Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyamaand H. Kunieda (Editors) c: 2001 Elsevier Science B.V. All rights reserved. ^ 1 /

Ionic partition to zwitterionic micelles

Kenji Iso and Tetsuo Okada

Department of Chemistry, Tokyo Institute of Technology, Meguro-ku, Tokyo 152-8551, Japan

Capillary electrophoresis has been used to measure the ^-potential of zwitterionic micelles in various electrolytes. Even for intrinsically neutral zwitterionic micelles, detectable ^-potential is induced by the imbalance between anionic and cationic partition. Its magnitude and sign are determined by the natures of ions (dominantly anionic ones) and the polarity of surfactant molecules. However, the former is a principal factor governing the <^-potential of zwitterionic micelles in aqueous solutions.

1. Introduction The natures of electrolytes are important factors to affect various

interfacial phenomena. Electrolytes play significant roles particularly in the solution and interfacial chemistry of electrically neutral micelles (e.g. zwitterionic and nonionic micelles) [l], and often cause drastic changes in micellar characteristics, such as cloud points, critical micelle concentration etc. Studies of electrolyte effects are thus expected to provide essential clues to solve puzzles involved in micellar chemistry.

The ^-potential of micelles is an efficient measure to characterize phenomena taking place at solution/micelle interfaces. Electrophoresis and its related methods are useful for evaluating (^-potential, because even rather small potential can be determined thereby [2,3]. In the present paper, the ionic partition to zwitterionic micelles has been studied by measuring their <^-potential with capillary electrophoresis. In previous papers, we indicated that the partition of ions into A^-dodecyl-A^-dimetyl-ammonio-propanesulfonate (DDAPS) micelles are governed by anionic nature, and that partition selectivity can be explained by solvation changes or ion-association of anions [3,4]. DDAPS micelles have inner positive and outer negative charges; anions are attracted by the inner cationic groups while cations are accumulated in the vicinity of the outer surface. It is the main purpose of the present paper whether the anion-dominated partition to DDAPS micelles originates from the molecular

118

polarity of DDAPS or from the natures of the hydration of ions. The ionic partition into zwitterionic micelles having different molecular polarities and charge densities is studied to elucidate this aspect, and results are discussed on the basis of (^-potentials evaluated by capillary electrophoresis.

2. Experimental Capillary electrophoretic apparatus was basically the same as used in the

previous research [3]. The length of a fused-silica capillary (0.05 mm i.d., 0.375 mm o.d.) was 600mm, and the effective length of the capillary was 400 mm. The applied voltage was ±12kV under the current lower than 80|iA, which depends on the nature and concentrations of electrolytes. A sample solution was introduced basically from the anodic end by siphoning; sample introduction was done from the cathodic end in case of the reversed electroosmotic flow. The detection wavelength was set to 280 nm. The entire system was set in an incubator thermostated at 25°C.

DDAPS was recry stalled from acetone twice. Octaethyleneglycol monododecyl ether(OGME) and hexadecyl phosphocholine(HPC) were treated with an ion-exchange resin, Amberlite EG-4, to remove ionic impurities. The structures of these surfactants are shown below.

OGME

3. Results and Discussion 3.1 Zeta potential determination from capillary electrophoretic data

Solution containing pjnrene and acetone was introduced as a sample into the capillary filled with a micellar electrolyte. The former, which is completely partitioned into the micelle, migrates together with micelles, while the latter, which is present in a bulk solution, acts as an electroosmotic flow marker. Thus, two peaks of spiked pyrene and the electroosmotic flow marker appear on an electropherogram.

The mobility of a micelle under a given condition (//) is represented by

119

/ = -V app

(1) ^eoJ

where ^ i s the magnitude of the applied electric field, L is the effective length of the capillary (the length between the injection end and the detection window), and tapp and teo are the migration times of the micelle and the electroosmotic flow. When the ^'•potential is not very high, the Henry's equation is applicable to the determination of f-potential based on electrophoretic mobility.

^ o < K^B) (2)

where rj is the viscosity of a medium andy(;d?5) is a Henry's coeflacient, thus determine the ^-potential of the micelle from //.

We can

3.2 DDAPS micelle Table 1 summarizes the ^'-potential of the DDAPS micelles measured in

various electrolytes. These values indicate that the (^-potential of the DDAPS micelles is dominantly determined by anionic nature, while cationic nature plays a minor role. There is a clear correlation between the (^-potential and the hydration energy of an anion, suggesting that the partition mechanism be related to the hydration of anions; the poorer the hydration of an anion the larger the partition. We indicated that there are two possible origins in ionic partition selectivity, and that large and polarizable anions (C104- and I) are likely to form ion-pairs with DDAPS molecules while small and well-hydrated anions (e.g. CI) undergo hydration changes upon going fi-om bulk to the palisade portion of the DDAPS miceUes [3]. These selectivity terms can be involved in the Poisson-Boltzmann equation for spheres, which provides the spatial distributions of electrostatic potential and ionic concentrations. The (^-potential of the DDAPS micelles was successfully interpreted on the basis of this model.

The charge density of the DDAPS micelles can be lowered by adding OGME, which forms micelles with almost the same size

Table 1 < -potential of DDAPS micelles induced by anion'dominated partition

salt cone /mM

^-potential /mV

NaC104

Nal

NaBr NaCl

LiCl BU4NC1

50

50

50 50 20 20 50

-57.2

-39.6

-19.0 -10.4 -5.8 -6.1 -8.8

120

and the same aggregation number as DDAPS [5]. The partition of ions to the mixed micelles is also governed by anionic nature, and the ^'-potential decreases with decreasing the relative concentrations of DDAPS. Interestingly, the ^'•potential of nonionic OGME micelle is also negative by predominant partition of an anion; e.g. -l lmV for 60mM NaC104, and the negative (^-potential was detected even for NaCl though quantitative evaluation was not possible because of a very small difference in migration between the micelle and an electroosmotic flow.

3.3 HPS micelles As stated above, the anion dominant partition occurs even in nonionic

micelles. This strongly suggests that the partition of ions be related to differences in the hydration nature between anions and cations rather than to the polarity of zwitterionic micelles. In order to elucidate this intrinsic aspect in the ionic partition into zwitterionic micelles, the ^-potential of HPS micelles was studied in various electrolytes. Results are summarized in Table 2. Negative <^-potential occurs for perchlorate salts, which is almost the same as induced in OGME micelles. If the polarity of the zwitterionic micelles were a principal factor to govern ionic partition, the potential should be more positive. In contrast, positive potentials are confirmed for tetrabutylammonium chloride, which induces negative potential for DDAPS. Thus, the polarity of surfactant molecules is a factor to determine the ionic partition into zwitterionic micelles, but less important than the hydration nature of ions. The Gibbs free energy of transfer for individual ions should be determined for more detailed discussions.

Table 2 <^-potential of HFC micelles in various electrolytes ^ , . , cone C-potential surfactant salt , T,, , \T /mM /mV HFC LiC104

NaC104 Bu4NCl Bu4NCl

160 160 80 80

-15.8 -13.9 28.4

1.9 HFC(5mM) + OGME(30mM)

OGME BU4NC1 80

References 1. Y.ChevaUer, N.Kamenka, M.Chorro, R.Zana, Langmuirl2il996)3225. 2. S.Terabe, KOtsuka, TAndo, Anal Chem. 57(1985)834 3. K.Iso, T.Okada, Langmuirheing submitted. 4. T.Masudo, T.Okada, Phys. Chem. Chem.Phys. 1(1999)3577. 5. M.Zulauf, KWeckstrom, J.B.Hayter. V.Degiorgio, M.Corti, J.Phys.Chem. 89(1985) 3411.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) c 2001 Elsevier Science B.V. All rights reserved. 121

Analysis of local structure of ion adsorbed on the gas/liquid interface

Makoto Harada,^ Tetsuo Okada^ and Iwao Watanabe^

^Department of Chemistry, Graduate School of Science and Engineering, Tokyo

Institute of Technology, Ookayama, Meguro-ku, Tokyo, 152-8551, J a p a n

^'Department of Chemistry, Gradua te School of Science, Osaka University,

Machikaneyama, Toyonaka, Osaka, 560-0043, J a p a n

The local structure of Br" in the vicinity of a surface film of a zwitterionic

surfactant, iV-dodecyl-iV,iV-dimecyl-ammonio-butane-sulfonic acid (DDABS),

was analyzed by X-ray absorption fine structure (XAFS) spectroscopy. It was

suggested tha t DDABS molecules are cross-linked by Br" attracting a few water

molecules due to the strong interaction between DDABS and Br".

1. Introduction

Surfactant molecules are adsorbed, and regularly aligned on the aqueous

solution surface (gas/liquid interface). On the solution site, the charged groups

of sur fac tants a t t ract water molecules as well a s ions to const i tute specific

surface s t r u c t u r e s . DDABS is a zwitterionic su r fac tan t wi th a cat ionic

qua te rna ry a m m o n i u m group and an anionic sulfoic group. The surface

s t ruc tu res of a t t rac ted ions depend on the concentrat ion and na tu r e of an

electrolyte added to the solution. In this work, the surface s t r u c t u r e s of

bromide ions adsorbed on the surface film of DDABS were analyzed by XAFS

technique, which can be an effective tool to probe such local atomic structures.

The total-refiection total-electron-yield (TRTEY) XAFS methodfl] allows us to

distinguish the information on the local s tructure of Br" at the solution surface

from that in bulk.

122

electron He Gas J L . - He Gas X-ray Electrode TlBias \r)lt:age

incident angle

t \\ypJJi ca.l()()A\i,, ' -y /

/ evanescent wave , / , . Sample Water BathJ_ Sample Cell

ahsorbmg atom ^=^ odinpit, VAU

Fig. 1. Principle of TRTEY-XAFS. Fig.2. TRTEY-XAFS Cell.

X-ray

A^indow

Sample Water BathJ_ 'smml^^^ Cel V

2. Experimental

2 .1 . TRTEY-XAFS Method

The principle of the TRTEY method is shown in Figure 1. X-ray was

in t roduced a t a grazing incident angle to the solution surface, which was

smaller t han the critical angle of total reflection for Br K-edge X-ray at water

su r face (ca. 2 m r a d for the energy of Br K-edge), Auger e l e c t r o n s a n d

photoe lec t rons a re induced from atoms absorbing evanescent wave(atoms

existing writhin ca. 100A of the solution surface). Since the n u m b e r of the

emitted electrons is proportional to X-ray absorption intensity, XAFS spectra

can be est imated from current intensity.

2.2 . XAFS measurement

Figure 2 shows a schematic diagram of the TRTEY-XAFS apparatus . The

cell was filled ou t w^ith helium gas. X-ray was introduced at about 1 mrad

grazing incidence angle. Ejected electrons successively ionize helium atoms.

Helium ions were collected by the cathode, to which an appropriate bias field

(ca. 1.5 kV-m'^) was applied; t h u s the amphfied current was detected. The

XAFS spectra were estimated from / / 4 , where 4 is the incident X-ray intensity

detected by 4cm ion-chamber filled with Nj gas and / is the X-ray absorption

intensity of bromine measured by the cell shown in Figure 2. All XAFS spectra

were recorded a t BL-7C of Photon Factory, High Energy Accelerator Research

Organization in Tsukuba.

123

Br iC-edge / \ 1

Kji^/f\w^y^ t/(b?v \j/ \ v / ^ F-~UcL- y/ Y rNidj^^-r-'T''^ , ^ ^ ' ^ 0 2 3

r/A

Fig.3. Fourier transforms for samples with Smmol-dm'^ DDABS and (a)5, (b) 10,(c)20,(d)40 mmol-dm-^ CuBr2 aq.

3 . Results and Discussion

Fig.4. The model for Br" and DDABS molecules on the solution surface.

Figure 3 shows the Fourier transforms of obtained XAFS spectra. A

parameter , r, cor responds to the distance between Br' and its coordinating

atoms. XAFS parameters for DDABS and CuBr2 aqueous solutions at Br K-

edge were determined by curve-fitting. Results are listed in Table 1. Nis the

n u m b e r of neighboring a toms and o is the Debye-Waller like factor, which

s tands for thermal fractuation. The parameters for 3mol-dm"^ KBr aqueous

solut ion de te rmined by the t r ansmis s ion method are also listed. These

parameters represent the local structure for Br' in water. It was reported that

Br' in water adopts the octahedral configuration, and thus the coordination

n u m b e r is six[2]. TV values for Br' adsorbed on DDABS surface films were

d e t e r m i n e d re l a t ive to t h i s so lva t ion n u m b e r in b u l k w a t e r . As the

concentration of CuBr2 decreases, the coordination distance becomes longer.

The bromide ions electrostatically interact with hydrogen atoms of the methyl

groups bonded to the quaternary nitrogen atom, as schematically illustrated in

Figure 4. The distance between Br' and the carbon atom of a methyl group can

be assigned the rvalue, 3.41 A. This interaction distance and the coordination

number suggest tha t DDABS molecules adsorbed on the solution surface form

the network s tructure bridged by bromide ions. Although Br" seems to connect

124

Table 1

Curve-fitting parameters for 5mmol-dm'^ DDABS and CuBr2 aq.

Concentration of CuBrj aq.

Smmol-dm'^

lOmmol-dm"^

20mmol-dni'^

40mmol-dm'^

3mol-dm-^ KBr aq.

r/k

3.41

3.42

3.20

3.18

3.20

N*

3.47

3.88

3.03

5.24

6.00

a/A

0.152

0.152

0.107

0.096

0.144

* N for KBr aq. is assumed to be six.

four DDABS molecules because the coordinat ion n u m b e r within the plane

parallel to the solution surface is four, this should be uns table due to steric

h indrance. XAFS analyses cannot completely dist inguish the coordination

distance of water molecule for B r (3.20A) from tha t of DDABS (3.41 A). The

hydrogen a toms of water molecules approach Br" and stabilize the network

structure illustrated in Figure 4; two methyl groups and two water molecules

coordinate Br'. Otherwise, Br' may be sandwiched by two DDABS molecules,

and surrounded by four methyl groups. In this case, it is expected that Br has

no coordinated water molecules. The result for 40 mmoldm"^ CuBra and 5

mmoldm"^ DDABS aq. is similar to tha t for KBr aq. because usually solvated

bromide ions (not interacting with DDABS) exist in the observable region.

Thus , the surface and bulk s t ruc tures of bromide ions are superimposed for

such high concentrations.

REFERENCES 1. I. Watanabe, H. Tanida, S. Kawauchi, M. Harada and M. Nomura, Rev. Sci.

Instrum., 68 (1997) 3307.

2. H. Ohtaki, N. Fukushima, T. Yamaguchi, in: H. Ohtaki, H. Yamatera (eds.),

S t ruc tu re and Dynamics of Solution. S tud ies in Physical and Theoretical

Chemistry 79, Elsevier, Amsterdam, 1992.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) •c 2001 Elsevier Science B.V. All rights reserved. 125

Adsorption of Nonionic Surfactants, Triton X and Triton N, on Hydroxyapatite after Surface Modification with Sodium Dodecyl Sulfate in an Aqueous Phase *)

Saburo Shimabayashi, Masashi Hoshino, Takahisa Ohnishi, and Tomoaki Hino

The University of Tokushima, Faculty of Pharmaceutical Sciences, Sho-machi 1-78-1, Tokushima 770-8505, Japan. E-mail: saburo@ph.tokushima-u.ac.jp

ABSTRACT Although Triton X-100 and Triton N, which are water-soluble nonionic surfactants, were scarcely adsorbed by HAP without SDS, these were adsorbed in the presence of SDS. On the other hand, methyl yellow, a water-insoluble dye, was also captured on the surface of HAP when HAP was treated with SDS. It was concluded that surface modification of HAP by SDS and hydrophobic interaction with SDS on the surface of HAP play important roles in the adsorption.

1. INTRODUCTION

Adsorption mechanism of a surfactant by a solid surface has been discussed so far[l, 2]. It has been concluded that the adsorbed surfactant ions were arrayed and oriented on the adsorbent surface after the adsorption, that is, the hydrophobic tails or head groups were protruded into an aqueous phase from the surface. Whether hydrophobic or hydrophilic group was protruded depends on the combination of an adsorbent and a surfactant. When the adsorption amount is high enough, hemimicelle/admicelle is formed on the surface.

Hydroxyapatite(Cai0(PO4)6(OH)2, HAP) easily adsorbs sodium dodecyl sulfate (Ci2H25S04Na, SDS) mainly through 2 mechanisms: (1) electrostatic attractive force between Ca2+ on the surface and dodecylsulfate anion(DS") of the added SDS, and (2) isomorphous substitution of the surface phosphate ion with a sulfate group of DS" [3, 4]. The surface of HAP becomes hydrophobic to some extent after the adsorption, because the Ci2 hydrocarbon tails of the adsorbed SDS were protruded to an aqueous phase from the surface. These tails interact with each other on the surface (i.e., lateral interaction and formation of hemimicelle), resulting in formation of the hydrophobic domain. It is expected that another hydrophobic compound should be captured by this modified surface through hydrophobic interaction[3, 4]. In the present paper, the interaction between the adsorbed SDS and hydrophobic or amphiphiUc compounds are discussed.

*T This study was supportedby Grant-in-Aid for Scientific Research (C)-(2) #11672143(1999-2000) of The Ministry of Education, Science, Sports and Culture, Japan.

126

2. EXPERIMENTAL

SDS, methyl yellow(MY), polyethylene glycol mono-p-isooctylphenyl ether (Triton X-100,TX-100,n = ca. 10), polyethylene glycol mono-p-nonylphenyl ether (Triton N, TN, n = ca.lO), and matured HAP were purchased from Nacalai Tesque Inc.(Kyoto). The concentrations were determined by an Epton method for SDS, a UV absorptiometry at 274 nm for TX-lOO and TN, and a colorimetry at 418 nm for MY. The experimental conditions are shown in a respective figure caption.

3. RESULTS AND DISCUSSION

3.1. Adsorption of Triton X-100 and Triton N

Triton X-IOO(TX-IOO) and Triton N(TN) are scarcely adsorbed to the surface of HAP in the absence of SDS. This is because their affinity for a raw surface of HAP is weak. However, it was found that these nonionic surfactants were adsorbed to the HAP in the presence of SDS. This fact suggests that the SDS adsorbed on HAP offers the adsorption site for TX-lOO and TN, as expected above.

Figure 1 shows the adsorption amounts of TX-lOO and TN in the presence of 0.50 mmol/dm3 SDS. The adsorption isotherms are sigmoidal, where the adsorption amount of TN was higher than that of TX-lOO and this is due to the fact that the hydrocarbon chain of TN is longer than that of TX-lOO. The adsorption amount of TX-lOO slightly decreased after attaining a maximum. The results for TX-lOO will be mainly discussed hereafter, because those for TN were qualitatively almost the same as those for TX-lOO.

An initial slope in the adsorption isotherm of TX-lOO became steeper while

30

^ 25

20 h

§ 15

I o 10

o

-^

5 h

h r y

200 400 600

[Nonionic surfiictant]free/( inioI/dm )

Fig. 1. Adsorption Amounts of Triton X-lOO(diamond) and Triton N(square). [SDS] = 0.50 mmol/dm3, [HAP] = 25.0 g/dm3,Temp = 30 OC. NaCl was not added.

127

the adsorption amount of TX-lOO at a plateau region decreased after attaining a maximum with a concentration of SDS added (data not shown). These facts suggest that the affinity of TX-lOO for HAP increases with a concentration of SDS when that of TX-lOO is low. On the other hand, when both concentrations becomes high enough, some molecules of TX-lOO tend to interact with SDS to form a mixed micelles in a mother solution but not on the surface. This results in an increase in the concentration of TX-lOO unbound to HAP and, therefore, in a decrease in the adsorption amount of TX-lOO.

The effect of NaCl on the adsorption amount of TX-lOO was also studied. Some results are shown in Fig. 2, where NaCl was added by 500 mmol/dm^. An initial slope of the adsorption isotherm increases in the presence of NaCl. This is the salting-out effect to increase in the adsorption of TX-lOO. On the other hand, the adsorption amount significantly decreases after attaining a maximum in the presence of 500 mmol/dm3 NaCl with a concentration of TX-lOO. This phenomenon might be explained as follows. Formation of the mixed micelle of SDS with TX-lOO in an aqueous solution of 500 mmol/dm^ NaCl is more preferable than formation of the mixed adsorption layer on the surface of HAP.

3. 2. Adsorption of MY on HAP in the Presence of SDS [5]

In the above section, it was shown that formation of the adsorption layer of

^ 30 o B 2t

§ ^

0 200 400 600 800 1000

[TX-100]fr^(/imol/dm3)

Fig. 2. Adsorption Amounts of Triton X-100 in the Presence of NaCl and SDS. [HAP] = 25.0 g/dm3. Temp. = 30 OQ (diamond): [SDS] = 0.50 mmol/dm3, (square): [SDS] = 0.50 mmol/ dm^ together with [NaCl] = 500 mmol/ dm^, (cross): [SDS] = 1.00 mmol/dm^, (triangle): [SDS] = 1.00 mmol/ dm^ together with [NaCl] = 500 mmol/dm3.

128

SDS on the surface of HAP is responsible for the adsorption of TX-lOO and TN. In order to confirm and expand this idea, the adsorption and/or capturing of MY to the surface of HAP was studied. It was found that MY, insoluble in distilled water and nonadsorbable to PiAP, was easily adsorbed to the surface of HAP in the presence of SDS. Some of the results are shown in Fig. 3. The adsorption amount of MY is decreased after attaining a maximum. These tendencies are quite similar to those mentioned above for the adsorption of TX-lOO and TN, and shows that the formation of hemimicelle/adsorption layer of SDS plays a significant role in the adsorption of MY to the HAP surface. Details are mentioned elsewhere[ 5 ].

lb

O 1-1

oT 10 < Di: bO >»

"o £ 5 £

^ >-S X I

(B)

.

" ^

3a

J

% _ «~.& „ a --a-

1 »

10 15

fSDSl^otal/(iniiiol/din3) [SDSl ot /(iiimol/din3)

Fig. 3. Adsorption Amount of MY on the Surface of HAP in the Presence of SDS andNaCl. [HAP]=25g/dm3, [NaCl]= 0(A) and 5 mmol/dm3(B). Data shown in (A) and (B) are for 2 runs. MY was adsorbed from its saturated solution at 30 C. Therefore, chemical potential of MY in the solution is equal to that of MY of the solid phase over the range of the SDS concentration studied [ 5 ].

REFERENCES

I. K. Shinoda, T. Nakagawa, B. Tamamushi, and T. Isemura(eds.), "Colloidal Surfactants", Academic Press, New York, 1963, pp.216-247 2.G. D. Parfitt and C. H. Rochester(eds.), "Adsorption from Solution at the Solid/ Liquid Interface", Academic Press, London, 1983, p. 116, p. 254, p.288, and p. 294. 3. S.Shimabayashi, S.Nishine, and T.Uno, "Adsorption of Hydroxypropylcellulose on Hydroxyapatite via Formation of Surface Complex with Sodium Dodecyl Sulfate", in Zahid Amjad(ed.), "Water Soluble Polymers: Solution Properties and Apphcations", Plenum Press, New York, 1998, Chap.9, pp. 105-116. 4. S. Shimabayashi and T. Uno, "Hydroxyapatite-Polymer Interactions", in J. C. Salamone (ed.), "Polymeric Materials Encyclopedia," vol.5(H-L), CRC Press, Boca Raton, 1996, pp. 3142-3147. 5. S. Shimabayashi, T. Hino, and T. Ohnishi, Phosphorus Research Bulletin, Vol. II, in press(2000).

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) (c 2001 Elsevier Science B.V. All rights reserved. 129

Miscibility of Dodecylpyridinium Bromide and Dodecylquinolinium Bromide in Adsorbed Films and Micelle

Takayuki Fujii, Katsuhiko Fujio, and Sumio Ozeki

Department of Chemistry, Faculty of Science, Shinshu University 3-1-1

Asahi,Matsumoto,Nagano 390-8621, Japan

Surface tension was measured for the dodecylpyridinium bromide (DPB) /

dodecylquinolinium bromide (DQB) / water system. Using the regular solution

approximation for nonideal mixing, the molecular interaction parameter, 0, and

composition were estimated in adsorbed films and micelle. The 0 values

obtained are slightly positive, which indicate that a weak repulsive force acts

between DPB and DQB in adsorbed films and micelle. It is also shown that

adsorbed films and micelle are richer in DQB than bulk solutions.

1.Introduction

In most applications surfactant mixtures are used rather than pure species.

Mixtures of different surfactant types often exhibit synergism in their effects on

various properties of a system. There is increasing interest in understanding the

structure and properties of mixed micelle and monolayer. The behavior of single

surfactant system has been widely investigated, but that of surfactant mixtures

has been investigated only to a limited extent.

In this work the surface tension of aqueous solutions of a mixture of

dodecylpyridinium bromide (DPB) and dodecylquinoUnium bromide (DQB) was

measured. On the base of regidar solution theory the miscibihty of these

surfactants in adsorbed films and micelle was investigated.

2.Experimental

DPB was synthesized by several recrystallizations of dodecylpyridinium

130

chloride in concentrated NaBr solution. DQB was synthesized from 1-

bromododecane and quinoline.

Surface tension was measured as a function of the total molality and fixed

composition of surfactant mixture at 298K by the drop-weight method. Water

doubly distilled from alkaUne permanganate was used.

S.Results and Discussion

Total molality of mixed surfactants is defined by

where mppg and m^Qg are molalities of DPB and DQB, respectively, and the mole

fraction of DPB in total mixed solute is given by

nir

Fig.l shows the concentration dependence of surface tension for the DPB/DQB

system at constant a jrjp^ . The surface activity of DQB is stronger than that of

DPB. CMC increases with increasing (2r ^ .

The molecular interaction parameter and composition can be estimated in

S

-5 -4 5 -4 -3 5 -3 -2 5 -2 -15

log m

Fig. 1. Surface tension vs log total molality for aqueous

solutions of DPB and DQB mixture at constant compositon.

131

adsorbed films and micelle by using the regular solution theory approximation for

nonideal mixing. In adsorbed films the molecular interaction parameter ,0\ is

given by

p" ln(a ,)

( i - ^ D . « r (1)

where ml^„ is the molality of pure DPB at a certain surface tension [1] . XQPQ is

the mole fraction of DPB in mixed adsorbed films which give that particular

surface tension, estimated by

^ DFB *^V^DPB mix,, yyinPH)

(1 - ^DPB f ln[(l - a J,,, )m /(I - X ^,^ )m]^^ J = 1. (2)

In Fig.2 the total molality of mixed surfactant in aqueous solution with the

surface tension of 45mN m" is plotted versus a ^p^ . The X^p^ obtained by

solving Eq.(2) indicates that adsorbed films are richer in DQB than bulk solutions.

The average oi J^"" values calculated by Eq.(l) is 0.24, which means that in

adsorbed films the interaction between DPB and DQB is sUghtly repulsive.

be

'o

s

1 nonideal aodtUfi^ 24)

1 O exper mental

1 —-o--"*' "'*' ^<i^^

be

OOIS

0 01

0 005

nonide*! Mdt 1(0=0 23)

O cxperMtntal

^^^^^^..^^x**^ ^^^^^

Fig. 2. Total molality vs mole fraction of

DPB in the total solute at the surface

tension of 45mN m'

Fig. 3. CMC of mixtures vs mole

fraction of DPB in the total solute

132

Similarly, the molecular interaction parameter in mixed micelles, 0^, is

expressed as

(^ ~ XopB )

where w ^ is the CMC of pure DPB and mf^ is the mixed CMC [2,3] . X'llpg

represents the mole fraction of DPB in mixed micelle, determined by

i^DFB) ^^(Q^DPB I ^DPB ^DPB) _ | (4)

(1 - Xl^.p, y ln[(l - a,,p, )m'^ /(I - X^,, )m^, ]

Fig. 3 shows the dependence of mixed CMC on a^PB • From the X^^p^ obtained

by solving Eq.(4), it is found that mixed micelles are richer in DQB than bulk

solutions. The average j3^ value of 0.23 estimated by Eq.(3) indicates the

weak repulsion between two surfactants in mixed micelles.

Reference

1. M.J.Rosen and X.Y.Hua., J. Colloid Interface SCL, 86,164(1982)

^.D.N.Rubingh iz? "Solution Chemistry of Surfactants", Vol.1,

KLMittalEd, Plenum Press, New York, 1979, p337.

3. P.M.Holland and D.N.Rubingh, J. Phys. Chem,. 87,1983(1984)

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) c 2001 Elsevier Science B.V. A l l rights reserved. 133

interaction and Complex Formation of Pluronic Polymers with Ionic Surfactants

Saburo Shimabayashi, Akihito Ichimori, and Tomoaki Hino The University of Tokushima, Faculty of Pliarmaceutical Sciences, Sho-machi 1-78-1, Tokushima 770-8505, Japan E-mail: saburo@ph.tokushima-u.ac.jp

ABSTRACT: Interaction of a pluronic po(ymer(Plu) with an ionic surfactant was studied. Specific behaviors of a solution of Plu mixed with an ionic surfactant, such as cloud point, relative viscosity, and solubilization of methyl yellow, were discussed, taking the binding isotherm into considerstion.

1. INTRODUCTION

Pluronic polymer(Plu) is a block coplymer of polyethylene oxide(PEO) and polypropylene oxide(PPO), (EO)A-(PO)B-(EO)A, where the PPO group is more hydrophobic than the PEO group. It is, therefore, known that the Plu Interacts more strongly with hydrophobic and amphiphilic compounds than the simple PEO [1]. In the present paper, the interaction and complex formation of Plu with an ionic surfactant were discussed, taking the published informations into consideration[2,3].

2. EXPERIMENTAL

Plu F68(A=76, B=29) and F127(A=100, B=65) were obtaind from Sigma Ltd.(USA). Sodium dodecylsulfate(SDS) and cetylpyridinium chloride (CPC) were products of BDH Ltd.(England) and Tokyo Kasei Ltd.(Japan). Other chemicals used were of the reagent grade from Nacalai Tesque Ltd.(Japan). Binding ratio of a surfactant to a Plu was determined by a dialysis method with a cellulose tubing at 30 00. Viscosity was measured with an Ubbelohde-type capillary viscometer also at 30 00. Oloud point of the solution was observed with the naked eye.

3. RESULTS AND DISCUSSION

Binding ratio(XsDS) of SDS to a Plu was determined at 30 oo over the concentration ranges of 0.025-0.2 g/dl Plu and 0-0.1 mol/dm^ Na2S04. The XSDS sigmoidally increased and then decreased after attaining a maximum, as shown in Fig. 1. The sigmoidal increase means that the binding of SDS to Plu is cooperative, while the decrease suggests that micelle formation of SDS on the polymer chain is competitive with that in the bulk solution[2-5]. The X S D S also decreased with a concentration of Plu, as shown from the top to the bottom of Fig. 1. On the other hand, it increased with a concentration of Na2S04 added, as shown from the left-hand side to the right-hand side in Fig. 1. These facts are showing that the space between the polymer chains, into which SDS could penetrate and bind to the polymer segments, becomes narrower with an increase in the polymer concentration due mainly to the steric hindrance of the polymer chain itself and its

134

<

< 0

Q

< <° ^

<o

o o<

1 1

1 r

Q

\ J j J J

-sU,

E

1 "*

2

g JQ

g

« O

«»

O

(6 / loujuj) X

O

^

OC

,o

o

<

2 ^

5> r

8 8

8 S

2 '

(6 / louitu) X

iVb

(6 / loujui) X

8 «

8 !£

2 (B

/ lOU

JUj) X

S 8

S £

S (6 / louiui) X

S

a

8 IP

2

(0 / lOU

JUi) X

8 2

2 (B

/ lOU

JUJ) X

CM

LL

<D

E

D)

L

m

'v.

•-^

O

o o CO

.«_• CO

00

CO

L

L

T3 c CO

F^

CO

U-3

OL

o U

) Q

C

O

o E

0) ^ *-* o C

CD

T-

o o x>

0)

o GL

O

—• E

0 CO

CV3 (D

C

) f"

3 Q

0) •n CO

T

J c en x

:

1_

(D

x:

•

o ~* CD

CO

o"o

L. .9

C

CO

x: 1 co*r r CD

O

c o CJ

CD

CX)

CD

II 'a?

o

a).ir ji)

CD

F 2

O

CO

C

\J CC

z "o ^^ C

Q

(LS^-B

135

hydration sheath[6]. When the salt, Na2S04' is added, the hydration layer of the polymer becomes thinner and the salting-out effect of the salt on SDS and Plu accelerates the binding of SDS to the polymer. The binding ratio of the surfactants to Plu F127 was almost the same as or slightly larger than that to Plu F68(Fig. 1). Similar tendencies were observed in the binding ratio of CPC to Plu. However, the binding ratio of CPC was remarkably smaller than that of SDS(data not shown).

