ca

daYam

aka

Primary amineMesoporousPore accessibility

es wondmiarapy,

crystallization easily occurred simply by centrifugation. The monodispersity was maintained even after

ted mucharaccernedropert

such as thermal hydrolysis of TiCl4 [2], hydrolysis of titanium alk-oxide in dilute alcohol solutions with hydroxypropyl cellulose [3],and slow hydrolysis of titanium glycolate precursors [4]. Thus farsynthesis of spherical titania particles with high monodispersibili-ty has proven difcult because of the fast and uncontrollable reac-tions of the hydrolysis, nucleation, and growth steps during the

face. It is noted that besides the specic surface area, thedevelopment of porous structures with facile mass transport tothe titania surface is a key prerequisite for the development of cat-alytic applications.

On the other hand, Mine et al. proposed a method using differ-ent combinations of solvents, i.e. acetone, ethanol, and acetonitrile[14]. The particle size and its distribution of titania can be well-controlled using this method. Monodisperse titania particles withthe lowest coefcient of variation Cv, 5.7%, were produced usinga ethanol/acetonitrile co-solvent with methylamine. They studiedthe photonic properties, such as refractive index and optical

* Corresponding author. Address: Department of Chemical Engineering, Facultyof Engineering, Kansai University, 3-3-35 Yamate-cho, Suita-shi, Osaka 564-8680,Japan. Fax: +81 6 6388 8869.

Journal of Colloid and Interface Science 334 (2009) 188194

Contents lists availab

Journal of Colloid an

.coE-mail address: shun_tnk@ipcku.kansai-u.ac.jp (S. Tanaka).titania as a system of photonic crystals is one of its most distinctiveapplications [1], and is largely determined by properties such asthe crystalline phase, particle size, and particle arrangement. Onthe other hand, the photocatalytic activity depends on the crystalphase, crystallite size, specic surface area, and porous structure.Therefore, optimization of the appropriate titania properties isessential to improving the performance of applied systems. How-ever, the morphology control of uniformly-sized titania particlesis a difcult task due to the high reactivity of titania precursors.Titania particles have been synthesized using several methods

use of various surfactants as templates [57]. However, the synthe-sis of mesoporous titania becomes much more complicated ascompared with mesoporous silica [810] because of the high reac-tivity of titanium alkoxide towards hydrolysis and condensation. Inaddition, it is difcult to prevent the collapse of the porous struc-ture caused by crystallization of the framework. Recently, meso-porous titania in lm morphology has been reported to beefcient for the photocatalytic decomposition [1113], becausetitania with a large surface area has many active sites so that sub-stances can be adsorbed in large quantities on to the titania sur-1. Introduction

Titania nanomaterials have attractheir unique chemical and physicalof the titania nanomaterials are conof the electrochemical and optical p0021-9797/$ - see front matter 2009 Elsevier Inc. Adoi:10.1016/j.jcis.2009.02.060crystallization of the particles by high temperature annealing. The titania particles prepared using DDAhad mesopores near the surface of the spheres, providing high pore accessibility to the sphere fromthe surfaceair interface. The particle size uniformity and photocatalytic reactivity of the titania preparedusing DDA were higher than those of the titania prepared using ammonia.

2009 Elsevier Inc. All rights reserved.

ch attention because ofteristics. The propertieswith the performance

ies. The use of colloidal

solgel processing. On the other hand, the specic surface area isa vital factor for the application of titania particles in catalysis. Inaddition, high photocatalytic activity is associated with crystallitesize, because small crystallite size can lead to quantum size effectsin the material. Research efforts to produce mesoporous titaniamaterials with well-tailored pore systems have focused on theTitaniaMonodisperse particles

DDA was effective for the synthesis of monodisperse titania spheres with low coefcient of variation.When the titania spherical particles with coefcient of variation less than 4% were obtained, the colloidalSynthesis of highly-monodisperse spheriin the submicron range

Shunsuke Tanaka a,b,*, Daisuke Nogami a, Natsuki TsuaDepartment of Chemical Engineering, Faculty of Engineering, Kansai University, 3-3-35bHigh Technology Research Center, Kansai University, 3-3-35 Yamate-cho, Suita-shi, Os

a r t i c l e i n f o

Article history:Received 7 November 2008Accepted 26 February 2009Available online 9 April 2009

