Colloids Elsevier

and Surfaces, 19 (1986) Science Publishers B.V.,

159-170 Amsterdam

159 Printed in The Netherlands

CORRELATION BETWEEN INTERFACIAL TENSION AND MICROEMULSION STRUCTURE IN WINSOR EQUILIBRIA. ROLE OF THE SURFACTANT FILM CURVATURE PROPERTIES

D. LANGEVIN, Laboratoire Paris Cedex (Received

D. GUEST

and J. MEUNIER Hertzienne accepted de l%.N.S., 16 January 24, rue Lhomond, 75231

de Spectroscopic 05 (France) 15 October 1985;

1986)

ABSTRACT We present here a correlation between measured interfacial tensions, dispersion sizes and surfactant film curvature properties in two model multiphase microemulsion systems. This correlation was performed with the help of recent microemulsion theories. The characteristic dispersion size L is found to be close to the spontaneous radius of curvature of the film when Ro is small, and to the persistence length of the film when Ro is larger (bicontinuous microemulsions). Interfacial tensions correlate only approximately with kT/L' . It is proposed that this occurs because film curvature contributions are dominant in two-phase systems. The situation is less clear in three-phase systems.

INTRODUCTION

Microemulsions are dispersions of oil and water prepared with surfactants emulsions, they are thermodynamically stable, because the characteristic size of the dispersion is very small, -100 A. It has been-shown theoretically by Ruckenstein and Chi [2] that thermodynamic stability arises from the fact that the surface energy, which is equal to the product of the interfacial tension, yaw and the total area between oil and water, can be compensated by the dispersion entropy, when yaw is sufficiently small: interfacial tensions are typically yaw 5 10m2 dyn cm- . Such ultralow commonly obtained by using a cosurfactant which is usually an alcohol. Note that typical interfacial tensions between oil and water without surfactant are about rtw - 50 dyn cm- . Depending on the spontaneous curvature of the surfactant film which covers oil-water interfaces, the structure of the dispersion may be, as for emulsions, of the oil-in-water type (O/W: oil droplets in a water-continuous phase) or of the water-in-oil type (W/O: water droplets in an oil phase). An evolution from the first type to the second can be obtained, for instance, by varying the temperature with non-ionic surfactants or the salinity with ionic surfactants. If a cosurfactant is used inversion may also be obtained by changing the amount of cosurfactant in the layer.

[l] . Unlike

0166-6622/86/$03.50

o 1986

Elsevier

Science

Publishers

B.V.

160

In this paper, we will restrict the discussion to ionic surfactants and present studies of two model systems where the curvature properties of the surfactant layers are different. We will present data at variable salinity (fixed amount of cosurfactant in the layers). At low water salinity (S) one obtains an O/W microemulsion which can coexist with excess oil. In such a case the radius of the droplets has almost reached the spontaneous radius of curvature: the droplets cannot accommodate more oil without a large increase in curvature energy and the excess oil is rejected in an excess phase [3]. The interfacial tension between the microemulsion (M) and the excess phase is very small, 70M N<10m2 dyn cm- . It happens to be equal to the interfacial tension between the excess oil phase and the water-continuous phase [4]. This is because the low tension is only due to the high surface pressure of the mixed surfactant-cosurfactant layer at the macroscopic flat interface yaw . At higher salinities one obtains a W/O microemulsion which can coexist with excess water. In such a case the radius of the droplets is again close to the spontaneous radius of curvature as above. The interfacial tension TwM is also equal to yaw , i.e. 5 10e2 dyn cm- . At intermediate salinities a middle phase microemulsion can coexist with both excess oil and water. The structure of this type of microemulsion is probably bicontinuous [5]. The surfactant film has been shown to have a very small mean radius of curvature [6] . Several theoretical models have been proposed recently to describe its properties: Voronoi tessellation of polygons by Talmon and Prager [7], and cubic cells by de Gennes and Taupin [8] and Widom [9]. When the oil or the water volume fractions, Go or Gw, are small, these models are equivalent to the droplet models. The interconnection of both oil and water cubes or polygons over macroscopic distances are predicted for 0.2 2 C#J~(or c$~ ) 5 0.8. The structure is then bicontinuous and evolves continuously from the O/W to the W/O structures between these limits. Recent X-ray and neutron scattering data are in agreement with some of the predictions of the models [lo--131 . The interfacial tensions between the microemulsion and the excess phases are still smaller than in two-phase equilibria. One of these tensions, the largest, is equal to the tension yaw between the excess phases [14]. Again, the low tension is associated with the high surface pressure of the surfactant layer. The second tension, the lowest, decreases to almost zero, close to the two-phase-three-phase boundaries. This was shown to be associated with the vicinity of critical end points in the phase diagram [ 141. In this paper we will mainly be concerned with the interfacial tensions associated with the surfactant film. These low tensions are the ones of interest in tertiary oil recovery [ 151. Several theories predict that such tensions are of the order of kTIL2, where L is the dispersion size, whereas the low tensions associated with critical phenomena are of the order of kT/g,2.,where [, is the correlation length for concentration fluctuations. In order to obtain information about the quantitative relationship between y and L, we have under-