Cloud point[4] of an aqueous solution of 0.5 g/dl Plu F127 was higher than 100 oc. It decreased with a concentration of Na2S04, while it increased again after attaining a minimum with a concentration of SDS in the presence of 0.15-0.30 mol/ dm^ Na2S04. On the other hand, it increased monotonously with a concentration of CPC(0-8 mmol/dm3) in the presence of Na2S04. Some of the results are shown in Fig. 2. Relative viscosity of a Plu solution decreased after attaining a maximum with a concentration of SDS, while it monotonously increased with that of CPC over

::: ^x>

iicr

locr

90

80

70 L

601

o o o o o o o

A A A A A A A ^ ^

A

D

50^ 2.5 5.0 7.5

[SDS] / mM 10.0 12.5 5.0 7.5

[CPC]/mM 12.5

Fig. 2. Cloud point of an aqueous solution of 0.5 g/dl Plu F127 + an ionic surfactant -hNa2S04. [Na2S04] /(mmol/dm^) = 0.15(circle), 0.25(triangle), and 0.30(square)

1.125

1.100

1.075

1.050'

o o o o

^ o o

22 A

2.5 5.0 7.5 [SDS]/mM

1.125h

1.100

f 1.075h

1.050' 10.0 12.5 5.0 7.5

[CPC]/mM 12.5

Fig. 3. Relative viscosity of 0.5 g/dl Plu F127 in an aqueous solution of an ionic surfactant+Na2S04 at 30 OC. [Na2S04] / (mmol/dm3) = O(diamond), 0.15(circle), 0.25 (triangle), and 0.30(square)

136

the concentration range studled(Fig. 3). Thus, SDS and CPC are considerably different in their effects on the cloud point and relative viscosity as well as their binding ratio. Data for Plu F68 are not shown In the present paper.

These results may be explained as follows, by taking the binding isotherms into consideration. At a low binding ratio of SDS, the SDS forms hydrophobic moiety on the polymer chain, which bridges between the adjacent polymer segments with each other.This fact corresponds to the decrease in the affinity of the polymer coil for water(i.e., lowering a cloud point) and to the increase in the volume of the flow unit of the polymer coil after the aggregation(l.e., an increase In relative viscosity). With an increase in the binding ratio of SDS. the intermolecular bridging collapses due to the increase in the intermolecular repulsion, resulting in the redlspersion of the polymer[2]. This fact appears as a high cloud point and a low relative viscosity. These effects were not observed in the system of CPC, probably because of its low affinity for the polymer or low binding ratio, as mentioned before.

Solubilization of methyl yellow(MY) by the polymer solution was studied. Results are shown in Fig. 4. MY was scarcely solubilized by a solution of 2 g/dl F68 in the absence of a surfactant. The solubilized amount, however, increased with the addition of SDS or CPC, which was higher than that solubilized by SDS or CPC alone. This effect is owing to the complex formation. On the other hand, 2 g/dl F127 showed a high solubillzing ability even in the absence of a surfactant. It increased with a concentration of a surfactant added. Thus, the solubillzing ability of F127 was higher than that of F68 owing mainly to the fact that the content and length of the PPO chain Is larger in F127 than in F68 and, therefore, the mixed micelles[3, 5, 6] of the surfactant and Plu F127 easily capture the MY molecules into them.

0.5,

2 0.4 E

o I E CD

"D , © 0.2/f n

o 0.1

^AA

20 0.5 1.0 1.5 2.0 5 10 15 [SDS]/mM [CPC]/mM

Fig. 4. Solubilization of MY(p-dimethylaminoazobenzene) in the system of Plu and an ionic surfactant at 30 oc. (circle): an ionic surfactant only, (triangle): an ionic surfactant + 2 g/dl F127, (square): an ionic surfactant + 2 g/dl F68

REFERENCES

1. P.AIexandridis and T.A.Hatton, Colloids and Surfaces A, 96(1995), 1. 2. S.Shimabayashi and T.Uno, Prog. Colloid Polym. Sci., 106(1997), 136. 3. M.AImgren, J.van Stam, and P.Bahadur, J.Phys.Chem., 95(1991), 5677. 4. R.Cardoso da Silva and W.Loh, J.Colloid Interface Sci., 202(1998). 385. 5. E.Hecht, K.Mortensen, and H.Hoffmann, J.Phys.Che.. 99(1995). 4866. 6. A.Caragheorgheopol and S.Schlick, Macromolecules, 31(1998), 7736.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) (Q 2001 Elsevier Science B.V. All rights reserved. 137

Formation of Chiral Aggregates of Acylamino Acids in Organic Solvents

H. Matsuzawa *, H. Minami', T. Yano *, T. Wakabayashi *, M. Iwahashi *, K. Sakamoto **

andD. Kaneko**

* School of Science, Kitasalo University, Sagamihara 228-8555, Japan

^ Ajinomoto Co. Kawasaki 210-8681, Japan

Through measurement of circular dichroism (CD), NMR and vapor pressure osmometry

for optically active acylamino acids in CH3CN, these acids were found to be present as

monomers in a polar solvent The decrease in CD intensity and amide proton NMR

chemical shifts with a rise in temperature were considered to arise from conformational

change in the acid molecule.

1. Introduction

Acylamino acids, abbreviation for amino and fatty acids, are mild to living organisms

and easily decomposed through the action of microorganisms. The salts of these acids are

thus often used as mild surfactants,^ whose properties in water, including chiral and racemic

modifications, have been studied in detail.^ The properties of less water-soluble acylamino

acids in organic solvents have been also studied. Optically active acylamino acids in organic

solvents show strong capacity for hydrogen-bonding and form three-dimensional gel-

structures able to trap numerous organic solvents.^ Sakamoto et al found that optically

active acylamino acids form lyotropic liquid crystals in aromatic solvents when suspended in

solvents that do not dissolve the acids and these crystals undergo swelling. During swelling,

optically active acid molecules assume an orientation that results in the formation of a helix.

"* Sakamoto et al found optically active N-lauroylglutamic acid (L- or D-LGA) and N-

lauroylvaline (L- or D-LVA) to completely dissolve in polar solvents such as methanol and

ethanol and show circular dichroism (CD) bands at 212 nm. They proposed the formation

of helix aggregates of acylamino acids.^ However, the size and structures of the aggregates

have not been determined.

For clarification of these parameters for aggregates of L-LGA, L-LVA, or methyl ester of

138

L-LVA (L-LVMe) in polar solvent, CD and NMR measurements were conducted in the

present study. Apparent mean aggregation numbers of L-LGA, L-LVA, and L-LVMe were

also obtained by vapor pressure osmometry (VPO) method. CH3CN was used, mstead of

alcohol, to prevent proton exchange that would make difficult IR and NMR measurements

for solvent-acylamino acid systems.

2. Experimental

L-LGA, L-LVA and L-LVMe (Ajinomoto Co. Ltd.) were used after being dried in vacuum.

CH3CN (spectroscopic grade, Dojin Chem. Co.) and CH3CN-flf3 (98 %, Aldrich Chem. Co.)

were used without further purification. VOP was conducted using molecular weight apparatus

(model 117, Corona Electric Co. Ltd.); diphenylethanedione (Junsei Co. Ltd.) was used for

apparatus calibration.^ NMR was carried out with a JEOL EX-400NMR spectrometer.

3. Results and Discussion

Figs. 1-A and B indicate

representative temperature dependency

of CD and UV absorption spectra for 1.0

X 10- mol dm- L-LVA in CH3CN. The

CD band intensity (Ae, difference in

molar absorptivity for left and right

circularly polarized light) was found to

decrease with temperature, as also noted

when using methanol as solvent.^ The

Steep portion of the absorption band in E 1200

Fig. 1-B at shorter wavelength is the 1

hem of the strong band due to C=0 n-n*

transition, the peak of whose band noted

to appear at 150 nm. The hem has a

shoulder attributable to the weak band

produced as a result of the C=0 n-K*

transition*^ at 215-220 nm. The CD

spectra thus is ultimately given by the

C=0 n-TT* transition.

20'C

245 250

205 210 215 220 225 230 235 240 245 250

Wavelength /nm

Fig. 1 (A) CD and (B) absorption spectra

for 1.0 X 10^ mol dm LVA in CH3CN at 20,

30,40, 50, and 60 °C.

139

u>

0.07

0.06

0.05

0.04

< 0.03

0.02

0.01

0

_ -g-,.

' t^

h

i ...,

•i .

•

• - e -

1 LVMe

LVA

! -

, . _ j ..«.„

10 20 30 40 50

T /x: 60 70

Fig. 2 AE / e VS. temperature for LVA (open

The CD spectra are generally influenced

by the intensity change of the absorption band

itself. Thus, for detailed comparison of Ae,

normalization of this parameter is necessary.

Accordingly, we divided Ae by the molar

absorptivity, e, at the same wavelength.

Fig. 2 shows the temperature dependence

of Ae /e for L-LVA and L-LVMe at 215 nm.

Ae /e obviously decreased with temperature

but did not depend on concentration of L-LVA

or L-LVMe within experimental error. The

decrease in Ae /e may possibly result from

dissociation of helix aggregates.^ To confirm

this point, apparent molar weights of L-LVA symbols) and LVMe (closed symbols):

and L-LGA in CH3CN were determined at Circles indicate 1.0x10^ mol dm ^ triangles,

various temperatures by VPO and then the 5.0 x 10*' mol dm ^ squares, 1.0 x lO"' mol

apparent mean aggregation number was dm' diamonds, 5.0 x 10* mol dm''

calculated.

Surprisingly, the mean aggregation number in nearly all cases was about 1.0. Acylamino

acids exist not as aggregates but as monomers in a polar solvent Namely, the decrease in CD

may possibly arise from the conformational change of the acid molecule and such change

would derive from diminished intramolecular hydrogen bonding (H-bonding) or acid-solvent

H-bonding (solvent H-bonding). To confirm H-bonding, NMR was carried out on L-LVA.

N-H proton signals of L-LVA shifted to a higher magnetic field with a rise in temperature,

with no dependence on L-LVA concentration, as neither did CD spectra. Conformational

change would thus appear due to decrease in intramolecular H-bonding or solvent H-bonding.

However, the intramolecular H-bonding between NH and COOH moieties is structurally

difficult; L-LVMe possessing no COOH moiety is similar in CD decrease to L-LVA, as

evident from Fig. 2

H, N, C and O atoms within the amidic group have resonance structures and are on a same

flat surface. Intramolecular rotation about the C-N bond in the group is somewhat restricted

since the bond has a nature of double bond. Solvent H-bonding between CH3CN and NH of

an acylamino acid enhances electron delocalization of the amidic bond and thus also

140

resonance with consequent restriction of rotation. Restricted rotation may serve to maintain

the chirality of the acylamino acid molecule. At a higher temperature, the solvent H-bonding

weakens, causing electrons to become localized. With the disappearance of the solvent H-

bonding the rotation should occur more easily about the C-N bond. Acylamino acid may

undergo conformational change with consequent reduction in chirality. NMR data for N-

methylacetamide and acylamino acid have shown electron localization with a rise in

temperature. '*^ Based on the present results, a polar solvent may be concluded to influence

rotation about the C-N bond, resulting in conformational change of acylamino acid *° and

lower CD intensity.

References

1. K. Sakamoto, J. Soc. Cosmet. Chem., 35 (1987) 353; J. Oleochem., 44 (1996) 256.

2. H. Staudinger, M. V. Bechker, Ber., 70 (1937) 889; M. Naudet, Bull. Soc. Chim. France,

(1950) 358; P. Heit-mann, European J. Biochem., 3 (1968) 346; M. Takehara et al. J. Am.

Oil Chem. Soc., 49 (1972)143; 50 (1973) 227; 51 (1974) 419; 49 (1972) 157; 51 (1974)

419; M. Shinitzky and R.Haimovitz, J. Am. Chem. Soc. 115 (1993) 12545; T. Imae, Y.

Takahashi and H. Muramatsu, J. Am. Chem. Soc. 114 (1992) 3414.

3. K. Hanabusa, K. Okui K. Karaki, T. Koyama, and H. Shirai, J. Chem. Soc., Chem.

Commun., (1992) 1371, K. Hanabusa, J. Tange, Y. Taniguchi, T. Koyama and H. Shirai,

/7?i ., (1993) 390.

4. K. Sakamoto, R. Yoshida, M. Hatano and T. Tachibana, J. Am. Chem. Soc. 100 (1978)

6998.

5. K. Sakamoto and M. Hatano, Bull. Chem. Soc. Jpn., 53 (1980) 339.

6. R. M. Silverstein, G. C. Bassler and T. C. Morrill, "Spectroscopic Identification of Organic

Compounds" John Wiley &Son, Inc. (1991)

7. K. Imabori "Introduction to the experimental biophysical chemistry 11" Baifukan (1972).

8. S.H. Gellman, G. P. Dado, G-B. Liang and B. R. Adams, J. Am. Chem. Soc. 113 (1991)

1164.

9. M. Akiyama and T. Ohtani, Spectrochimica Acta 50A (1994) 317.

lO.J. Manzur and G. Gonzale, S. Naturfousch B36 (1981) 763.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) P 2001 Elsevier Science B.V. All rights reserved. 141

Formation and Structure Control of Reverse Micelles by the Addition of Alkyl Amines and their Applications for Extraction Processes of Proteins

K. Shiomori^*, T. Honbu^, Y. Kawano^, R. Kuboi^ and I. Komasawa^

^Department of Applied Chemistry, Miyazaki University, 1-1 Gakuenkibanadai-nishi, Miyazaki 889-2192 Japan

^Department of Chemical Science and Engineering, Graduate School of Engineering Science, Osaka University, 1-3 Machikaneyama-cho, Osaka 560-8531 Japan

Extraction of water and lysozyme in mixed micellar systems of sodium bis(2-ethylhexyl) sulfosuccinate and various long chain alkyl amines were investigated. Extraction of water in the mixed micellar systems decreased rapidly with decreasing pH in the aqueous phase. The pH dependence on water extraction is affected by the structure of the amines. Extraction of lysozyme is also controlled by the formation of the reverse micelles, and does not completely occurr in the pH range in which no water is extracted into the organic phase. Lysozyme extracted into the mixed micellar systems can be successfully back-extracted with high activity yield by destruction of the micelles at acidic pH. By increasing the amine concentration, the pH values initiated by the back-extraction of lysozyme are raised, and the activity of the back-extracted lysozyme is decreased.

1. INTRODUCTION

Reverse micelles, which are self-organization molecular aggregates of surfactants in apolar media, have been extensively investigated as media for the extraction of proteins [2, 5]. One intensively studied surfactant is sodium bis(2-ethylhexyl) sulfosuccinate (AOT). The extraction of proteins occurrs by interaction between the protein surface and the reverse micelles. Extraction control of the protein in the usual reverse micellar system is carried out using changes of the surface characteristics of proteins and the size of reverse micelles effected by the pH and salt concentration. However, proteins are known to suffer from denaturation, and the efficiency of protein separation using reverse micelles decreases significantly when the interaction between proteins and micelles is too strong. If it is possible to greatly change the characteristics and structure of reverse micelles by means of external stimulation, it will be possible to construct a more effective separation process of proteins. Some studies have reported on the control of the formation of reverse micelles by pH [3,4], pressure [6] and temperature [1]. Previously, the addition of either tri-«-octylamine or di-2-ethylhexylamine to the AOT reverse micellar system was found effective to control the formation of reverse micelles by pH change [7]. In this paper, various long chain alkyl amines are used for the additives to the AOT reverse micelles. The effect of the amine structure on the formation of reverse micelles and the extraction characteristics of lysozyme by the mixed micellar systems are investigated.

142

2. EXPERIMENTAL

AOT was purchased from Nacalai Tesque Co. The long chain alkyl amines used are shown in Fig. 1. AOT was dissolved in isooctane, then the amine was solubilized in the solution. A buffer solution containing NaCl (0.1 - 1.0 M) was used as the aqueous phase [7]. Egg white lysozyme was purchased from Wako Pure Chemical Co. Extractions of water and proteins were carried out by the phase transfer method [7]. The water concentration in the micellar organic phase was determined by Karl-Fisher titration. The concentration of lysozyme in the aqueous and organic phases, respectively, was measured by the adsorption at 280 nm. Lysozyme activity was measured by the lysis rate of Micrococcus lysodekticus.

3. RESULTS AND DISCUSSION 3.1 Extraction behavior of water

• H

n-OctyUiinlne (OA)

Oi-ff-Octylamlne (DOA)

/CH3 CsH.. . ^

^CH3 Dimethyl-n-Octyiamlne

(OMOA)

CeHi7^ CeH,7-N CeH,7

Trl-n-OctyWimlne aOA)

0139^2 , H

CH3CHiCH2CH2CH CH2-r<^

2-Cthylhexylamine

(EHA)

CH39H2

CHaCHaCHzCHaCHCHav

CHgCHaCHzCHaCHCHa^

CH3CH2

Di(2-Ethylhexyl) amine

(DEHA)

C8Hi7^

C8H,7^

D(-/M>ctyl methylamlne (OOMA)

.N-CH3

Fig.l

/ CH3 CH3 \ CH3CCH2CHCH24N

V CH3 A Trl-/!$o-Octyl«mlne

CnOA)

Structures of long chain alkyl amines used.

The effect of pH on the water extraction in the mixed micellar systems is shown in Fig. 2. Extraction of water in the AOT system is almost constant regardless of pH. In the mixed micellar systems, the water extracted into the organic phase decreased with pH lower than a specific pH value, and was negligibly small at lower pH. This change of the water extraction effected by pH corresponds to the formation of the reverse micelle. The pH values at which the water extraction begins to decrease are dependent on the kind of amine used, and they are raised in the following order: tertiary amines < secondary amines < primary amines, branched-alkyl chain < straight alkyl chain, trialkylamines < dialkyl methylamines < alkyl dimethylamines. These tendencies agree well with the tendency demonstrated by the p/Ca value of amines, determined in the research of extraction behavior of acids by alkyl amines, to change in association with the alkyl group [8]. The control mechanism of the micellar formation in the mixed micellar systems is illustrated in Fig. 3. The amines, B, undergo protonation and form a cationic ammonium ion in the acidic pH range (Eq. (1)). This reaction

2.5

cr2.o

E

E - ?1.0 o

O CM

I 0.5

0.0

Key Amines

• OA O EHA • DOA A DEHA A DMOA n DOMA ^ TIOA O TOA

AOT system

(AOTlo^g = 50 (molAn3l

(Aminel^g =50[mol/m3]

(NaCI]^ = 0 1 [kmolAn3]

_ S 2 a ^ An°^°A

A

vvl s«L

Fig.2

3 5 7 9 11 13 pH H

Effect of pH on water extraction in the long chain alkylamine-AOT system

Fig.3

pH>pKa P' ^^P^^

Long chain alkyl amine (B) OT- (ion form of AOT)

B + H+ ;=rBH+ (1) AOT :^ OT- (2) OT-+BH+;::OT-BH (3) OT-BH + (n-1 )OT- 71 (AOT)n-BH (4)

Schematic illustration of formation control of reverse micelles in the mixed micellar system

143

will be affected by the basicity of the amine added. The ammonium ion reacts with the ion form of AOT (OT") to form the intermolecular ion complex by electrostatic interaction (Eq.(3)). It is considered that the complex has lower surface activities and no ability to form reverse micelles. Further, the higher-order complexation of the ammonium ion with 0T~ expressed by Eq. (4) has been implied when the amine concentration is lower than that of AOT [7]. The decrease in the water extraction is considered to be due to the consumption of AOT, which forms reverse micelles and extracts water, induced by these complexation reactions.

3.2 Forward extraction of lysozyme

Effects of pH on the extraction of lysozyme in the mixed and the AOT single systems, respectively, are shown in Fig. 4. In the AOT system, lysozyme extraction effectively occurred at a wide pH range around neutral pH. In the pH range lower than 5, the extracted fraction of lysozyme into the organic phase, £f, was decreased, whereas the removed fraction of lysozyme from the aqueous phase, Rf, is very high and a large number of aggregates was observed at the interface and in the aqueous phase. Lysozyme is considered to be denatured by strong electrostatic interaction with AOT.

In the DEHA-AOT and DOA-AOT mixed systems, £f was also decreased with pH values lower than a certain pH, a pH higher than that in the AOT system. Under this condition, /?f decreased with pH and the aggregates were not formed. The decrease in both £f and /?f almost corresponded to the decrease in the water extraction. Given that the reverse micelles do not form at acidic pH by the interaction with the alkyl ammonium ion, it is considerd that lysozyme was not extracted, and that the interaction between lysozyme and AOT was supressed.

3.3 Back-extraction of lysozyme from the micellar phase

Back-extraction of lysozyme extracted into the micellar phase at pH 9 and 0.1 M NaCl was carried out by contact with a new aqueous phase. The effects of pH on the back-extracted fraction of lysozyme into the aqueous phase, E\y, the removed fraction of lysozyme from the organic phase, R\^, the concentration of water in the organic phase, and the residual activity of the back-extracted lysozyme, RSA, are shown in Fig. 5. In the AOT system, lysozyme could not be back-extracted using the aqueous phase containing 0.1 M

8 10^12 PHH pi

Fig.4 Effect of pH on extraction of lysozyme in the mixed and the AOT systems

1.0

e e 10 i pHH P»

Fig.5 Effect of pH on back-extraction of lysozyme in the mixed and the AOT systems

144

NaCl, regardless of the pH values. By increasing NaCl concentration up to 1.5 M, lysozyme could be back-extracted in the pH range higher than the pi. Because lysozyme has a very high pi (pl=ll.l), back-extraction of lysozyme in the AOT system requires that the aqueous phase have a at very high pH and high salt concentration. In the mixed micellar system, lysozyme was effectively back-extracted at a more acidic pH range than that of the forward extraction, in which the water extraction is very low and the reverse micelles are not formed. RSA was high and decreased moderately with pH. Back-extraction of lysozyme in the mixed system is considered to be carried out by the destruction of the reverse micelles that are caused by the complex formation between AOT and the cationic ammonium ion formed in the acidic pH range.

By increasing the DOA concentration at a fixed AOT concentration, the pH values, at which the back-extraction of lysozyme began, were raised, and the activity of the back-extracted lysozyme decreased (Fig. 6). However, the concentration change of both AOT and DOA, sustaining an equimolar ratio, was unaffected by the back-extraction behavior (Fig. 7). These results suggest that a quantitative relation among the concentrations of the amine, AOT, and lysozyme is present; this would explain the back-extraction mechanisms of proteins in the mixed micellar system. It is also obvious that lysozyme is denatured by addition of the amine to excess.

REFERENCES

1. M. Dekker, K. Van't Riet, J. J. Van Der Pol, J. W. A. Baltussen, R. Hilhorst and B. H. Bijsterbosch, Chem. Eng. J.,46(I991)B69. 2. M. P. Pileni (ed.). Structure and Reactivity in Reverse Micelles, Elsevier Co., Amsterdam, 1989. 3. M. Goto, K. Kondo and F. Nakashio, J. Chem. Eng.,Japan, 23 (1990) 513. 4. T. Kinugasa, A. Hisamatsu, K. Watanabe and H. Takeuchi, J. Chem. Eng.Japan, 27 (1994) 557. 5. M. E. Leser and P. L. Luisi, Chimia, 44 (1990) 270. 6. J. B. Philips, H Nguyen and V. T. John, Biotechnol ?rog.,7(1991)43. 7. K. Shiomori, Y. Kawano, R. Kuboi and I Komasawa, J. Chem Eng, Japan, 32 (1999) 177. 8. A. M. Eyal, B. Hazan and R. Bloch, Solv. Extr. Ion Exch.,9 (\99\)2\\.

1.0

_ 0.8

— 0.6

0.2 ( A O T l ^ = 5 0 moJAn3

Fig.6 Effect of pH on back-extraction of lysozyme at various concentrations of DOA.

80

£ 60

i 40

20

0 100

80

P E D T C X B X K X B O

i^^lAOTlorg [DGAlofg • '^[molAn3lImolMi3l

^ O 50 50 L ^ 30 30 L n 20 20 r o 10 10

V-

6 8 10 12 PHH

Fig. 7 Effect of pH on bac^ extraction of lysozyme at various concentrations of DOA and AOT.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) (.o 2001 Elsevier Science B.V. All rights reserved. 145

Preparation and surface-active properties of cotelomer type surfactants of alkyl acrylate and acrylic acid

T. Yoshimuraa Y. Koide^, H. Shosenjia and K. Esumi^

aDepartment of Applied Chemistry and Biochemistry, Faculty of Engineering, Kumamoto University, Kumamoto 860-8555, Japan

bDepartment of Applied Chemistry and Institute of Colloid and Interface Science, Science University of Tokyo, Tokyo 162-8601, Japan

Cotelomer type surfactants of w-hexyl acrylate, «-octyl acrylate, 2-ethylhexyl acrylate or «-dodecyl acrylate and acrylic acid having hydropholic groups of 2-3 were prepared by the cotelomerization and examined for surface activity. Surface tensions of aqueous solutions of cotelomers with alkyl chain length of 6-8 were 27-32 mN m'\ The addition of 300 ppm of Ca ^ to aqueous solutions of cotelomers reduced the surface tension. Cotelomers with shorter alkyl chains had high foam stability, while those with branched alkyl chains showed poor stability. Highly stable oil-in-water type emulsions, which were brought about by shaking toluene with aqueous solutions of cotelomers, were formed by using cotelomers having alkyl chains of 2-3 and carboxylate functions of 2-3.

1. INTRODUCTION

Dimeric (gemini) surfactants possessing two alkyl chains and two or three hydrophilic groups and trimeric surfactants possessing three alkyl chains and two or three hydrophilic groups have been developed [1-4]. Dimeric and trimeric surfactants are known to exhibit enhanced surface-active properties such as better water solubility, lower critical micelle concentration (cmc) and more efficient in lowering surface tension of water than conventional monomeric surfactants. These excellent properties may be derived from the enhancement of the interfacial density of surfactants having multi-alkyl chains and multi-hydrophilic groups. However, relations between structures of multi-alkylated surfactants and their properties have not been elucidated in detail. Telomers (oligomers), a type of polymer with polymerization degree (Pn) of 5-20, are obtained by polymerizing monomers in solvents with large chain-transfer constants. Telomers are considered to be suitable for introducing the multi-alkyl chains of the determined number. Recently we have synthesized vinylpyridine telomer type surfactants having multi-alkyl chains and multi-pyridinium functions [5-7]. Many of them showed abilities to lower surface tension and properties of emulsification superior to the corresponding monomeric telomers.

In this study, we prepared multi-alkylated surfactants of cotelomers of «-hexyl acrylate (R^A), «-octyl acrylate (R8(No)A), 2-ethylhexyl acrylate (R8(Eh)A) or w-dodecyl acrylate (R12A) and acrylic acid (AA) with Pn of 3-6 and investigated their surface activities such as surface tension, foaming property and emulsification.

146

2. EXPERIMENTAL

2.1. Preparation Cotelomer type surfactants having alkyl chains of 2-3 (xR^A-yAA, x, y and m mean

number of alkyl chains, number of hydrophilic functions and alkyl chain length, respectively) were prepared by the radical cotelomerization of AA with R^A, RgACNo), RgACEh) and RjA respectively, using 2-aminoethanethiol hydrochloride as a chain-transfer agent in the presence of a,a'-azobisisobutyronitrile as an initiator at 60 °C for 6 h under nitrogen atmosphere. After cooling to room temperature, mixture solutions of sodium hydroxide and methanol were added to the NHoCoH.S^CHg-CHV(cHo-CH4f^ solution and the precipitates were collected by ^ c-0 6=cy filtration. The precipitates were dissolved in water, and the precipitates were obtained by adding ethanol into the solutions. The products were purified by a repeated dissolution-precipitation process and dried under reduced pressure. Structures of cotelomers xR„A-yAA were confirmed by means of IR (Shimadzu IR-408) and 'H-NMR spectra (JEOL JMN-EX-400) and elementary analyses. Fig.l shows structure and abbreviation of xR^A-yAA.

<!)-Rrr 6l Na+

xRn^A-yAA

R6= -CH2(CH2)4CH3 R8(No)= ~CH2(CH2)6CH3 R8(Eh)= -CH2CH(CH2)3CH3

4H2CH3 R12 = -CH2(CH2)ioCH3

Fig.l Structure and Abbreviation

2.2. Measurements Surface tension of aqueous solutions of xR„A-yAA was measured on a Shimadzu Surface

Tensometer (Shimadzu Corp., Kyoto, Japan) using a glass plate. Foaming property was measured by the height of foaming after shaking of cotelomer solutions in vitro with a Shaker machine (Ikemotorika ltd., Tokyo, Japan). Emulsification was determined by the height of emulsion layer after shaking of cotelomer solutions and organic solvent in vitro with a Shaker machine. All measurements were carried out at 25 °C.

3. RESULTS AND DISCUSSION

3.1. Surface tension Cotelomers were soluble easily in water. Surface tensions decreased with increasing

concentration of the cotelomers reaching clear break points, which were taken as cmc. The cmc shifted to lower concentration with increasing alkyl chain length and number of alkyl chains in R^ homologous series. Table 1 lists the values of cmc, surface tension at cmc (Vcmc) and cross-sectional molecular area (A) of some cotelomers. Surface excess concentration (r) at the interface of water and air can be calculated by applying the Gibbs adsorption isotherm equation [8].

r = -( l / iRT)(dY/dlnC) where R is the gas constant, T is the absolute temperature, y is the surface tension and C is the concentration of surfactant. Here, i was taken to be

Table 1 Cmc, Ycmc ^^^ ^ ^^ cotelomers. Cotelomer

2.9R^A-2.3AA

2.1Rg(No)A-12AA

2.8Rg(Eh)A.2.5AA

2.7R12A-2.9AA

^^^ Ycmc .3

/mol dm AriN m 5.4X10"^ 27.7

2.2X10'^ 32.3

7.2X10'^ 29.7

1.8X10'^ 37.8

A

' / A mol'

558

120

284

175

147

y+1 for xR„A-yAA since ionic surfactants are dissociated into positive and negative ions. The A was calculated from A=l / Nr, where N is Avogadro's number.

Cotelomer molecules are considered as bundles of sodium alkanoate of conventional surfactants. Hence surface activities of cotelomers were compared with those of sodium alkanoate. The cmcs of cotelomers were smaller by one to two orders of magnitude than those of monomeric surfactants v^th the same alkyl chain length. 2.1R8(No)A-1.2AA and 2.8R8(Eh)A-2.5AA gave lower cmc than sodium «-dodecanoate [9], which has longer alkyl chain. Cotelomers having shorter alkyl chains gave lower surface tension than the monomeric surfactants, while that having dodecyl chains gave somewhat higher surface tension than sodium w-dodecanoate. The increase of the number of alkyl chains in cotelomers reduced the surface tension and the increase of the hydrophobic chain length rendered the cotelomers less surface active. The values of A of cotelomers were greater than that of sodium Ai-dodecanoate (92A). Cotelomers seem to be adsorbed at surface between water and air by orienting their alkyl chains along with skeletal hydrocarbon chains to air. The cross-sectional molecular areas per one alkyl chain of 2.1R8(No)A-1.2AA and 2.7Ri2A-2.9AA were smaller than that of sodium «-dodecanoate. Cotelomers having 2-3 alkyl chains seem to orient themselves so as to cause effective surface activities due to good balance between hydrophobic functions and hydrophilic functions.

Cotelomers dissolved in hard water containing 300 ppm of Ca ". Cotelomers in the presence of Ca * gave lower cmc, lower surface tension and smaller cross-sectional molecular area than those in the absence of Ca^^ The addition of Ca ^ ion to aqueous solutions of cotelomers improved their surface activities by reducing the static repulsion among carboxylates which prevented the dense packing of cotelomers at surface of water.

3.2. Foaming property The foaming abilities of 2.9R^A-2.3AA, 2.1R8(No)A-1.2AA, 2.8R8(Eh)A-2.5AA and

2.7Ri2-2.9AA in water were as high as that of sodium «-dodecanoate. The foam stabilities of cotelomers were significandy influenced by alkyl chain length and nature of hydrophobic functions. Fig.2 shows a relation between HLB (hydrophile-lipophile balance) and foam stabilities after 60 min of standing. The values of HLB of cotelomers were calculated by Oda's equation [10, 11]. The foam stabilities of cotelomers were correlated to their HLB. Cotelomers with HLB of about 20 gave high foam stabilities in homologous series. xR^A-yAA and xR8(No)A-yAA, which have shorter alkyl chains, had high foam stabilities. xR8(Eh)A-yAA, which has branched alkyl chains, showed poor foam stabilities. Conventional surfactants having branched alkyl chain are known to reduce the foam stabilities [12]. Cotelomers having branched alkyl chains seem to orient themselves at interface less easily than those having straight alkyl chains. It is important to control the balance of hydrophobic functions and hydrophilic functions of cotelomers in order to form the high foam stability.