Keywords:

a b s t r a c t

Monodisperse titania sphersized by hydrolysis and cusing ammonia or dodecylaature. The samples were chmission electron microsco

www.elsevierll rights reserved.l titania particles with diameters

a, Yoshikazu Miyake a,b

ate-cho, Suita-shi, Osaka 564-8680, Japan564-8680, Japan

ith particle diameters in the range 380960 nmwere successfully synthe-ensation of titanium tetraisopropoxide. The preparation was performedne (DDA) as a catalyst in methanol/acetonitrile co-solvent at room temper-cterized by powder X-ray diffraction, scanning electron microscopy, trans-dynamic light scattering, and nitrogen sorption measurement. The use of

le at ScienceDirect

d Interface Science

m/locate / jc is

2.2. Synthesis of monodisperse titania spheres using NH3 or DDA ascatalyst

dn X

nidi=X

ni 1Cv

Xdi dn2=

Xni

q=dn 2

Average particle size was also determined using a ELS-8000 (Ot-suka Electronics) DLS system. TEM images were recorded on a JEM-2010S JEOL microscope at an acceleration voltage of 200 kV. PXRDpatterns were recorded on a RINT-TTR III Rigaku diffractometerusing Cu Ka radiation with k = 1.5418 . The copper anode wasoperated at 40 kV and 20 mA. The nitrogen adsorption/desorptionisotherms of the titania powder were measured at 77 K using aBELSORP 28 instrument (Bel Japan) and the surface area foundusing the BrunauerEmmetTeller (BET) formalism. The pore sizedistribution was derived from the adsorption branches of the iso-therms by the BarrettJoynerHalenda (BJH) model. TGA was car-ried out on a DTG-60H (Shimadzu) with a heating rate of 10 C/

d Interface Science 334 (2009) 188194 189Titania particles were prepared by hydrolysis and condensa-tion reaction of TTIP with ammonia or DDA as a catalyst in aco-solvent of methanol/acetonitrile. The molar compositions ofthe reaction mixtures were in the range of 714 methanol/271acetonitrile/2.8 water/1 TTIP/00.17 ammonia or 714 methanol/271 acetonitrile/0.816 water/1 TTIP/06.4 DDA. In a typicalpreparation, 0.18 ml of water was added to 150 ml of metha-nol/acetonitrile solution. Then 0.28 g of DDA was dissolved inthe mixture. After stirring for 10 min, 1 ml of TTIP was addedand stirred for 12 h. Hydrolysis and condensation reaction oc-curred, and a suspension of titania particles was obtained. Thesuspension was centrifuged at 1500 rpm for 10 min. The parti-cles obtained were washed with methanol and centrifugedagain; the process was repeated three times. No other sedimen-tation treatments were carried out for colloidal crystallization.The products were then dried at 60 C followed by calcinationat 400 C for 5 h.

2.3. Characterization

SEM images were recorded on a VE-8800 Keyence microscopeat an acceleration voltage of 5 kV. Samples for SEM analysis wereprepared by carefully scratching the powder onto a mortar, wherethey were crushed before depositing them onto carbon adhesivereectance, of colloidal crystals, but did not examine the internalpore structure of the particles. We noted that the introduction ofnanoscale pores in the titania is an efcient way to continuouslycontrol the refractive index of the titania spheres. To date meso-porous titania in uniform sphere morphology has not been fullysynthesized and characterized in detail. Further development is re-quired toward fabrication of monodisperse titania spherical parti-cles with high pore accessibility. In this study, we report thepreparation and characterization of spherical titania particles bya solgel method using a primary amine catalyst. The hydrolysisand condensation of titanium isopropoxide (TTIP) was performedin a methanol/acetonitrile co-solvent by reference to the ethanol/acetonitrile co-solvent reported by Mine et al. [14]. It is a novel as-pect in this study is the use of dodecylamine (DDA) of which thecarbon chain is longer than that of amines used by Mine et al.Our expectation was that DDA would also act as a moleculartemplate to introduce the mesopores into the titania particles.We focus on the effects of the DDA on mean particle size, particlesize distribution, pore structure, and photocatalytic activity. Themorphological characteristics of titania spheres were investigatedusing scanning electron microscopy (SEM), transmission electronmicroscopy (TEM), and dynamic light scanning (DLS). Thestructural characteristics were investigated by the combinationof powder X-ray diffraction (PXRD), TEM, nitrogen sorption mea-surements, and thermogravimetric analysis (TGA).

2. Experimental

2.1. Chemicals

Methanol, acetonitrile, ammonia solution (10 wt%), and TTIPwere purchased from Wako Pure Chemical Industries and used asreceived. DDA was purchased from Tokyo Kasei Kogyo Co. and usedas received.