161

taken a study of model systems with different surfactants: one which forms micelles in pure water and one which forms lamellar phases. According to the theory of de Gennes and Taupin, the elastic curvature modulus must be higher for the surfactant film-forming lamellar phases in water [8]. Consequently, the characteristic dispersion size L in the second system should be higher and the tensions lower. In the following, we will briefly recall the existing theories of inter-facial tensions. We will then present the data obtained on the model systems and compare them with the theories.
THEORETICAL BACKGROUND

Structural

models

The Talmon-Prager-de Gennes-Widom model can be used to describe droplet dispersions and bicontinuous structures as well. In the simplest version given by de Gennes, the microemulsion volume is divided into consecutive cubes of linear size l randomly filled with either oil or water. The interface between cubes of different types is covered by a surfactant film, the volume of which is neglected. If $o and & are the oil and water volume fractions, the average total area between oil and water per cube is 6 @o Gw t2 and the corresponding surface The entropy of dispersion per cube is k (tie In Go energy is 6 @O 4~ t2 YOW . + Gw In rjw ). The statistic of the interface is thus reduced to a lattice-gas model. When

a single-phase microemulsion is thermodynamically stable. With t - 100 A, it follows that: yaw 2 10m2 dyn cm- . The area per surfactant molecule, z1, is generally reasonably constant whatever the spontaneous curvature may be. It follows then that t is determined when the total number of surfactant molecules ns is given. The total area is given by ns C, and with Cs = ns /V, where V is the total volume, we have:

A similar expression holds for the average size of the polygons in the TalmonPrager model: & = 5.82 $. +w /Cs C. In the limit of isolated oil or water droplets of radius R. or Rw one would have:
R. = -

340 CSC

(24

162

or
Rw = CSE

3&v

(2b)

which are limiting cases of Eqn. (1) for small Go or &. with R = t/2.
Curvature energy

In the above model, the occurrence of W/O or O/W structures is related only to the relative amounts of oil and water. Only single-phase and twophase equilibria are found. A more complete description of the microemulsion is obtained by introducing the curvature energy of the surfactant layer [16,17] :
F, = (3)

where K and x are the elastic moduli of curvature, RI and R2 are the principal radii of curvature of the surfactant film and R, its spontaneous radius of curvature. The integral is over the film surface A. R. describes the tendency of the interface to curve towardssither the water (R, > 0) or the oil (R, < 0). The saddle-splay term with K favors structures with equal radii of curvature. In the following we will assume K = 0 and R 1 = R 2 = R. If K is large, R will be close to R,,, irrespective of the amount of oil and water. This is probably the situation found in several types of W/O microemulsions that can be concentrated up to large water volume fractions (-50%) without becoming water continuous. If K is smaller, R will begin to depend on $J (&, or &,, ). Its value will be found by minimizing the total free energy in the following way [17] . As soon as the microemulsion is formed, we have seen that yaw no longer varies. The surfactant layer at the interface is now incompressible, and the area per surfactant X is constant (saturated interface [8] ). The surface energy terms being irrelevant, R is determined by minimizing the sum of curvature energy and dispersion entropy:

