Phase separation in the system with sodium silicateand sodium dodecyl sulfate under acidic conditions

Ryoji TAKAHASHI,³ Satoshi SATO,* Yasuhide KOJIMA,* Toshiaki SODESAWA,*

Ikuya YAMADA, Daisuke NISHI, Katsuhiro MURAMATSU, Hidenori YAHIRO,Hiroyuki YAMAURA and Naoki MIKAMI**

Graduate School of Science and Engineering, Ehime University, 2–5, Bunkyo-Cho, Matsuyama 790–8577*Department of Applied Chemistry and Biotechnology, Graduate School of Engineering, Chiba University,1–33, Yayoi, Inage, Chiba 263–8522

**Tokuyama Corp., 1–1, Mikage-cho, Shunan-shi, Yamaguchi 745–8648

Phase separation in an acidic aqueous solution with sodium silicate and sodium dodecyl sulfate (SDS) was investigated. Thesolution separated into silica-rich and SDS-rich phases during condensation of silica components. During the phase separation,the transition structure was fixed by gelation, resulting in mutually continuous morphology. The addition of a small amountof water-miscible organic solvents, such as methanol, ethanol, 1-propanol, and 1-butanol, into the reacting solution greatlydecreases both viscosity and phase separation tendency of the solution. In the system with high concentration of strong acid, SDSmolecules form wormlike micelles, which not only increase viscosity of the solution but also play an important role in the phaseseparation in the solution because the formation of wormlike micelles increases apparent molecular weight of SDS and decreasesmixing entropy. The decrease in the viscosity by the addition of organic solvents suggests that wormlike micelles become small inthe presence of organic solvent molecules. Consequently, the addition of appropriate amount of organic solvent is effective forcontrolling phase separation tendency and size of the resulted macropores in the system with SDS through the change in thestructure of the micelles.©2010 The Ceramic Society of Japan. All rights reserved.

Key-words : Spinodal decomposition, Wormlike micelles, Macropore, Bimodal porous silica gel, Sodium dodecyl sulfate, SDS,Water glass

[Received November 16, 2009; Accepted February 18, 2010]

1. Introduction

Sol­gel preparation of bimodal porous silica gel, which hasboth macropores and mesopores, is widely reported, where themacropores are formed by fixing transitional structure of phaseseparation induced during sol­gel transition.1),2) In the process,various water-soluble additives are known to have a potential forinducing phase separation in a homogeneous reacting solution.Because the polymerization degree of silica component contin-uously increases during the sol­gel reaction, mixing entropy inthe solution decreases. When mixing free energy becomespositive from negative in the homogeneous solution by thedecrease in the mixing entropy, the solution separates into twophases. The additives have an effect to increase initial mixingfree energy so that the solution becomes unstable before gelation.For example, the addition of organic polymers with acidicfunctional groups, such as sodium polystyrene sulfonate (NaPSS)and poly(acrylic acid) (HPAA), decreases the initial mixingentropy.3)­5) Then, the solution separates into two phases beforegelation under appropriate reaction conditions at a specific solu-tion composition. The transitional structures of phase separationcan be fixed as permanent morphologies by gelation of silica.Recently, we have reported that sodium dodecyl sulfate (SDS)

has ability for inducing phase separation in an aqueous sodiumsilicate solution.6) The aqueous solution of sodium silicate is

generally called water glass (WG), and used in the industrialproduction of silica gel. The formation of macroporous silica gelsfrom a mixture of SDS and WG is promising in industrialproduction because of low raw materials cost, although themacropore size have been controlled only by the change ingelation temperature.6) On the other hand, surfactants werewidely used in sol­gel reaction for the formation of periodicmesophases, such as hexagonal, cubic and lamella, by utilizingself-assemble ability of surfactant molecules.7)­10) In the for-mation process of mesopores, hydrophilic functional groups insurfactants interact with condensed silica gel surface throughelectrostatic interaction or hydrogen bonding. Because silicasurface is acidic, either cationic or nonionic surfactant with basicor neutral nature is used in the formation of mesophase in thesilica gel. These surfactants are also used in the macroporeformation through phase separation, where interaction betweensurfactant molecules and silica gel decreases solubility ofsurfactant in the solution, and leads phase separation betweensolvent and surfactant associated with silica gel.11),12)