In the presence of 300 ppm of Ca *, the foaming properties of cotelomers were also influenced by alkyl

60 20 30 40 HLB

Fig 2 Relation between HLB and foam volume after 60 min.

• : R6, • : RgCNo), A: RgCEh), Q: R ^

148

chain length. xRjjA-yAA in the presence of Ca * showed higher foam stabilities than those in the absence of Ca " . In the presence of Ca " , cotelomers were packed closely at surface between water and air due to interaction between carboxylates and Ca^^

3.3. Emulsiflcation Emulsifications of organic solvents were formed

by shaking vigorously with aqueous solutions of cotelomers. The degrees of emulsifications were in the order : toluene > hexane > kerosene > ligroin > chloroform. Fig.3 shows the percentage of emulsion phases prepared with aqueous solutions of xR^A-yAA and toluene. Number of hexyl chains and carboxylate functions of cotelomers affected the stabilization of emulsions. 2.9R6A-2.3AA showed good stability, with more than 60% emulsion phases after 24h. Emulsions that were produced with xR^A-yAA dissolved in water led to oil-in-water (o/w) type emulsions. 2.1R8(No)A-1.2AA, 2.8R8(Eh)A-2.5AA and 2.7Rj2A-2.9AA as well formed highly stable o/w type emulsions. In the present cotelomer molecules having alkyl chains of 2-3 and hydrophilic functions of 2-3, carboxylates as well as alkyl chains are combined by chemical bonds so tightly that they easily give rise to efficient in stabilizing emulsions.

REFERENCES

30 40 Time /min

Fig 3 Relation between elapse of time and volume of emulsion layer.

0:1.1R6A-3.9AA, A: 2.3R^-3.0AA, • : 2.9R6A-2.3AA, • : 3.2R5A-6.5AA.

orient themselves at the interface to

1. M. J. Rosen, CHEMTECH, (1993) 30. 2. M. J. Rosen, D. J. Tracy, J. Surfact. Deterg., 1 (1998) 547. 3. K. Esumi, K. Taguma, Y. Koide, Langmuir, 12 (1996) 4039. 4. E. Onitsuka, J. Beppu, T. Yoshimura, Y. Koide, H. Shosenji, K. Esumi, J. Jpn. Oil Chem. Soc, 49 (2000) 929. 5. Y. Koide, T. Yoshimura, H. Shosenji, K. Esumi, J. Jpn. Oil Chem. Soc, 48 (1999) 123 6. T. Yoshimura, Y. Koide, H. Shosenji, K. Esumi, J. Jpn. Oil Chem. Soc, 48 (1999) 1297. 7. K. Esumi, H. Mizutani. K. Shoji, M. Miyazaki, K. Torigoe, T. Yoshimura, Y. Koide, H. Shosenji, J. Colloid Interface. Sci., 220 (1999) 170. 8. M. J. Rosen, "Surfactants and Interfacial Phenomena", 2nd ed, John Wily & Sons, New York (1989). 9. Nihon-Yukagaku-Kyokai, "Yushi-Kagaku-Binran", Maruzen, Tokyo (1990) 480. 10. A. Fujita, Kagaku No Ryoiki, 12 (1957) 719. 11. R. Oda, K. Teramura, "Kaimen-Kasseizai No Gosei To Sonooyo", Maki Shoten, Tokyo (1962)501. 12. T. Yoshida, S. Shindo, T. Ogaki, K. Yamanaka, "Kaimen-Kasseizai-Handobukku", Kougakutosyo, Tokyo (1987) 156.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) c 2001 Elsevier Science B.V. All rights reserved. 149

Iridescent and coloured colloidal phases in highly dilute systems containing decyl a and |3-D-glucopyranosides, decanol or octanol and water

B. Hoffmann and G. Platz

Department of Physical Chemistry I, University of Bayreuth, D-95440 Bayreuth, Germany

The phase behaviour of decyl a-D-glucopyranoside / decanol or octanol / water is extremely influenced by traces of sodium decylsulfate. No swollen lyotropic phases are present in the pure ternary systems. Milky emulsions which belong to an extended liquid -liquid miscibility gap are obtained above the solution temperature of the surfactant. The addition of traces of sodium decylsulfate to the ternary systems of decyl a-D-glucopyranoside or decyl a/p-D-glucopyranoside / octanol or decanol / water is necessary to induce the formation of highly swollen phases. Iridescence can be obtained with extraordinary bright colours reaching from blue to red. These phases remain far stable below the Kraffi boundaries of the binary systems.

1. INTRODUCTION

Alkyl a/p-glucopyranosides are obtained from Fischer syntheses in an 7:3 ratio together with higher glucosides [1,2]. The alkyl a-D-glucopyranosides are characterized by substantially higher Krqfft points and higher crystallisation energies than those of the P-anomers [2]. Therefore there are various applications for the alkyl p-D-glucopyranosides but until now none for the alkyl a-D-glucosides. Highly diluted lyotropic single phase regions are of special interest for surfactant applications. It is well known that dilute micellar solutions of surfactants in water can be transformed to vesicular dispersions, lamellar and sponge phases when increasing amounts of alcohols with medium chain lengths are added [3]. Such a behaviour is known for polyglucopyranosides [4] but not for alkyl a- or P-D-glucopyranosides. For example octyl p-D-glucopyranoside / octanol / water forms an extended L2-phase but no swollen lyotropic phase [5]. Further more the high Krqfft points of the a anomers hinder the phase formation in the room temperature region. We show that highly dilute lamellar and brilliant iridescent phases can be obtained with decyl a-D-glucopyranoside (a-CioGi) or decyl a/p-D-glucopyranoside / decanol or octanol and water when traces of ionic surfactant like sodium decylsulfate (SDeS) are present.

2. MATERIALS AND METHODS

Decyl a-D-glucopyranoside was obtained by Fischer synthesis and crystallisation from water [2]. Sodium decylsulfate was synthesised according to Dreger et al. [6]. Decanol and octanol were purchased from Fluka Chemical. Water was HPLC quality. - 2 ml samples were

150

prepared in 10ml test tubes with screw caps. After heating and homogenisation the samples were stored at 30 "^C. The iridescent phases develop within several hours or days.

3. RESULTS AND DISCUSSION

Fig. 1 explains the phase behaviour of a-CioGi and water in the dilute region (cf. [7]). The solution temperature of the surfactant remains constant at 45 °C (Krqfft boundary). The crystalline surfactant is found at the bottom of the test tube. On heating above the solution temperature of the surfactant clear isotropic solutions or, at locations within the miscibility gap, turbid emulsions were obtained. The two phase region transforms to clear and optical isotropic solutions above the upper solution temperature. When the turbid emulsions are cooled dovm, clear and isotropic solutions are found from all points within the gap. The crystallisation of the undercooled surfactant solution takes place more than 10 °C below the lowest solution temperature of the two phase region. Thus a closed boundary loop exists between 40 and 93 °C which separates the coazervate from the isotropic liquid single phase region.

100

0 2 4 6 8 10 12 14

aClOGl wt%

0.002 0.004 0.006 decanol wt%

0.008 0.01

Fig 1: phase diagram of a-CioGi Fig 2a Influence of traces of SDeS on the phase • solution temperature on heating diagram of a-CioGi (2.6 wt%) o recrystallisation on cooling 2b Influence of decanol on the phase behaviour A 1(|) to liquid-liquid transition of 0.93 wt% a-CioGi which contains 0.07 %

SDeS, Symbols see fig 1

151

The extension of this miscibility gap decreases strongly when small amounts of sodium decylsulfate are present (fig 2a, cf decyl p-D-glucopyranoside [8]) and increases when decanol (fig 2b) or octanol are added. Only milky dispersions are obtained when the ternary systems of a-CioGi / octanol or decanol / water are heated above the solution temperature of the surfactant. No single phase regions are formed in the investigated temperature region up to 100 °C. On the other side a-CioGi and SDeS form low viscous isotropic micellar solutions above the solution temperature of the surfactant mixtures.

The phase behaviour becomes much more interesting when traces of SDeS are present in the a-CioGi / decanol / water system. At 30 °C 0.9 wt% a-CioGi and 10' wt% SDeS (that means 1.1 % in a-CioGi) form a crystalline dispersion. The phase volume intersection (fig 3a) shows that the crystals are transformed to coloured phases when decanol is added. With 0.15 - 0.25 % decanol an iridescent phase grows from the bottom of the tube below a turbid isotropic phase. This region disappears by about 0.30 wt% decanol and a blue iridescent single phase region is obtained. At 0.73 wt% the system becomes turbid and colourless. Higher amounts of decanol remain undissolved and form a decanol rich L2 phase which separates as concentrated upper phase above the iridescent region. Blue iridescence is found above green, red and colourless which indicates sedimentation effects.

The phase diagram intersection fig. 3b elucidates the transformation from ternary system with decanol to the swollen lamellar phase in dependence of the SDeS concentration. Below 0.6 % SDeS only milky dispersions are found. With concentrations between 0.6 ~ 0.9 % SDeS in a-CjoGi a turbid phase which display blue to green colours is observed. Above 1% a blue iridescent single phase region appears . The iridescent disappears when the amount of SDeS is increased above 1.7 %, however a bluish scattering remains. It should be emphasized that the weight ratio of ionic surfactant to decyl a-D-glucopyranoside which is necessary for the swollen lamellar phase is in the order of 0.010 - 0.017. This means that it is sufficient that only about one of 100 surfactant molecules in the lamellar structure is an ionic one. Increasing the fraction of the sulfate decreases the optical birefringence. This is an indication for a continuous transformation from a planar lamellar structure to a system with charged vesicles.

0.5 4- 0,5 i T ^

1 • • • • < !

C B b

0.2 0.4 0.6 0.8 decanol wt%

1.2 0.5 1 1.5 2 2.5

(mass ratio SDeS / a^^^G^riOO

Fig 3) phase volume intersection, T= 30 °C a) 0.92 wt% a-CioGi, 0.0092 wt% SDeS b) 0.92 wt% a-CjoGi, 0.48 wt% decanol t = slightly turbid, R,G,B = red, green or blue iridescence, m = milky dispersion, c = coloured milky dispersion , b = colourless region with flow birefringence and bluish scattering.

152

Strongly iridescent phases can be obtained with pure decyl a-D-glucopyranoside or with mixtures of decyl a/p-D-glucopyranosides and octanol or decanol. The system a-CioGi/decanol/SDeS in a weight ratio of 97/66/0.97 displays linear swelling on dilution for a total volume fraction > 0.009. This corresponds to a maximum interlamellar distance of 248 nm with a Bragg Peak at 660 nm. The bilayer thickness is 2.30 nm. Some more examples are presented in table 1.

The appearance of two 5r<afgg-peaks (m=l, m=2) in UV-VIS measurements prove the planar lamellar structure and strong wall orientation. The iridescence phases are only little influenced by changes of the temperature. Iridescence was found within the whole investigated temperature region (20 °C - 50 °C). The detailed temperature dependence has not been investigated until now. It is also of interest that a Ls-phase was never found. So the systems behave like typical ionic surfactant systems. The elastic bend moduli of the lowly charged lamellae should be too high for stabilising a highly dynamic sponge phase structure.

Table 1 Iridescent phases

decyl a-D-glucopyranoside decyl a/p-D-glucopyranoside (a:p = 85:15)

colour wt% decanol (wt%) SDeS ^ wt% octanol (wt%) SDeS ^ blue 0,93 0,64 green 0,78 0,53 red 0 64 OM

^ (mass ratio SDeS / a-CioGO* 100 ~~~"

ACKNOWLEDGEMENTS

The authors thank Deutschen Forschungsgemeinschaft (DFG) for financial support.

REFERENCES

1. A.J.J. Straathof, H. van Bekkum, A.P.G. Kieboom, starch/Starke, 40 (1988) 229. 2. V. Adasch, B. Hoffmann, W. Milius, G. Platz, G. Voss, Carbohydr. Res., 314 (1998) 177. 3a) H. Hoffmann, C. Thunig, U. Munkert, H.W. Meyer, W. Richter; Langmuir, 8 (1992)

2629; b) U. Munkert, H. Hoffmann, C. Thunig, H.W. Meyer, W. Richter, Prog. Colloid Polym. Sci.,93(1993)137.

4. G. Platz, J. Policke, W. Kirchhoff, D. Nickel, Colloids Surfaces A, 88 (1994) 113. 5a) S. Kunugi, Y. Hayashi, A. Koyasu, N. Tanaka and M. Shiraishi, Bull. Chem. Soc. Jpn., 68

(1995) 1012; b) J. Chopineau, M. Ollivon, D. Thomas, M.-D. Legoy, Pure&Appl. Chem., 64 (1992) 1757; c) J. Chopineau, D. Thomas and M.-D. Legoy, Eur. J. Biochem., 183 (1989)459.

6. E.E. Dreger, G.I. Keim, G.D. Sheldovsky, J. Ross, Ind Eng. Chem., 36 (1944) 610. 7a) L.D. Ryan, E.W.Kaler, J. Colloid Interface Sci.,201 (1999) 251 ;b)L.D. Ryan, E.W. Kaler,

Langmuir, 13 (1997) 5222; c) M. Kahlweit, G. Busse, B. Faulhaber, Langmuir, 11 (1995) 3382; d) H. Kahl, K. Quitzsch, E. H. Stenby, Fluid Phase Equilibria, 139 (1997) 295.

8) L.D.Ryan, E.W. Kaler, J. Phys. Chem. B, 102 (1998) 7549.

1,01 1,01 1,01

0,92 0,77 0,70

0,58 0,48 0,44

0,66 0,66 0,66

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) c 2001 Elsevier Science B.V. All rights reserved. 153

Complex formation between water-soluble calixarenes and dodecylpyridinium chloride

K. Murakami Faculty of Education, Yamaguchi University, Yoshida 1677-1, Yamaguchi 753-8513, Japan

The complex formation between water-soluble calixarenes (p-sulfonatocalix[n]arenes, n = 4, 6, 8, CALXSn) and dodecylpyridinium chloride (DPC) has been studied by the potentiometric titration method using a surfactant-selective electrode, at [CALXSn] = 1 X lO"" mol dm•^ 1=0.01, pH=2.0, 4.0, and 6.9, and 25°C. The observed binding isotherms showed that the binding of DPC to CALXSn occurs in two stages. The first stage is the strong binding to one site, and the second stage is the cooperative binding to the residual sites. White precipitates have been observed at the beginingof the second stage. The values of the binding constant and the coop erativity parameter have been found to change with an increase in pH, depending on CALXSn species. These results were discussed from the view point of small cooperative binding system compared to large systems.

1. INTRODUCTION

The binding of amphiphilic substances such as surfactants and dyes to macromolecules such as Unear polymers and proteins occurs frequently in concerted manners and sometimes leads to the reduction of their solubility and the conformational changes of the macromolecules including protein denaturation. '"* This cooperative binding phenomenon has been of strong interest for many scientists. Although several theories of statistical mechanical analysis have been formulated^"^ and successfully appHed to linear polymer and globular protein systems, the detailed nature of the interactions involved seems not to be fully recognized in molecular level, since the systems studied are too large and too complex to be discussed in detail. It is therefore desirable to study small cooperative binding systems which are composed of small number of binding sites, as the models of local structures of the macromolecules.

This paper describes the complex formation between /7-sulfonatocalix[n]arenes (n = 4, 6, 8, CALXSn) and dodecylpyridinium chloride (DPC) to discuss the nature of the small cooperative binding system as the model of the binding of amphiphilic substances to protein local structures.

154

2. EXPERIMENTAL

Sodium/?-sulfonatocalix[n]arenes (n = 4, 6, 8) gifted from S u ^ Kagaku Co. were purified by two reciystallizations from an aqueous methanol solution. DPC purchased from Tokyo Chemical Industry Co., Ltd. was purified from three recrystallizations from acetone solution. These were dried in a vacuum at 110 ""C for 24 h. AH the other chemicals used were of reagent grade. The sample solutions were prepared in the three buffer solutions of 1=0.01: pH=2 (KCl- HCl buffer), pH=4 (acetate buffer), and pH=6.9 (phosphate buffer). The extent of DPC binding was measured by the potentiometric titration^ using a surfactant-ion-selective electrode at 25.0±0.1 **C. The potentiometric measurements were made with the electrochemical cell: Ag^AgCl, KCl | salt bridge | reference solution | PVC manbrane | sample solution | salt bridge i A^AgCl, KCl. The slope of the plot of emf. vs. log(surfactant concentration) showed a good Nernstian slope, i.e., 58.9 mV / decade.

3. RESULTS AND DISCUSSION

Figure 1 shows the Scatchard plot ^ for the binding of DPC to CALXSn at pH=2.

15

s

S

1^

10

^ 5

Q

a

-A.

o^'-o-Oo A D

-^^ooX ' •Bj 0 1 2 3 4 5 6 7 8

V

Fig. 1. The Scatchard plots for the binding of DPC to CALXS4(0), CALXS6(A), and CALXS8(n) at pH=2, I-O.Ol, and 25^C.

155

This figure shows that the binding of DPC to CALXSn occurs in two stages; the first stage is the strong binding to one site and the second stage is the cooperative binding to the residual sites. Assuming that these binding stages are independent each other, the number of binding sites (wi) and the binding constant (Ki) for the first stage were at first evaluated from the data in the low concentration region where the second stage does not appear. Next the binding isotherms for the second stage have been calculated by subtracting the contribution of the first stage from the overall binding number. The number of binding sites (W2) and the cooperativity parameter (w) for the second stage were evaluated from the binding isotherms. Here, u is defined, at the half saturation point, by ' ^

u={4(d6/d\nLf)y^,,, (1)

where 6 is the degree of binding calculated as the binding number devided by the number of binding sites. The values of these parameters are listed in Table 1.

Table 1 Binding parameters and cooperativity parameters for the binding of DPC to CALXSn at 1=0.01 and 25°C.

pH

2 4 6.9

CALXS4 CALXS6 CALXS8

wi 10%/morMm^W2 w Wi 10%/morMm^ «2 w Wi 10%/mor'dm^ «2 w

0.88 4.3 3 9.9 1 0.68 4 105 1 5.3 6 106 0.78 7.6 3 71. 1 1.88 5 69 1 14. 6 237 0.84 13. 4 18. 1 18.8 5 68 1 80. 7 144

The values of Ki are comparable to or larger than the values of the binding constants reported for the stilbene dyes and alkylammonium ions binding to CALXSn. ^ The large values ofKi suggest that the hydrophobic interaction between the alkyl chain of DPC and the cavity of CALXSn as well as the electrostatic interaction between the charged groups are incorporated in this binding stage. Ki for each CALSXn has the tendency to increase with an increase in pH, i.e., the deprotonation of the hydroxyl groups of CALXSn.

The values of the intrinsic binding constant of the second stage, K2, were estimated to be of the order of 100 mol'Mm^ this suggests that the main force of this step is electrostatic in nature. «2 tends to increase with an increase in pH. The fact that the values of/72 are smaller than the numbers of the residual ne^tively-charged groups, which are estimated from the pKa values ' "* of CALXSn, shows that not all of these groups serve as the cooperative binding sites. The values of A'2 and u were found to be comparable to those for the other

156

polymer-surfactant systems in the similar conditions. ' ^ The pH dependences of u for

CALXS4 and CALXS8 systems have the similar tendency of having maximum values at

pH=4, but u for CALXS6 tends to deaease with an increase in pH. These results suggest

that the differences in the binding bahavior arise from the differences in the conformation of

CALXSn and the manner of its change with pH.

For all of the present systems, white precipitates have been observed at the begining of

the second stage, i.e., around v - 1 . This means the presence of neutral complexes to form

precipitates at the low level of binding and suggests that there are two kinds of complex

species; one is the species which bound one DPC molecule due to the first stage and the

other is the species in which almost all of the binding sites are occupied by DPC molecules.

The latter neutral complexes tend to aggregate. This all-or-none type binding behavior which

is typical of the strong cooperative binding may be attained by the small system size. This

is in contrast to the case of the large sy stems such as sodium dextransulfate-surfactant

systems, in which the precipitates appear only at the high binding level where almost all of

the sulfate groups are bound by the surfactant molecules.^^ This difference in the binding

behavior between the small and the large systems may be due to the difference in the

magnitude of the cluster size of bound surfactants relative to the system size. ^

REFERENCES

1. J. Steinhardt and J. A. Reynolds, Multiple Equilibria in Proteins, Academic Press, New

York, 1969.

2. S. Lapanje, Physicochemical Aspects of Protein Denaturation, John Wiley & Sons, New

York, 1978.

3. K. Murakami, Bull. Chem. Soc. Jpn., 71 (1998) 2293.

4. K. Murakami and K. Tsurufuji, Bull. Chem. Soc. Jpn., 72 (1999) 653.

5. G. Schwarz, Eur. J. Biochem., 12 (1970) 442.

6.1. Satake and J.T. Yang, Biopolymers, 15 (1976) 2263.

7. J.D. McGhee and P.H. Von Hippel, J. Mol. Biol., 86 (1974) 469.

8. K. Murakami, Langmuir, 15 (1999) 4270.

9. K. Shirahama, Y. Nishiyama, and N. Takisawa, J. Phys. Chem., 91 (1987) 5928.

10. G. Scatchard, Ann. N. Y Acad. Sci., 51 (1949) 660.

11. T.L. Hill, Cooperative Theory in Biochemistry, Springer, New York, 1985.

12. M. Nishida, D. Ishii, I. Yoshida, and S. Shinkai, Bull. Chem. Soc. Jpn., 70 (1997) 2131.

13.1. Yoshida, N. Yamamoto, F. Sagara, D. Ishii, K. Ueno, and S. Shinkai, Bull. Chem. Soc.

Jpn., 65 (1992) 1012.

14. M. Sonoda, K. Hayashi, M. Nishida, D. Ishii, and I. Yoshida, Anal. Sci., 14 (1998) 493.

15. Y. Moriyama, K. Takeda, and K. Murakami, Langmuir, in press

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) © 2001 Elsevier Science B.V. All rights reserved. 157

Influence of Oil Droplet Size on Flocculation/Coalescence in Surfactant-Free Emulsion

Toshio Sakai\ Keiji Kamogawa''' , Fuminori HarusawaS Nobuyuki Momozawa*^ Hideki Sakai"' and Masahiko Abe'^

Taculty of Science and Technology, Science University of Tokyo, 2641, Yamazaki, Noda, Chiba 278-8510, Japan **Ele. & Sec Ed. Bureau, the Ministry of Education, Sports, Culture and Science, Kasumigaseki, Chiyoda, Tokyo 100-0013, Japan ""Institute of Colloid and Interface Science, Science University of Tokyo, 1-3, Kagurazaka, Shinjuku, Tokyo 162-8601, Japan

We investigated the influence of oil droplet size on the growth processes (flocculation and/or coalescence) of benzene and fluorobenzene droplets ultrasonically dispersed in water by freeze-fracture electron microscopy (FFEM). Oil droplets with the size of 100-1000 nm (M class) and those with the size of >1000 nm (L class) were formed through flocculation of small droplets ( 100 nm, S class) and coalescence of M class droplets, respectively. Furthermore, FFEM observation verified the presence of benzene and fluorobenzene droplets with diameters below lOOnm and appearance of discrete distributions of S and M class droplets.

1. INTRODUCTION

Emulsions are mixtures of two immiscible liquids, such as oil and water, stabilized by an emulsifier. Emulsions are thermodynamically unstable systems that generally break down over a short time through a variety of physicochemical destabilizing processes, e.g., gravitational separation, flocculation, coalescence, and the Ostwald ripening [1-4]. These destabilizing processes cause changes in the spatial distribution and/or size of droplets.

We have investigated surfactant-free O/W emulsions (SFEs), which can be prepared without addition of surfactant with ultrasonicator at extremely low oil concentrations (-0.1 vol%) slightly above oil solubility [5-7], to elucidate the mechanism of stability and growth processes of emulsions. SFE is the simplest system of emulsions capable of helping us clarify the evolution and growth processes, and stabilizing factors of emulsion systems.

In our recent investigation, discontinuous growth of oil droplets and discrete droplet size distributions were observed for SFEs by dynamic light scattering (DLS) measurements. For example, SFE using benzene as the oil phase had three size distributions appearing one by one with the lapse of time; 20-100 nm (S class), 200-1000 nm (M class) and 3000-4000 nm

158

(L class)[5]. But DLS couldn't easily distinguish growth processes, for example, flocculation and coalescence.

In this paper, we focus on the growth processes of benzene and fluorobenzene droplets in SFEs and examine the processes by a combination of dynamic light scattering (DLS) and freeze fracture electron microscopy (FFEM). FT EM is very useful to discuss the growth processes because it can yield direct imaging of size, aggregates, and shape of oil droplets dispersed in water.

2. EXPERIMENTAL SECTION

Materials and Preparation of dispersions Benzene, and fluorobenzene (Tokyo Kasei Co., Ltd.) were GR-grade and used as

received. Distilled and deionized water of injection grade (Ohtsuka Pharmacy Co., Ltd.) was used without further purification. Oil (weighed by volume with a syringe) was mixed with water in a flask and the mixture was kept at 25 "C. The concentrations of oils (benzene and fluorobenzene) added to water were 30mM and 17mM, respectively, which were determined from the solubility of the oils in water. The mixture was then sonicated for 2 min in a cleaning bath (Bransonic 220, 125 W, Smith Kline Company).

Freeze Fracture Electron Microscopy Immediately after the dispersion treatment, samples for electron microscopy were

prepared by freeze-fracture replication. A small volume (-10 ^L) of each sample was placed on a small holder plate (0=3 mm; Hitachi) and the samples were frozen in liquid nitrogen at -190 T . The specimens were transferred to a freeze fracture device (Hitachi, FR-7000A), fractured at -120 "C and < 10^ Torr. The fractured surfaces were immediately replicated by evaporating platinum-carbon mixture from an electrode at an angle of 45"* onto the fractured surface, followed by carbon film at normal incidence, to increase the mechanical stability of replica. The replicas were washed in acetone (Tokyo Kasei Co., Ltd.), rinsed with distilled and deionized water of injection grade (Ohtsuka Pharmacy Co., Ltd.), and collected on 300 mesh copper electron microscope grids (OKENSHOJI Co., Ltd.). The replicas thus prepared were examined in a transmission electron microscope (JEM1200EX, JEOL) in conventional transmission mode using 80-kV electrons. Images were recorded on an electron-microscopic film (FUJI PHOTO FILM Co., Ltd.). Shadows (absence of platinum) appear light in the prints.

3. RESULTS AND DISCUSSION

Figures 1 and 2 show the oil droplet size distribution obtained by DLS and FFEM images of benzene and fluorobenzene droplets, respectively, immediately after sonication. For benzene SFE, the number based-droplet size distributions (diameter) at 20^-200 nm (S class) was observed by DLS [5]. Furthermore, spherical droplets with diameters at

159

>«..

f' #^

Figure 1 FFEM image of benzene droplets (30mM) dispersed in water immediately after sonication.

30-100 nm (S class) were also observed by FFEM as

shown in Fig. 1. For fluorobenzene SFE, the droplet

size distribution at 60-900 nm (M class), which is different

from that of benzene SFE, was monitored by DLS

immediately after preparation as shown in Fig. 2(a).

Figure 2(b) is FFEM of fluorobenzene droplets, in which

droplets at 100-1000 nm are observed. These sizes are

in agreement with those obtained by DLS. Furthermore,

droplets at 20-60 nm (S class) could be observed by FFEM

in another spot of the replica as shown in Fig. 2(c). The amount of droplets with these

size, however, was less than that of M class droplets, which could be hardly detected by DLS.

The size of S class droplets observed for both benzene and fluorobenzene is one order of

magnitude smaller than those of M class droplets.

10 Dropiet size / nm

(b) Figure 2 (a) Size distribution measured by dynamic light scattering (Sub-miaon particle analyzer system 4700 Malvern Instrument Co.), (b) FFEM image at M class droplets, and (c) at S class droplets of fluorobenzEne (l7mM) dispersed in water immediately after sonication.

Figures 3 and 4 show FFEM images of benzene and fluorobenzene droplets 60min after

sonication, respectively. We were able to observe those images which show benzene

droplets at S class size flocculated keeping their shape and size as individual droplets in Fig.

3, and two or more fluorobenzene droplets at M class size merged into a bigger one in Fig. 4.

We also obtained similar results, the coalescence process, for benzene droplets at M class.

Thus, these processes (flocculation and/or coalescence) are independent of the kinds of oil,

but depend on droplet size (S and/or M class droplets). These results proved that M and L

class droplets are formed through flocculation of S class droplets and coalescence of M class

droplets, respectively. These findings are explained in terms of the classical Derjaguin-

Landau-Verwey-Overbeek (DLVO) theory [8, 9]. When the surface-to-surface distance between two droplets

V * %

{ >

; • >

Figure 3 FFEM image of aggregaes of benzene droplets (S class)60min after sonication.

SL Figure 4 FFEM image of coalescence of fluorobenzene droplets (M class) 60min after sonication.

becomes less than twice the

thickness of the electrical

double layer (K"') around

the droplets there arises a

strong electrostatic

repulsion between the

droplets [10-12]. The

electrostatic repulsion

160

between droplets increases with decrease in droplet size because of the overlapping of the electrical double layers surrounding the droplets. In fact, benzene droplets in surfactant-free condition had a ^-potential around -35 mV, because hydroxyl ions in water adsorb preferentially on the droplet surface due to the difference in dielectric constant between the oil and aqueous phases [13], and the value decreased with increase in droplet size [7, 14]. Weiss and McClements reported that 25 wt% n-octadecane oil-in-water emulsions stabilized with sodium dodecyl sulfate (SDS) underwent a liquid -to- solid change when the oil droplet size decreased below 85 nm (radius) because of the overlapping of the electrical double layers surrounding the droplets [10].

As has been mentioned so far, FFEM can provide valuable information on fine liquid droplets and their growth. The technique verified the presence of benzene and fluorobenzene droplets with diameters below lOOnm and appearance of discrete distributions of S and M class droplets.

REFERENCES

1. D. F. Evans, H. Wennersrrom, THE COLLOIDAL DOMAIN SECOND EDITION, Wiley-VCH, Inc.: New York, 1999.

2. D. J. McClements, Food Emulsions: Practice and Techniques; CRC Press: Boca Raton, FL, 1998.

3. R. J. Hunter, Foundations of Colloid Science; Oxford University Press: Oxford, 1986; Vol. 1. 4. P. C. Hiemenz, R. Rajagopalan, Principles of Colloid and Surface Chemistry, 3rd ed.;

Marcel Dekker: New York. 5. K. Kamogawa, T. Sakai, N. Momozawa, M. Shimazaki, M. Enomura, H. Sakai, M. Abe,

J. Jpn. Oil Chem. Soc. 47 (1998) 159. 6. K. Kamogawa, M. Matsumoto, T. Kobayashi, T. Sakai, H. Sakai, M. Abe, Langmuir 15

(1999) 1913. 7. K. Kamogawa, H. Akatsuka, M. Matsumoto, S. Yokoyama, T. Sakai, H. Sakai, M. Abe, to

be submitted. 8. B. V. Derjaguin, L. D. Landau, Acta Physicochim. URSS 14 (1941) 633. 9. E. J. W. Verwey, J. Th. G. Overbeek, Theory of the Stability of Lyophobic Colloids,

Elsevier, Amsterdam, 1948. 10. J. Weiss, D. J. McClements, Langmuir 16 (2000) 2145. 11. J. W. Goodwin, A. M. Rhider, Colloid and Interface Science, Kerker, M., Ed.; Academic Press: New York, 1976; Vol. 4, p 529. 12. R. Buscall, J. W. Goodwin, M. W. Hawkins, R. H. Ottewill, J. Chem. Soc., Faraday Trans. 78 (1982) 2889. 13. R. S. Schechter, A. Garcia, J. Lachaise, J. Colloid Interface Sci. 204 (1998) 398. 14. K. Kamogawa, N. Kuwayama, T. Sakai, H. Sakai, M. Abe, to be submitted.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) o 2001 Elsevier Science B.V. All rights reserved. 161

Morphology of Microemuision Droplet Confining a Single Polymer Chain

K. Nakaya", M. Imai", I. Miyata^ and M. Yonese*'

'Department of Physics, Faculty of Science, Ochanomizu University, 2-1-1 Otsuka, Bunkyo, Tokyo 112-0012, Japan

''Faculty of Pharmaceutical Sciences, Nagoya City University, 3-1 Tanabe-dori, Mizuho, Nagoya 467-8603 Japan

We have investigated morphological change of a microemuision droplet induced by confinement of a polymer chain using a small angle neutron scattering (SANS) technique. The confinement induces two morphological changes, 1) increase of the mean droplet size and 2) increase of the droplet size polydispersity. The increase of mean droplet size can be described by a simple scaling concept on the basis of membrane rigidity and entropy loss of polymer chain by the confinement.