S. Tanaka et al. / Journal of Colloid antape. The diameters of more than 200 particles in the SEM imageswere measured to determine average particle size, dn, and Cv, de-ned by the following equations:min.The photocatalytic activity was measured through the forma-

tion rate of I3 due to the oxidation photo-reaction of I to I2 in ex-

cess I conditions [15]. A 50 mg sample of titania was added to10 ml of 0.2 M KI aqueous solution and stirred. The mixture wasirradiated with 365-nm light (UV-lamp 20W) for 60, 120, and180 min. The reactant solution was collected by centrifuge and di-luted with deionized water, and then the absorbance of the solu-tion was measured with a UV-2450 UVvis spectrophotometer(Shimadzu). No I3

formation was observed when the experimentswere performed in the dark or in the absence of the titania sam-ples. For comparison, commercially available titania nanoparticlesP25 (Nippon Aerosil) were tested in the same manner.

3. Results and discussion

SEM images of titania particles prepared with different ammo-nia concentrations were shown in Supporting Information (FiguresSI1af). The titania particles prepared at an ammonia/TTIP molarratio less than 0.1 were highly dispersed in size and aggregated.On the other hand, at an ammonia/TTIP molar ratio of 0.15, thetitania particles were isolated and of more uniform size, with dnof 450 nm and Cv of 10.6%. Fig. 1 shows the variation in the dnand Cv as a function of ammonia/TTIP molar ratio. The dn was foundto decrease with increasing ammonia/TTIP molar ratio. Moreover,the dn was constant for ammonia/TTIP molar ratios above 0.1.

Fig. 2 shows dn and Cv of titania particles as a function of DDAconcentration. The dn and Cv values varied from 380 to 960 nmand from 2.6% to 7.2%, respectively. Compared to the particles pre-pared using ammonia, a lower Cv was achieved using DDA. Fig. 3ak show SEM images of titania spheres prepared with different DDAFig. 1. The variation in the dn and Cv as a function of the ammonia/TTIP molar ratio.Black circles and bar indicate the average particle size and coefcient variation,respectively.

concentrations. For DDA/TTIP molar ratios ranging from 0.43 to 4.3,the Cv values were less than 4%. The products prepared at the molarratio of 0.43 and 4.3 exhibit ordered close-packed arrangement oftitania colloids over a sample area of about 10lm. On the otherhand, at the DDA/TTIP molar ratios of 0.043, no ordered close-

packed arrangement of spherical particles, for which the Cv was7.0%, was observed. An increase in DDA/TTIP molar ratio to 6.4led to an increase in the dn to 960 nm and the Cv to 6.0%. In thesesamples the titania particles were isolated. When the titania spher-ical particles with Cv of less than 4% were obtained, the colloidalcrystallization easily occurred simply by centrifugation.

Fig. 4 shows dn and Cv of titania particles as a function of waterconcentration. The dn was observed to decrease with increasingH2O/TTIP molar ratio. Hydrolysis rate of TTIP was increased withwater concentration. At high H2O/TTIP molar ratios, the numberof nuclei generated by the hydrolysis reaction seems to be large,which consequently led to a decrease in the dn. On the other hand,the Cv was observed to increase with increasing H2O/TTIP molarratios.

Fig. 5af shows SEM images of titania spheres prepared withdifferent volume ratios of methanol and acetonitrile. In all runs,the volumes of the methanol/acetonitrile co-solvent were heldconstant. The volume ratio of methanol to acetonitrile was variedwhile the molar ratios of water/TTIP and DDA/TTIP were held con-stant at 2.8 and 0.86, respectively, in the reaction mixtures. Withmethanol the only solvent in the reaction mixture, the titania par-ticles were highly dispersed in size. Higher monodispersity wasachieved by combining methanol and acetonitrile. At the metha-nol/acetonitrile volume ratio of 7:3, ordered arrangement of titaniacolloids was observed. The methanol/acetonitrile volume ratio of

Fig. 2. The variation in the dn and Cv as a function of the DDA/TTIP molar ratio. Themolar compositions of the reaction mixtures were in the range of 714 methanol/271 acetonitrile/2.8 water/1 TTIP/06.4 DDA. Black circles and bar indicate theaverage particle size and coefcient variation, respectively.