F,+F,

47rR* + n,hT

[@IIn r$ + (1 - 4) In (1 - $)]

subject to the condition n,YZis constant, where n, is the number of surfactant molecules ( ns = Cs V). For spherical droplets, R = 3$/C, C and [ 17 ] :
R -=1+%

R0

ln$+p

1.5 - $I 4

ln (I-

@)

ForK-

kTand(g-

0.2, R is only 10% smaller than Ro.

163

If the spontaneous radius of curvature is large and K small, the thermal fluctuations of the interface might become the dominant effect in determining the characteristic size of the dispersion. This has been predicted by de Gennes and Taupin [8]. This size will be determined by the persistence length of the interface gk: at scales smaller than ,$k the interface is essentially flat, whereas at scales larger than $k it is strongly wrinkled. gK is related to the elastic modulus of curvature by: tK = a exp (27rK/kT) length. and an excess phase containing spherical (5)

where a is the molecular

Inter-facial tension between a microemulsion Most existing calculations droplets of radius R. Israelachvili [18] provides tributing terms. Let us call molecule in an aggregate of N
PN =

apply

to microemulsions

a simple description of the nature of the conpN the chemical potential of the surfactant molecules:

&

E lnXh
N N

where &, is a standard chemical potential containing the interaction terms between the surfactant molecules in the droplet; XN is the surfactant concentration fraction present in the droplet. The interface between the microemulsion and the excess phase is covered by a surfactant layer which can be considered as an infinite aggregate. This aggregate is subjected to a tension. The surfactant chemical potential is therefore: Ps = Pm --yc = &-rc tension is finally obtained by putting pN equal to ps :

The interfacial

(f-3)
The first term, Ye, is an entropic contribution arising from mixing. It is analogous to the expression derived by Ruckenstein on a different thermodynamical basis [ 191. The second term yC, is a curvature contribution, which has been calculated explicitly by other authors [ 20-221. With:

164

and

these calculations
Yc =

are equivalent

to: (7)

R+ 0

The orders of magnitude of the two terms are the same: NC = 47rR*, R - R. and K - kT, thus yc - 7e - kT/R* . It can be seen that the tension depends ultimately on the droplet radius, although it does not change when the microemulsion is replaced by its continuous phase, i.e. as the droplets are removed [4]. In fact both y and R are determined by the properties of the surfactant layer and it is not surprising to find a relation between them. Equation (6) does not contain any contribution from interaction between different droplets. This contribution is expected to be negligible as soon as one is sufficiently far from a critical point. This is certainly the case when y is equal to yaw , which does not contain any interaction term. When the microemulsion is bicontinuous, it is not so easy to calculate its interfacial tension with an excess phase. The curvature energy is small (small K, large R,) and comparable to the dispersion entropy. A calculation has been performed recently by Borzi et al. [23] : the separation between curvature and entropy contributions is less straightforward than in Eqn. (6). However, it may be expected that the curvature contributions will be small [they sould tend to zero according to Eqn. (7)] and that the entropy contributions will be dominant, although the curvature term in the free energy is needed to obtain the three-phase equilibria [9].
EXPERIMENTAL PROCEDURE

Sample composition We will now discuss the results obtained with two model systems. The first is a mixture of brine (37 wt%), toluene (37 wt%), butanol (4 wt%) and sodium dodecyl sulfate (SDS, 2wt%). The brine is an aqueous solution of Swt% sodium chloride. The salinity was varied between 3.5 and 10. The salinities of phase separation have been found to be S1 = 5.4 and S2 = 7.4 at 20C. The second model system is a mixture of brine (56.83 wt%), dodecane (38.19wt%), butanol (3.32wt%) and sodium hexadecylbenzenesulfonate (SHBS or Texas No. 1, 1.66wt%). This composition corresponds to equivalent volumes of oil and brine. The brine is an aqueous solution of S wt% sodium chloride. The salinity was varied between 0.4 and 0.9. The salinities of phase separation have been found to be S, = 0.52 and S2 = 0.61 at 20C.