1. INTRODUCTION

Microemulsions are thermodynamically stable structured fluids consisting of water, oil, and surfactants. By changing compositions or external fields, morphologies of the surfactant membrane shows a variety of meso-structures having order of hundreds of angstroms, such as droplet (sphere), lamellar, and bicontinuous structures. Among these morphologies, water-in-oil droplets provide isolated meso-scale water region separated by continuous oil phase.

Recently behaviors of surfactant membrane confining polymer chains have received much attention. For example, the confinement of polymers between lyotropic lamellar structure brings change of elastic nature of the membrane [1]. If we introduced water soluble polymers into the water-in-oil microemuision droplet system, polymer chains can be confined in the spherical meso-space surrounded by the surfactant membrane. Lai and Auvray [2] investigated perturbations of microemuision droplets by confining poly(ethylene glycol) chains and observed increase of the droplet size distribution. Quellet et al. [3] and Hearing et al. [4] introduced concentrated associating polymer (gelatin) into dense microemuision droplet system (isooctane/Aerosol OT/water) and found that the droplet system transforms to three dimensional network structure (formation of transparent stable gels). Thus the confined polymers affect behavior of surrounding membrane deeply.

In this study we introduced a single associating polymer chain into a microemuision droplet, where the size of the droplet is comparable to the radius of gyration of the confined polymer chain and investigated morphology change of the surfactant membrane using the SANS method.

162

2. EXPERIMENTS

In this experiment, we used water-in-oil microemulsion droplets consisting of isooctane, Aerosol-OT, and water. For the neutron scattering experiments, deuterated isooctane and deuterated water were mixed with corresponding hydrogenated components in order to adjust the scattering length densities. As confined polymers we used gelatin (type A, Bloom 300) having molecular weight of 9.6x 10\ The radius of gyration (Rg) of this gelatin is 81 A, which is obtained by the SANS measurement.

We varied the droplet radius by changing the water to surfactant ratio (COQ) in the vicinity of the Rg of the gelatin. The volume fraction of droplets (water+AOT) was fixed to 7%. The microemulsion droplets confining polymer chains were made as follows: gelatin was allowed to swell in water at 60 °C and then mixed with solution of AOT in isooctane at 60 °C. After well stirring, the samples were transferred to SANS cells and hold for 1 hour at 30 °C before the measurements.

The SANS measurements were performed using a SANS-U instrument of Institute for Solid state Physics, the University of Tokyo at JRR-3M reactor of Japan Atomic Energy Research Institute at Tokai [5]. All measurements were carried out at 30 " C.

3. RESULTS AND DISCUSSIONS

First we show behavior of microemulsion droplets without polymers. In Fig. 1 we plot the scattering profile for coo(=[H20]/[AOT])=57.6. The scattering profile shows a broad peak at ^=0.03 A ' and q'^ decay in ^>0.1 A"' region, which is typical behavior for microemulsion droplets under film contrast condition. The q'^ decay indicates sharp and smooth membrane interface of sphere droplets. The obtained scattering fimction can be described by a scattering fimction for spherical shell model [6] taking account of size polydispersity given by

I(q)-Np(r)P{q,r)dr (1)

P(q)'l6n'{p, -pJ{Rlf,{qR,)-AI^fo{qRr)f (2) with /o= (sinx- Jccos;c)/x^

^'{Ps-p^er)/{ps''Pou)

where ^N is the number density of droplets, P(q) is form factor, and p , Po,,, and p^,„ are the scattering length density of surfactants, oil, and water, respectively. For the polydispersity function we adopted a Schultz distribution /(r). Here we ignore the influence of the structure factor because we use the dilute system. The shell model without the structure factor well describes the observed profiles and as an example, the fitted curve is given as solid line in Fig. 1. From the fitting we extracted two parameters, mean radius of the droplet (R) and the polydispersity parameter p {p^=<R^>/<R>^'\). We obtained good linear relationship between R and CDQ, and the samples with a)o=57.6, 65.4, and 74.0 had /?= 78.0, 86.0, and 98.0 A, respectively.

Next we examined the microemulsion droplets confining a polymer chain system. The average number of polymer chain in a droplet (/ip) is calculated from the number of

163

0.01 n ni ^

q [A-»] Fig.I. SANS profiles for the system of AOT-iso-octane-water microemulsion droplets of a)Q=57.6 without(circle) and with(trianglc) gelatin at 30*C. The solid lines arc curves fitted with the shell model.

Fig.2. Changes of droplet radius (open) and polydispersity (closed) of microemulsion without (circle) and with (triangle) gelatin.

droplet and the number of gelatin chains in the system. In this study we prepared three samples a)o=57.6 (/Zp=0.82), 65.4 (0.88), and 74.0(0.88). We plot the SANS profile for the droplet ((Oo=57.6 and /jp=0.82) with corresponding SANS profile without polymer in Fig. 1. The features of scattering profiles for the droplets confining a polymer chain are smearing of the characteristic peak of the droplet and asymptotic q^ behavior in high q region. The latter indicates that the droplets are not deformed significantly by the contmement. It should be noted that increase of scattering intensity in low q region is due to unmatching of the scattering length densities between inner water pool containing hydrogenated gelatin chain and outer oil matrix.

We fitted the experimental profiles with the shell model described above. The obtained R and/? without gelatin and with gelatin are plotted as a ftmction of COQ in Fig. 2. The R and p are increased by the confinement of a polymer chain. Taking account that the gelatin chain has i^g=81 A, the smaller droplets confine a polymer chain, the larger increase of the droplet size and polydispersity are observed. Unfortunately at present, we can not make clear origin of observed increase of the polydispersity. Here we focus our attention on the change of the mean droplet size.

In order to explain observed increase of the mean droplet size, we consider a simple scaling law describing free energy of droplet membrane confining a polymer chain. The droplet membrane with the spontaneous curvature HQ (or sphere radius R^ confines a polymer chain having Flory radius of chain R^, resulting in the formation of sphere droplet having the radius of R, In this treatment we assume that the polymer adsorption at the interface and electrostatic interaction are negligible. The total free energy of the system may be composed of polymer confinement term [7] and membrane elasticity term (Helfrich expression) [8] as

F ^ F +F ^tot •• conf ^ ^mt

~T(^y^f[2K(H-H,y^KK]dS (3)

where T is temperature, K is the bending modulus, K is the saddle-splay modulus, H is mean curvature, and K is the Gaussian curvature. For simplicity we assume that the

164

o

? 1 s 0.9

0 8 0 7

•

oT"

• • 0.45

09 Rg/R

1

Fig.3 A log-log plot of R/RQ versus R /R. The slop represents the power law behabior (R>T^h(Rg/R)0*5.

Then we can express F,o, using three observable

(4)

(5)

droplet radius without polymer is RQ, parameter, R^, R, and R^ as follows

Fu. ~ n j p ^ +4;I[2K - 4 K ( - | ) + 2 K ( - | ) ^ + K ] .

Minimizing the free energy against R gives a simple scaling law

In Fig. 3 we compared the experimental data with the scaling law, where we replaced Rj: by ^g of the gelatin. Roughly speaking, the experimental data obeys this simple scaling law, although the data is limited in the narrow region. This is because that if we confine the polymer chains in a smaller droplet, large deformation of droplet is observed. This issue will be discussed forthcoming paper.

ACKNOWLEDGEMENTS

This work was supposed by Sasakawa Scientific Research Grant from the Japan Science Society, Grant of ihe Sumitomo Foimdation (No. 990575), and Grant-in-Aid for Exploratory Research (No. 12874047) from the Ministry of Education, Science, and Culture of Japan. This work was done under the approval of the Neutron Scattering Program Advisory Committee, Japan (No. 9564).

REFERENCES

1. C. Ligoure, G. Bouglet, G. Porte, and O. Diat, J. Phys. II (France), 7 (1997) 473. 2. J. Lai, and L. Auvray, J. Phys.II (France), 4 (1994) 2119. 3. C. Quellet, and H.-F. Eicke, CHIML\, 40 (1986) 233. 4. G. Haering, and P.L. Luisi, J. Phys. Chem., 90 (1986) 5892 . 5. Y. Ito, M. Imai, and S. Takahashi, Physica B, 213&214 (1995) 889. 6. M.Gradzielski, D.Langevin, and BFarago, Phys. Rev., E 53 (1996) 3900. 7. P.G. de Gennes, Scaling Concepts in Polymer Physics, Cornell I niversity Press, New York, 1979. 8. W.Helfrich: Z.Naturforsch, 28 (1973) 693.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) c 2001 Elsevier Science B.V. All rights reserved. 165

AOT microemulsion structure depending on both apolar solvent and protein concentration

Rika Kawai-Hirai,t Mitsuhiro Hirai, * Hisae Futatsugi,^ Hiroki Iwase,§ and Tomohiro Hayakawa.S

tMeiwa Gakuen Junior College, Maebashi, Gunma 371-0034, Japan.

^Department of Physics, Gunma University, Maebashi 371-8510, Japan.

By using synchrotron radiation small-angle X-ray scattering (SR-SAXS) method we studied w/o AOT (sodium bis(2-ethylhexyl)sulfosuccinate) microemulsion systems (water/AOT/n-hexane, /i-heptane, and n-octane) entrapping proteins. We foimd that the structure of the w/o AOT microemulsion clearly depends on the protein concentration and the length of hydrocarbon chain of apolar solvent.

1, INTRODUCTION

Enhancement of catalytic activity, so-called super-activity, of enz3nnes entrapped in w/o microemulsions has attracted significant interest concerning not only with future practical applications such as microreactors but also with biophysical catal3rtic mechanisms of enzymes at a limiting condition such as a water pool of w/o microemulsion [1, 2]. By using SR-SAXS and enzymatic activity measurements we clarified that the catalytic activity of a-chymotrypsin entrapped in the water/AOT /isooctane microemulsion is enhanced at low WQ (= [H201/[A0T]) range of 8-16 and that the three different phases (oligomeric phase, transient phase and monomeric phase) appear successively with increasing WQ value [3]. Our neutron spin echo (NSE) study of water/AOT/heptane system showed that the effective diffusion coefficient relating to the bending fluctuation of the microemulsion is significantly enhanced at the transient phase [4]. As shown in other SR-SAXS study [5], there exists the penetration limit of apolar solvent depending on the linear hydrocarbon chain length, which results in the shift of the WQ value of the above phase boundaries. These previous studies suggest that the presence of the transient phase and the enhancement of the bending fluctuation of the microemulsion would induce the increase of an effective surface area of enzymes for the contact with substrates, which would result in the acceleration of the metabolic turnover. Then, we have carried out further SR-SAXS experiments to examine more precisely the structural features of the various AOT microemulsions entrapping enzymes.

Corresponding author.

166

2. EXPERIMENTAL

AOT was purchased from Nacalai Tesque Inc. Apolar solvents used were 96.5+ % n-hexane, 99.9+ % n-heptane and 97+ % n-octane, which were purchased from Wako Pure Chemical Industries Ltd. The enzyme used was a-chymotrypsin from bovine pancreas, type 11 produced by Sigma Chemical Co. Water was purified by a Millipore system. The AOT microemulsions were prepared by using an ii jection method. The wo values of the samples were varied from 0 to 50. The AOT molar concentrations were 0.1 M for all samples. a-Chjrmotrypsin was solubilized in 10 mM Hepes buffer adjusted at pH 8.0. The molar concentrations of o-chymotrypsin in the samples were varied from 2.4x10*5 M to 2.8x10"^ M depending on WQ. SR-SAXS experiments were carried out by using a small-angle X-ray scattering equipment installed at the synchrotron radiation source (PF) at the High Energy Accelerator Research Organization (KEK), Tsukuba, Japan. The measurement condition was the same as the previous experiments [5]. We carried out the following standard analyses for the obtained scattering data as follows. To estimate structural changes of the AOT microemulsion depending on both a-chymotrypsin concentration and apolar solvent, we calculated the distance distribution functions p(r) by the following Fourier transform of the scattering intensity Kq).

p(r) = -2-J''^/(^)sin(r<7)^ (1) 0

where q is the magnitude of the scattering vector defined as q- (4;r/A)sin(^/2), 6 and Xy are the scattering angle and the X-ray wavelength. The radius of gyration -Rg of the solute particle was obtained from the p(r) function as follows.

R^^kjll (2)

3. RESULTS AND DISCUSSION

Fig. 1 shows the wo dependence of the scattering curves Kq) of the AOT microemulsions with three different apolar solvents (n-hexane, n-heptane and /i-octane) imder the constant protein concentration of 2.4xl0"5 M. The decrease of the scattering intensity below q = 0.02 A' is attributed to the beam stopper. With increasing the wo value, the scattering curve shifts to a small-^ region with the rise of the scattering intensity below q = -0.04 A* for all cases of the three different AOT microemulsion systems, suggesting the enlargements of the microemulsion structure. The p(r) functions in Fig. 2 using Eq. (1) show this process more clearly. Namely, the position of the maximimi value p(r)niax of the p(jr) function shifts to a long distance side with the increase of the maximum diameter Dmax estimated from p(r) = 0 for r

167

| 1 0 *

-6 IAS

1, gio*

10^

^N, > ~" ^ ?"

1 — i _ i 1 1 J

rT) r—r-^

(a)1

n-hex ' S^^v^ 1

>\ - |

it 1 1

0.06

I 0.04

CL 0.02

0

0.06

= 0.04 (0

'S 0.02

0

0.06

rk-

(a)

n-hexane 1

' •

r r

(b) j

n-heptane

0 50 100 150 200 r(A)

Fig. 1. wo dependence of the scattering curves of ^^' 2. Distance distribution functions p(r) of the w/o AOT microemulsions (water/AOT//i-hexane, scattering curves of AOT microemulsions shown (a); n-heptane. (b); n-octane, (c)) occluding a- " ^ - 1./i-hexane, (a);/i-heptane, (b); n-octane, chymotiypsin (2.4x10-5 M). «>- V«"°^^ ^ " « ^^^ ^ " ^^- •

>^max. Whereupon, the change of the p(r) occurs in the different manner for each system, especially at low water contents. The tailing of the p{r) function to a long distance direction can be seen at wo = 4 for the n-hexane system, at wo = 4 - 8 for the n-heptane system, and wo = 4 - 12 for the n-octane system, respectively. As explained previously [3], the tailing of the p(r) function suggests the presence of a certain amount of oUgomeric AOT microemulsion particles at tiie low water contents. Thus, Pig. 2 shows that the lengthening of the hydrocarbon chain of the apolar solvent extends the transient region from the oligomeric phase to the monomeric phase. Figs. 3(a) and

168

10 20 30 40 50 [HgOMAOT] (M/M)

0.1 0.2 [El (mM)

Fig. 4. Protein concentration dependence of p(r) function (a) andp(r)inax (b). In (a), water/AOT/n-hexane at wo = 20.

Pig. 3. wo dependence of the J7g (a) and p{r)max (Jb) of w/o AOT microemulsions occluding a-chymotiypsin (2.4xl0-^ M).

3flb) show the WQ dependence of the Rg andp(r)max, respectively. The relation between the WQ andp(r)niax values shows a good linearity for all systems. Whereas, with the lengthening of the hydrocarbon chain the wo vs. Rg relation becomes to deviate from a simple linearity and to separate into three different regions with different slopes. The slope above wo = 16 becomes smaller with the lengthening of the hydrocarbon chain. These results depending on the hydrocarbon chain length are essentially the same as those observed in the w/o AOT microemvdsion systems without proteins [5]. In Fig. 4 the protein concentration dependence of the pir) function and p(r)max shows that the occlusion of the proteins tends to decrease the microemulsion radius, which is more clearly seen with increasing water content or with shortening the hydrocarbon chain length. This would result from the attractive electrostatic interaction between the polar head of AOT and the basic residues of the protein surface.

REFERENCES

1. R. Hilhorst, In Structure and Reactivity in Reversed Micelles; Pileni, M. P. (ed.), Elsevier, Amsterdam, (1989) 323. 2. R. H. Pain (ed.). Mechanisms of Protein Folding, IRL Press, New York, 1994. 3. M. Hirai, et al., J. Chem. Soc. Faraday Trans., 91 (1995) 1081; J. Phys. Chem., 99 (1995) 6652. 4. M. Hirai, et al., J. Phys. Chem. Solids, 60 (1999) 1297. 5. M. Hirai, et al., J. Phys. Chem. B, 103 (1999) 9658.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) (C) 2001 Elsevier Science B.V. All rights reserved. 169

Phase transition in Gibbs monolayers of mixed surfactants

Md. Mufazzal Hossain,* Tomomichi Okano^ and Teiji Kato**

* Department of Applied Chemistry, Faculty of Engineering, Utsunomiya

University, Yoto 7-1-2, Utsunomiya 321-8585, Japan

^ Lion corporation, Tokyo, Japan

Abstract

Phase transitions in mixed monolayers of 2-hydroxyethyl laurate (2-HEL) and Na-salt of 3,6,9,12-

tetra oxa octacosanoic acid (TOOCNa) formed by co-adsorption fiom their mixed bulk solutions have

been studied. The presence of cusp points followed by plateau regions in tiie 7c-t curves, vs4iich is

accompanied by two phase coexistent state indicates a first-order phase transition. The domains have

circular shape with internal segments whereas those of pure 2-HEL at the same temperature are of

fingering pattem with uniform brightness all over the domains. With increasing the fi:action of

TOOCNa, the domains show more and more expanded behavior which favor easy fusion of tiiem.

1. INTRODUCTION

In recent years, we as well as several other research groups have demonstrated the existence of

first-order phase transition fiom gas or liquid expanded to condensed phase in Gibbs monolayers of

highly purified amphiphiles.[l,2] Extensive research has been performed on mixed surfactant systems,

since they can show superior performance as compared to single surfactant alone. When a trace of

dodecanol is added to sodium dodecyl sulfate solution die interfacial properties of the aqueous

solutions are markedly altered. This is attributed to an increase in packing and ordering of the

monolayers at the air-water interfece due to the co-adsorption of dodecanol.[3] Shah et aL [4]

proposed that when two surfactants are mixed with a molar ratio of 1:3, the properties of the surfactant

systems are changed strikingly due to 2D hexagonal packing of tiie molecules. Recently, Bain et aL [5]

reported tiie phase transition in the mixed monolayers of cationic surfactants and dodecanol at the air-

water interface by sum fiiequency spectroscopy.

In this paper, we evidently report that two water soluble surfactants show a first-order phase

transition and form liquid condensed (LC) domains in an appropriate mixture. We have chosen 2-

hydroxyethyl laurate (2-HEL) and Na-sak of 3,6,9,12-tetraoxa octacosanoic acid (TOOCNa) as

amphiphiles (Fig.l).

170

2-HEL: CH3(CH2),OCCXX:H2CH20H

TOOCNa: CH3(CH2)i50" ^ ^ ^ ^ ^ O '

Fig. 1. Chemical structures of the surfactants used in this study

/ COONa

2. EXPERIMENTAL

The material, 2-HEL was synthesized with a purity of > 99.5% and TOOCNa was obtained

fix)m Lion corporation, Japan with a purity of > 99%. The solutions were prepared separately in ultra

pure water of resistivity 18 MQ -cm and then mixed in appropriate volume ratio to obtain the desired

molar ratio. To specify the ratio of a given mixture in the latter part of this paper, we always follow the

order 2-HEL: TOOCNa.

The surfece pressure-time (u-t) curves were measured in a home buik Langmuir trough of very

shallow type. The experimental procedure [6] was detailed elsewhere. The surface pressure was

measured by the Wilhehny metfiod and the domain morphology was characterized simultaneously by

Brewster angle microscopy (BAM) [7] using 20 mW He-Ne laser as a light source.

3. RESULTS AND DISCUSSION

Fig. 2 shows 7i-t adsorption kinetics of

2.0 X10'^ M aqueous solutions of both the

surfactants separately and in mixed solutions

v^th different concentration ratios at 15°C.

This concentration is suflBcient to form

condensed domains by only 2-HEL but not

sufficient to do that by TOOCNa. Condensed

domain formation in pure TOOCNa

monolayer is not possible even with more

OHicentration solutions because these are

above its cmc. Under the present conditions,

the rate of adsorption for pure TOOCNa is

higher than that of 2-HEL but in the mixed

system this rate is in between the pure

systems. WrAi increase in the fraction of

Fig. 2.7c-t adsorption kinetics at 15°C of 2.0 X10'^ M

aqueous solution of 2-HEL (I), TOOCNa (V), and

their mixtures with diffoent molar ratios of 2-

HELrTOOCNa; 3:1 (II), 1:1 (ffl), 2:3 (iV). The

vertical arrows indicate the position of the cusp points

in the reflective curves.

TOOCNa in the mixture, the overall rate of adsorption increases indicating the co-adsorption of the

surfactants. However, the 7c-t curves in the figure show the cusp points followed by plateau regions up

to die ratio 2:3. For spread monolayers of some otiier amphiphiles, a true cusp point in the surface

171

pressure-area (TC-A) isothenns indicates a discontinuity in 5G/57C (G, Gibbs fiee eneigy) because

5G/57C =A. This is the characteristic of a first-order phase transition. Since, A decreases with time in

Gibbs monolayers, a cusp point in the n-t curve also indicates a first-order phase transition.[l,2] hi the

7c-t curve of ratio 3:1 (curve II), the concentration of 2-HEL is only 1.5 X10' M which is not suflScient

to form condensed domains at 1 5 t when it is used alone.[8] However, phase transition is possible in

the mixed system of ratio 2:3 where the concentration of 2-HEL is only 0.8 X10'^ M. These results

demonstrate tiiat the existence of phase transition in these systems are due to both of the surfactants.

The critical surfece pressures {n^ necessary for the phase transition are almost Ae same except in one

case where the fiction of 2-HEL is relatively low (curve IV). hi the latter case, tiie TC, is higher than

those of the other systems. This should be due to the rapid adsorption of mainly TOOCNa molecules

which cover most of the surfaces before a considerable amount of 2-HEL can accumulate to initiate

condensation. However, once the concentration of 2-HEL becomes sufficient LC domain formation

starts, but before this can happen the surfece pressure becomes almost close to the equilibrium value of

TOOCNa. With further decrease in the fiaction of 2-HEL beyond 2:3, phase transition vanish.

The in situ BAM observation for all the

curves with cusp points shows condensed

domains which become larger with time and

finally solution surface is covered with them.

Fig. 3 presents the typical shape and texture of

the domains of pure 2-HEL and the mixed

surfactants at 15°C. For pure 2-HEL, the

domains are of fingering pattern with uniform

brightness all over the domains (image A) at

this temperature. The shape of tiie domains for

mixed monolayers is circular with inner

segments (images B-D) at and above this

temperature. For the mbced surfactant system of

ratio 3:1, the most of the domains have stripe

pattem. With increase in the fiaction of TOOCNa, the texture of the domains becomes irregular with a

variety of pattem. Few examples of different defects are given in the Fig. 3 (images B-D). The fusion

of tiie domains is rarely observed during the formation process of the domains in pure 2-HEL and

even in mixed surfactants containing higher fiaction of 2-HEL (ratio 3:1). However, fusion becomes

more and more favorable witii increase in the flection of TOOCNa. Fig. 3D presents an elliptical

domain which is formed by the fusion of two circular domains. This type of fusion is rather a common

phenomenon during the monolayer formation in the mixed system of ratio 2:3.

Fig. 3. Shapes and textures of the domains formed

in the monolayers of 2-HEL (A) and the mixed

systems of ratio 2:3 (B-D). Size: 400 X 300 pml

172

All these results can be explained considering formation of the mixed monolayers. With tfie ratio

3:1 of the surfactants, the monolayers is dominated by tiie 2-HEL, but considerable extent of

TCXXnSIa molecules is included into tiie domains. For other cases, the extent of 2-HEL molecules is

comparatively small. Since, TOOCNa contains sixteen carbon chain, it is expected to have higher line

tension if it is introduced into the domains. This high line tension causes circular shape domains

although the pure 2-HEL show fingering pattem.[8,9] At 15°C, the 2-HEL molecules remain almost

normal to tiie surface, wiiich causes uniform brightness in these monolayers.[8] Nevertheless, the

TOOCNa molecules containing longer carbon chain as well as larger hydrophilic group should be

tilted. Thus, overall balance among these molecules should cause such complex pattern in the

monolayers. With decrease in the fraction of 2-HEL, the monolayers have a tendency to show more

irregular textured domains. This is a clear evidence in favor of the formation of domains by both of the

surfactants because for pure surfactants regular pattern is ahvays observed. The irregular pattern most

probably due to tfie uneven distribution of the components in the domains. We must take consideration

of the repulsive forces e.g. electrostatic, dipolar, hydration etc. of the long hydrophilic head groups of

TOOCNa molecules. W ith decrease in the fraction of 2-HEL, tiie tendency of incorporation of

TOOCNa molecules into the domains increases. When the concentration of 2-HEL is sufficiently low

to initiate condensed domain formation, phase transition does not occur.

4. CONCLUSIONS We provide evidence for the first-order phase transition in mixed monolayers of two water

soluble amphiphiles, 2-HEL and TOOCNa. It is clear fix)m the BAM images that the formed domains

contain both of the component, although the exact composition is still unknown. The circular domain

formation in these monolayers is a direct evidence for the effect of line tension on domain shape. The

irregular textured domain formation may be attributed to an uneven distribution of the surfactants at

the diflFerent part of the domains.

REFERENCES

l.D. Vollhardt, V. Melzer, J. Phys. Chem. B 101 (1997) 3370.

2.M. M. Hossain, M. Yoshida, T Kato, Langmuir 16 (2000) 3345.

3. B. D. Casson, C. D. Bain, J. Phys. Chem. B102 (1998) 7434.

4. A. Patist, S. Devi, D. O. Shah, Langmuir 15 (1999) 7403.

5. B. D. Casson, C. D. Bain, J. Phys. Chem. B 103 (1999) 4678.

6. M. M. Hossain, M. Yoshida, K. Iimura,N. Suzuki, T Kato, ColloidSmf.A 171 (2000) 105

7. S. Henon, J. Meunier, Rev. Sci. Imtnrn. 62 (1991) 936.

8. M. M. Hossain, T. Kato, Langmuir 2000 (in press)

9. S. Siegel, D. Vollharxit, Thin Solid Films 284/285 (19%) 424.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) c 2001 Elsevier Science B.V. All rights reserved. 173

Mesoscopic structures of J aggregates of organic dyes at a solid/liquid interface and in solution: spectroscopic and microscopic studies

Hiroshi Yao,* Sadaaki Yamamoto,** Noboni Kitamura*' and Keisaku Kimura"

^ Faculty of Science, Himeji Institute of Technology, Hyogo 678-1297, Japan ^ Material Science Laboratory, Mitsui Chemicals, Inc., Chiba 299-0265, Japan " Graduate School of Science, Hokkaido University, Sapporo 060-0810, Japan

Supramolecular structures in J aggregate systems are characterized. J aggregates of a pseudoisocyanine dye (PIC) at a mica/solution interface were in situ observed by tapping-mode atomic force microscopy (AFM). The single aggregates possessed a three-dimensional disk-like island structure in a mesoscopic scale. The island size ranged -400-600 nm long, -100 nm wide, and -3-6 nm high. Morphological differences can be observed between the J aggregates at a solid/liquid interface and those in bulk solution. Mesoscopic string structures of 5,5'-dichloro-3,3'-disulfopropyl thiacyanine (TC) J aggregates were detected in an aqueous solution for the first time by both fluorescence microscopy and microspectroscopy. The length of the J string was several tens of \im while the width was very narrow.

1. INTRODUCTION

Extensive research has been directed toward a better understanding of supramolecular aggregate systems and their interesting optical and electronic properties (!]. 7 aggregates are specific dye supramolecular assemblies characterized by a narrow and intense absorption band that shows a bathochromic shift compared to the relevant monomer band. Since the aggregate structure reflects highly on its spectroscopic properties, detailed investigations of the structures and/or morphologies of the single aggregates are of primary importance. Thus, we examined morphologies and optical properties of different types of mesoscopic J aggregate systems: J aggregates at a solid/liquid interface and in a solution phase.

2. EXPERIMENTAL

l,r-Diethyl-2,2'-cyanine chloride (pseudoisocyanine; abbreviated as PIC) and 5,5'-dichloro-3,3'-disulfopropyl thiacyanine (abbreviated as TC) were purchased from Nippon Kankoh-Shikiso Kenkyusho Co., and used as received.

Conventional absorption and fluorescence measurements were carried out on a Hitachi U-3300 spectrophotometer and an F-4500 spectrofluorometer, respectively.

174

In order to investigate J aggregate formation at a mica/water interface, cationic PIC dye was used. A sample for spectroscopic measurements was prepared by placing an aliquot of an aqueous PIC solution between mica and hydrophobic glass plate.

A TC dye was used for examining the structure of J aggregates produced in a solution phase. A solution sample was prepared by dissolving TC sodium salt in an aqueous NaCl solution (5.0 mM).

AFM images were recorded on a Nanoscope Ilia (Digital Instruments) operating at a tapping-mode in a liquid phase. Triangular Si3N4 microcantilevers (Nanoprobe; NP-S, Digital Instruments) possessing a spring constant of 0.58 Nm*' were used.

Fluorescence microscope images were obtained by using a CCD camera (Hitachi, Remote Eye) set on an optical microscope (Nikon, Optiphoto-2). Fluorescence microspectroscopy was conducted by using a polychromator-multichannel photo-detector set (Hamamatsu Photonics, PMA-11) equipped on the microscope. A monochromatic beam (454.5 nm) was used as the excitation sources. A sharp-cut filter (Y-47) was mounted in front of the CCD camera and photodetector set.

3. RESULTS AND DISCUSSION

3.1. Mesoscopic island structures of PIC single J aggregates at a mica/water interface

Figure 1 shows an optical path length dependence of the absorption spectrum of an aqueous PIC solution (2.0 mM). The spectrum showed a sharp and intense J band (580 nm). It is worth noting that no J band can be observed when using a hydrophobic glass cell. Moreover, the figure indicates clearly that the J band is independent of the path length while the bulk monomer (525 nm) or dimer band (480 nm) increases with increasing the path length. The results indicate that J aggregate formation is concluded to be confined to the vicinity of the mica/solution interface.

400 500 600 Wavelength / nm

Fig. 1. Optical path length dependence of the absorption spectrum measured for an aqueous PIC solution (2.0 mM) between mica and hydrophobic glass plate.

2 Mm

Fig. 2. AFM top-view image of the PIC J aggregates at a mica/solution interface (|PIC| = 0.2 mM). Arrows show the periodic orientation of negative holes on mica surface.

175

Thus, atomic force microscopy (AFM) was conducted to examine the microstructures of the J aggregates at the interface. Figure 2 shows AFM image at [PIC] = 0.2 mM. Since the mica surface was unchanged and atomically flat until the J band appeared (< 0.1 mM), the observed mesoscopic leaf-like islands were considered to be the J aggregates. The size of these islands ranged -400-600 nm long, -100 nm wide, and -3-6 nm high. Interestingly, our AFM image revealed that the J aggregates have a three-dimensional disk-like structure but not a two-dimensional monolayer structure. In addition, morphological changes of the islands were observed with changing the PIC concentration: The number density of the J islands increased with increasing PIC concentration, and then, they coalesced into larger domains. In contrast, the height of the islands was independent of the PIC concentration [3]. The constant height of the islands would be determined by the balance between the adsorption/aggregation and dissolution energies.

Furthermore, the long axis of the islands are anisotropically oriented relative to the alignment of the negative holes on mica surface formed by dissociating K" ions, which was shown as white arrows (three directions) in Figure 2. The results suggest the existence of epitaxial interaction between PIC molecules and the lattice of a mica substrate. The highly probable epitaxial interaction is the one that the positively charged N atoms of the dye are placed at the negative holes on mica [4,5]. jisland _ _ - - - - _ According to this epitaxial interaction, there are two possible alignments of / \ ^ the dye molecules in the islands: the long axes of the dye are either parallel ^ ' ^/ or 60** relative to the long axes of the islands. In terms of energy, however, the dye molecules may grow so that O negative holes on the mica surlace the long axis of dye molecules is cziizj PIC molecules parallel to the long axis of the islands as shown in Figure 3, which shows an Fig. 3. Schematic model of the alignment energetically stable brick-stonework of PIC molecules in a J aggregate island. alignment of dye molecules.