The results in the morphology of silica gel prepared inSDS­WG system suggest that phase separation in the systemproceeds between SDS phase and silica phase in contrastto other surfactant systems.6) This is reasonable in a sensebecause we cannot expect attractive interaction between acidicsilica surface and acidic surfactants. However, it is difficult tounderstand that the addition of small ionic SDS moleculesdecreases the mixing entropy in the solution or increases themixing enthalpy enough for separating phases. Then, we have

³ Corresponding author: R. Takahashi; E-mail: rtaka@chem.sci.ehime-u.ac.jp

Journal of the Ceramic Society of Japan 118 [4] 295-299 2010 Paper

©2010 The Ceramic Society of Japan 295

proposed a hypothesis that SDS molecules form some kinds ofassociates such as ramified aggregates or linear micelles in acidicsolution, and that the SDS associates enlarge viscosity of thesolution with SDS under high concentration of nitric acid.Formation of such associates increases apparent molecularweight of SDS, which decreases the mixing entropy, and inducesphase separation.In this work, we investigated the effect of the addition of

organic solvents in the acidic SDS­WG solution in order toconfirm the hypothesis mentioned above. If the formation ofSDS associates plays an important role in phase separationin the SDS­WG system, addition of a cosolvent with variousnatures will affect the structure of the SDS associates and phaseseparation tendency of the solution. In addition to such scientificinterest, it is quite useful to control macropore structures only bythe small change in the solution composition instead of thechange in the reaction temperature.

2. Experimental

Silica gel samples were prepared using a commerciallyavailable WG (No. 3 in Japanese Industrial Standards (JIS),containing 29 wt% SiO2 and 10 wt% Na2O, Toso Sangyo Co.Ltd.) as a silica source. SDS was used to induce phase separation.For neutralization of WG and catalysis of polycondensation ofsilica, 60 wt% aqueous solution of nitric acid (conc.HNO3)was used. Various organic solvents, such as methanol, ethanol,1-propanol, 1-butanol, 2-propanol, and acetone, were used as anadditive for controlling phase separation tendency. All thereagents except for WG were purchased from Wako PureChemical Industries, Ltd., Japan.The silica gel was prepared by typically the same method as

reported previously.6) Firstly, SDS was dissolved in distilledwater. After the SDS had been dissolved, conc.HNO3 and anorganic solvent were added in the order into the SDS solution.Then, WG, which was preliminarily diluted with distilled water,was added slowly into the acidic SDS solution with stirring toavoid inhomogeneous precipitation of silica by rapid condensa-tion. Because some organic solvents are easily oxidized by nitricacid, the mixing of the raw materials must be done quickly.The starting compositions adopted in the present work wereWG:SDS:organic solvent:water:conc.HNO3 = 45:y:z:97:37 inweight (g), where y and z varied in the ranges of 0­3, and0­10, respectively. The resultant homogeneous solution wassealed in a plastic container, and kept at 25°C for 24 h. Theobtained monolithic wet gel was soaked 3 times in a water bathfor 24 h for removing sodium salt. Then, the wet gel was dried at50°C till no weight change was observed.Viscosity of SDS solution was measured by using capillary

viscometer. Scanning electron microscope (SEM, SM200,Topcon) was used for the observation of micrometer-scalemorphology of the obtained samples. A fresh fractured surface ofa sample was coated with gold before observation for avoidingcharge-up. The pore size distribution in the size range of >50 nmwas measured for selected samples by mercury porosimetry(POREMASTER33P, Quantachrome).

3. Results

Previously, we have reported that aqueous SDS solutionshowed high viscosity by increasing concentration of a strongacid.6) Figure 1 shows the change in viscosity of acidic SDSsolution with the addition of organic solvent. The addition of awater-miscible organic solvent into the acidic SDS solutiondecreased the viscosity of the solution. This effect depends on

both molecular weight and amount of additives. The viscositysimply decreased with increasing the ethanol content. At thesame content, the viscosity became smaller with the increase inthe carbon number of primary alcohol.Figure 2 shows the fractured surface of silica gels obtained at

different ethanol contents. At low ethanol contents, sphericalparticles with the size of ca. 10 ¯m were obtained (A). With theincrease in the ethanol content, the morphology changed tomutually continuous one and the macropore size decreased (B)­(E). Finally, gels with no-macropores were obtained at highethanol contents (F). In the mutually continuous morphology, themacropores were filled with small particles when their size islarge. This is one of the features of the macroporous gelsobtained by fixing transitional structure of phase separation.1)