165

Experimental

techniques

Inter-facial tensions have been measured using surface light scattering experiments. The frequency broadening of the scattered light was measured with a wave analyzer after photon beating detection. The width of the spectrum is directly proportional to y [24]. The characteristic sizes in microemulsion phases were measured with elastic and quasi-elastic bulk light scattering techniques in the two-phase regions where the microemulsion structure is that of droplets dispersed in a continuous phase [ 14, 251. In the three-phase region, dilution is not possible. The characteristic length in the dispersion has been deduced from X-ray and neutron experiments [ 11-131. Ellipsometry experiments have been performed to investigate the elasdomain, the ticity properties of the surfactant layer: in the three-phase interfaces between excess phases have been studied; and in the two-phase domain the interfaces between the excess phase and the microemulsion continuous phase have been studied. All these interfaces show the same features: they exhibit ultralow tensions yaw [4, 141, and the phases in equilibria contain no microstructures. The ellipsometry signal is then dominated by the roughness of the interface due to thermal motion. The measured ellipticity is in this case [26] :

e. and +, are the dielectric length of the light.

EXPERIMENTAL DATA

constants

of the two phases, and X is the wave-

AND DISCUSSION

Sizes in microemulsion

phases

The normalized sizes LC,C/6, with L = 2R for droplets and L = t in middle phases, are plotted in Fig. 1, together with the theoretical predictions for spheres (Eqn. (2), lines) and bicontinuous structures (Eqn. (l), parabola). It is seen that in the two-phase region, sizes L vary as r#~~or #w, as predicted by Eqn. (2). In the three-phase region there is a remarkable continuity between droplets radii and characteristic size [. [ varies as tie @I~, as predicted by Eqn. (1). It was observed that the [ values were unexpectedly constant over the three-phase domain. The surfactant concentration C, and the area per surfactant C are also constant, which occurs because the product Go&, is approximately constant. This result is strongly in favor of the

166

Fig. 1. Characteristic length in the microemulsion phases LC,c/G versus oil volume fraction &,. SDS system: (0) light scattering data [9] ; (a) X-ray data [20] ; (x) neutron data [ 221. SHBS system: (0) light scattering data [19] ; (A) X-ray data [21]. Lines corresponds to Eqn. (3) with L = 2R, the parabola to Eqn. (2) with L = gK.

existence of bicontinuous structures, and demonstrates that droplet structures are unlikely. A large difference between the L values for the two systems is oberved. In the three-phase region, assuming [ N {x (Eqn. (5)], the difference between g values should be associated with larger elastic moduli in SHBS films than in SDS ones. This happens although the SHBS films contain more alcohol than the SDS ones: respectively three and one alcohol molecules per molecule of surfactant. This confirms that pure SHBS films are much more rigid than pure SDS ones, thus favoring the formation of lamellar phases in water as observed [ 271. In the two-phase region, the droplet radius is close to the spontaneous radius of curvature R, . For the SDS system the difference between R and R. is of the order of 10% in the O/W microemulsions and 20-30% in the W/O microemulsions as obtained with Eqn. (4) and K values deduced from ellipsometry data (Table 1). For the SHBS system, K has not yet been

TABLE 1

Correlation between the largest interfacial tensions (-y > -y*) and the characteristic lengths. Comparison with curvature and entropy In o, L = R or t contributions: yc = ZKIRR, and ye = -(kT/4nRZ) Y (dyn cm-
-1

System S YL/kT Yc (dyncm ) K (erg)

R (8) Yom 7.5 1o-2 2 1o-2 1.4 1o-2 1.35 0.85 0.85 5 lo-l4 3 lo-l4 3 lo-l4 2 lo-l4 2.1 1o-2 3.6 1O-2 0.65 1.86 2.93 1.1 10-l 5.9 10-z 4 1o-2 230 Y 2Ylo-2 7.5 1o-2 4.5 1o-3