3.2. Mesoscopic string structures of TC single J aggregates in solution On the other hand, structural and/or morphological differences are expected

between the J aggregates at a solid/liquid interface and those produced in bulk solution. We demonstrate that fluorescence microscopy and microspectroscopy enable a direct observation of single J aggregates in solution. Here, we examined the microstructures of 5,5'-dichloro-3,3'-disulfopropyl thiacyanine (TC) J aggregates in an aqueous solution.

Figure 4 shows a fluorescence microscope image at [TC] = 0.05 mM, above which the J band appears, and mesoscopic string structures were clearly observed. Since a characteristic fluorescence image was not detected below this concentration.

L Q Cl C Q O C .

r^crc" o :, s o o

176

the strings distributed in solution were considered to be 7 aggregates of TC. To the best of our knowledge, this is the first observation of mesoscopic J aggregates in a solution phase.

The length of the string was several tens of \km while the width was very narrow; sub-^m. This string structure is probably due to anisotropic interactions between TC molecules in solution (i.e., quasi-one-dimensional stacking interactions), different from that of PIC J aggregates observed at a mica/water interface. It is noteworthy that a single string is likely to bend in an arc form, suggesting that the mesoscopic J aggregate is flexible and polycrystalline-like.

Figure 5 shows fluorescence spectra observed for a single string of J aggregates and at the periphery of the string. The excitation beam diameter was -10 jim, and the measurements were conducted at various position in the string. However, the spectral shape did not change with the observation position in the string. Since the spectrum was quite similar to that measured in bulk solution, the strings detected in Figure 4 were concluded to be TC 7 aggregates. It is worth noting that fluorescence was scarcely observed at the outer periphery of the string, indicating that TC in an aqueous solution produces exclusively mesoscopic-size J aggregates.

30f,im

Fig. 4. Fluorescence microscope image of the TC J aggregates in an aqueous solution ([TC] = 0.05 mM). The strings correspond to the TC J aggregates.

i 200 >,0 c «

S 100

u

periphery of the string

^m^^^ 600 650 400 450 500 550 Wavelength / nm

Fig. 5. Fluorescence spectra observed for a single string of TC J aggregates and at the periphery of the string.

REFERENCES

1. P. W. Bohn, Annu. Rev. Phys. Chem., 44 (1993) 37. 2. H. Schmidt, J. Vac. Sci. Technol. A, 8 (1990) 388. 3. H. Yao, S. Sugiyama, R. Kawabata, H. Ikeda, O. Matsuoka, S. Yamamoto and N.

Kitamura, J. Phys. Chem. B, 103 (1999) 4452. 4. V. Czikkely, H. D. Forsertling and H. Kuhn, Chem. Phys. Lett., 6 (1970) 11. 5. S. S. Ono, H. Yao, O. Matsuoka, R. Kawabata, N. Kitamura and S. Yamamoto, J.

Phys. Chem. B, 103 (1999) 6909.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) i£' 2001 Elsevier Science B.V. All rights reserved. 177

Energy of Breaking of Aqueous GEMINI Surfactant Film

Tadahiko Kidokoro and Junichi Igarashi

Department of Chemistry, Faculty of Science, Tokai University

1117 Kita-Kaname, Hiratsuka-shi, Kanagawa-ken 259-1292, Japan

l.Introduction A number of methods have been proposed for measuring the surface tension of a liquid.

Single liquid film is considered to be a very simple colloid system, and is useful for

elucidating various colloidal and interfacial phenomena. Interest has been shown by many

investigators, particularly upon the effect of aqueous ionic surfactant for its drainage of the

film. However, there are few such reports on ionic Gemini surfactant films.

Surface active substance usually forms a stable adsorption film in the surface or interface

of liquids. As the typical methods for studying the mechanical behavior of such a surface

film, equilibrium method of measuring surface free energy( Op), namely the surface tension

measurement and dynamic film detachment method of measuring surface film detachment

followed by its breaking( a p) are listed. In the case of pure liquid, a p= a p is assumed to

hold, since pure liquid rarely produces a stable foam. However, in the case of aqueous

surfactant solution, excess energy corresponding to a F ~ cr p = a pp is confirmed to detach

the aqueous surfactant thin film from its aqueous surface. Thus, the value of o pp in such a

case is considered as the measure of the film stability.

In the present study, a reliable measurement of this energy of surface film detachment is

attempted. For this purpose, we adopted the frame method that enabled the process of

dynamic film detachment to occur in a well-defined condition. The measurement of 0 pp is

carried out for the aqueous solutions of Gemini type surfactant, and the stability of aqueous

thin film is discussed by taking into account of the structure of Gemini molecule. For this

purpose, frame method is original and especially suited for the measurement of surface

tension in the present case. Here, we attempt to establish an empirical equation available

for the calculation of a p, since there are no empirical equations of Harkins and Brown, and

Harkins and Jordan in drop weight method and ring method, respectively.

The difference of the interfacial tension of several mNm' is obtained in ionic surfactant

and nonionic surfactant surface, which exists qualitative values.

178

Here, the Gemini style surfactants that were synthesized in recent years were investigated

sufficiently the surface properties especially for the properties mentioned above.

2.£xperimental

Z.l.Materials

The sample of Gemini surfactant used here is disodium N-N - bis [(2-carboxyethyl)

lauroylamide] ethylene diamine corresponding to monomer, made by KANEBO, Ltd.,

Cosmetics Lab. Product.

The water used was distilled from a solution prepared by dissolving potassium

permanganate and sodium hydroxide in ion-exchanged water, using borosilicate glass

equipment immediately before each measurement. Pure liquids (water, cyclohexane,

toluene, and chlorobenzene) were obtained by twice simple distillation. These solutions

were freshly prepared just before the use.

2.2.Apparatus

The cell was submerged in a thermostated water bath of 303.15±0.2K, and the whole

apparatus was set in a chamber of 303.15 ±0.2K for about 15 hours before the measurement.

Air room (gas phase) is filled with nitrogen gas.

2.3.Method

In the present measurement, the glass frame used for frame method was shown in Fig.l- i .

A microscopic cover glass with a

perimeter of 21=8.030cm was used as a

plate, to which a thin glass rod carrying

the hook was attached as shown Fig.l-

ii. The frame used was made of glass

rod of 0.51mm in diameter and

2L=8.856cm in width as shown in Fig.l-

i .

Pure water and the aqueous solution,

whose surface tension is considered to

be constant at a constant temperature is

put in vessel, which is slowly raised

until the surface just touches the low end

of the plate.

Gemini solution used was put in the

vessel, o p and o p were measured after about 15 hours aging cither by Wilhelmy method

or frame method. The solutions were gradually ra sed. le frame was vertically

withdrawn at the rate 0.75~2.5mmsec"^ from the solution by electric motor. The vessel was

= 0 . 5 1 Kill!

F i g . 1 - i

21-8.()3()cni

F i g . l - i i Glass f r a n i e s ( i ) and P l a t e ( i i ) used in t h i s s tudy.

179

contained about 300ml of the solution.

The surface tension of both the plate and frame are converted to the electric signal by strain

gauge UR-2GR of MINEVEA Co., Ltd. and is put into the recorder connected.

The plate and frame were cleaned before each measurement by immersing it in chromic

acid mixture for about 8 hours , followed by rinsing with water.

3.Result and Discussion

The frame method of measuring downward pull of liquid is considered as the more

idealized method of measuring the force of film detachment. Since we have no equations

applicable for the frame method, as it

exists in the case of drop weight JQQ

method of Harkins and Brown, and ring

method of Harkins and Jordan, we

attempted first to establish an empirical

equation applicable for the calculation

of a p, by using pure liquids such as

water, cyclohexane, chlorobenzene and

toluene.

In the case of film extension rate of

0.75"^2.5mmsec'^ range, a straight line

relation was obtained as shown in Fig.2

between film tension (P/2L) and inverse

of film extension rate V" for pure

liquid. From the extrapolated value of

(P/2L) to V^=0 as expressed by (P/2L)oo, and a p, A a is obtained as shown equation(l).

(P/2L)oo-ap=Aa (1)

Figure 3 shows plots of A a in each pure liquid against (P/2L)c», showing a straight line of

equation (2).

A a =0.0899(P/2L)«,-0.6018 (2)

In the present study, surface detachment energy a p measured by the frame method is defined

as Eq.(3),

aF=(P/2L)oo-A a (3)

From Eqs. (2) and (3), Eq. (4) is obtained,

a F=0.9101(P/2L)oo+0.6018 (4)

79J [•

79.6 ^

79.4 -

' e 79J r z I 79.0 •' 3

78.6 ^

78.4 [

782 j

78.0 —

0.0 0.4 0.6

V7(n»n-»ec")''

Fig.2 Surface tension as a function of V'^

Here mNm' is used as the unit of surface tension. As examples, the detachment process of

the film in the frame method at regular rate of film expansion is shown in Fig.4. The pattern

180

showed maximum B, and shows a linear rise in according to the change in expansion rating of

the film in the case of Gemini systems A.B.C and D. As shown in Fig.4 surface cut energy

increased with downward speed. Similar variation was also observed with an increase in

concentration.

b 30

<

aoo 4Q0 eao 8Q0

(P/aj./fTWm'

Fig.3 Aa-(P/2Lio curves for pure liquids.

B D

Fig.4 Chan of dynamic surface tension by frame meihod.

Energy of detaching and breaking thin

aqueous Gemini film o p according to Eq.4

are shown in Fig.5. This energy of

breaking surface film, which is considered

as an indication of the stability for an

adsorbed film, is seen to change with

concentration of surfactant and shows a

maximum in the range of the surface

saturation concentration of the solute. The

difference of detachment energies between

the monomer and dimer Gemini compound is

considered to be explained by the difference

in the hydrophobic interaction of the

respective compounds. Also, the special

characteristic connection effect that of

Gemini should be taken into account.

50.0

45.0

40.0

35.0 -

30.0

25.0

-115 -10.5 -9.5

ln(C/mol-dn"')

-8.5

Fig.5 The static and dynamic surface tension

of Gemini surfactant solution plotted against

inc.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) c 2001 Elsevier Science B.V. All rights reserved. 181

Molecular Aggregation States and Polymerizability of Potassium and Calcium 10-Undecenoates in Aqueous Systems

Yoshio Shibasaki, Hideki Saitoh, and Atsuhiro Fujimori Faculty of Science, Saitama University, Urawa, 338-8570, Japan

Tel:81 -^8-858-3381, Fax:81 -^8-858-3700, Email:saitoh(achem.saitama-u.ac.ip

Molecular aggregation states of potassium and calcium 10-undecenoates in aqueous systems with different water contents were studied by differential scanning calorimetry (DSC) and a temperature controlled X-ray diffraction measurement in relation to the polymerizability of these amphiphilic monomers by a r -ray-initiated polymerization in different aggregation states.

1. INTRODUCTION Amphiphilic compounds in aqueous systems exhibit various states of mole

cular aggregation, such as coagel, gel, liquid crystal, and disk or rod-like micelle, depending on the water content and temperature, as reported previously [1,2]. In the case of amphiphilic monomers, their polymerizability is expected to be influenced by the aggregation states through differences in the separation, orientation and mobility of the functional groups [2-4]. Polymerization in lamellar structure, thin films, micelle, etc. is an important problem in the field of colloid and surface science. In this work, we studied the molecular aggregation states in relation to the polymerizability of potassium and calcium 10-undecenoates in aqueous systems with different water contents. The aggregation states were investigated by thermal analysis using a differential scanning calorimeter and a controlled temperature X-ray diffractometer. Polymerizabilities in various aggregation states were examined by the low temperature r -ray irradiation post-polymerization.

2. EXPERIMENTAL

2.1 Materials Potassium 10-undecenoate [CH^ = CH(CH^)8C(= 0)0K:Abbr.KUD] was synthesized

by the reaction ol 10-undecenoic acid (mp. 2 3 . 5 — 2 4 . 1 0 with potassium eth-oxide in ethanol, and was purified by recrystallization from methanol solution. Calcium 10-undecenoate [ (CH, = CH(CH J 8 C ( = 0)0)2 Ca:Abbr.Ca(UD)2] was synthesized by the dropwise addition of 10-undecenoic acid in a cone, solution of calcium hydroxide and used after purification with n-hexane. Carefully dried undecenoates were mixed with given amounts of water in a mortar and were homogenized by storing in a refrigerator for 1 week at ca. 5 X).

2.2 Apparatus and Procedures Thermal analysis were carried out by a Seiko Instruments model DSC 6200.

Aggregation mode of molecules was examined by a Mac Science MXP18VA dif-

182

fractmeter (CaKa radiation) equipped with a refrigerator using liquid Ng. Polymerization in various states were carried out at various temperatures

( - 2 0 ' - 5 0 t ; ) . The monomer samples in aqueous systems of about Ig were sealed in Pyrex tubes (id. 10mm) in nitrogen atmosphere, and irradiated with ®°Co r -rays (30 kGy at desired temperatures in a Dewer flask). The r -ray irradiation was carried out at the Japan Atomic Energy Institute at Takasaki. The polymerization process was followed by a Perkin Elmer system 2000R.

3. RESULTS AND DISCUSSION

3.1 Aggregation States of Monomer Molecules DSC curves for the samples of KUD and Ca(UD)2 with different water con

tents of 10-90wt% are shown in Figure 1. In the case of KUD-water system, an endothermic peak at O'X: appeared from the water content of 20wt% and increased with water content, while a peak at \5V decreased. In addition, three small peaks were observed at ^0, 85 and 120t:. In the case of Ca(UD)2-water system, a clear endothermic peak appeared at 30'€, except for a peak at OV, and two small peaks were observed. Moreover, in a range of water content of 10-50wt%, two broad peaks were observed at 140 and 180t:. To estimate aggregation states of monomers in these systems. X-ray diffraction patterns of the samples at various temperatures were observed. In the case of KUD (water content 30wt%, Fig.2A), the coagel state at - 30t: was changed to a gel-1 state at 5V, in which some free water penetrate into the head-to-head layer of KUD molecules. It can be expected that the aggregation state of KUD changed progressively to a gel-B (at 20t:) , a lamellar liquid crystal (at 30t:), a disk-like micelle (at 80T:). In the case of Ca(UD)2-water system (water content 10wt%, Fig.2B), the regular arrangement of long alkyl chains are almost destroied between 30T: and 50T:, because a peak at 20 =r diminished at 50't;, and above 50X: their aggregation state changed

(A) KUD-water system (B) Ca(UD)2-water system

I

1

-20 0 20 40 60 80 100 120 140 t /Op -30 0 30 60 90 120 150 180

Fig.l DSC curves of KUD and Ca(UD)^-water systems.

183

(A) KUD (B) Ca(UD)2

30 "C

(water content 10 wt%)

tf. r. ^^^.•^r7r^>^rAJ:

30 "C

10 X,

- 30 "C -30 r

2 0 ' ' 26 Fig.2 X-Ray di f fract ion patterns of KUD and Ca(UD)g-water systems

at various temperature.

to a disk-like micelle. The exothermic peaks at 1 0 and ISOt: in DSC curve of Ca(UD)2 (water content 10%) were examined by FTIR spectral change of the samples, those were heated up to ]30V and 190T:. Band intensities at 3090 cm" ( ] ^ C - M ) , 16^0 cm- ' (yc-c) , and 990 and 910 c m - ' ( 6 c - i i ) corresponding to - C H = C H 2 were clearly decreased indicating the thermal polymerization.

Phase diagrams of KUD-water and Ca(UD)2-water systems constructed from thermal analysis data are shown in Figure 3. In the case of KUD-water system

(A) KUD-water system (B) Ca(UD)2-water system

180

160

140

120

£ 100 0)

% 80

I 60

40

20

0

-20

-

- "N Disk-like ^ micelle \

r iiiiiiiin \ iiiiiiiHi

r Lamellar j liquid crystal i

m i 1 Gel-n 1

©xl Micellar solution x

^ Liquid crystal

(hexagonal) |

1 • G C I - I " • • • • . .

Coagel mvmjtt www

20 40 60 80 100

Water Content (%)

180

160

140

120

Pioo 0)

ro 80 0)

i 60

40

20

0

-20

(Thermal Dolymerization)

•

(Micellar solution)

Disk-like micelle

- . • "Lamellar" " .^

Liquid crystal fjd^.

' Gel S ^ ^

Coagel WA XM

20 40 60 80 100 Water Content (%)

Fig.3 Phase diagrams of KUD and Ca(UD)^-water systems.

184

(A) KUD-water syste (B) Ca(UD) ..-water system

60

50

^ 4 0 h

3 30 h

20 h

H20

70X

^ 80%

L 30%

20%

10%

L 0%

90%

i.

Coagel

"A

Gel- 1

50%

w

V^ ; yeo*

K\' ; Gel- n

--15

-40 -20 0 20 40 Polymerization Temperature (tT)

-40 -20 0 20 40 60 Polymtrization Temperature (°C )

Fig .4 Saturated conversion-temperature relationships of KUD and Ca(UD)2-water systems at various temperatures.

(Fig.3A), in a range of water content (10-30wt%), aggregation state changes as follows coagel-* gel-I -> gel-H^ lamellar L.C.-* disk like micelle, while in a region of ^0-70wt% the state changes at 15'X:gel-I-^ hexagonal L.C., and above 70% the hexagonal L.C. changes to a micellar solution at 100-130X:. In the case of Ca(UD)2-water system (Fig.3B), aggregation state changes as coagel^ gel-* lamellar L.C.-* disk-like micelle, and a region of 50-80wt% a transition expected to be a disk-like micelle-* micellar solution transition was observed. Above ]50V thermal polymerization occurred gradually.

3.2 7 -Ray-Irradiation Polymerization in Various States In the 7 -ray-irradiation polymerization of KUD-water and Ca(UD)z-water

systems, polymerization proceeded rapidly and saturated after 3-^ hours. The polymerizability of KUD was highest in the coagel state at water contents of 70'-80wt% and the maximum conversion reached to 60 % as shown in Figure ^A. On the contrary, the polymerizability of Ca(UD)2 was highest in the coagel at water contents of ]0---20wt%, and the maximum conversion was ca. 23 % (Fig. ^B). In both cases, the polymerizability decreased in the order of coagel, gel-1, gel-B and lamellar liquid crystal. It can be concluded that the regular arrangement of the monomer molecules together with mobility of terminal vinyl groups is an important factor for the polymerization, although details of the results obtained in this work are not consistent with the results of sodium and zinc 10-undecenoates in aqueous systems[2].

REFERENCES 1. M.Kodama and S.Seki, J.Colloid Interface Sci., 117(1987)^85. 2. Y.Shibasaki and K.Fukuda, Colloids and Surfaces, 67(1992)195. 3. A.Fujimori, H.Saitoh and Y.Shibasaki, J.Therm.Anal.Calori.,57(1999)631. ^. A.Fujimori, H.Saitoh and Y.Shibasaki, J.Polym.Sci.,A:Polym.Chem.,37( 1999)3845.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) c 2001 Elsevier Science B.V. All rights reserved. 185

Effects of Shear Flow on the Structure of the Lamellar Phase Formed in Nonionic Surfactant-Water System

K. Minewaki^, T. Kato^ and M. Imai^

^Department of Chemistry, Tokyo Metropolitan University, Minamiohsawa, Hachioji, Tokyo, 192-0397, Japan

^Department of Physics, Ochanomizu University, Ohtsuka, Bunkyo-ku, Tokyo 112-0012, Japan

Small-angle neutron scattering has been measured on the lamellar phase formed in a nonionic surfactant (Ci6E7)-water system under shear flow at shear rates in the range 10~' -10^ s"" From the dependence of the peak position and the peak intensity on the shear rate it has been suggested that the lamellar domain is disrupted into small fragments at shear rates about 0.1^1 s~ and that the original microstructure is reconstructed at higher shear rates.

1. INTRODUCTION

In recent years, the effects of shear flow on the structure of the lamellar phase formed by nonionic surfactants have been investigated by using microscopy[1], viscometry[l], small-angle neutron scattering (SANS) [1-3], light scattering[l], and NMR[1,4] and attention has been paid to the formation of multilayered vesicles by shear flow and the orientation of the sample for structural analyses. In a previous paper, we studied the effects of shear flow on the structure of the lamellar phase in Ci6H33(OCH2CH2)70H (abbreviated as Ci6E7)-water system by using SANS in the range of shear rate 7 = 10"^-10 s~ which is much lower than that of other studies reported so far. It has been shown that significant changes in the peak position and the peak intensity were observed at 7 c 0.1 s~^[5]. In these measurements, however, we used the neutron beam only along the gradient direction by which the perpendicular and transverse orientations are observed (orientation of the layer normal along the vorticity, flow, and velocity gradient directions are referred to as perpendicular, transverse, and parallel, respectively). In this study, measurements have been made by using a neutron beam along both gradient (radial) and flow (tangential) directions alternatively at each shear rate. In addition, effects of the

186

0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0

q/nm- q / nm~

Fig. 1. Scattering intensities for the vor-

ticity direction at 328 K (a) and 343 K

(b) integrated over a sector of ±10° at

different shear rates. The results at 343 K

have been already reported [5].

0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0

q I nm-^ q I nvar^

Fig. 2. Scattering intensities for the

flow (a) and vorticity (b) directions at

328 K integrated over a sector of ±10° at

different shear rates (different run from

that of Figure 1(a)).

shear history have been examined.

2. E X P E R I M E N T S

Measurements of SANS were carried out at the instrument SANS-U of Institute for

SoHd State Physics of University of Tokyo in JRR-3M at Tokai with a Couette shear

cell[6]. All the measurements were made for the sample containing 55 wt% of CieEy at

328 K and 343 K.

3. R E S U L T S A N D D I S C U S S I O N

Figures l a and l b show the scattering intensities for the vorticity direction at 328 K

187

1.4

« 1.3

1.2

vorticity direction —o— gradient direction - -o—

vorticity direction —o— p 500 I- gradient direction -o---S 400 [ 0-'-<^

0 0.0010.01 0.1 1 10 100

Fig. 3. Shear rate dependence of the peak position (a) and the peak intensity (b) of the first reflection for the vorticity direction at 328 K an 343 K.

0 0.0010.01 0.1

Fig. 4. Shear rate dependence of the peak position (a) and the peak intensity (b) in the vorticity and gradient directions at 328 K.

and 343 K, respectively. SANS patterns were recorded twice (10 min each) at each shear rate. The peak positions and the peak intensities are plotted against the shear rate in Figure 3. As reported before[5], for 7 < 0.1 s"• the peak intensities for the vorticity and flow directions decrease with increasing shear rate without the change in the peak position. At the shear rates around 0.1 s~\ a new peak appears at the higher scattering angles. As the shear rate increases further, the peak position shifts to the lower scattering angle while the peak intensity increases.

Taking into account the results of small-angle x-ray scattering measurements[5,7], we proposed the following scenario[5]. At lower shear rates, the lamellar domain contracts into smaller domains and some of them become disordered states. At the shear rates around 0.1 s~^, a new domain is formed composed of bilayer sheets with water-filled defects. As the shear rate increases further, the fraction of the water-filled defects decreases and the original microstructure is reconstructed.

In the present study, we performed five independent runs which gave different shear histories to the sample. Results for one of them are shown in Figures 2 and 4. Below

188

^-N

3

^ Ul

^ ^ w* c« a

c

350 300 250

200 150 100 50 0

r -I

r ^ 1 1 ....1 1

go

— 1 1

328 K j

^

1 .-,...

0.9 1.0 ^max

1.1 /nm-1

1.2 1.3

Fig. 5. Peak intensity vs. peak position for the vorticity direction at 328 K and 343 K. Different symbols indicate different runs.

1 s~\ the scattering intensity for the gradient direction increases while that for the vorticity direction decreases, which may result from the fact that a reorientation from perpendicular to parallel alignment is enhanced. Above about 1 s~\ on the other hand, the intensities for both the vorticity and gradient directions reach a minimum and then increase with increasing shear rate.

The specific shear rate where the peak intensity reaches a minimum depends on the shear history (0.1 ~ 1 s~^). However, a systematic relation can be found between the peak position and the peak intensity regardless the shear history as can be seen from Figure 5 where the results of different runs are shown. These results in the present study confirm the scenario reported before [5].

REFERENCES

1. S Miiller, C. Borschig, W. Gronski, C. Schmidt, D. Roux, Langmuir, 15 (1999) 7558. 2. C. E. Fairhurst, M. C. Holmes, M. S. Leaver, Langmuir, 12 (1996) 6336. 3. J. Penfold, E. Staples, A. K. Lodhi, I. Tucker, G. J. T. Tiddy, J. Phys. Chem. B, 101 (1997) 66. 4. M. Lukaschek, S. Miiller, A. Hansenhindl, D. A. Grabowski, C. Schmidt, Colloid Polym. Sci., 264 (1996) 1. 5. K. Minewaki, T. Kato, H. Yoshida and M. Imai, J. Thermal Analysis Calorimetry, 57 (1999) 753. 6. Y. Takahashi, M. Noda, M. Naruse, H. Watanabe, T. Kanaya, T. Kato, M. Imai, submitted to J. Rheol. . 7. K. Minewaki, T. Kato, H. Yoshida, M. Imai and K. Ito, submitted to Langmuir.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) G 2001 Elsevier Science B.V. All rights reserved. 189

Solubility behavior of benzylhexadecyidimethylammoniuin salts in oils.

Noritaka Ohtani

Department of Materials-process Engineering and Applied Chemistry for Environments, Faculty of Engineering and Resource Science, Akita University, Akita, 010-8502 Japan

The solubility of benzylhexadecyldimethylammonium salt (BHDAX) in aromatic hydrocarbons was examined as a function of temperature. The solubility curve afforded characteristic features, which corresponded to Krafft point and critical micelle concentration of ionic surfactants in water. When the molecular sizes of oil and counter ion were small, the solubility was high because of the low critical temperature. Above the critical concentration, BHDAX was assumed to aggregate in benzene based on NMR analyses.

1. INTRODUCTION

Benzylalkyldimethylammonium salts have been used as the catalysts of phase-transfer catalysis (PTC)' or as the catalytic moieties of the corresponding polymer-supported phase-transfer catalysis." Our early studies on polymer-supported phase-transfer catalysts suggested the presence of a specific microstructure like reversed micelles within the insoluble catalyst polymer matrix based on the analyses of oil and water, which were imbibed under the reaction conditions."^ The solution behavior and reactivity of the corresponding linear polymer attached by quaternary salts (cationic ionomers) have shown that the quaternary salts aggregate in nonpolar oils.^

The recent development of studies on both theoretical and experimental aspects of microemulsions has provided a better understanding of the parameters controlling the phase behavior of microemulsions, though the conditions dealt with have been far from those of phase-transfer catalytic conditions.^ In this article, we examine the phase behavior of BHDAX (X = CI, Br, I, OAc, OMs). We will show that simple solubility measurements afford valuable information about how the parameters, such as temperature, organic solvent, or counter-anion, influence the formation of reversed micelles or w/o microemulsions.

2. EXPERIMENTAL

2.1. Materials and Equipments BHDACl was obtained from Tokyo-Kasei and used without further purification.

BHDAX with other counter ions was prepared through the usual ion-exchange method from BHDACl and dried in vacuo. The water content was determined by 'H analysis. The degree of the ion-exchange was ascertained by the GLC analyses of decyl derivatives that were formed by the reaction of decyl methanesulfonate with the BHDAX in benzene.^ The purities of all the prepared BHDAX were over 98%. 'H(270.05MHz) NMR spectra were recorded at probe temperature of 60 °C on a Varian Mercury 300. Chemical shifts are referenced to proton impurities of C D (8 7.15) and are reported downfield of TMS.

190

60

p50i

2 40

2 30,

i 20

10

hmm I « BHDAI : mem

I I I !

0-Q " 1 1 1

0 0.1 0.2 0.3 0.4 0.5 0.6

weight fraction of BHDAX

(a) Solubility of BHDAX in benzene

60|

O 50 1 < gHbACl '

1 1 1 1

S 2of _,_-•—-"•"^ ' 1

0 0.1 0.2 0.3 0.4 0.5 0.6

weight fraction of BHDAX

(b) Solubility of BHDAX in water

Fig. 1 Solubility of BHDAX in benzene and water

2.2. SolubUity To a 10 mL screw-capped test tube with Teflon-coated stirring bar, given amounts of

BHDAX and oil were added. The tube was transferred to a temperature-variable oil or water bath, and the temperature of the system was raised with stirring at a rate of TC / min '. The temperature was read when the BHDAX crystalline solid disappeared. The uncertainty of the temperature measurement was ±0.2 °C. Then, the solution was progressively diluted with the oil, and the measurement of the solubility temperature was repeated in the same way. When the BHDAX concentration was low, the system tended to form a stable supersaturated solution. In this case, the mixture in a tube was cooled to a dry ice-methanol temperature. The tube was then transferred to a water bath maintained at 6 °C, and the BHDAX crystalline was separated out.

3. RERULTS AND DISCUSSION

The solubility curve of BHDAX in benzene is shown in Fig. la. The curves bend at certain BHDAX concentrations. At low temperatures, solid BHDAX (Q phase) coexists with a liquid solution (O phase). The concentration of BHDAX in the O phase is very low. Within a very limited range of temperatures, however, the amount of the solid BHDAX is sharply decreased as if the quaternary salt melts at the temperatures. The solubility suddenly increased, and BHDAX was soluble freely in benzene up to a considerably high concentration above the temperature. This type of solubility behavior resembles to those observed for a number of ionic surfactants in water. In fact, the solubility of BHDAX in water affords unique features characterized by Krafft point and cmc as shown in Fig. lb. At low temperatures, most of BHDAX (X = CI, Br, or OMs) is present as a solid phase (Q phase) that coexists with an aqueous solution (W phase). The concentration of BHDAX in the W phase is similarly very low. BHDAOAc was very soluble in water. Its Krafft point was not observed because BHDAOAc gave a homogeneous solution irrespective of its concentration even at 5 °C. On the other hand, BHDAI was hardly soluble in water. The solid BHDAI merely melt at 56.5 °C in the presence of water to give a liquid-liquid two-phase separation at higher temperatures. The bottom phase was an aqueous phase that contained few amounts

4.0

«o 3.0

*^2.0

<1 .0

(f(PhC

-o -J

H7 6

(UPhCH^)- (5(NCH3)

h^ ^ " ""V:

5

13 2 1 0

0 0.5 1.0 1.5 2.0 2.5 3.0

1 / [BHDACI] (mmo|- dm )

o benzene D toluene 0 p-xylene • ethyl benzene 1 I I

191

0.1 0.2 0.3 0.4 0.5 0.6

weight fraction of BHDACI

Fig. 2. Influence of BHDAQ concentration on the chemical-shift difference between benzyl methylene protons and A -methyl protons in benzene-4,. Dashed lines indicate predicted set of two straight lines when the micelle aggregation number is constant.

Fig. 3. Solubility of BHDACI in aromatic oils.

of BHDAl. The upper phase was a surfactant-rich liquid phase. It is said that surfactants are able to dissolve as an aggregate above Krafft point and their solubility increases sharply at a temperature a few degrees higher than the Krafft point. At the temperatures higher than Krafft point, surfactants form a micelle (Wm phase) if the concentration is higher than cmc. Below cmc (W phase), they are present as monomer or small aggregate.

In the same way, the results in Fig. la suggest that it is possible to define a critical solubility temperature and a critical solubility concentration, which correspond to the Krafft point and cmc of ionic surfactants in water, respectively. In benzene, therefore, it is assumed that BHDAX forms an aggregate like a reverse micelle instead of a micelle above the critical concentration. We represent the phase as Om, where BHDAX is assumed to form aggregates. The Om phase is able to imbibe a large amount of water to form microemulsions (M phase) in the same way as the Wm phase can imbibe benzene to form microemulsions. Fig. la also shows that, with an increasing volume of halide counter-anion, Krafft point and cmc of BHDAX increase. The sharp bending of solubility curve diminished for the BHDAX whose counter-ion was large, particularly for BHDAOMs. These indicate BHDAX with small counter-anions is rather soluble in benzene probably due to the feasible formation of compact reverse micelles. BHDAOAc is the most soluble in benzene among the BHDAX used, suggesting that acetate ions assist to form aggregates.