Figure 3 shows pore size distribution of the samples. A sharppeak was observed in all the samples at the size corresponding tothat of continuous pores observed in SEM.Figure 4 shows the change in morphology with the SDS

content. As has been reported for the system without ethanol,6)

pore connectivity was affected by the SDS content: the increasein the SDS content led to the change in macropore morphologyfrom closed macropores to particle aggregates through mutuallycontinuous morphology. The ethanol addition to the SDS­WGsystem did not alter this trend.Figure 5 summarizes relation between the gel morphology

and solution composition. We can see two trends in Fig. 5. Thefirst one is that the increase in the SDS content decreases theconnectivity of silica gel phase as described in Fig. 4. Thesecond one is that ethanol addition has two effects: one is todecrease the macropore size and the other is to decrease the poreconnectivity. The decrease in pore size with the increase inethanol content is also clearly shown in Figs. 2 and 3. Regardingpore connectivity, morphology changes from particle aggregatesto mutually continuous one, and to closed macropores withincreasing ethanol content. In the region at high ethanol and lowSDS contents, as a result, silica gels with closed macropores wereobtained. In contrast, in the region at low ethanol and high SDScontents, silica particle aggregates were obtained.

0 2 4 6 80.5

1

5

10

50

100

Additive content /g

Dyn

amic

vis

cosi

ty, µ

/mPa

s

Fig. 1. Change in viscosity of acidic SDS solution with additiveamount at 30°C. SDS and HNO3 concentrations were 4 and 20 wt%,respectively, and no WG was added. Closed circle, ethanol; open triangle,methanol; open square, 1-propanol; open circle, 1-butanol.

Takahashi et al.: Phase separation in the system with sodium silicate and sodium dodecyl sulfate under acidic conditionsJCS-Japan

296

Table 1 summarizes the change in the morphology with thecontent of various organic solvents. All the additives have thesame effect: addition of the organic solvents decreased macro-pore size. That is, the phase separation tendency is decreasedby the addition of any alcohols. At the same content, largermolecules have larger effects. Difference in the effect of

cosolvent on the pore structures among 1-propanol, 2-propanoland acetone was small.

4. Discussion

4.1 Macropore formation in sol–gel system byphase separation

When phase separation proceeds through spinodal decompo-sition, compositional fluctuation grows as superposition of

(A) (B)

(C) (D)

(E) (F)

Fig. 2. SEM images of silica gel prepared at the composition ofWG:SDS:ethanol:water:conc.HNO3 = 35:2.7:z:97:37 (unit: g) at 25°C.(A) z = 2.0, (B) 3.0, (C) 4.0, (D) 5.0, (E) 6.0, and (F) 7.0.

10-1 100 101

0

2

4

6

8

10

Pore diameter, D / µm

dVP/d

logD

/ cm

3 g-1

(A)

(B)

(C)

Fig. 3. Pore size distribution of silica gel prepared at the compositionofWG:SDS:ethanol:water:conc.HNO3 = 35:2.7:z:97:37 (unit: g) at 25°C.(A) z = 4.0, (B) 5.0, (C) 6.0.

(A) (B)

(C) (D)

(E) (F)

Fig. 4. SEM images of silica gel prepared at the composition ofWG:SDS:ethanol:water:conc.HNO3 = 35:y:4:97:37 (unit: g) at 25°C. (A)y = 2.1, (B) 2.4, (C) 2.7, (D) 3.0, (E) 3.3, and (F) 3.6.

1.2 1.5 1.8 2.1 2.4 2.7 3 3.3 3.6 3.9

0

2

4

6

8

SDS, y / g

Eth

anol

, z /

g

55

25

35

10

35

6

1.5

30

12

6

1

25

6

2

100

100

100

100

100

100

100

100 100

1.5

8

Fig. 5. Morphologies of gels prepared at the composition ofWG:SDS:ethanol:water:conc.HNO3 = 35:y:z:97:37 (unit: g) at 25°C.Double circle: isolated macropores as shown in Fig. 4(A), circle withcross: mutually continuous morphologies as shown in Fig. 2(B)­(E), opencircle: particle aggregates as shown in Fig. 2(A), closed circle: withoutmacropores. Numbers on the symbols indicate macropore size (¯m)estimated from SEM images of the mutually continuous morphology.