190 S. Tanaka et al. / Journal of Colloid and Interface Science 334 (2009) 188194Fig. 3. SEM images of titania particles prepared at DDA/TTIP molar ratios of (ac) 0.043, (df) 0.43, (gi) 4.3, and (j and k) 6.4. Scale bar; 2 lm.

tion is shown in Fig. 6b. Fig. 6c shows nitrogen adsorption/desorp-tion isotherms of the particles prepared with different DDAconcentrations. An increase in the amount of adsorbed nitrogenwas observed at the relative pressure of 0.20.5. No large differ-ence in isotherm type was observed between titania samples pre-pared with different DDA concentrations. The pore sizedistributions are shown in Fig. 6d. The DDA-based samples pre-pared with DDA/TTIP molar ratios of 0.43and 4.3 have smaller poresizes than that prepared with a DDA/TTIP molar ratio of 0.043. TheBET surface area, pore size, and average particle size data of thetitania spheres are summarized in Table 1. There is no large differ-ence in the BET surface areas between the sample prepared withammonia and those prepared with DDA/TTIP molar ratios of 0.43and 4.3. On the other hand, the spheres prepared using DDA havesmaller pore sizes overall than did those prepared using ammonia.

Fig. 7a shows TGA and DTA curves of titania spheres preparedusing ammonia. The TGA trace exhibited a weight loss of 20.5%at temperatures below about 300 C, beyond which no further lossin weight was observed. The exothermic peak at 280 C corre-

Fig. 4. The variation in the dn and Cv as a function of the H2O/TTIP molar ratio. Themolar compositions of the reaction mixtures were in the range of 714 methanol/271 acetonitrile/0.816 water/1 TTIP/0.86 DDA. Black circles and bar indicate theaverage particle size and coefcient variation, respectively.

S. Tanaka et al. / Journal of Colloid and Interface Science 334 (2009) 188194 1917:3 was suitable for preparing monodisperse titania spheres andregular, close-packed arrays. Of particular interest is that thestubbed marks were found on the sphere surface, as indicated bythe arrows in Fig. 5e, suggesting that high adhesion between thetitania spheres.

XRD patterns of calcined titania particles prepared with ammo-nia or DDA were shown in Supporting Information (Figure SI2). Thediffraction peaks corresponding to the anatase phase can be ob-served. The crystallite diameter, D, of titania particles preparedwith ammonia was 45.0 nm from the measured (101) diffractionpeak using the Scherrer formula,

D Kk=b cos h 3where K is the Scherrer constant, 0.94, k is the X-ray wavelength, bis the full width at half maximum of the diffraction curve and h isthe Bragg diffraction angle. On the other hand, the D of titania par-ticles prepared with DDA was 21.4 nm, smaller than that of parti-cles prepared using ammonia.

Nitrogen adsorption/desorption measurements were performedto investigate the pore structure of the titania spherical particles.Fig. 6a shows nitrogen adsorption/desorption isotherms of the par-ticles prepared using ammonia. An increase in the amount of ad-sorbed nitrogen, corresponding to the lling of mesopores, was

observed at the relative pressure of 0.50.8. The pore size distribu-

Fig. 5. SEM images of titania colloids prepared at the different methanol/acetonitrile volu(f) is a portion of the box in (d). Scale bar; 1 lm. The volumes of the methanol/acetonitsponds to the decomposition of the hydrated oxide, the condensa-tion of OH groups and of non-bonded oxygen. The secondexothermic peak at 440 C is due to the crystallization of the amor-phous to the anatase phase. Finally, the last exothermic peak cen-tered at 670 C corresponds to the phase transition of anatase torutile that was accompanied with no weight loss in the TGA curve.Fig. 7b shows TGA and DTA curves of titania spheres preparedusing DDA. The DDA-based sample exhibited a higher weight loss(25.0%) than that of the ammonia-based sample. The rst exother-mic peak shifted toward lower temperature as compared with thatof the ammonia-based sample. This shift of exothermic peak couldbe derived from the high degree of condensation of the inorganicframework. This shift of exothermic peak suggested that the con-densation of the inorganic framework occurred at lower tempera-ture due to higher molar ratio of DDA/TTIP than that of ammonia/TTIP. In addition, two exothermic peaks at higher temperatureshifted toward higher temperature as compared with the of ammo-nia-based titania. The DDA-based samples turned brown after cal-cination at 400 C for a short time within 3 h. In contrast, the colorof the sample prepared with ammonia was white after calcinationunder the same conditions. These results indicate that carbonizedproducts derived from DDA remained in the titania particles. Con-sequently, the crystallization and phase transition seem to havebeen suppressed by the carbonized residue derived from DDA.These results are consistent with the results reported by Inagakime ratios of (a) 10:0, (b) 9:1, (c) 8:2, (d and e) 7:3, and (f) 4:6. The magnied imagerile co-solvent were held constant 150 ml.