The aggregation of BHDACI is proved by the measurement of 'H NMR in benzene-J^ at 60 °C. As shown in Fig. 2, the chemical-shift difference of two singlet peaks, benzyl methylene proton and N-methyl proton, is plotted against the inverse BHDACI concentration in a similar way as the method of Lindman.*" At low concentrations, the chemical-shift difference was independent of BHDACI concentration but roughly at I mM it started to decrease because of downfield shift of A -methyl proton and upfield shift of benzyl methylene proton. The two straight lines intersect around 10 mM, which may be taken as cmc of BHDACI in benzene. The transient concentration region from 1 mM to 10 mM probably indicates the presence of pre-reversed micelle aggregates and/or the polydispersity of reversed micelles.^ The assumed cmc from this NMR measurement is coincident well to the bending

192

point of the solubility curve in Fig. la. The concentration range from the pre-reversed micelles to reversed micelles is also in good agreement with the results obtained by fluorescence quenching method.*

In Fig. 2 is also shown the chemical shift of the residual water proton; water to surfactant molar ratio was adjusted to 2.0. Water peak shifted with BHDACl concentration in concert with the peaks of BHDACl. Water peak was very close to that of free water in benzene (0.41 ppm) at low BHDACl concentrations and moved downfield with an increase in BHDACl concentration. This indicates that, at low BHDACl concentrations in the absence of reversed micelles, most of the water is freely dissolved in bulk benzene and that, at high BHDACl concentrations, water resides in the BHDACl aggregates.^ In fact, an incremental addition of water to a benzene solution of BHDACl, of which concentration is higher than the assumed cmc, induces further downfield shifting of water peak, indicating an increase in the unbound water in the core of microemulsions.

Fig. 3 shows the influence of oil on the solubility of BHDACl. The solubility curves resemble each other except in ethylbenzene. The phase transition from an 0-Q to a Wm was observed for all oils. Introducing a methyl group to benzene ring heightened the Krafft point of BHDACl: p-xylene > toluene > benzene. This suggests that the oil with a small volume easily penetrates into the surfactant tail, leading to the formation of more compact aggregates. The solubility temperature in ethylbenzene approached the value in p-xylene when the BHDACl concentration was high. However, the bending of the solubility curve was rather gentle compared with other oils, suggesting a slow increase in the aggregation number of reversed micelles and/or the wide polydispersity of the aggregation number.

REFERENCES

1. C M . Starks and C. Liotta, 'Thase Transfer Catalysis Principles and Techniques." 1978, New York: Academic Press; E. V. Dehmlow and S. S. Dehmlow, 'Thase Transfer Catalysis. 2nd Ed. ed." 1983, Weinheim, Veriag Chemie.

2. S. L. Regen, Angew. Chem.. 91, 464(1979); F. Montanari, D. Landini, and F. Rolla, Top. Curr. Chem., 101, 147(1982).

3. N. Ohtani, C. A. Wilkie, A. Nigam, and S. L. Regen, Macromolecules, 14, 516(1981); N. Ohtani and S. L. Regen, Macromolecules. 14, 1594(1981).

4. N. Ohtani, Y. Inoue, H. Mizuoka, and K. Itoh, J. Polym. Sci., Polym. Chem. Ed. 32, 2589(1994); N. Ohtani, Y. Inoue, Y. Kaneko, and S. Okumura, J. Polym.Sci., Polym. Chem. Ed., 33, 2449(1995); N. Ohtani, Y. Inoue, Y. Kaneko, A. Sakakida, I. Takeishi, and H. Furutani, Polymer J. 28, 11(1996).

5. A. Jada, J. Lang, and R. Zana, J. Phvs. Chem. 94, 381(1990); A. Jada, J. Lang, and R. Zana, R. Makhloufi, E. Hirsch, and S. J. Candau, J. Phys. Chem. 94,387( 1990).

6. B.-O. Persson, T. Drankenberg, and B. Lindman, J. Phys. Chem., 83,3011( 1979). 7. Y. Moroi and R. Matuura, Bull. Chem. Soc. Jpn., 61,333( 1988). 8. C. D. Borsarelli, C. M. Previtali, and J. J. Cosa, J. Colloid Interface ScL 179, 34( 1996). 9. F. Heatley, J. Chem. Soc. Parody Trans. 1, 84, 343( 1988); M. Seno, K. Swada, K. Araki,

K. Iwamoto, and H. Kise, J. Colloid Interface Sci., 78, 57( 1980).

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyamaand H. Kunieda (Editors) c 2001 Elsevier Science B.V. All rights reserved. 193

Deswelling kinetics of freeze-dry-treated poly(A^-isopropylacrylainide) gel in sugar solution

Norihiro Kato, Shinichiro Yamaguchi, Fujio Takahashi

Department of Applied Chemistry, Faculty of Engineering, Utsunomiya University

7-1-2 Yoto, Utsunomiya, 321-8585 JAPAN

The kinetic parameters of deswelling on a freeze-dry-treated polyiN-

isopropylacrylamide) gel were investigated using glucose and maltose as

hydi-ophilic probes. The activation entropy change at 313K (ASs^y) shifted from

negative (-190 J K^mol'O in water to positive (+410 J K' mol" ) in 0.2 M glucose.

This is explained by the structural similarity between /3 -D-glucopyranose and

water clusters. jS-D-Glucopyranose retains the normal CI conformation (chair

form). On the other hand, a water cluster possesses a tridymite-structure.

Consequently, the equatorial OH of glucose can easily form hydrogen bonds with

the surrounding water in the gel. The apparent activation energy (EJ and AS i.

are higher in 0.2 M glucose than in water, which may reflect the presence of larger

clusters of the iceberg around isopropyl groups in the presence of hydrophihc

glucose possessing many hydroxyl groups in equatorial positions.

1. INTRODUCTION

Our previous report described that poly(A^-isopropylacrylamide) (NlPAAm) gel

turned fast-responsive concerning thermal deswelling after freeze-drying (FD)

and hydration in water [1]. The deswelling rate accelerated 10' times as compared

with the conventional gel (without the FD-treatment). The hydrophobic

interaction between isopropyl groups on the side chains of polymers was

considered to play an important role during the deswelling process. The kinetic

parameters for deswelling were discussed in the viewpoint of the dehydration

process from gel in connection with the bond-formation between polymer chains in

the gel [2]. The structured water called "iceberg" (ice-like clusters formed around

the hydrophobic isopropyl gi'oups in the gel) was particularly important

concerning a lower critical solution temperature (LCST) for the poly(NIPAAm) gel.

194

The structure of sugars retaining the chair conformation was reported to resemble

stereospecifically that of the ice (tridymite structure) [3].

It remains, however, to explain how the structure of the hydrated water

changes during the deswelUng process. In this report, we present the kinetic

parameters of desweUing of the FD-treated poly(NIPAAm) gel in mono- or

disaccharide solution as hydrophihc probes. The kinetic parameters obtained

are discussed in connection with the stereospecific structure of sugars.

2. METHODS

2.1 Preparation of poly(NIPAAm) gel

Cylindiical poly(NIPAAm) gel ( 0 2mm) crosslinked by 4 mol% A';A^ '-

methylenebisacrylamide was prepared in water and then freeze-dried in the same

way as described in our previous reports [1,2]. The dried gel rods were re-swoUen

in 0-1.0 M glucose or maltose solutions at 22 °C.

2.2 Kinetic parameters

The initial desweUing rate is equal to the desweUing rate constant (k) when

the decrease of the volume for the FD-treated gel can be regarded as a zero-order

reaction. The apparent activation energy (EJ and the activation entropy change

at 313 K (AS313) were calculated according to the Arrhenius equation.

3. RESULTS

The representative deswelling-

curves in glucose solutions are shown

in Fig. 1. The gel rods equilibrated in

0.25 M glucose solution at 22°C (T )

were transferred to freshly prepared

glucose solution (C=0.25 M) at T2

(above the LOST). The desweUing

of the FD-treated gel was accelerated

with increasing T2. The initial

desweUing rate, -d(L/LJ^/dt, was

determined from the slope of the

straight Une in Fig. 1. The desweUing

rate constants (k's), obtained from the

desweUing profiles in glucose

solutions, were plotted against C (Fig.

- J

- J

1.0

0.9

0.8

0.7

0.6 400

Fig. 1. DesweUing profiles of FD-

treated poly(NIPAAm) gel.

195

2). In addition to glucose, Fig. 2

showed the plots of desweUing rate

constants against the maltose

concentration. The minima of de-

swelling rates appeared at around 0.15

M glucose and around 0.01 M maltose,

respectively. Arrhenius plots with

respect to various concentrations of

glucose solutions fell on straight lines

(data not shown). The E,'s and ASgis's

were calculated from the slopes and

intercepts of these straight Unes.

Similarly, the de-swelling kinetics of

the FD-treated gel was investigated

using maltose solutions. The E^'s and

ASsis's calculated were plotted against

the concentrations of glucose and

maltose (Fig. 3). The maxima of E,

and AS313 appeared at around 0.2 M

glucose and 0.02 M maltose,

respectively. The cuives of E,—C and

^Ssio—C showed a similar tendency

for each sugar solution.

4. DISCUSSION

Figure 3a showed that E 's

reversed to decrease when glucose

concentration exceeds 0.2 M. This is

because the desweUing rate increases.

The maximum E^ was 230 kJ mol^ in

0.2 M glucose. This is almost twice

larger as compared with around 130 kJ

mol" obtained by using alkaUmetal

halides or the NIPAAm monomer as

additives [2]. It was considered that

collapse of the iceberg during the

0.08

0.06

1

^ 0.04 ^

0.024

4

Glucose # '

Maltose

^ ^ 1

) 0.2 C/M

1 1

0.4

Fig. 2. Dependence of the

desweUing rate constant (k) with

sugar concentration.

300

Glucose

<-200

Fig. 3. (a) Dependence of the

apparent activation energy (EJ

with sugar concentration,

(b) Dependence of the activation

entropy change at 313 K (AS313)

with sugar concentration.

196

dehydration process was caused by the strong interaction between glucose

molecules and the iceberg around the isopropyl groups. Excess glucose might

destroy the water clusters from the utmost surface of the iceberg at

concentrations higher than 0.2 M. This may be deduced from the fact that E^

and AS313 decrease with increasing glucose concentration.

The kinetic parameters obtained should reflect the rate-determining steps of

the deswelling process. The sign of AS313 is considered to be important forjudging

the rate-determining steps of deswelling. There should be two rate-determining

steps in 0 and 0.15 - 0.20 M glucose: One for a bond formation process between

polymer chains (AS;:.i3 < 0) and another for a dehydration process from hydrated

polymers in the gel (AS313 ^ 0)- It is plausible that the additive glucose stabilizes

the water clusters around polymer chains. It means that the collapse process of

the water clusters may become the rate-determining step.

Maltose concentration for the maximum E^ or AS313 (around 0.02 M) was

smaller than that for glucose solution (C = 0.2 M). These facts are explained by

the fact that twice numbers of hydroxyl groups in a molecule of maltose are able to

interact with the iceberg.

The oxygen atom of each water molecule is linked tetrahedi-ally by two

hydi-ogen bonds and two covalent bonds in clusters. The conformation of water

clusters is similar to that of tridymite SiOo [3]. The distance between two

adjacent oxygen atoms bound to the D-glucopyranose hydroxyl groups is known to

be 0.286 nm. This is consistent with the distance between two adjacent oxygen

atoms of water molecules in the tridymite water (0.290 nm) [3]. Equatorial

hydi'oxyl gi'oups on/3-D-glucopyranose retaining the normal CI conformation

(chair form) are capable of forming hydrogen bonds with the adjacent water

molecules in the tridymite cluster. The molecule of glucose might substitute the

adjacent water molecules in the tridymite cluster of the gel.

ACKNOWLEDGEMENTS

This work was partly supported by Utsunomiya University SateUite Venture

Business Laboratory (SVBL).

REFERENCES

1. N. Kato, F. Takahashi, BuU. Chem. Soc. Jpn. 70 (1997) 1289.

2. N. Kato, H. Hasegawa, R Takahashi, Bull. Chem. Soc. Jpn. 73 (2000) 1089.

3. M. A. Kabayama, D. Patterson, Can. J. Chem. 36 (1958) 563.

Studies in Surface Science and Catalysis 132 Y. Iwasawa. N. Oyama and H. Kunieda (Editors) >c 2001 Elsevier Science B.V. All rights reserved. 197

NMR Study on the Effect of Added Salt on Aikylpyridinium Bromide Micelles

Shinya Kobayashi, Katsuhiko Fujio, Yuhei Uzu, and Sumio Ozeki

Department of Chemistry, Faculty of Science, Shinshu University, 3-1-1 Asahi, Matsumoto,

Nagano 390-8621, Japan

The self-diffusion coefficients, D^^^, of dodecylpyridinium bromide (DPB), l-dodecyl-2-

propylpyridinium bromide (2ProDPB) and tetradecylpyridinium bromide (TPB) were

measured in aqueous NaBr solutions by the pulse-gradient spin echo NMR technique. The

self-diffusion coefficients of micelles of these surfactants, D^, determined from their D^ 's

increase with increasing NaBr concentration. These increments of D^ would be attributed to

the decrease in an electrostatic drag due to the interaction between a micelle and an electrical

double layer (the relaxation effect).

1. INTRODUCTION

Added salts change the size and shape of ionic surfactant micelles. For example, the

aggregation number of dodecylpyridinium bromide (DPB) micelle increases from 46.0 in

water to 71.0 in 6 mol dm' NaBr solution without a significant shape change[l ]. Adding

NaBr makes tetradecylpyridinium bromide (TPB) micelle grow from a spherical micelle with

the aggregation number of 77 at 0 mol dm' NaBr to a rod-like micelle with the aggregation

number of 499 at 0.5 mol dm' NaBr [2].

In this study, we measured the self-diffusion coefficients of DPB, 1-dodecylpyridinium

bomide (2ProDPB) and TPB in solutions of NaBr in DjO by the pulse-gradient spin echo

(PGSE) NMR technique in order to investigate the effect of added salt on the self-diffusion

coefficient or hydrodynamic radius of micelles of these surfactants.

2. EXPERIMENTAL

DPB, 2ProDPB and TPB were synthesized from the corresponding pyridine and 1-

bromoalkane.

198

The PGSE NMR was measured on protons of methyl group of hydrophobic alkyl chain of

surfactants at 399.6 MHz by a JEOL JNM-LA400 spectrometer at 25±0.11:. The absolute

magnitude of the field-gradient was calibrated against the self-diffusion coefficient of pure

water [3]. Deuterium oxide purchased from Merk (99.8%) was used as a solvent.

3. RESULTS AND DISCUSSION

Fig. 1 shows the concentration dependence of the self-diffusion coefficient, D^, of DPB

in aqueous NaBr solutions. Due to the fast exchange between the monomeric and micellized

surfactants, the observed self-diffusion coefficient, D^^^, is a weighted average of self-

diffusion coefficients of the monomer, Df, and the micelle, D^, which is expressed as

_{m-m,)D^-^m,D, 1

on the assumption that above the critical micelle concentration (CMC) the monomer

0.02 0.04 0.06 0.08 0.1 -1

E o

O

JS o

m /mol kg '

100

m' /kg mol*

200

Fig. 1. Concentration dependence of the

self-diffusion coefficient of DPB at NaBr

concentrations: 0(0), 0.10(#), 0.30(A)

mol dm'\

Fig. 2. Self-diffusion coefficient vs

reciprocal of concentration of DPB at NaBr

concentrations: 0(0), 0.10(©), 0.30(A)

mol dm\

199

concentration is equal to the CMC, TTIQ [4]. m represents the total surfactant concentration.

Because of low surfactant concentrations measured, we assume that Df and D^ are almost

independent of m. This assumption is supported by the fact that the D^^ vs /w"' plot is

linear above the CMC as shown in Fig. 2. Therefore, the micellar self-diffusion coefficient

can be obtained from the intercept of the D^^ vs m~^ plot.

Fig. 3 shows D^ of DPB, 2ProDPB and TPB micelles as a function of NaBr concentration,

Wg. Hydrodynamic radii, /?„, of these micelles calculated from D^ by Stokes-Einstein

equation are listed in Table 1, together with values of D^ . It is found that for all surfactant

micelles studied D^ increases and /?„ decreases with increasing NaBr concentration.

But the increment of D^ can not be attributed to the decrease in micellar size because of the

static light-scattering result that the aggregation numbers of DPB and TPB micelles increase

with NaBr concentration [1,2], and /?„ values without NaBr are too large compared with the

lengths of the surfactant molecules with a completely stretched alkyl chain, which are 2.18,

2.17 and 2.43 nm for DPB, 2ProDPB and TPB. For charged colloidal particles, in addition to

Q 0.6

Ws / mol kg"

Fig. 3. Dependence of micellar self-diffusion coefficient

on NaBr concentration for DPB(O), 2ProDPB(#) and

TPB( A) . Dashed curves represent calculated values.

200

Table 1

Self-diffusion coefficients and hydrodynamic radii of DPB, 2ProDPB and TPB micelles

m^

mol kg"'

0

0.10

0.30

DPB

^ . 1 0 ' V s '

0.530

0.877

0.931

nm

3.74

2.26

2.13

m^

mol kg*'

0

0.10

0.30

2ProDPB

^ n ,

10'Ws'

0.696

0.953

0.980

nm

2.85

2.08

2.02

/Ws

mol kg"'

0

0.02

0.10

TPB

D^ 10'Ws'

0.497

0.834

0.840

RH

nm

3.98

2.37

2.36

the hydrodynamic drag (the contribution of particle size), the electrostatic drag due to the

interaction between a charged particle and small ions in the electrical double layer

surrounding it (the relaxation effect) are predicted by some theories[5, 6]. According to

Ohshima et al.[5] and Tominaga and Nishinaka[6], the diffusion coefficient, D, of a charged

particle is expressed as

D==D,{\-cQl^)

where D^ is the diffusion coefficient of an uncharged particle, Q^^ is the particle charge, and

c is a complicated function of particle radius and ionic strength. Dashed curves in Fig. 3

show the values calculated from the above equation taking the maximum D^ values as D^

and using Q^^ values estimated from aggregation number [1,2] and degree of counterion

binding[7, 8]. Although calculated D^ values without NaBr are larger than observed ones,

the electrostatic drag can qualitatively explain the dependence of /),„ on NaBr concentration.

To quantitative discussion further studies are necessary.

REFERENCES

1. K. Fujio and S. Ikeda, Langmuir, 7 (1991) 2899

2. K. Fujio, Bull. Chem. Soc. Jpn., 71 (1998) 83

3. R. Mills, J. Phys. Chem., 77 (1973) 685

4. P. Stilbs, J. Colloid Interface Sci., 87 (1982) 385

5. H. Ohshima, T. W. Healy and L. R. White, J. Chem. Soc. Faraday Trans. 2, 80 (1984) 1299

6. T. Tominaga and M. Nishinaka, J. Chem. Soc. Faraday Trans., 89 (1993) 3459

7. W. A. Wan-Badhi, R. Palepu, D. M. Bloor, D. G. Hall and E. Wyn-Jones, J. Phys. Chem.,

95(1991)6642 8. M. BeSan, M. MalavaSi and G. Vesnaver, J. Chem. Soc. Faraday Trans., 89 (1993) 2445

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) 201 'c 2001 Elsevier Science B.V. All rights reserved.

Small-angle neutron scattering study of w/o AOT microemulsion entrapping proteins

Mitsuhiro Hirai,t* Rika Kawai-Hirai,§ Hiroki Iwase,t and Tomohiro Hayakawa.t

tDepartment of Physics, Gunma University, Maebashi 371-8510, Japan.

§Meiwa Gakuen Junior College, Maebashi, Gunma 371-0034, Japan.

We have studied the structure of the water/AOT/heptane microemulsion containing a-chymotrypsin by using the solvent contrast variation method in small-angle neutron scattering. The protein entrapped in the microemulsion is clearly indicated to locate at water pool by the scattering data analyses. The present results agree well with the previous ones obtained using X-ray scattering.

1. E^RODUCTION

A solvent contrast variation method in small-angle neutron scattering is known to have an advantage for determining an internal scattering density distribution of a solute particle (contrast meaning an effective excess average scattering density of a solute particle in comparison with an average scattering density of a solvent). A w/o microemulsion occluding proteins consists of different components (surfactant, water, and proteins) which have distinct contrasts. In neutron solution scattering method we can easily change the scattering density of the solvent by using a mixture of protonated and deuterated apolar solvents. Hence, we can measure scattering data in a variety of contrast-profiles of solute particles to determine an internal structure of a w/o microemulsion. On the other hand, native structures of water-soluble proteins are ordinarily considered to be maintained under biological solvent conditions. Studies of unfolding-and-folding of proteins show that the addition of apolar solvent or surfactant greatly affects native structures of proteins and induces unfolding of proteins in many cases. Thus an appearance of super-catalytic activity of proteins entrapped in w/o microemulsions is a very attractive phenomenon not only from the viewpoint of a possible practical application of w/o microemulsions such as microreactors but also from that of biophysical mechanisms of protein structure stability.

By using synchrotron radiation small-angle X-ray scattering (SR-SAXS), small-angle neutron scattering (SANS), and neutron spin-echo (NSE), we have been studying

202

the relation between w/o microemulsion structure and enzymatic activity of proteins, where we found the following evidences. 1) Hydrolysis of some esters catalyzed by a-chymotrypsin entrapped in the water/sodium bis(2-ethylhexyl)sulfosuccinate (AOTVisooctane microemulsion is enhanced only at a low wo range (WQ S [water]/[AOT]) [3]. 2) At this low wo range the microemulsion is in a transient phase between the oligomeric and monomeric phases [4]. 3) A bending fluctuation of the microemulsions is significantly enhanced at this low wo range [5]. 4) Structural properties and phase boimdaries between different phases (oUgomeric, transient and monomeric phases) of w/o AOT microemulsions are defined by the penetration limits of apolar solvents depending on the linear hydrocarbon chain lengths of those solvents [6]. In the present study, to clarify the location of proteins in w/o AOT microemulsions we have carried out SANS experiments using the solvent contrast variation method.

2. EXPERIMENTAL

AOT was purchased from Nacalai Tesque Inc. Solvents used were 99.9+ % n-heptane from Wako Chem. Co., die-n-heptane (99+ atom % D) and deuterium oxide (100 atom % D) from Aldrich Chem. Co. Normal water purified by a Millipore system was used. Type-II o-Chymotrypsin firom bovine pancreas produced by Sigma Chemical Co. was used without further purification. The a-ch)rmotrypsin was solubilized in D2O/H2O mixed water (98 % D2O v/v) whose scattering density is matched to that of di6-n-heptane. The microemulsions were obtained by using an injection method. The fractions of di^-n-heptane in the apolar solvent were 100 % v/v, 85 % v/v, and 70 % v/v. The [waterMAOT] molar ratios ( s wo) were selected to be 12 and 20. The AOT concentrations of all samples were 0.1 M. The concentrations of a-chymotrypsin [E] were 0 M and 1.32 x 10-4 M for wo = 12, and 0 M and 2.21 x 10-4 M for WQ = 20.

SANS experiments were carried out by using a SANS instrument of JRR-3M of JAERI, Tokai, Japan. The sample solutions were contained in a quartz cell controlled at 25 ""C. The following standard analyses of SANS data were executed. By using the Guinier plot (iTiIiq) versus q^) on the data sets in a defined small q range (0.03-0.05 A-^), we determined the values of the zero-angle scattering intensity 1(0) and the radius of gyration Rg of the solute particle by using the following equation

/(^) = /(0)exp(V/?//3) (1)

where q = {4n/X)sin{d/2), 6 and X are the scattering angle and the neutron wavelength. The distance distribution function p(r) was obtained by the following Fourier transform of the scattering intensity Hq).

P(r) = '^]rql(q)smirq)dq (2)

203

Thep(r) function depends both on the geometrical shape and on the internal scattering density distribution of the solute particle.

3. RESULTS AND DISCUSSION

Fig. 1 shows the change of the scattering curve I(q) depending on the ratio of [di6-n-heptane]/[n-heptane] and on the protein concentration,. The significant decrease of the scattering intensity below q = 0.03 A'l is attributed to the presence of a beam stopper. In Fig. 1(a), the scattering curve Kq) varies with decreasing the die-n-heptane concentration, which results from the change of the contrast ofthe AOT

microemulsion. The protein concentration dependence of the scattering curve is shown in Fig. 1(b). Due to the occlusion of the proteins in the microemulsion, the internal scattering density distribution greatly varies, which results in the change ofthe scattering curve. As is well known in the solvent contrast variation method, a heterogeneity in the internal scattering density distribution of a solute particle is more clearly reflected with approaching the solvent scattering density to the solute scattering density. In the present case, the effect of the presence of the proteins on the microemulsion structure can be seen most evidently at the low contrast solvent condition, namely, at 70 % v/v di6-n-heptane since the contrast matching points of the AOT and protein molecules are lower than 70 % v/v die-n-heptane.

The distance distribution function pir) using Eq. 2 in Fig. 2 directly shows the change ofthe internal structure ofthe microemulsion in real space by the presence of the proteins. The peak position of thep(r) function shifts from 55 A to 49 A with holding the bell-shaped pir) profile, indicating that the occluded proteins are located at the center ofthe microemulsion.

The zero-angle scattering intensity

J I I L

10' 1 1 1 1 1

in 70 % d-heptane (b)

[E1 = 0

[E]«2.21x10-^M

j _

0 0.05 0.1 0.15 0.2 0.25 0.3

Fig. 1. Change of the scattering curve of AOT microemulsion (wo = 20) depending on the d-heptane fraction in the solvent (a) and on the protein concentration (b).

204

€ CO

1.0

0.5

1 1 1 1 in 70 % d-heptanej

/ T 1 / A / [E]«2.2x10-^M*\ y \ 1 1 ' \ 1

25 50 75

(A) 100

Fig. 2. Difference between the distance distribution functions p(r) obtained from the scattering curves in Fig. 1(b).

- e — w=20, [EJsO

0 20 40 60 80 d-heptane fraction (%v/v)

1500

' 0)1250 cc

1000

T

( r - . /

r 1

•

1 i 1 -0.02 -0.015 -0.01 -0.005

1/contrast (% d-heptane)

Fig. 3. (a). Square root of the zero-angle scattering intensity plotted against d-heptane fraction, (b), Change of Stuhrmann plot due to the occlusion of the proteins. Marks in (b) are as same as in (a).

7(0) and the radius of gyration Rg of the solute particle were determined by the Guinier plot using Eq. 1. The contrast matching point of the solute particle was obtained from the plot of [7(0)] 1 2 against d-heptane concentration, as shown in Fig. 3(a). As we know the contrast matching point and can calculate the average scattering density of the heptane solvent at each d-heptane fraction, we can plot the Rg^ against 1/contrast. This plot, so-called the Stuhrmann plot, well reflects a heterogeneity of internal scattering density distribution of a solute particle [7]. The negative slope in Fig. 3(b) indicates that the solute particle consists of a high density core surroimded by a low density shell. This agrees with the present microemulsion structure since its water pool contains 98 % D2O v/v. The decrease of the absolute value of the slope by the occlusion of the proteins also shows that the proteins locate at the center of the microemulsion to descend the scattering density of the water pool.

REFERENCES

1. R. Hilhorst, In Structure and Reactivity in Reversed Micelles; Pileni, M. P. (ed.), Elsevier, Amsterdam, (1989) 323. 2. R. H. Pain (ed.), Mechanisms of Protein Folding, IRL Press, New York, 1994. 3. M. Hirai, et al., J. Chem. Soc. Faraday Trans., 91 (1995) 1081. 4. M. Hirai, et al., J. Phys. Chem., 99 (1995) 6652. 5. M. Hirai, et al., J. Phys. Chem. Solids, 60 (1999) 1297. 6. M. Hirai, et al., J. Phys. Chem. B, 103 (1999) 9658. 7. H. B. Stuhrmann, and A. Miller, J. Appl. Cryst., 11 (1978) 325.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) 205 cc 2001 Elsevier Science B.V. All rights reserved.

Neutron Spin Echo Investigations on Slow Dynamics in Complex Fluids Involving Amphiphiles

Takayoshi Takedaa, Youhei Kawabatab, Hideki Setoa, Shigehiro Komurac and Michihiro Nagao^

aFaculty of Integrated Arts and Sciences and ^Graduate School of Bio-Sphere Science, Hiroshima University, Higashi-Hiroshima 739-8521, Japan

cDepartment of Physics, ChuoUniversity, Bunkyo-ku, Tokyo 112-8551, Japan

^Institute for Solid State Riysics, University of Tokyo, Tokai, Naka Ibaraki 319-1106, Japan

In order to elucidate the self-assembling mechanisms in complex fluids involving amphiphiles, we have studied slow dynamics in complex fluid systems such as the non-ionic surfactant C12E5 / n-octane / D2O system and lipid DPPC / E)20 / CaCla system using neutron spin echo spectroscopy (NSE). The intermediate functions I(Q, t) obtained from the NSE experiments were well fitted to a stretched exponential function in time l(Q,t) = / (Q,0)exp[- (17 )2/3]. The relaxation rate F increased as (?3. The NSE results support the theory presented by Zilman and Granek [Phys. Rev. Letters 77 (1996) 4788.]. We estimated the bending modulus x of the interfacial membrane using their theory.

1. INTRODUCTION

In order to elucidate the self-assembling mechanisms in complex fluids involving amphiphiles, microscopic parameters like the bending modulus x on the local scale of the interfacial membrane deduced from dynamical experiments using neutron spin echo spectroscopy(NSE) are important. The wave vector Q and time t dependent intermediate functions /((?,/) were obtained from neutron spin echo (NSE) experiments on the following two systems; (a) film and bulk contrast samples in the non-ionic surfactant H(CH2)i2(OCH2CH2)50H(=Ci2E5)/n-octane / D2O ternary system at equal volume fraction of octane and water with 0.2 volume fraction of C12E5, which shows a sequence of low temperature micro-emulsion(LTM)/lamellar(MTL)/high temperature microemulsion (HTM)phase with increasing temperature[l, 2], (b) the lipid dipalmitoylphosphatidyl-choline (= DPPC)/D20/CaCl2) system. In the C12E5 /n-octane/D20 system, the LTM and HTM phase have a bicontinuous structure. The lamellar repeat distance 1 in the liquid crystalline phase La varies greatly upon addition of salt in the DPPC/water system[3]. In order to study the dynamics in undulation of lipid bilayers, NSE experiments were carried out on the dilute lamellar phase of the DPPC / water / CaCl2 system with d\ longer than 300 A in order to avoid interactions between neighbouring sheets of membranes.

206

2. EXPERIMENTAL

For sample preparation, 99.7%pure C12E5 was purchased from Tokyo Chemical Co., 99% n-octane from Aldrich Chemical Co., 99 % deutrated n-octane from Isotec Inc., 99.9 % D2O from Isotec Inc., 99% a-L-DPPC from Sigma Chemical Co. and 99.5% CaCl2-2H20 from Wako Pure Chem. Ltd. These materials were used without further purification. In the ternary system C12E5/ n-octane / D2O at equal volume fraction of octane and D2O with volume fraction 0.2 of C12E5, protonated n-octane was used for the bulk contrast samples and deutrated n-octane for the film contrast samples. To prepare the lipid samples, 7.7wt% DPPC was dispersed in D2O solutions with 6.8mM CaCk for DSl sample and 6.3wt% DPPC in D2O solutions with 6.6mM CaCk for DS2 sample. In small angle neutron scattering patterns, the peak corresponding to d\ = 430 A was observed for DSl sample. The NSE experiments were performed using ISSP- NSE at C2-2 port of JRR-3M, JAERI[4-6]. The experiments were also carried out as the performance test of the spectrometer. Silica gel and Grafoil were used to measure the resolution function of the spectrometer. Neutron beams with wavelength X = 5.9 A (FWHM of its resolution AX/X = 15%) and X = 7.14 A (AX/X = 18%) were used.

3. RESULTS AND DISCUSSION

The intermediate correlation functions I(Qj) obtained from the NSE experiments were well fitted to the following equation,

/((?,O=/((?,0)exp[-(rO2/3 ] (1)

in both the bicontinuous microemulsion and the lamellar phases of the Ci2E5/n-octane/D20 system and also in the lamellar phase of the DPPC/D20/CaCl2 system. The relaxation rates r obtained from the fitting to Eq.(l) increased as (? 3 over the range of Q from 0.08 A-l to 0.17 A-l for the C12E5 / n-octane/D2O system and Q from 0.05 A-l to 0.14 A-l for the DPPC / D2O / CaCl2 system as shown in Fig. 1. In our previous papers concerning mainly the Ci2E5/n-octane/D20 system[7-9], we reported that the NSE results supported the theory presented by Zilman and Granek[10]. They predicted a stretched exponential relaxation of / (Q/) as follows Eq.(l) where the relaxation rate F is given by

r = 0.025y (/CBT/X )l/2(ter/rj )Q 3 . (2)

Here, x is the bending modulus of the membrane and rj the viscosity of the surrounding medium. The factor y originates from averaging over the angle between Q and the plaquette surface normal in the calculation of / {Qj). Figure 2(a) shows x estimated in the bicontinuous microemulsion and the lamellar phases of the C12E5 / n-octane / D2O system using Eq.(2) where we put y = l a n d used 3 times the value of average solvent (n-octane and D2O) viscosities for rj (rj = 3rjsoivent) taking the local dissipation at the membrane into consideration[ll]. The values of x are nearly same as that estimated from dynamical light scattering in the C12E5 / hexanol / water system at room temperature[12]. x in Fig.2(a) decreases monotonically with increasing temperature independently of the mesoscopic structure and the scattering contrast. The rate F deviated from Eq. (2) at Q higher than

207

3.5 ! - ' • ' . ' ' • •

tbicontinuous

1

^ 0.1

'in •b T—

>- u

0.01

0.001

- ^ - T

r =

-

1 1 1

/Jk//^

1 1 1 1 1 1 1

O / J

c/^/ 1 Qm/ \

Rale 1 d i ^ ^A -1

n /v/ 1

z-y i 3

•

-d

1 — • — » • 1 — - — , 1

0.04 0.06 0.080.1

Q (A-^)

0.2

Fig. 1. The dependence of the relaxation rate F on Q obtained using the fitting to Eq.(l) in the LTM(full circles) and HTM phase (open circles) of the film contrast sample of the CiaEs/n-octane/water system, and DSl sample at 46*C ( full triangles) and at 52oC ( open triangles). The lines indicate Q3 proportionality.