Journal of the Ceramic Society of Japan 118 [4] 295-299 2010 JCS-Japan

297

infinite number of sinusoidal waves with constant wavelengthbut oriented randomly in three-dimensional space.1),2) Thewavelength of the compositional fluctuation becomes longerwith the progress of the phase separation. When onset of phaseseparation is much earlier than gelation, therefore, macroporesbecome larger till heavy phase precipitates or till the growingstructure is fixed by gelation. The early onset of phase separationis observed for the system with low stability in one phasesolution. Then, domain size fixed by gelation becomes large.After drying, the phase without silica becomes macropores. Thus,pore size of macropores is controlled by the timing of phaseseparation and gelation. In other words, macropores becomelarger for the system with higher phase separation tendency.In the macropore formation by phase separation, volume ratio

of the two phases affects the connectivity of the resultant gelmorphology. During the growth of domains, a minor phasefragments into spherical particles. Mutually continuous mor-phologies become when the volume of the two phases is almostthe same. If volume fraction of a phase containing silicapolymers is low, the silica phase fragments and morphology ofparticle aggregates results. In contrary, if volume fraction of aphase without silica is low, macropores fragment into sphericalmacropores.In the previous work, we proposed that phase separation in the

SDS­WG system generates conjugate phases; one rich in SDSand the other in silica, based on the change in the morphologywith SDS content.6) The morphology changed from closedmacropores, through mutually continuous morphologies, toparticle aggregates with an increase in the SDS content. Theconnectivity of the silica-rich phase decreased with an increase inSDS concentration accompanied with the increase in the macro-pore connectivity. Therefore, it is simply concluded that SDS andsilica are the major constituents of the respective conjugatephases, and that the SDS-rich phase becomes macropores afterdrying. In the system with ethanol in the present work, similarchange in the morphology with SDS content was observed assummarized in Fig. 5. Unambiguously, the phase relation in theSDS­WG system can be consistently depicted by a quasi binaryphase diagram with SDS and silica as end-members.

4.2 Micelle structures of sodium dodecyl sulfate inaqueous solution

Surfactants form micelles in aqueous solution, when theconcentration of a surfactant exceeds critical micelle concen-tration. It is confirmed experimentally that SDS forms sphericalmicelles with ca. 60 molecules in an aqueous solution.13)

Theoretical calculation also supports the spherical micellestructures of SDS with ca. 60 molecules.14) On the other hand,

micelle structures of surfactants depend on the conditions such astemperature and solution compositions. In the system with SDS,transition of micelle structure from spherical to rod or wormlikeis frequently observed by adding various components, such ascationic surfactants,15) neutral surfactants with ethylene oxidechain,16) organic salts,17),18) urea19) and inorganic aluminumnitrate salt,20) in aqueous solution of SDS. Such change occurs asa result of modification of hydrophilic characters of SDS by theadditives.The spherical-to-wormlike transition of micelle structure

accompanies usually the increase in the viscosity of the solutionbecause the formed wormlike micelles are entangled each otheras if long polymer chains are dissolved. To our knowledge, thereis no report on the occurrence of spherical-to-wormlike transitionunder highly acidic conditions. However, it is reasonable toattribute the increased viscosity of SDS solution at highconcentration of strong acid to the formation of wormlikemicelles because modification of hydrophilic characters of SDSat extensively low pH conditions is expected through protonationof sulfate group in SDS. The decrease in the viscosity of the SDSsolution by the addition of cosolvents shown in Fig. 1 suggeststhe change in the micelle structures. The length of wormlikemicelles would decrease with the increase in the ethanolconcentration.In our previous work, we proposed a hypothesis that SDS

molecules form some kinds of associates such as ramifiedaggregates or linear micelles in acidic solution in order to explainthe phase separation behaviors and high viscosity in acidic SDS­WG system.6) This hypothesis is reinforced by the literatures. Inthe next section, we will show that the results in the present studyare clearly explained based on the formation of wormlikemicelles of SDS under highly acidic conditions as well as onthe change in the structures of the micelles by the reactionconditions.