Ye (dyn cm-)

SDS S<S*

3.5 5 5.2

85 130 155

8 1O-3 3 1o-3 2 1o-3

s=s*

s>s* 8 10

180 125

2 1o-3 5 1o-3

SHBS S < S* 0.4 0.5 0.52

112 203 250 3 1o-3 2.9 1o-3 400 YWIl 1.1 1o-2 3.5 lo+
0.95

Yom 8.9 1O-3

0.28 0.31 0.46 0.35

4.2 1O-3 1o-3 6 1O-4

s=s* 8.7 1O-4

s>s* 0.7 0.9 1.46

192 129

1O-3

3 1o-3

168

measured but according to the discussion relating to &, it is expected to be larger, so that R/R,, will be closer to unity. The spontaneous radius of curvature is largely controlled by simple steric constraints [28] . The hydrophobic part of the SHBS surfactant is more bulky than that of the SDS surfactant. R, is then likely to be larger in the SHBS microemulsions, as observed. For the SDS system, ellipsometry measurements give a value for K of (3 f 1) lo-l4 erg in the three-phase domain [26]. With t = 240 A [ 11, 121, and by taking $ = & and a - 10 A, one gets from Eqn. (5): K - 2 lo-l4 erg, in remarkable agreement with the experimental value. Measurements of K in the SHBS system are in progress in order to determine whether, as expected K is larger in this system. An earlier X-ray study of a model system containing SHBS [lo] has been performed. The composition of the samples was the same as here except for the alcohol which was isobutanol instead of butanol and the temperature was 25C instead of 20C. The sizes were found to be about 30% smaller. Although there is no information about the composition of the surfactant layer, it is expected that branched alcohols lower the value of K; K/kT and $k should indeed be smaller in this model system.

Relation

between

inter-facial tensions and sizes

The measured interfacial tensions versus salinity are plotted in Fig. 2; y* = 4.5 10m3 dyncm- for the SDS system, y* = 8.7 10e4 dyncm- for the SHBS system. We have seen that the tensions y > y* are expected to be of the order of kT/L2. The ratios yL2 /kT have been plotted in Table 1. As predicted, these ratios are of the order of unity, but they systematically increase with salinity. Moreover, they are systematically larger in the SDS system than in the SHBS system. It has therefore to be concluded that the ratio yL2 /kT is not universal. As explained above, in the two-phase region y contains two kinds of terms: an entropy term 7e which correlates with kT/R2 (the factor log@ being a slowly varying function of concentration) and a curvature term which correlates with KIR2. In order to see if this second term is responsible for the variation of yR2 /kT with salinity, a numerical comparison was performed with the SDS system in Table 1. It must be noted that this comparison cannot be very accurate, in view of the large experimental uncertainties: 20% on R, 30% on K. yC is larger than y for the O/W microemulsions. This may indicate that the saddle-splay term cannot be neglected here [Eqn. (3)]. Indeed, the droplets have been shown to be elongated [ 141. Their ellipticity increases as salinity increases, and as the discrepancy between yc and y increases. yc is expected to be the most significant contribution to y, since ye is typically 10 times smaller than y (Table 1).

169

1U

10

10-4L
0.2

0.3

0.4

05 tensions

0.6

Cl,for SHBS system, and SDS system (insert).

Fig. 2. Interfacial

versus salinity

The agreement between the calculated values of yC and y is better for the W/O microemulsions. The difference between R and R, becomes significant here (20-30%) and has been taken into account.
CONCLUSION

We have performed a detailed study of multiphase microemulsion model system where successively an O/W microemulsion coexists with excess oil, a middle phase microemulsion coexists with excess oil and water and a W/O microemulsion coexists with excess water. The surfactants were SDS and a pure alkylbenzenesulfonate, SHBS. The interfacial tensions between the microemulsion and the excess phases, as measured with surface light scattering methods, are ultralow, with the SHBS system being even lower. The sizes of the droplets were obtained in the O/W and in the W/O microemulsions using bulk light scattering methods. They are very close to the