C12E5/D20/n-octane

bicontinuousi

295 300 305 310

T(K) 315

14

^ 12

^ 10

>i 8

6

4

L ' ' '

i V. r(b) 1.J-J 1 1 1 1

•^-..

DPPC/DjO/CaClj :

J • 0 * — . . •

^... _1 .. .—._ . 1 . _. . . 1 315 320 325

T(K) 330 335

Fig. 2. The dependence of bending modulus X of the membrane on the temperature T obtained from NSE experiments for the two systems; (a) the bulk( open circles) and film sample(full circles) in the C12E5 / n-octane / water system , and (b) DSl sample(full circles) and DS2 sample(open circles) in the DPPC/DaO/CaCb system. The lines which are a guide for the eye are fitting curves to an exponential function x - a exp(- bIT ), where a and A are constants.

0.18 A-l and at Q lower than 0.08 A-l as shown in Fig. 1. At Q lower than 0.08 A-l, the dependence of F on Q depended on the mesoscopic structure and the scattering contrast. The collective motions of the membranes that are not considered in the Zilman and Granek's theory may play an important role in the dynamics at Q lower than 2x1^ where ^ is a typical size of the mesoscopic structure, while the Zilman and Granek's model is considered to be applicable to dynamics of a single membrane at higher Q, The deviation at Q higher than 0.18 A-l suggests that the effect of the thickness of the membrane which is neglected in the the Zilman's model plays a significant role in the dynamics at higher Q. In the case of the DPPC /DaO/CaCb system, the Zilman and Granek's model is considered to be applicable, since <? » 1 is satisfied. Figure 2(b) shows x estimated in the DPPC/DaO/CaCk) system using the similar procedure to the case of the CiaEs/n-octane/DaO system; y = 1 and ^ = 4 r7soivent. ^ decreases roughly with increasing temperature. The values of x are the

208

same order as that estimated from the microscopic observation of giant flaccid vesicles[13]. The values of x seem to depend strongly on d\ though d\ is longer than 300A.

The estimated values of x decrease with increasing temperature and seem to be reasonable over the range oi x from 0.86 ArBr to 13 iteT'in the present study. This result indicates that F depends on « in the manner predicted by Zilman and Granek which shows an anomalous dependence of T on p , F'^x -i/2, in contrast to other theories of membrane undulations and that their theory describes well membrane undulations in these complex fluids involving amphiphiles.

ACKNOWLEDGEMEl^TS

This experiments at JRR-3 were done under the approval of the Neutron Scattering Program Advisory Committee. One of the authers(T. T.) was financially supported by the Grant-in-Aid for Scientific Research (No. 06640505, No. 07236103, No. 08044089, No. 09640466) from the Japanese Ministry of Education, Science, Sports and Culture.

REFERENCES

1. M. Kahlweit, R. Strey, D. Haase, H. Kunieda, T. Schmeing, B. Faulhaber, M. Borkovec, H. F. Eicke, G. Busse, F. Eggers, T. Funck, H. Richmann, L. Magid, O. Soederman, P. Stibs, J. Winkler, A. Dittrich, W. Jahn, J. Colloid & Interface Sci. 118 (1987) 436. 2. S. K. Ghosh, S. Komura, J. Matsuba, H. Seto, T. Takeda, M. Hikosaka: Progr. Colloid Polym.Sci. 106(1997)91. 3. T. Takeda, S. Ueno, H. Kobayashi, S. Komura. H. Seto, Y. Toyoshima : Physica B 213&214 (1995) 763. 4. T. Takeda, S. Komura, S. Seto, M. Nagai, H. Kobayashi, E. Yokoi, C.M.E Zeyen, T. Ebisawa, S. Tasaki, Y. Ito, S. Takahashi and H. Yoshizawa : Nucl. Instr. and Methods in Phys. Research, A364 (1995) 186. 5. T. Takeda, H. Seto, S. Komura, S. K. Ghosh, M. Nagao, J. Matsuba, H. Kobayashi, T. Ebisawa, S. Tasaki, C. M. E. Zeyen, Y. Ito, S. Takahashi, H. Yoshizawa; J. Phys. Soc. Jpn. 65 SuppLA (1996) 189. 6. T. Takeda, H. Seto, Y. Kawabata, D. Okuhara, T. Krist, C. M.E. Zeyen, I.S.Anderson, P. Hfghfj. M. Nagao. H. Yoshizawa, S. Komura. T. Ebisawa, S. Tasaki and M. Monkenbusch, J. Phys. Chem. Solids 60 (1999) 1599. 7. T. Takeda, Y. Kawabata, H. Seto, S. Komura, S. K. Ghosh and M. Nagao, AIP CP-469 (1999) 148. 8. T. Takeda, Y. Kawabata, H. Seto. S. Komura, S. K. Ghosh and M. Nagao, D. Okuhara, J. Riys. Chem. SoUds 60 (1999) 1375. 9. T. Takeda, Y. Kawabata, H. Seto, S. K. Ghosh. S. Komura and M. Nagao, AIP CP-514 (2000) 190. 10. A. G. Zihnan and R. Granek, Phys. Rev. Letters 77 (1996) 4788.4. 11. B. Farago, M. Monkenbusch, K. D. Goecking, D. Richter, J. S. Huang: Physica B 213&214 (1995) 712. 12. F. Nallet, D. Roux, J. Prost: J. Phys. France 50 (1989) 3147. E. Freyssingeas, D. Roux, F. Nallet: J. Phys. H France 7 (1997) 913. 13. H. Engelhardt, H. P. Duwe, E. Sackmann: J. Phys. (Paris) Lett. 46 (1985) L-395.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) 209 (O 2001 Elsevier Science B.V. All rights reserved.

NEUTRON SPIN ECHO STUDIES ON EFFECTS OF TEMPERATURE AND PRESSURE IN DYNAMICS OF A TERNARY MICROEMULSION

Y. Kawabata^ M. Nagao^ , H. Seto" and T. Takeda"

^Graduate School of Bio-Sphere Science, ^Faculty of Integrated Arts and Sciences, Hiroshima University, Higashi-Hiroshima 739-8521, Japan

^Institute for Sohd State Physics, The University of Tokyo, Tokai 319-1106. Japan

In order to understand dynamical features of amphiphile membranes, we performed neutron spin echo (NSE) experiments in a dilute droplet system consisting of AOT(dioctyl sulfosuccinate sodium salt), D2O and C10D22 at room temperature, at high temperature, and at high pressure. The bending modulus of membranes was obtained from the model fitting to a theory proposed by Milner and Safran[9] describing diffusive dynamics of droplets. We found that membranes at high temperature was more flexible than those at room temperature, while, those at high pressure was more rigid. This result was consistent with our previous study in a dense droplet microemulsion system.

1. Introduction

Ternary complex fluids, consisting of water, oil, and amphiphile, form various structures depending on its composition, temperature, and/or pressure. We have investigated static and dynamical features of the A0T/D20/n-decane system by means of small angle X-ray(SAXS) and neutron scattering(SANS), and neutron spin echo(NSE) so far[l]-[5], and showed that temperature and pressure induced a phase transition from one-phase dense droplet to two-phase coexistence with an ordered lamellar phase and a disordered bicontinuous phase. Our results indicated that the static features of the high-temperature and the high-pressure phases were similar, however, the dynamical features were different. We concluded that amphiphihc membranes between oil and water became flexible with increasing temperature, while rigid with increasing pressure. However, the correlation between droplets cannot be ignored in the dense droplet system, and it was difficult to understand the dynamical features of the membranes only with the dense droplet system.

In this study, we focused on the dynamical features of amphiphilic membranes in a dilute droplet system (0 ~ 0.1, where 0 is the droplet density). In this system, observed dynamics was successfully explained in terms of the shape and size fluctuation and the translational diffusion of droplets, as shown by Huang et al[6]. We also calculated the ratio w = ro/vs of a radius of droplet ro to a spontaneous radius of droplet rg. We found that the behavior of the n was the same with the previous results[4,5].

210

2. Exper iment

The NSE experiments were performed at the ISSP-owned NSE spectrometer at JRR-3M in JAERI, Tokai[7]. The measured momentum transfer q ranged over 0.04 < ^ < 0.14 [A~^] and Fourier time t over 0.15 <t< 15[ns]. The volume fraction of each composition of the used sample was 5.4% of AOT(99% purity, purchased from Sigma purified with extracting from benzene), 4.6% of D20(99% purity, purchased from Isotec Inc.) and 90% of C10022(99% purity, purchased from Isotec Inc.). The structure of this system was characterized to be the dilute water-in-oil droplet structure with the droplet radius of about 30A at ambient temperature and pressure. The experiments were carried out under the following three conditions; (a)25°C under ambient pressure (RTF), (b)43°C and O.lMPa (HT), (c)25°C and 45MPa (HP). For the high pressure experiment, a high pressure cell made with non-magnetic stainless steel coupled with sapphire windows was used.

3. Analys i s

The present NSE data were analyzed using a theory proposed by Milner and Safran[9]. In their theory, the normalized intermediate correlation function I(q^t)/I(q,0), which is directly observed from the NSE experiments, decays with single exponential function as follows,

I{q,t)/I{q,0) = exp[-D,ff{q)qH], (1)

where Deff{q) is the effective diffusion coefficient and q is the scattering vector. Taking the shape deformation of droplets into account, the Deff(q) is a sum of two terms,

DefM) = Dtr^DdeM). (2)

where A r is the translational diffusion coefficient and Ddef{q) the shape deformation diffusion coeflficient. The Ddef{q) can be written as,

D (n) = 5A2/2(gro)(|a2p) .3 . '^^^"^^ q^inryoiqro) + bf2{qro){\a2\')]

with the second mode(n = 2) of the shape deformation is taken into consideration. The n = 2 mode can be mainly observed using the ISSP-NSE spectrometer, because the damping frequency of the deformation of the other modes except the second mode is out of the dynamic range of the spectrometer. Here, A2 is the damping frequency of the second mode, ro the radius of droplet, w the ratio of ro to a spontaneous radius of droplet r^, and fn the n - th weighting factor including the spherical bessel function. The mean square of the second mode fluctuation amphtude, (|a2p), and the A2 can be expressed as,

(|a,P) = A^^^^rg, (4)

K 9 6 X . . A2 = —^ir^w, (5)

rjr^ 55

21]

t [nsl

Figure 1. Observed intermediate correlation function from NSE experiments at HP. Lines are fitting results to Eq.(l) .

Figure 2. The effective diffusion coefficient Deff obtained from the fitting to Eq.(l) . Full circles correspond to the RTF phase, open triangles the HT phase and full squares the HP phase. The lines are fitting curves to Eq.(2).

where T] is the average of the viscosity inside and outside the droplet, n the bending modulus of amphiphilic membranes. A2 was obtained from the fitting to the observed Deff{q). In order to calculate TS and u^ we used the following expression[9],

r{qro = 7r) y ^ Q.3A2

To J W

7r/2 4-

0.15 (6)

This equation is reduced by taking polydispersity of droplet due to the averaging numerator and denominator over the different radii in the sample into account[6,9].

4. Resul ts and Discussion

The observed I{q,t)/I{q,0) at HP and De/f are shown in Figure 1 and Figure 2 with the fitting curves to Eq.(l) and Eq.(2) respectively. In Figure 2, a peak position and peak height are corresponding to the characteristic wave number of the droplet radius {Qpeak = 7r/ro) and shape fluctuation diffusion coefficient Ddef. respectively. The qpeak shifts to larger q and the Deff at <? = 0 becomes larger with increasing temperature and pressure. This means that ro become smaller than that at RTF and Dtr also become larger than that at RTF, with increasing temperature and pressure. These evidences indicate that the translational diffusion of spheres become faster with decreasing of the size of spheres.

In the previous paper, we estimated the K and (|a2p)/ro with w = 1 and //(the average of viscosity of water and oil)[5]. In the present analysis, we evaluated the K and (|a2p)/ro with w ~ 0.3 and 4r). The value of w can be obtained from Eq.(6), and that of Ar] is the viscosity which is taken the local dissipation of membranes into account. The calculated

212

Table 1 The obtained fit parameters for each condition from the NSE experiments.

K[kBT] ll£2[!l

^0

Ar[xlO-^cmVs]

-o[A]

RTP

1.7

0.10

3.9

36

HP

2.0

0.08

4.7

32

HT

1.4

0.13

5.6

31

parameters from the fitting resuhs to Eq.(2) and (5) at each condition are summarized in Table 1.

The fractional displacement against droplet radius. (|a2p)/'"o' decreased with increasing pressure and increased with increasing temperature. This indicates that the shape fluctuation is suppressed with increasing pressure and vice versa with increasing temperature. The K at HP was larger, and that at HT was smaller than that at RTP. This result shows that the membranes become rigid with applying pressure, and flexible with increasing temperature. The value of K is consistent with the result obtained by Farago et al[8].

These evidences are consistent with our previous results in the dense droplet system[2.3]. and we concluded that the dynamical features at HT are different from those at HP. This difference between the effects of temperature and pressure is due to the interaction around surfactant molecules as discussed in our previous articles[l]-[3]. The tail-tail attractive interaction increases with increasing pressure, while the head-head repulsive interaction increases with increasing temperature because of the dissociation of counter-ion from head group.

REFERENCES

1. M. Nagao and H. Seto, Phys. Rev. E 59 (1999) 3169. 2. M. Nagao H. Seto Y. Kawabata and T. Takeda, J. Appl. Cryst 33 (.2000) 653. 3. H. Seto, D. Okuhara, Y. Kawabata, T. Takeda, M. Nagao, J. Suzuki, H. Kamikubo

and Y. Amemiya, J. Chem. Phys. 112 (2000) 10608. 4. M. Nagao, Y. Kawabata, H. Seto and T. Takeda MP Conference Proceedings 469

(1999) 154. 5. Y. Kawabata, M. Nagao, H. Seto and T. Takeda AIP Conference Proceedings in press. 6. J. S. Huang, S. T. Milner, B. Farago and D. Richter, Phys. Rev. Lett. 59 (1987) 2600. 7. T. Takeda, H. Seto, S. Komura, S. K. Ghosh, M. Nagao, J. Matsuba, H. Kobayashi,

M. Nagai, H. Kobayashi, T. Ebisawa, S. Tasaki, C. M. E. Zeyen, Y. Ito, S. Takahashi and H. Yoshizawa, J. Phys. Soc. Jpn. 65, Suppl. A (1996) 189.

8. B. Farago, D. Richter, J. S. Huang, S. A. Safran, and S. T. Milner, Phys. Rev. Lett. 65, (1990) 3348.

9. Milner, S. T., and Safran, S. A., Phys. Rev. A 36, (1987) 4371.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) 213 c 2001 Elsevier Science B.V. All rights reserved.

Dimerization of penicillin V as deduced by frontal derivative chromatography

Seiji Ishikawa, Saburo Neva, and Noriaki Funasaki

Kyoto Pharmaceutical University, Misasagi, Yamashina-ku, Kyoto 607-8414, Japan

The aggregation pattern of penicillin V (PCV) is investigated by frontal derivative chromatography on Sephadex G-10 columns. The height and position of the trailing peak, computer-simulated on the basis of the stepwise aggregation model, are in excellent agreement with the observed data, whereas those simulated on the basis of the trimer -dodecamer model remarkably differ from the observed data, particularly in dilute solutions. Thus, PCV forms dimers and self-associates stepwise.

1. INTRODUCTION

Many drugs form aggregates in water by hydrophobic interactions. This causes significant changes in bioactivity, chemical stability, and osmotic pressure of the drugs in aqueous solutions [1]. Dimerization must take place in all the self-associating systems being considered. However, the formation of higher multimers may often overshadow it and may lead to difficulties involved in even detecting it. Frontal derivative chromatography is an excellent method for detecting dimerization and determining the dimerization constant, because the frontal derivative chromatogram of the dimerizing system exhibits a characteristic pattern [1 ]. This characteristic pattern of dimerization can be used to detect dimer even in the presence of other multimers. Generally, dimerization does not cause any inflection, corresponding to a cmc, in the concentration dependence of physical properties. This often leads to the misinterpretation that dimmers are absent in micellar systems.

In 1994 we showed by frontal gel filtration chromatography (GFC) that penicillin V (PCV) self-associates stepwise in a 150 mM potassium chloride solution at 298.2 K [1]. In particular, we determined the dimerization constant of PCV on the basis of GFC data, such as the centroid volume and the position and height of the trailing peak. Very recently, however, it was reported that PCV forms only two species, trimer and dodecamer, in remarkable contrast with our model [2].

In the present work we provide further chromatographic evidence for our stepwise aggregation model of PCV to resolve the above disagreement on the aggregation pattern of PCV.

2. EXPERIMENTAL SECTION

A specimen of PCV (potassium phenoxymethylpenicillinate) was received from Sigma. Sephadex G-10 (Pharmacia) columns were treated as suggested by the manufacturer. The

214

double-distilled water was degassed just before each GFC experiment. All GFC experiments were carried out in a 150 mM potassium chloride solution under

a flow rate of ca. 0.36 cm^ min'\ Two columns, B and C, were used for PCV. The columns were jacketed in order to maintain them at a constant temperature of 298.2 ± 0.2 K. A large volume of sample was applied so that the plateau region appeared on the elution curve. The concentration of PCV in eluate was monitored continuously with a differential refractometer. The GFC experiments were carried out at a number of concentrations. For the case of two concentrations, Co = 0.2002 and 0.2600 M, on column C, 2.00 cm^ of eluate was taken in a test tube with a fi-action collector, because the refractometer did not respond to such high concentrations. Simulations of chromatograms were carried out by a plate theory. The number of the plate (AO and the void volume (Vo) of the column are the same as those already reported [1 ].

3. RESULTS AND DISCUSSION

3.1. Frontal chromatogram of PCV A large volume of a dilute PCV solution was applied on the Sephadex G-10 column,

so that the plateau region appeared on the chromatogram (Fig. la). This chromatogram is termed the frontal chromatogram. Because the concentration C of PCV in the plateau is the same Co as that applied, this chromatogram affords us quantitative information on the self-association of PCV. We may assume that the equivalent sharp boundary for the leading or trailing edge of the solute zone is approximately the initiation or termination of the plateau region (centroid) of the elution profile. In Fig. 1, S denotes the applied volume of the sample. Because the centroid volumes, Fc' and Fc, at the leading and trailing boundaries were equal to each other within experimental errors, we took the average of them as Fc. Frontal derivative 5 chromatograms (Fig. lb) reflect o aggregation patterns.

As we have shown [3], we can determine the monomer concentration C] from

Cy=(Vc-Vrn)C/iV,-Vm) (1) <r^ 0.08

Here V\ and F^ denote the centroid volumes of the monomer and aggregates of PCV, respectively: Fi = 24.00 cm^ and Vm = 7.03 cm^ on column B and Fi = 21.84 cm^ and Vm = 6.35 cm" on column C. From Eqn. 1 the monomer concentration of PCV was determined as a function of the total PCV concentration. According to multiple equilibrium theory for self-association, the micellar weight average aggregation

0.04

0.00 80 40

V{cr[9) Fig. 1. (a) Frontal chromatogram of PCV at CQ = 0.2002 M and 5 = 42.50 cm^ and (b) its derivative with definitions of chromatographic parameters.

215

number w is calculated from

«w = dlog(C-C,)/dlogC, (2)

Our data showed that PCV forms dimer at low concentrations and larger multimers at higher concentrations [1].

Figure 2a shows the observed derivative chromatograms of C'o 0.04177 mM at the leading and trailing boundaries on column B. The volume coordinate for the trailing boundary is shown as F - S\ the volume coordinate is assigned a zero value when the trailing boundary of the applied sample leaves the column bed. The shape of the derivative chromatogram at the trailing boundary reflects the aggregation pattern. The peak volume decreases with increasing concentration. According to asymptotic theory, the volume, Fp, of the trailing peak for the dimerization system approximately obeys the following equation [3]:

{{V,;-F2p°^)/(Kp - K2p*)} = 1 + 9.6/:2C^x (3)

Here Cmax denotes the concentration at the peak volume, Fp, at the trailing boundary (Fig. 1) and Ki stands for the dimerization constant. The Fip"" value may be estimated from extrapolation of the Fp values to zero concentration, and the F2p'' value may be set as the Fp value of blue dextran; Fip"" = 23.50 cm^ and Vi^ = 6.35 cm^ for column B. As Fig. 3 shows, Eqn. 3 holds true at dilute concentrations. The dimerization constant of PCV evaluated from the slope of the linear portion in Fig. 3 is 4.7 M* and is close to 4.38 M " obtained from the centroid volume data [1].

3.2. Two self-association models of PCV We have proposed a stepwise aggregation model for PCV. According to this model, the

total PCV concentration can be written as

oo

C = Ci + IK iC^ + I /CVexp(a/ - ^/'^' - ci^^') (4) 3

Here Ai stands for the overall aggregation constant for the formation of i-mer from / monomers. The term a corresponds to the driving force of micellization due to the transfer of the hydrophobic group from water to the micelle. The term h expresses the reduction of hydrophobic surface area caused by spherical micelle formation. The term c denotes the electrostatic repulsion between the hydrophilic groups at the micellar surface. These aggregation parameters were determined to fit the calculated Fc value to the observed one for a given concentration. The best fit parameters were determined to be ^ 2 = 4.38 M\ a = 78.2, h = 75.8, and c = 19.3.

On the other hand, Valera et al. proposed that PCV forms trimers and dodecamers alone [2]. According to their model, the total concentration of PCV is written as

c = c, + 3A:3Cr'+i2 i2cV^ (5)

216

0.010

0.008

I 0.006 h

^ 0 . 0 0 4

c3 •9. 0.002 h

0.000

(a)

5 •• • • r •

• L • • 1 •

f rf'"^o« L o ° ; °4

82__« 1 • • ^ n

(b) '

/\ i

/' \ '• /•' \' f— / ' \ '

^ /' \' / ' \ 'i

/lp\J,

-/1\: 10 20 10 20 30

V (cml \/ (cm') Fig. 2. (a) Observed derivative chromatograms of PCV of Co = 0.04177 M at the leading (closed circles) and trailing (open circles) boundaries and (b) simulated on the basis of the stepwise aggregation model (solid lines) and the trimer - dodecamer model (dashed lines).

0.00

Fig. 3. Plots of trailing peak positions according to Eqn. 3: O; observed, solid line; simulated on the basis of the stepwise aggregation model, dashed line; simulated on the basis of the trimer -dodecamer model.

Using our observed Vc data, we determined the best fit trimerization and dodecamerization constants of A'3 = 71.93 M"' and AT^^ 19.53x10^ Wr^\ respectively.

3.3. Simulations of derivative chromatograms We simulated the derivative chromatograms at Co = 0.04177 M on the basis of the two

aggregation models for PCV. As Fig. 2b shows, our stepwise model is clearly better fit to the observed chromatogram (Fig. 2a) at Co = 0.04177 M than the trimer - dodecamer model. In particular, the peak positions of the derivative chromatogram simulated on the basis of the trimer - dodecamer model are distant from the observed ones. This disagreement is mainly ascribed to the neglect of dimer in the trimer - dodecamer model.

As Fig. 2 shows, the peak heights simulated on the basis of the stepwise aggregation model are in an excellent agreement with the observed ones, but those simulated on the basis of the trimer - dodecamer model are not so. Figure 3 depicts that the peak volume data simulated on the basis of the stepwise aggregation model are in an excellent agreement with the observed ones, but those simulated on the basis of the trimer -dodecamer model are far from the observed ones. This is a very useful plot to show dimerization.

In conclusion, our stepwise aggregation model for the self-association of PCV is better than the trimer - dodecamer model.

REFERENCES

I.N. Funasaki, S. Hada, and S. Neya, Chem. Pharm. Bull., 42 (1994) 779. 2. L. M. Varela, C. Rega, M. J. Suarez-Filloy, J. M. Ruso, G. Prieto, D. Attwood, F. Sarmiento, and V. Mosquera, Langmuir, 15 (1999) 6285. 3. N. Funasaki, Adv. Colloid Interface Sci., 43 (1993) 87.

Studies in Surface Science and Catalysis 132 Y. Ivvasawa, N. Oyama and H. Kunieda (Editors) 'o 2001 Elsevier Science B.V. All rights reserved.

217

Two-dimensional Clusters of Magnetic Fine Particles at the Surface of Magnetic Colloidal Suspension

N. Tanaka, S. Doi and I. Takahashi

School of Science^ Advanced Research Center of Science, Kwansei Gakuin University (ARCS-KGU), 2-1 Gakuen, Sanda 669-1337, Japan

We report the in-plane structure of clusters of magnetic nanoparticles condensed at the free surface of a magnetic fluid (ferrofluid). Strong diffuse X-ray reflectivity which could not be explained without assuming the clusters was observed. Fractal dimension of the cluster and the surface tension of the magnetic fluid were evaluated. The fractal dimension strongly suggests a chain-like cluster of the colloidal spheres lying beneath the surface which is fluctuating due to capillary waves. The average direction of the magnetic dipole moment is also suggested to be parallel to the surface. They are consistent with the stability of such a surface-induced clusters observed by specular X-ray reflectivity at various temperatures and under AC magnetic fields.

Specular RetJecitvitv

'1:,

pifjuse Rcfiecfiviir

Sample Surface

l.INTRODUCnON The aggregation process of particles has

attracted a great deal of interest in recent years. Not a few systems are known to have strong tendency to spontaneously form highly concentrated regions or clusters. From X-ray reflectivity (XR), we found a stable, concentrated layer of the magnetic nanoparticles at the free surface of a magnetic fluid (the coverage of the particles was estimated as about 70-80% [1]). Although the magnetic fluid can be regarded simply as an ensemble of disordered magnetic dipole moments, the mechanism of such an aggregation phenomenon has not yet been understood. In this paper, we report diffuse XR from the surface of the magnetic fluid. After having the structure, we discuss the magnetic structure of the aggregates.

7> Fig.l Schematic drawing of X-ray surface

scattering. Notations are explained in the text.

Dotted curve represents a transverse scan.

218

2. EXPERIMENTAL In order to measure the XR

from the free surface of liquids, several difficulties should be overcome. At least, the sample surface must be kept horizontal, and the vibration from the rotating anode X-ray generator should be sufficiently insulated. The details on our diffractometer and experimental setup were described elsewhere [1,2]. The sample was a conunercially available magnetic fluid (LS-40, Taiho Industry Inc.). The superparamagnetic nanoparticles, about 100 A in diameter, were dispersed in alkyl naphthalene. Due to appropriate surfactant molecules which coat the particles, the bulk colloid remains stable until the volumetric packing fraction of

5

LOOE'28

LOOE-31

-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.6

a -sin '^ (q A /4 Ji) (deg,)

Fig.2 Transverse scans for different q values. For clarity all

the curves are displaced by three order of magnitude.

-TV

•^ so

5

\ D

' A" T|-l= -0.1084\

\

]

q = 0.1138

^^at—. j

B

\ %

I.OE-06 LOB-^ l.OE-04 l.OE-03 l.OB-02

qs(A-')

Fig.3 Typical Log-Log plot of diffuse XR after a correction of the

illuminating area of the incident X-ray beam. The slope of the

diffuse scattering at ^=0.1138[A"^] was fitted to be-0 .1084 (the

solid line). The calculated XR with Y=0.026[N/m] is also plotted.

the nanoparticles exceeds O.I.

3. RESULTS Figure 1 shows the scattering

geometry of specular and diffuse XR; whole the measurements were carried out by a transverse scan, namely the sum of the incident and outgoing angles (denoted as a and x in the figure) remains constant throughout the scan. Figure 2 indicates the transverse scans for different q (= \ q \ ) values (the momentum transfer q is defined as ko-ki, where ki, ko are the incident and scattered wavevectors of X-ray (\ki\ = \ko\ =1/X)) .hi the

219

middle, a narrow specular XR is seen. The broad diffuse XR component is also visible for all the scans.

0.00

•0.04 Y

I -0.08 h

'0.12

•0.16

4. ANALYSIS

Figure 3 is a Log-Log plot of the typical scattered intensity for fixed ^=0.1138[A-^]. A power-law region is clearly seen, leading us to free capillary waves are present on the surface [3]. However, the surface fluctuations of the magnetic fluid could not be explained only by the capillary waves of uniform liquids: Neither intensity nor slope of the diffuse XR for kerosene

agreed with observation, although the surface tension of kerosene was thought to be ahnost the same as that of the magnetic fluid (notice the theoretical diffuse XR from kerosene plotted in Fig. 3). Therefore, an alternative stmctural model is necessary to understand the diffuse XR from the magnetic fluid.

To analyze the diffuse XR, we consider fractal clusters of colloidal particles lying beneath the surface which is fluctuating due to capillary waves. When the amplitude of the capillary waves is much smaller than the radius of the nanoparticles rnp, the electron density may be written as

0.00 0.02 0.04 0.06 0.08 0.10

Fig.4 Ti-1 vs. q^. The line is a fit to the data.

^{x, y, z) ^ H[z - h(x, y)] { Apo g(x, y) + p(alkyl naphthalene) }, (1)

where H is the step function ( / / ( z ) = 0 for z<0, and / / ( z ) = 1 for z> l ) , h represents the height of the surface at {x, y), Apo is the density contrast (= p(nanoparticle) - p(alkyl naphthalene) ), and g represents a function on occupancy of the nanoparticles ( g = 0 if there is not the nanoparticlc at (x; y, h(x, y) + rnp), and g = 1 if there is the nanoparticle at (x; y, h(Xy y) + r„p) ). For fractal aggregation with fractal dimension D, the pair correlation ftinction is given by

< g(0, 0) g(x, y) >^( x'^y') (^-^, (2)

where d is the space dimension set to be 2 in the present study. Since the instrumental resolution in the direction qy out of the scattering plane was rather loose, qy must be integrated over. Thus, the diffuse scattering 7 , g ) can be given by

« q ) - / J J P( O P{ r' ) exp{ iq'(r-r')}6r dr'd^y. (3)

Inserting Eq. (1) into Eq. (3), and assuming that there is no correlation betweenH{ r)H(r') and

g( ^ )g(''') statistically, we obtain

^diffi q )-q: .^nJ^-' (4)

220

Ti= d-D + ^ 9 ^ (5)

where ks, T, y are the Boltzmann constant, temperature[K] and surface tension coefficient, respectively. The third term of Eq. (5) represents the contribution of capillary waves [3]. Figure 4 represents the fitted T as a function ofq^'. The r\ revealed to be well reproduced by Eq. (5); the solid line in Fig. 4 is the fitted line. From the slope and the intercept at the ordinate, D and y were readily evaluated as 1.13 and 0.05 [N/m], respectively.

5. DISCUSSION In the present study we have obtained the reasonable value of surface tension (y at IS'C =

0.073(water), 0.026(kerosene) [N/m]). It might be one reason that our structural model is a plausible one.

The fractal dimension of diffusion limited aggregation (DLA) in two dimensional system is known to be D^l.6. Since obtained fractal dimension of D^l . l for the magnetic fluid is rather small compared to 1.6, the surface induced clusters must not originate solely from the accepted processes, like DLA. Furthermore, fractal clusters with D^ l . l can be regarded roughly as one-dimensional rather than two-dimensional. Therefore, the clusters would not have a net-like or two-dimensional crystalline (or porous) stmcture. Consequently, we can conclude that a chain-like structure with few branches is the most plausible. In Ref.[l], two types of arrangements of dipole moments, i.e. surface normal and surface parallel arrangements in which the average direction of the dipole moment is surface noraial and surface parallel respectively, were pointed out, although we could not decide which arrangement was reasonable (Fig.6 in [1]). The chain-like in-plane particle structure obtained in the present work clearly allows us to support the surface parallel arrangement.