d In192 S. Tanaka et al. / Journal of Colloid anet al. [1618], who showed that a carbon coating on commerciallyavailable anatase titania ST-01 (Ishihara Sangyo) can stabilize ananatase structure and suppress the transition from anatase to ru-tile. The crystallite diameter of DDA-based samples was smallerthan that of samples made with ammonia, as estimated usingScherrer formula, indicating that the DDA suppressed the crystal-lite growth. It is conjectured that the change in phase transitiontemperature of anatase to rutile may be attributed to the carbon-

Fig. 6. Nitrogen adsorption/desorption isotherms and pore size distribution of the partiH2O/TTIP molar ratios were 0.17 and 2.8, respectively.

Table 1BET surface area, pore size, and average particle size of the titania samples. The H2O/TTIP

BET surface area (m2/g) Pore size (nm)

Ammonia (ammonia/TTIP = 0.17) 43 7.1DDA (DDA/TTIP = 0.043) 18 4.1DDA (DDA/TTIP = 0.43) 42 2.8DDA (DDA/TTIP = 4.3) 41 2.9Aerosil P25 56 terface Science 334 (2009) 188194ized residue, mesoporosity, and/or spherical morphology. A moredetailed study on this question will be reported in due course.

The titania structures were elucidated by the TEM observations.Fig. 8ad show TEM images of titania particles prepared with DDAand ammonia. The shape of the titania particles prepared usingDDA was maintained after calcination at 400 C. In many regionsof the samples prepared using DDA, ordered close-packed arrange-ment could be observed, as shown in Fig. 8a. On the other hand, no

cles prepared using (a and b) ammonia and (c and d) DDA. The ammonia/TTIP and

molar ratio was 2.8.

Particle size (nm) before calcination Particle size (nm) after calcination

452 396381 334546 507620 568 2050

Monodisperse anatase titania spheres were prepared with a

Fig. 7. TGA and DTA curves of titania spheres prepared using (a) DDA and (b)

S. Tanaka et al. / Journal of Colloid and Interface Science 334 (2009) 188194 193examples of such arrangement were observed from the productsprepared using ammonia. In addition, when ammonia was used,the spherical morphology was slightly collapsed after calcination.The contrast of the micrograph represents the voids between thecrystallite particles, as shown in Fig. 8c and e. The contrast insidethe particles prepared using DDA was lower than that of the parti-cles prepared using ammonia, which agreed with magnitude corre-lation of crystallite size estimated using Scherrer formula. Inaddition, only when DDA was used, nano-scaled pores could be ob-served on the edges of calcined titania spheres, as indicated by thearrows in Fig. 8c.

Fig. 9 shows the photocatalytic activity of titania particles. TheI3 concentrations of KI solutions with titania particles prepared

with ammonia and DDA after 180 min of irradiation were about0.34 104 M and 1.62 104 M, respectively. The I3 concentra-tion of P25 (BET surface area is 56 m2/g), which is commerciallyavailable titania material, was higher than that of the titania parti-cles prepared in this study. On the other hand, from the standpointof I3

concentration per unit surface area, photocatalytic activity oftitania prepared using DDA was comparable to commercial P25.The I3

concentration per unit surface area of titania prepared

ammonia. The ammonia/TTIP, DDA/TTIP, and H2O/TTIP molar ratios were 0.17, 0.43,and 2.8, respectively.using DDA, that prepared with ammonia, and the commercialP25 were calculated to be 8.1 105, 1.6 105, and 7.1 105 M/m2-surface area, respectively. In spite of no large difference

Fig. 8. TEM images of titania particles prepared using (ac) DDA and (d and e) ammoniarespectively. Scale bars; (a) 1 lm, (b) 200 nm, (c) 20 nm, (d) 400 nm, and (e) 50 nm.in the BET surface area of the titania prepared using DDA andammonia, titania particles prepared using DDA exhibited higherphotocatalytic activity than did titania particles prepared withammonia. These results including TEM images imply that DDAwould greatly affect the pore morphology and photocatalytic activ-ity of the resulting titania. Besides the surface area, pore size, andparticle size of the titania materials, the open mesostructure ofsuch material should be useful for providing high pore accessibilityto guest substances from the surfaceair interface of the particles.The activity of ammonia-based samples was lower than that ofsamples produced with DDA because of the lack of pore accessibil-ity of the particles, in other words, low effective surface area. TheDDA-derived introduction of mesopores into the near-surfaces ofthe titania spheres substantially increased their photocatalyticperformance.