4.3 Phase separation in the system with sodiumdodecyl sulfate and water glass

SDS is a small molecule with molecular weight of 288. In anaqueous solution, it forms spherical micelles with ca. 60molecules resulting in apparent total molecular weight of ca.17000.13) When it forms wormlike micelles, however, the totalmolecular weight becomes much larger than that of sphericalone. It is reported that the wormlike micelles can grow to severalmicrometers with radii of 2­3 nanometers under appropriateconditions.21) Formation of such large supramolecular structuresaffects not only the viscoelasticity of the solution but also thephase separation tendency.16) The wormlike micelles can act aslarge molecules instead of individual small SDS molecules in thesolution. Increase in molecular weight of a solute in polymersolution decreases the mixing entropy and increases the volumefraction of the solute phase. By forming wormlike micelles,therefore, SDS obtains ability of inducing phase separation in thesol­gel system, and a small amount of SDS becomes enough forthe formation of mutually continuous morphology.6)

However, the phase separation tendency of SDS solution is toostrong to control macropore size: gel preparation at ambienttemperature resulted in silica gel with large macropores, typically>50 ¯m, which were filled with small silica particles. Anexample of such morphology is shown in Fig. 2(B). It is difficultto expect as a high-performance porous material for the gels withthis kind of morphology. We have controlled the macropore sizeby increasing gelation temperature in the previous work.6) In thiswork, we found that the addition of organic compounds affects

Table 1. Relation between macropore size (¯m) and the amountof organic solvent added. WG:SDS:additive:water:conc.HNO3 =

45:3:z:97:37 in weight (g)

z (g) 2.0 3.0 4.0 5.0 6.0 7.0

methanol ® ® 40 15 7 2.5ethanol PAa 30 12 6 1 NMc

1-propanol ® ® 4 0 NMc ®

2-propanol ® ® 7 1 NMc ®

acetone ® ® 6 1 NMc ®

1-butanol PAa CPb CPb CPb CPb ®

a. PA: morphology with particle aggregates.b. CP: morphology with closed macropores.c. NM: no macropores.

Takahashi et al.: Phase separation in the system with sodium silicate and sodium dodecyl sulfate under acidic conditionsJCS-Japan

298

phase separation tendency in the SDS solution. It is reported thatthe addition of various cosolvents affects the formation ofwormlike micelles.18)

In the present system, the addition of ethanol in the SDSsolution decreases the viscosity of the solution (Fig. 1). This canbe interpreted as that the presence of ethanol molecules decreasesthe length of wormlike micelles of SDS in the present system.The generation of shortened wormlike micelles leads to theincrease in mixing entropy and the decrease in phase separationtendency. Thus, the decrease in the macropore size with increas-ing ethanol content in Figs. 2 and 5 is reasonably understood. InFig. 5, ethanol seems to be distributed preferentially in silicaphase during phase separation, because the silica phase becomesmajor phase with the increase in ethanol content. However, it isdifficult to consider that cosolvent ethanol is distributed only toone phase. This change in volume fraction would be related tothe decrease in the volume fraction of SDS phase by the decreasein apparent molecular weight of wormlike micelles of SDS. Asdescribed above, volumes occupied by polymer chain in asolvent increases with the increase in molecular weight.1) Theaddition of ethanol disturbs the formation of wormlike micellesof SDS. Therefore, it is natural that the volume ratio of silica andSDS phases in phase separation depends on the ethanol content.Thus, the present results are well explained by the presence ofwormlike micelles of SDS.The same effects were also observed in the addition of various

water-soluble additives. That is, the viscosity of SDS solutiondecreases with the addition of various cosolvents (Fig. 1), andmacropore size decreased with increasing the amount of addedcosolvent (Table 1). The change in the morphology from particleaggregates to closed macropores was also observed by theincrease in the amount of added cosolvent. This means that mostof cosolvents have an effect to decrease the length of wormlikemicelles of SDS and to decrease phase separation tendency. Here,the extent of the effect depends on the kind of added cosolvent.A larger molecule seems to disturb significantly the formation ofwormlike micelles. On the other hand, difference in molecularstructure of additives affects little in comparison among theresults of 1-propanol, 2-propanol, and acetone (Table 1). Prob-ably, different hydrophilic characters among additives wouldaffect the structure of SDS micelles.Matsui et al. obtained results similar to ours in the system with

tetramethoxysilane (TMOS) and alkyl sulfonates with variedalkyl chain length, C14, C16, and C18.22) In their work, becauseof the formation of methanol by hydrolysis of TMOS, increase in1 M nitric acid solution means the decrease in the alcoholconcentration. Therefore, phase separation occurs at high solventcontents. Their results that fragmentation of gel skeleton occursby the increase in the amount of alkyl sulfonates and 1 M nitricacid are essentially the same as those shown in Fig. 5. Althoughthey did not mention about the effect of alcohol concentration,their results are well understood by considering the formation ofwormlike micelles of alkyl sulfonates as we discussed above. Inaddition, they have reported that phase separation tendencyincreased in the order of C14 < C16 < C18.22) This result is alsoreasonably understood by the formation of wormlike micelles,because the increase in the length of alkyl chain promotes theformation of wormlike micelles.23)