170

spontaneous radius of curvature of the surfactant film and are larger in the SHBS system. The difference remains in the three-phase domain where sizes were taken from recent X-ray studies. In this domain, the sizes are close to the persistence length of the surfactant film. As K and R, are expected to be larger in SHBS films, this explains the observed size difference. The relationship between the largest interfacial tensions (y > y*) and characteristic sizes has been quantitatively investigated. It has been shown that y correlates only approximately with kT/L2 . The ratio yL2 /kT varies with salinity and with the nature of the surfactant. It is proposed that these differences arise from curvature contributions to the interfacial tensions in the two-phase region. The situation is less clear in the three-phase region.
REFERENCES L.M. Prince, Microemulsions, Academic Press, New York, 1977. E. Ruckenstein and J. Chi, J. Chem. Sot. Faraday Trans. 2, 71 (1975) 1690. S.A. Safran and L.A. Turkevich, Phys. Rev. Lett., 50 (1983) 1930. A. Pouchelon, D. Chatenay, J. Meunier and D. Langevin, J. Colloid Interface Sci. 82 (1981) 418. L.E. Striven, in K.L. Mittal (Ed.), Micellization, Solubilization and Microemulsions, Vol. 2, Plenum, New York, 1977, p. 877. L. Auvray, J.P. Cotton, R. Ober and C. Taupin, J. Phys. Chem., 88 (1984) 4586. Y. Talmon and S. Prager, J. Chem. Phys., 69 (1978) 2984. P.G. de Gennes and C. Taupin, J. Phys. Chem., 86 (1982) 2294. B. Widom, J. Chem. Phys., 81(1984) 1030. E.W. Kaler, H.T. Davis and L.E. Striven, J. Chem. Phys., 79 (1983) 5685. L. Auvray, J.P. Cotton, R. Ober and C. Taupin, J. Phys. (Paris), 45 (1984) 913. A. de Geyer and J. Tabony, Chem. Phys. Lett., 113 (1985) 83. D. Guest, L. Auvray and D. Langevin, J. Phys. Lett. (Paris), 46 (1985) L-1055. A.M. Cazabat, D. Langevin, J. Meunier and A. Pouchelon, Adv. Colloid Interface Sci., 16 (1982) 175. D.O. Shah (Ed.), Surface Phenomena in Enhanced Oil Recovery, Plenum, New York, 1981. W. Helfrich, 2. Naturforsch., 28 (1973) 693. S. Safran, in S.H. Shen (Ed.), Statistical Thermodynamics of Micelles and Microemulsions, Springer Verlag, in press. J. Israelachvili, in K.L. Mittal and P. Bothorel (Eds), Surfactants in Solution, Plenum, New York, in press. E. Ruckenstein, in K.L. Mittal and B. Lindman (Eds), Surfactants in Solution, Vol. 3, Plenum, New York, 1984, p. 1551. M. Robbins in K.L. Mittal (Ed.), Micellization, Solubilization and Microemulsions, Vol. 2, Plenum, New York, 1977, p. 713. D.J. Mitchell and B.W. Ninham, J. Phys. Chem., 87 (1983) 2996. C. Huh, J. Colloid Interface Sci., 97 (1984) 201. C. Borzi, R. Lipowski and B. Widom, Faraday Discuss. Chem. Sot., 20 (1985). D. Langevin, J. Meunier and D. Chatenay, in K.L. Mittal and B. Lindman (Eds), Surfactants in Solution, Vol. 3, Plenum, New York, 1984, p. 1991. D. Guest and D. Langevin, J. Colloid Interface Sci., in press. J. Meunier, J. Phys. Lett. (Paris), 46 (1985) L-1005. E.I. Franses, H.T. Davis, W.G. Miller and L.E. Striven, ACS Symp. Ser., 91 (1979) 35. D.J. Mitchell and B.W. Ninham, J. Chem. Sot. Faraday Trans. 2, 77 (1981) 601.

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28