers, G. Vanhoyland, H. Van Den Rul, J. DHaen, M.K. Van Bael, J. Mullens, L.C. Van Poucke, Mater. Res. Bull. 37 (2002) 901. [14] M. Singhal, V. Chhabra, P. Kang, D.O. Shah, Mater. Res. Bull. 32 (1997) 239. [15] D. Kaneko, H. Shouji, T. Kawai, K. Kon-No, Langmuir 16 (2000) 4086. [16] International Centre for Diffraction Data, Joint Committee on Powder Diffraction Standards, Powder Diffraction File, 1601 Park Lane, Swarthmore, PA, USA. [17] S. Hayashi, N. Nakamori, H. Kanamori, Y. Yodogawa, K. Yamamoto, Surf. Sci. 86 (1979) 665. [18] M. Andr s-Verg s, A. Mifsud, C.J. Serna, J. Chem. Soc. Faraday e e Trans. 86 (1990) 959. [19] M. Andr s-Verg s, C.J. Serna, J. Mater. Sci. Lett. 7 (1988) e e 970. [20] T. Tanigaki, S. Kimura, N. Tamura, C. Kaito, Jpn. J. Appl. Phys. 41 (2002) 5529. [21] T.C. Damen, S.P.S. Porto, B. Tell, Phys. Rev. 142 (1966) 570. [22] J.M. Calleja, M. Cardona, Phys. Rev. B16 (1977) 3753. [23] G.J. Exarhos, S.K. Sharma, Thin Solid Films 270 (1995) 27. [24] J. Zuo, G. Xu, L. Zhang, B. Xu, R. Wu, J. Raman Spectr. 32 (2001) 979. [25] M. Risti , M. Ivanda, S. Popovi , S. Musi , J. Non-Cryst. Solids c c c 303 (2002) 270.

Fig. 4. TEM photographs of ZnO powders: (a) S1, (b) S2, (c) S5.

increase in the Sb-doping content up to 2.15% the band at 528 cm1 appeared, whereas the band at 580 cm1 shifted to 568 cm1 . Heating of Sb-doped ZnO at high temperatures caused a gradual decrease in the relative intensities of the bands at 528 cm1 and 568 cm1 in relation to the band at 433 cm1 . Raman spectra of samples S1S6, shown in Fig. 3, indicated distinct changes in relation to the Raman spectrum of big ZnO crystals. The Raman band at 392 cm1 , observed for samples S1S4, was located at 381 cm1 for big ZnO crystals. Also, the band at 407 cm1 , observed for big ZnO crystals, shifted to 418 cm1 for nanocrystalline ZnO. These effects, as well as the increase in relative intensities of the band at 580 cm1 and the shoulder at 542 cm1 and their broadening, can be assigned to the effect of the very small size of ZnO particles produced in the present work. The effect of very small size SnO2 particles, as well as the aggregation effect of these particles were considered in the interpretation of the Raman spectrum of SnO2 in our previous work [25]. TEM measurements have shown that very small ZnO particles are present in all powders. Fig. 4 shows TEM photographs of selected ZnO powders S1, S2 and S5. The size of the majority of ZnO particles in these powders varied between 20 and 50 nm. Taking into account the results of crystallite size measurements by XRD, it can be concluded that the crystallite size is approximately equal to the particle size in ZnO powders prepared in the present work.