5. Conclusion

We investigated the effect of the addition of cosolvent in the

phase separation system of WG and SDS. In this system, phaseseparation is induced by the incompatibility between wormlikemicelles of SDS formed in strongly acidic solution and polymer-izing silica oligomers. The addition of cosolvents, such asalcohol and acetone, strongly disturbed the formation of worm-like micelles of SDS, and led to decrease in apparent molecularweight of SDS micelles, increase in mixing entropy, and decreasein the phase separation tendency. As a result, macropore size inmutually continuous morphology formed by fixing transitionalstructure of phase separation became small with increasing theamount of added cosolvents. Therefore, the macropore structurescould be controlled by changing solution composition withoutchanging gelation temperature.

Acknowledgement This work was supported by IndustrialResearch Grant Program in 04A25503c from New Energy andIndustrial Technology Development Organization (NEDO) of Japan.

References1) K. Nakanishi, J. Porous Mater., 4, 67­112 (1997).2) K. Nakanishi, Bull. Chem. Soc. Jpn., 79, 673­691 (2007).3) K. Nakanishi and N. Soga, J. Am. Ceram. Soc., 74, 2518­2530

(1991).4) K. Nakanishi and N. Soga, J. Non-Cryst. Solids, 139, 1­13

(1992).5) R. Takahashi, S. Sato, T. Sodesawa and A. Yachi, J. Ceram.

Soc. Jpn., 109, 577­579 (2001).6) R. Takahashi, S. Sato, Y. Kojima and N. Mikami, J. Ceram.

Soc. Jpn., 115, 882­887 (2007).7) C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli and

J. S. Beck, Nature, 359, 710­712 (1992).8) 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 and J. L. Schlenker,J. Am. Chem. Soc., 114, 10834­10843 (1992).

9) S. Inagaki, Y. Fukushima and K. Kuroda, J. Chem. Soc., Chem.Commun., 1993, 680­682 (1993).

10) H. Yang, A. Kuperman, N. Koombs, S. M. Afara and G. A.Ozin, Nature, 379, 703­705 (1996).

11) K. Nakanishi, T. Nagakane and N. Soga, J. Porous Mater., 5,103­110 (1998).

12) K. Nakanishi, Y. Sato, Y. Ruyat and K. Hirao, J. Sol-Gel Sci.Technol., 26, 567­570 (2003).

13) N. Turro and A. Yekta, J. Am. Chem. Soc., 100, 5951­5952(1978).

14) N. Yoshii and S. Okazaki, Chem. Phys. Lett., 425, 58­61(2006).

15) H. Yin, Y. Lin and J. Huang, J. Colloid Interface Sci., 338,177­183 (2009).

16) D. Acharya, T. Sato, M. Kaneko, Y. Singh and H. Kunieda,J. Phys. Chem. B, 110, 754­760 (2006).

17) P. Hassan, S. Raghavan and E. Kaler, Langmuir, 18, 2543­2548 (2002).

18) K. Din, Z. Khan, S. Kumar and T. Ahmad, Colloid Polym. Sci.,286, 335­341 (2008).

19) S. Kumara, D. Sharma, G. Ghoshb and K. Din, Colloids Suf. A,264, 203­206 (2005).

20) D. Angelescu, A. Khan and H. Caldararu, Langmuir, 19, 9155­9161 (2003).

21) T. M. Clausen, P. K. Vinson, J. R. Minter, H. T. Davis, Y.Talmon and W. G. Miller, J. Phys. Chem., 96, 474­484 (1992).

22) T. Matsui, K. Nakanishi, K. Kanamori and T. Hanada, Mater.Res. Soc. Symp. Proc., 1056, 418­423 (2008).

23) P. J. Missel, N. A. Mazer, G. B. Benedek and M. C. Carey,J. Phys. Chem., 87, 1264­1277 (1983).

Journal of the Ceramic Society of Japan 118 [4] 295-299 2010 JCS-Japan

299