4. Conclusions

Fig. 9. Photocatalytic activity of the (open triangle) commercial titania P25 andcalcined titania particles prepared using (closed circle) DDA and (open circle)ammonia, and (closed triangle) no sample. The ammonia/TTIP, DDA/TTIP, and H2O/TTIP molar ratios were 0.17, 0.43, and 2.8, respectively.hydrolysis-based method using ammonia and DDA. The use ofDDA as a catalyst was effective for the synthesis of monodispersetitania spheres with low Cv. In addition, mesopores, which are lo-

. The ammonia/TTIP, DDA/TTIP, and H2O/TTIP molar ratios were 0.17, 0.43, and 2.8,

cated on the surface of the titania spheres, could be introduced byusing DDA, although the amount of mesopore was small, resultingin relatively-high photocatalytic activity. The hydrolysis and con-densation method using DDA should be useful in preparing mono-disperse spherical titania particles with high pore accessibility.Further developments in this synthetic strategy with differentamines should lead to the preparation of mesoporous titaniaspheres with well-tunable particle size. The control of particle sizeand porosity is an efcient way to continuously control the surfacearea and refractive index of the titania spheres. Such titaniaspheres should perform very well as photocatalyst, photonic crys-tals, and sensors.

Acknowledgment

The authors thank Associate Prof. J. Hayashi (Chemical ReactionEngineering System Laboratory at Kansai University) for the nitro-gen sorption measurements.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.jcis.2009.02.060.

References

[1] C. Lpez, C. Adv. Mater. 15 (2003) 1679.[2] M. Visca, E. Matijevic, J. Colloid Interface Sci. 68 (1979) 308.[3] J.H. Jean, T.A. Ring, Colloids Surf. 29 (1988) 273.[4] A. Jiang, T. Herricks, Y. Xia, Adv. Mater. 15 (2003) 1205.[5] D. Khushalani, G.A. Ozin, A. Kuperman, J. Mater. Chem. 9 (1999) 1491.[6] P.C.A. Alberius, K.L. Frindell, R.C. Hayward, E.J. Kramer, G.D. Stucky, B.F.

Chmelka, Chem. Mater. 14 (2002) 3284.[7] G.J.D.A. Soler-Illia, E.L. Crepaldi, D. Grosso, C. Sanchez, Curr. Opin. Colloid

Interface Sci. 8 (2003) 109.[8] C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, J.S. Beck, Nature 359 (1992)

710.[9] J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge, K.D. Schmitt, C.T.W.

Chu, D.H. Olson, E.W. Sheppard, S.B. McCullen, J.B. Higgins, J.L. Schlenker, J. Am.Chem. Soc. 114 (1992) 10834.

[10] D. Zhao, Q. Huo, J. Feng, B.F. Chmelka, G.D. Stucky, J. Am. Chem. Soc. 120 (1998)6024.

[11] J.C. Yu, X. Wang, X. Fu, Chem. Mater. 16 (2004) 1523.[12] J. Tang, Y. Wu, E.W. McFarland, G.D. Stucky, Chem. Commun. (2004) 1670.[13] H. Wang, J.J. Miao, J.M. Zhu, H.M. Ma, J.J. Zhu, H.Y. Chen, Langmuir 20 (2004)

11738.[14] E. Mine, M. Hirose, D. Nagao, Y. Kobayashi, M. Konno, J. Colloid Interface Sci.

291 (2005) 162.[15] K. Nakagawa, F. Wang, Y. Murata, M. Adachi, Chem. Lett. 34 (2005) 736.[16] T. Tsumura, N. Kojitania, I. Izumi, N. Iwashita, M. Toyoda, M. Inagaki, J. Mater.

Chem. 12 (2002) 1391.[17] M. Inagaki, Y. Hirose, T. Matsunaga, T. Tsumura, M. Toyoda, Carbon 41 (2003)

2619.[18] M. Inagaki, F. Kojin, B. Tryba, M. Toyoda, Carbon 43 (2005) 1652.

194 S. Tanaka et al. / Journal of Colloid and Interface Science 334 (2009) 188194

Synthesis of highly-monodisperse spherical titania particles with diameters in the submicron rangeIntroductionExperimentalChemicalsSynthesis of monodisperse titania spheres using NH3 or DDA as catalystCharacterization

Results and discussionConclusionsAcknowledgmentSupplementary dataReferences