By using X-ray diffraction, the in-plane particle structure of the clusters of the magnetic fine particles was discussed on the standpoint of fractal aggregation. Computer simulation studies of which results can be compared with those of the present study are required so as to understand such a unique aggregation at the free surface of magnetic fluids.

Acknowledgments We are greatly indebted to Prof. H. Terauchi and Prof. J. Harada for fruitful discussions and

criticisms. Part of this study was supported by a special research grant from Kwansei Gakuin University, and by grant-in-aid for scientific research 09740252,11874055 from the Ministry of Education, Science and Culture in Japan.

REFERENCES [1] I. Takahashi, K. Ueda, Y. Tsukahara, A. Ichimiya, J. Harada, J. Phys.: Condens. Matter 10 (1998) 4489-4497. [2] S. Doi and I. Takahashi, Philos. Mag. A 80 (2000) 1889-1899. [3] M. K. Sanyal, S. K. Sinha, K. G. Huang, B. M. Ocko, Phys. Rev. Lett. 66 (1991) 628-631.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) 221 c; 2001 Elsevier Science B.V. All rights reserved.

Colloid-chemical properties of chitosan

Bratskaya S.Yu., Shamov M.V., Avramenko V.A., Chervonetskiy D.V.

Institute of Chemistry, Far East Department of the Russian Academy of Sciences, 159, Prosp. 100-letya Vladivostoka, Vladivostok 690022, Russia

Flocculation properties of chitosan in solutions of humic substances were investigated at various pH and ratios chitosan to humic acid. Effect of additions of Fe " on the flocculation effectiveness was also studied. Experimental results on chitosan interaction with carbonic acids of low molecular weight have shown that hydrophobic internal domains of chitosan helices along with chitosan amine groups play an important role in chitosan interaction with organic substances.

INTRODUCTION

In the recent years, chitosan has been attracting the researches interest as a very promising material with widely ranging applications in water treatment, food and pharmaceutical industries, cosmetology, and biotechnology. Chitosan is a basic polymer of helix structure having reactive amine groups that gives a lot of possibilities of modification and ionic interactions. Combination of chitosan reactive amine groups and helix structure with internal hydrophobic domains determines such colloid-chemical properties of chitosan as organic substances sorption, flocculation, and stabilization of colloid systems that are of great interest both in theoretical terms and in industrial application possibilities. In this paper we discuss use of chitosan for humic substances flocculation. Some results on carbonic acids sorption on chitosan are also presented to illustrate the mechanism of chitosan interaction with organic substances.

EXPERIMENTAL

Chitosan flakes (dacetylation degree=75%) obtained from crab shells were used in all experiments. Humic acids were isolated from peat [1], amide of humic acids was obtained by heating humic acid solution with NH3 at 140 ° C and high pressure for 5h.

For testing chitosan flocculation properties, 1 g of chitosan flakes were dissolved in 100 ml of O.IN HCl solution. Fixed amount of prepared chitosan solution was added to solutions of humic substances with concentration of 50-500 g/L and pH was adjusted to 7. After 24h solutions were filtered and optical density was measured at 413nm. Effectiveness of flocculation was calculated as a percentage of color removal.

Interaction of chitosan with carbonic acids was studied as follows: 40mg of dry chitosan flakes and 20 ml of acidic solutions with concentrations from 1 up to 100 ^mol/L were shaken for 24h at controlled temperature 20''C. When equilibrium was established, the equilibrium acid concentration was determined by potentiometric titration with "Radelkis 0P211/1" pH-meter and "ORION 96-21" combination pH-electrode as described in [2].

222

The same method of titration was used to determine pK-distribution in humic acids and their amide.

RESULTS AND DISCUSSION

Flocculation of humic substances is of great interest for decontamination of natural and waste water from humic substances and metals bound to them as well as for recovery of valuable metals from technological solutions. Conventional method of humic acids removal from solutions is flocculation at pH=l-2 but it does not allow obtaining color removal higher than 80-90%. Our experimental results have shown that flocculation effectiveness of 96-100% can be achieved at pH=6.5-7.5 when chitosan is used as a flocculant in humic acid solutions -Fig.l. This figure also illustrates that too high content of chitosan is resulted in the flocculation effectiveness reduction because of the stabilization of humic colloids by chitosan. Effect of chitosan concentration on effectiveness of humic acids and their amides flocculation is shovm in Fig.2.

Fig. 1. Effectiveness of humic acids flocculation at ratios chitosan to humic: • -1:5, •-1:2, A- 1:1, T .2 :1 , • -4 :1

30 40 50 60

Cchitosan'lO^.^

Fig.2. Effect of chitosan concentration on flocculation of humic acids: •-lOOmg/1,. •-500mg/l; amide of humic acids: A-lOOmg/1, T-500mg/l.

It is obvious that chitosan concentration required for effective flocculation depends on humic acid concentration in solution and the nature of their functional groups. As a result of modification, amide of humic acids does not contain functional groups with pK less than 6.8 while original humic acids have carboxylic groups with average pK equal to 4 and 5.5 - Fig.3. Thus, amide of humic acids has a lower charge density in comparison vsdth original humic acids, and, therefore, less amount of chitosan is required to satisfy "cationic demand" of negatively charged colloid particles of humic acid amide. Fig.4 illustrates flocculation of humic acids by chitosan when Fe3+ is added. It is seen that in this case effective flocculation was obtained at significantly lower concentrations of chitosan. It should also be mentioned that even very high concentrations of Fe3-»- used as a coagulant without chitosan addition do not provide effective flocculation of humic acids. Thus, charge neutralization is not the only reason of chitosan effective work as a flocculant.

223

0.020

0.015 i

0.010

0.005 i

0.000

I- humic acids

humic acids amide

Fig.3. pK-distribution of functional groups of humic acids and their amide

100

0 1 2 3 4 5 6

Cchitosan-103.%

Fig.4. Effect of Fe^* on humic acids (HA) and their amide (AHA) flocculation:

• -HA(100mg/l)+ Fe^*(50mg/1), • - H A ( 5 0 0 m g / l ) +Fe^*n00mg/1), A- AHA(100mg/l)+ Fe^*(50mg/I),

W e assume that chi tosan internal hydrophobic domains aside from its amine groups play an important role in interaction o f chitosan with organic substances including humic acids. T o confirm this assumption w e have studied interaction o f chitosan with homologous series o f carbonic acids - acet ic , propionic , butyric, n-valeric and isovaleric acids in aqueous and water-ethanol solut ions . Strong correlation w a s found b e t w e e n m a x i m u m sorption o f acid on chitosan and its p K va lue in aqueous solut ion with the only except ion - isovaleric acid -Fig.5.

2.2

2.0

r 1.2

1.0

aceticacid

J • 1 •

isQK/afericacid butyric a d d ^ \ ^

vatericadd \ ^

propionic add

2.2

2.0

I 1.8

c 1.6 o

I 1.4 CO

1.2

1.0

\ acetic aa

]

propionic add

V"" * \ \ butyric add

\ \ /

\ / \ / propionic add

y^valericadd

butyric add I

isovaleric add •

\ \ \

valeric add

4.74 4.76 4.78 4.80 4.82 4.84 4.86 4.8 i 2 3 4 5 6

pK the number of carbon atonns Fig. 5. Correlation between maximum Fig.6. Correlation between hydrocarbon radical sorption on chitosan and pK of carbonic acids: length of carbonic acids and their maximum

sorption on chitosan: • - in aqueous solution • - in water/edianol (2:1) solution

224

Taking into account that this acid has a branched hydrocarbon radical we suggested that structure and length of the radical can also effect carbonic acids sorption on chitosan. This effect should be more explicit in solvent of polarity less than polarity of water, where hydrophobic interactions play more important role, for instance, in water/ethanol solutions. Our experimental results have shown that there is no strong correlation between pK and maximum sorption of carbonic acids on chitosan in water/ethanol solutions. But if we plot maximum sorption versus hydrocarbon radical length, we observe increase of sorption with increase of radical length (Fig.6) that was not observed for aqueous solutions. In water/ethanol solutions internal hydrophobic domains of chitosan helices become more accessible for interaction with hydrocarbon radical of carbonic acids that can result in acid trapping inside chitosan helices. It is obvious that hydrophobic interaction between chitosan and carbonic acids will be stronger with increase of the acid hydrocarbon radical length. In our experiments valeric acid has shown the highest values of sorption in water/ethanol solutions. Nevertheless, isovaleric acid with the same number of carbon atoms has the smallest value of sorption. This effect can be explained by space hindrances during penetration of a branched hydrocarbon radical inside chitosan helices.

CONCLUSION

We conclude that chitosan can be used as a very effective and biodegradable [3] flocculant of humic substances. Amount of chitosan required for effective flocculation depends on humic acid origin and concentration as well as on content of metal ions, especially Fe"'"*". Chitosan shows the best flocculation properties in pH range from 6.5 up to 7.5, where the highest values (up to 100%) of color removal were obtained.

Summarizing the results obtained on sorption of low molecular weight carbonic acids on chitosan, we can conclude that the structure of hydrocarbon radical of the acid effect its sorption on chitosan due to possibility of hydrophobic interaction with internal domains of macromolecular helices. This effect becomes stronger with increase of the hydrocarbon radical length. Thus, we can suggest that for high molecular weight humic acids containing fragments with aromatic and long aliphatic radicals [4] hydrophobic interaction with chitosan should play a very important role. Most likely this type of interaction, aside from ionic interaction with chitosan amine groups, determines high effectiveness of chitosan as a flocculant in humic acids solutions.

REFERENCES

1. Kemdorff H., Schnitzer M., Geochim. et Cosmochim. Acta, Xe 11 (1980) 1701. 2. Bratskaya S.Yu., Golikov A.P., J.Anal.Chem., XeS (1998) 234. 3. Boryniec S., Ratajska M., Fibers and Text. East.Eur. Xo4 (1995) 60. 4. Orlov D.S. Soil Chemistry, Moscow State University, Moscow, 1992.

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) 225 (c 2001 Elsevier Science B.V. All rights reserved.

Science and art of fine particles

Egon Matijevic

Center for Advanced Materials Processing, Clarkson University, Potsdam, New York 13699-5814, USA

Without science we should have no notion of equality, without art no notion of liberty. W.HAuden

INTRODUCTION This presentation deals with an area of materials science, which is causing a great deal of

excitement for academic and practical reasons, i.e. with the so called monodispersed colloids. While the subject is an old one, dating back to Faraday's gold sols described in 1857, it has become a topic of wide-spread interest only relatively recently.

There are some intriguing aspects of this field of science and technology which justify some comments. First, it is obvious fi*om the title, that the size of the "fine particles" to be discussed is missing. Presently we have two groups of scientists interested in fmely dispersed matter, some dealing in the nanometer and the other in the micrometer range. In many cases these systems are treated as separate worlds, yet it will be shown that they are closely connected, especially when these dispersions are prepared by chemical reactions in solutions.

Another interesting aspect is that the uniformity of particles has seldom been encountered in natural environments. Indeed, there are only a few examples of indigenous monodispersed systems, opal being probably the best known. Yet scientists have produced a large number of such solids over a broad range of modal sizes, of simple or mixed (internally or externally) compositions, of different structures, and in a variety of shapes. The reasons for the discrepancy between the natural occurrences and research achievements will be addressed in the presentation.

It will also be shown that, despite this progress, many questions remain to be resolved, which represent major challenges to the workers in the field. One of the problems is the actual mechanism (or mechanisms) by which uniform particles are formed. A recent development on the subject will be described. Another aspect is the predictability of shapes and structures of monodispersed colloids, to which there is still no answer.

Finally, the title of the lecture requires further explanation. The word "art" has two meanings

226

in the English language: it is used to describe skilled craftsmanship or something of beauty. In dealing with well defined particles there is room for both of these interpretations.

EXAMPLES OF UNIFORM PARTICLES There are many different techniques, which may yield particles of uniform size and shape,

but none can be used to produce every kind of dispersions of desired properties. Due to the simplicity, versatility, and practicality, the precipitation fi-om solutions is the method of choice. In the past this procedure yielded a number of monodispersed systems, although mostly by serendipity. About a quarter of a century ago systematic studies were initiated, which resulted in a multitude of well defined dispersions as reviewed in several articles (1-3). In most instances the processes have involved either mixing reactants or decomposing complexes, normally under mild temperatures and in moderate concentrations. A few examples of imiform particles of simple chemical compounds are illustrated by the electron micrographs in Figure 1.

Using the same technique, it is also possible to precipitate composite particulates. The latter can be homogeneous of exact stoichiometry, as exemplified by pure or doped barium titanates. To achieve these conditions rapid mixing is required, such as by using the controlled double jet precipitation process. In contrast, slow precipitation results, as a rule, in internal inhomogeneity, i.e., the composition changes fi'om the center to the periphery, although the particles may still be perfectly spherical, as observed with mixed alumina/silica, or copper/lanthanum, or copper/yttrium oxides.

Another major area of interest are coated particles. Again, it is possible to produce uniform surface layers of varying thickness on inorganic cores with either inorganic or organic coatings or organic cores with inorganic coatings. It is interesting that the shell of the same material can be deposited on different cores. For example, yttrium basic carbonate coatings were produced on silica, hematite, and latex!

It is essential to note that conditions needed to obtain a given material as a uniform dispersion are sensitive to a great degree to the experimental parameters, such as the temperature, concentration of reactants, pH, ionic strength, solvent composition, etc. In some cases even a small change in these conditions can not only affect the particle uniformity, but it may yield solids of different chemical composition, structure, or morphology. This sensitivity explains the paucity of naturally occurring monodispersed particles. It should also be noted that the laboratory preparations require minutes or hours, while the processes in nature extend over geological times.

A final comment in this section refers to scaling up the production of well defined powders. Interestingly, the two materials first obtained in large quantities are the polymer latex and silica. The first is not a subject of this presentation, while silica will be exemplified in connection with some applications

227

One important aspect of scaling up precipitation processes, which yield uniform particles, is the necessity that any engineering design must consider the optimum conditions established in small batches. Much success has been achieved by using a plug-flow type of a reactor for continuous precipitation of a variety of uniform colloid dispersions, such as yttria, silica, aluminum hydroxide, and barium titanate (4). The schematic presentation of this equipment is given in Figure 2.

r

0.5/cm 0.5/cm

2\ITX\ S^m

Figure 1. Transmission electron micrographs (TEM) of (a) zinc sulfide; (b) manganese(II) phosphate, (c) and (d) hematite (Fe203) particles.

228

RESERVOIR 1 RESERVOIR

2

PERISTALTICI PUMP

TEMP com.

SCR U

' PRODUCT

Figure 2. Schematic presentation of the apparatus for continuous flow precipitation.

MECHANISM OF THE FORMATION OF MONODISPERSED PARTICLES As one would expect, the understanding of the mechanism of the formation of

monodispersed colloids by precipitation has been of major concern to workers in the field. For a long time the concept developed by LaMer was generally accepted; i.e., such dispersions should be generated, if a short lived burst of nuclei in a supersaturated solution is followed by controlled diffusion of constituent solutes onto these nuclei, resulting in the final uniform particles. This mechanism is indeed operational in some, albeit limited cases, and more often only at the initial stage of the precipitation process.

For a long time the writer of this article has been puzzled by some experimental observations, one of which is illustrated by the electron micrograph of zinc sulfide in Figure la. These perfectly spherical particles, obtained by precipitation in ionic solutions, exhibit X-ray characteristics of a known mineral (sphalerite). Obviously, it is not easily understood why would such homogeneously precipitated perfect spheres have crystalline characteristics. Importantly, low angle X-ray measurements showed these particles to be made up of essentiality identical nanosized subunits. Electron microscopy and other methods of evaluation demonstrated on numerous other dispersions, that particles of different morphologies and chemical compositions clearly exhibited particulate substructures.

229

Based on the illustrated samples and many others, it is now firmly established that the prevailing mechanism in the formation of monodispersed particles proceeds in several stages: (7) nucleation, (2) growth to nanosize particles, and (3) aggregation of nanosize particles to uniform fmal colloids. Extensive studies have indeed documented the existence of the aggregation stage (5). For example, the electron micrographs, X-ray analysis, and independently prepared precursor

particles all yielded the average diameter of crystallite subunits of 35^ nm of spherical gold particles displayed in Figure 3(6).

Figure 3. Colloidal gold particles

It was then necessary to derive a mechanism, which would account for the aggregation of a huge number of nanosized particles into identical larger colloids. Recently a kinetic model has been developed that explains this size selection (7). The latter is based on the assumption that the nucleation is followed by a rapid formation of singlets, i.e. primary (nanosized) particles, which are sufficiently sparsely populated, and once formed are not further generated. Thus, their concentration decays by aggregation, when the conditions in the system eliminate repulsion between them. The latter can be due either to an increase in the ionic strength or to a change in the pH in course of the process. It is also assumed that the diffusion constant of singlets is larger than that of aggregates. The dominance of the irreversible singlet capture in the growth process can, under certain set of conditions, result in the size selection (i.e. uniformity) of the fmal particles.

230

60

oc

"O 40

[

f=0.1 sec

1 sec

I a = 0.57 HJm

10sec

0.1 0.2 0.3

Figure 4. Distribution of secondary particles by their sizes at 0.1, 1, and 10 sec, calculated for the precipitation of spherical gold particles, using the model described in (7).

The calculated size distributions of the secondary (final) spheres for three reaction times, using the parameters for the precipitation of the displayed gold sol, are given in Figure 4. The model describes reasonably well the experimental observations, considering the simplification used in the calculations.

An interesting consequence of the newly established mechanism, is the fact that precipitation in solutions yields, as a rule, nanosized particles. If the process is arrested at this stage by additives, such as surfactants or microemulsions, one obtains stable nanosystems. However, in the absence of stabilizers, these particles aggregate (rather than to grow) into larger final products, which under appropriate conditions consist of monodispersed colloids. Thus, here we have found the *'bridge" between the two ''worlds" mentioned earlier!

NANOSIZED PARTICLES The understanding of the mechanism of the formation of colloids throws new light on the

possible preparation procedures of nanosized particles. Normally, wet routes involve precipitation in the presence of large amounts of surfactants (e.g. microemulsions, vesicles, etc), which affect the nucleation stage, but also stabilize the resulting finely dispersed matter. These processes make it difficult to separate solids from the additives.

Obviously it is of interest to produce monodispersions in quantities with a minimum amount of stabilizers. One such process is based on the recognition that larger particles are aggregates of much smaller precursors. Should it be possible to peptize such colloids, one would have a new avenue of approach in the generation of nanoparticles. This method was indeed proven possible in some cases. Thus, monodispersed colloidal indium hydroxide was prepared in ethylene glycol. On

231

addition of water the original organic solvent was leached out and the precipitated solids fell apart into constituent subunits of nanometers in size.

Another useful technique proves to be the controlled double jet precipitation (CDJP), which can yield in larger quantities nanparticles in the presence of moderate amounts of surfactants, as demonstrated on a variety of systems, including ZnO, PbS, BaTiOa, etc.

"ART" AND SCIENCE The new developments in the understanding of the mechanisms of formation of

monodispersed colloids have greatly advanced the scientific aspects of this area of materials. Yet in actual preparations the art, i.e. skills, still play an essential role, especially since we do not know how to predict and control some properties, such as the shape or even the composition of the resulting particles.

The finely dispersed matter offers much in terms of the other aspect of art, i.e. the beauty. The latter can be affected by shape or color or both. Examples of such artistic impressions will be offered using electron micrographs of monodispersed particles and their surfaces. Even more importantly, pigments, marbles, metals, etc. are made of fine particles, without which we would have no paintings, sculptures, and other works of art, which so much embellish our lives.

REFERENCES

1. E. Matijevic: Formation of Monodisperse Inorganic Particulates. In Controlled Particle, Droplet and Bubble Formation (D.J. Wedlock, Ed.), Butterworth-Heinemann, London, 1994, pp. 39-59.

2. E. Matijevic: Uniform Colloid Dispersions - Achievements and Challenges. Langmuir, 10, 8-16(1994).

3. E. Matijevic: Preparation and Properties of Uniform Size Colloids. Chem. Mater., 5,412-426(1993).

4. Y.-S. Her, S.-H. Lee, and E. Matijevic: Continuous Precipitation of Monodispersed Colloidal Particles. II. SiOz, A1(0H)3, and BaTiOs- J. Mater. Res., 11, 156-161 (1996).

5. S.H. Lee, Y.-S. Her and E. Matijevic: Preparation and Growth Mechanism of Uniform Colloidal Copper Compounds by the Controlled Double-Jet Precipitation. J. Colloid Interface Sci., 186, 193-202 (1997).

6. D.V. Goia and E. Matijevic:Colloids Surf, 146, 139-152 (1999). 7. V. Privman, D.V. Goia, J. Park and E. Matijevic: Mechanism of Formation of

Monodispersed Colloids by Aggregation of Nanosize Precursors. J. Colloid Interface Sci., 213,36-45(1999).

Studies in Surface Science and Catalysis 132 Y. Iwasawa, N. Oyama and H. Kunieda (Editors) 233 c 2001 Elsevier Science B.V. All rights reserved.

Hydrothermal synthesis of nano-size ZrOi powder, its characterization and colloidal processing

0. Vasylkiv and Y. Sakka National Research Institute for Metals, 1-2-1, Sengen, Tsukuba, Ibaraki Pref., 305-0047, Japan

Nano-size 3Y-TZP powder with controlled secondary particles size was synthesized by hydrolytic coprecipitation. Hydrous-zirconia gel produced by urea precipitation results in a nano-powder with a primary particles size of 5 - 8 nm, and secondary aggregates size of 45 - 65 nm. After calcination powders exhibit a single tetragonal phase. Determination of the optimal formation parameters, post-synthesis treatment for fabrication of agglomerate-free zirconia powder with finest primary crystallites uniformly packed into the secondary aggregates, as far as determination of the best suspensions properties allowed preparation of the uniformly densified green body with high packing density of 58 % after slip casting and CIP.

1. INTRODUCTION

Fine ceramic particles, uniformly agglomerated are generally desirable for producing dense ceramic, because of the close packmg and uniform densification [1-12]. Powders produced by wet chemical methods are usually polydispersed and consist of: primary crystallites, aggregates of primary crystallites, which are formed during reaction time, and powders agglomerates [2-11]. The crystallites size is dependent of the nature of powder processing i.e. technique or methodology of synthesis. The rate of crystallites growth is negligible low up to the temperature of 900 °C [5].

Tetragonal zirconia polycrystal (TZP) based ceramics have attracted special attention because of its excellent mechanical properties and attractive possibility of obtaining the nano-grained ceramic with controllable microstructure [2-10]. Nano-grained 3Y-TZP ceramics are expected to show the excellent mechanical properties. Nano-scale zirconia-based ceramic is a good start for post-sintering heat treatment, the grains in the final microstructure can be grown to any desired size, provided the grain growth can be controlled [6].

The purpose of our study is to obtain a finest possible powder with narrowest particle size distribution in order to prepare the uniform slurry and slip cast for obtaining the uniform green microstructure, witfi high reactivity during sintering.

2. EXPERIMENTAL

ZrOCl2*8H20, Y2O3, urea (High Purity Chemicals, Japan), and HCl (Kosochemical, Japan), were used for this investigation. The mixed sols of composition Zr02 + 3 mol% Y2O3 at different cations concentration with urea, were hydrothermally treated at 155 °C. The urea decomposed into NH3 and CO2 through reaction with H2O and pH changed to 9. The homogeneous precipitate formed was hydrous Zr02 with Y2O3 and it is crystallized under hydrotiiermal conditions.

Several techniques were used for powder washing and drying in order to minimize the agglomeration of the powder [5]. Powders were washed with deionized water to remove CI"

234

ions and ammonia, and with ethanol. Subsequently ethanol was evaporated (T= 65 °C). After pH overreached 8 the micro tip ultrasonication for 10 min, using 20 kHz and 160 W (Shimazu, USP-600) was used to destroy the powder agglomerates in suspension [5, 11].

Phase identification was determined from X-ray diffractometry data (JEOL JDX-3500). The primary crystallite size was determined through an X-ray diffraction line-broadening method, from surface area data (Coulter SA 3100) and from TEM observation (JEM-100-CX, Japan) data. The aggregates and agglomerates size have been analyzed using a Laser Particle Size Analyzer (LSPZ-lOO, Otsuka Electronics).

Aqueous suspensions were prepared from the powders with different morphology and from the same powder by changing the amount of additional dispersant (anmionium polycarboxylate, Toaghosei Co., ALON A-6114). The suspensions were ultrasonically dispersed for 10 min and stirred under same conditions of mixing with a magnetic stirrer [11, 12]. The rheological behavior of the suspensions was studied witfi a viscometer (Toki-Sangyo Co., RE500L). Consolidation of the suspensions by slip casting and subsequent CIP at 400 MPa were applied. The densities of the green bodies were measured by the Archimedes method using kerosene. Relative density was based on a 6.02 g/cm .

3. RESULTS AND DISCUSSION

3.1. Powder preparation To prepare powders with homogeneous composition and uniform morphology metal-

chlorides and urea-contained sol tiiermal hydrolysis have been used. Each of the produced particles of hydrous zirconia powder was an aggregate of many small primary particles with calculated diameter of 5 - 8 nm. The initial solution concentration influences the powders properties. With an increasing solution concentration, the size of the resulting secondary aggregates and degree of agglomeration became larger. The powder's surface area decreases only from 96 to 87 m /g for calcination at 600 " C, or from 53 to 47 m /g for calcinations at

800 °C (for 1 h holding) if the concentration of the stock solution increases from 0.1 to 0.5 mol/1. At the same time the size of the primary particles obtained from TEM observation was nearly the same for all concentrations at the same conditions of calcination. We clarified the decreasing of the surface area to the primary particles bonding and appearance and thickening of the solid bridges between the crystallites packed into the agglomerated secondary aggregates. The size and morphology of the aggregates of primary crystallites is strongly depends on the solution concentration, temperature-time conditions of hydrolysis, conditions of washing and drying, and time-temperature conditions of calcination.

Ehiring hydrolytic precipitation, dense aggregates of hydrous zirconia primary particles forms when the capillary forces overcome the interparticle repulsive forces and bring as formed oxide particles closely to each other. The fine particles are bound together into the aggregates by

van der Waals forces and such bonds forms necks between the particles, the initial reagents react into the solution and then precipitate at the torroidal region between the particles. Figure 1 shows the TEM photograph of the 3Y-TZP aggregates synthesized from urea-chlorides aqueous solution with initial concentration of 0.2 mol/1 and calcinated at 800 °C for 1 h.

uw

Fig. 1. TEM photograph of hydro-thermally derived and calcinated 3YTZP powder.

235

No tendency appears for bridging between neighboring aggregates for the powder washed with ethanol. However, after drying of washed powders the surface of fine powder can absorb up to ~ 3 wt.% of water. Such water amount is enough for particles bounding. In addition, the particles can be densely packed during initial stage of calcination, and solid necks between primary crystallites and inter-aggregate bridges forms during subsequent calcinations.

Stabilizing of oxide powders against coagulation in an aqueous solution required a low initial solution concentration (0.05 - 0.2 mol/1) and pH value - several units above isoelectric point (lEP). Micro-tip ultrasonic treatment for breaking up the appeared agglomerates has been used at the final stages of washing [5] and after calcination. The average size of secondary aggregates could be reduced close to the value of 45-65 nm. The bonds between the crystallites into the aggregates with such average size has to be rather strong because ultrasonication was unable to break up them and fiirther reduction in size was found to be impossible. It is clear that the powders produced with different wet chemical techniques are in fact agglomerates of a small, strongly bond aggregates of the crystallites [2-9]. Zirconia nano-powder (Fig. 1) with primary crystallites size of 5 - 10 nm, with average secondary particles size of 50 nm, and without tertiary agglomerates has been prepared. The optimum inintial sol's concentration was found to be 0.05 - 0.2 mol/l.

3.2. Colloidal processing Aqueous suspensions were prepared by changing the solid content and the amount of

additional dispersant. The suspensions witii the minimum viscosity i.e. best flowability [10-12] were prepared fi-om the powders with different degree of aggregation-agglomeration. The

appropriate solid content of the suspension was 25 vol% for the finest agglomerate-fi-ee 3Y-TZP powder. For coarser, agglomerated powder the solid content of 30 vol% was applied. Figure 2 shows the changes of the suspensions viscosities (at the share rate of 100 s") over the amount of dispersant for suspensions with different average secondary particle size and degree of agglomeration. The viscosity of the suspensions could be reduced by the addition of water. However, subsequent slip casting of the suspensions with solid content lower than 25 vol% leads to non-uniform densification of green body and cracking. The amounts of dispersant 4.5 wt% for 50 nm powder, 3.5 wt% for 110 nm powder, and 2.5 wt% for coarser powder were found to be appropriate for preparing well-dispersed suspensions. Suspension's stability as far as good dispersion is the main trend of the uniform green microstructure. The high viscosity of suspensions implies the rapid flocculation of the particles in non-

stirred suspension. Viscosity of the suspension prepared with 25 vol% of solid and 4.5 wt% of dispersant changed ft"om 8 mPas to 12 mPas during first 2 hours (without stirring) and after 7 hours reached the value of only 27. Such time is enough for preparing of uniform green body. At the lower or higher amount of dispersant the suspension became to stiff for slip casting immediately or during the first 2 hours. 4.5 wt% of dispersant and 25 vol% of solid content have been chosen as the best conditions for preparation of the uniformly densified green body.

Schematic of influence of agglomeration on the densification and green microstructure is depicted in Fig. 3. The agglomerates in zirconia powder prepared from the sols with concentration of more than 0.2 mol/1, water-washed, dehydrated with ethanol, dried and

1 2 3 4 5 Amount of dispersant, wt%

Fig. 2. Suspensions viscosity versus the amount of dispersant.

236

concentration of more than 0.2 mol/1, water-washed, dehydrated with ethanol, dried and calcinated cannot be completely breaking up during subsequent processing. It can be indicated (Fig. 4) from lower green density of as-slip casted compacts (Relative density D = 35 - 43 % in comparison with density of compacts from the powder with narrow size distribution of secondary particles (D= 45 - 52 %). Non-uniform packing from agglomerates lead to localized, inhomogeneous densification during subsequent sintering. The highest green density was obtained for the sample prepared from the powder with average aggregate size of 110 nm. The green densities for tiie finest uniformly aggregated powder after slip casting and slip casting with CIP were 46 % and 52 % respectively.

Unifomily aggregated primary partictes

uniform green microstructure Q\p

from sip casting best packing densrty

high packing density The size range of irtraaggregale and interaggregate pores is same

V 80 -o « 60 -

Circuits (agglomerates) of the secondary aggregates

- slip casting - slip casting + CIP • sintering at 1300.2 h

nonuniform microstnicture low packing density

CIP poor packing density

Large interagg^on)erate pores

Fig. 3. Schematic of influence of agglomeration on the densification and green microstnicture.

IV. CONCLUSION

0 100 200 300 400 500 600 :

Mean aggregate (agglomerate) size, nm

Fig. 4. Effect of aggregate (agglomerate) size on compacts density.

The mam potential of nano-powders - increasing reactivity and reducing of sintering temperatm'es can't be realized unless powder agglomeration will be minimized during preparation or agglomerates will be eliminated, and uniform densification of green body ^yill be achieved. This study showed that determination of the optimal powder formation parameters, and post-synthesis treatment for fabrication of non-agglomerated zirconia powder with finest primary crystallites uniformly packed into the secondary aggregates with narrow size distribution, and determination of the best suspension's properties allowed preparation of the uniformly densified green body with density high for nano-size powders.

REFERENCES

1. W. H. Rhodes, J. Am. Ceram. Soc. 64, (1981) 19. 2. W. Luan, L. Gao, and J. Guo, NanoStructured Materials 10, (1998) 1119. 3. S. Theunissen, A. Winnubst and A. Burggraaf, J. Eur. Ceram. Soc. 11, (1993) 315. 4. O. Vasylkiv and Y. Sakka, J. Am. Ceram. Soc. submitted (2000). 5. P. Duran, M. Villegas, F. Capel, C. Moure, J. Mater. Sci. 15, (1996) 741. 6. S. Lawson, J. Eur. Ceram. Soc. 15, (1995) 485. 7. N. Enomoto, S. Maruyama, and Z. Nakagawa, J. Mater. Res. 12, (1997) 1410. 8. O. Vasylkiv and Y. Sakka, J. Am. Ceram. Soc. 83 [9] (2000). 9. O. Vasylkiv H. Borodians'ka and Y. Sakka, J. Am. Ceram. Soc. accepted (2000). 10. F. F. Lange, J. Am. Ceram. Soc., 72, (1989) 3. 11. T. Suzuki, Y. Sakka, K. Nakano and K. Hiraga, Materials Transactions, JIM. 39, (1998) 689.