J . Chem. SOC., Faraday Trans. I , 1985, 81, 2947-2958 Lower Consolute Boundaries of a Poly(oxyethy1ene) Surfactant in Aqueous Solutions of Monovalent Salts BY KRISTIAN WECKSTROM* AND MARTIN ZULAUF European Molecular Biology Laboratory, 156X, 38042 Grenoble Cedex, France Received 8th January, 1985 Lower consolute boundaries up to 50 vol% surfactant have been determined for n- octylpenta(oxyethy1ene glycol) (C,E,) in the presence of the monovalent salts NaF, LiC1, NaCl, KC1, CsCl, NaBr and NaI. We find that the lower consolute boundaries are shifted, with only small changes in shape, to lower or higher temperatures (salting-out or salting-in, depending on the salt used). The amount and sign of the miscibility shift is determined almost solely by the anion, and the shift is related to the surface charge density of the ion.The shifts in the lower consolute boundary cannot be explained by structural changes in the bulk water structure due to the addition of salts. We have applied the Flory-Huggins lattice theory to relate the shifts of the lower consolute boundaries to changes in micelle-solvent interactions. This analysis shows that small changes in the free energy of interaction between oxyethylene and water can explain the shifts in temperature of the lower consolute boundary. We discuss the use and limitations of the Flory-Huggins theory, as applied to micellar phase separation. The salt effects are explained in terms of salt-deficient or salt-rich regions around the oligo-oxyethylene chains. Non-ionic surfactants of the poly(oxyethy1ene) type [CH3(CH,),-,(OCH,CH,),0H, abbreviated as C,E,] exhibit a variety of solution phenomena in water.lt2 When a single homogeneous phase is observed, the surfactant molecules form micellar or mesophase structures with widely different properties.At room temperature the first structured phase is often the micellar phase, which forms above the critical micelle concentration (c.m.c.). At higher concentrations liquid-crystalline mesophases may occur, formed by large anisotropic aggregates arranged in regular array^.^ Depending on temperature and concentration there are solubility limits beyond which phase- separated solutions are found. Upon heating the isotropic micellar solution, a critical temperature is eventually reached at which the solution suddenly becomes turbid ('cloud point).After some time demixing into two transparent liquid phases occurs, one of which contains practically all of the surfactant. The temperature at which the demixing is reached depends on concentration, and the whole boundary is called the lower consolute boundary (1.c.b.). This consolution phenomenon shows features characteristic of transition^^-^ observed with binary liquid mixtures, and it implies that micelles interact attractively to bring about concentration fluctuations preceding the phase separation. The increased interactions, and subsequent demixing, are most probably due to a decrease in the oxyethylene-water interaction, i.e. a decrease in hydration.s Previous neutron-scattering measurements showed that micelles formed by the non-ionic surfactant octylpenta(oxyethy1ene glycol) (C,E,) remain small and approximately spherical, even in the neighbourhood of the demixing b ~ u n d a r y .~ The dehydration of the pentaoxyethylene chains thus has practically no effect on the micellar geometry. Furthermore, the surfactant shows no mesophases above 5 "C and therefore develops a broad micellar It is interesting with this background to 29472948 POLY(OXYETHYLENE) SURFACTANTS study how added inorganic salts change the demixing process. The main effect of the salts is to shift the position of the binary phase boundary in terms of temperature. After presenting the experimental results we discuss ion specificity and apply the Flory-Huggins lattice theory9--13 to analyse the shifts in the lower consolute boundary with temperature.EXPERTMENTAL MATERIALS The non-ionic surfactant C,E,, octylpenta(oxyethy1ene glycol), was used as received from Bachem Feinchemikalien AG (CH-4416 Bubendorf, Switzerland). It has been shown previously that this surfactant is of high purity as judged by elemental composition, thin layer chromato- graphy, infrared spectra and ten3i0rnetry.l~ Also, it is monodisperse with respect to alkyl chain length as well as with respect to its hydrophilic part. No ageing effects were found even after several months' storage in a desiccator at room temperature (cloud points remained unchanged). The salts NaCl, KC1, LiCI, CsCl, NaBr, NaI and NaF were from Fluka AG (puriss) and were used without further purification. The water used was freshly double-distilled.METHODS To determine the clouding temperatures, samples contained in seven quartz cuvettes, (2 x 10 x 40 mm, Hellma) mounted in a Plexiglas box and connected to a thermostat, were observed while slowly raising the temperature (0.1 "C min-' near the critical point) until clouding occurred. The temperature was measured by a platinum resistance thermometer placed in an additional cuvette. Samples were prepared from a stock solution of 50 vol C,E, in water containing various amounts of salt. This solution was allowed to equilibrate overnight. Dilution of the stock solution was then made in order to obtain the final concentrations in quantities of 5@-100 mm3. These were allowed to equilibrate for an additional 2 h before measurement. 'The solutions were prepared with micropipettes on a volume-percent basis, and it was confirmed by weighing that satisfactory accuracy was obtained.Since the density of C,E, is 1.0080 g cm--3 at 20 "C, the weight-percent and volume-percent concentrations are the same within errors of preparation. It was also verified that the measured cloud-point values did not depend on sample volume; volumes of 50, 100, 200 and 400 mm3 gave the same cloud-point values. The readings of cloud-point temperatures obtained by visual inspection were checked by recording absorbance at 450 nm with a spectrophotonieter equipped with a temperature cell and a programmable temperature controller. A temperature increase rate of 20 "C h-l was used. The absorbance showed a clear break at the clouding temperature for C,E, concentra- tions above 0.50/,, and good agreement was found (k0.1 "C) between the two methods and by repeating measurements immediately after clouding.RESULTS It is convenient to represent the different phases formed by non-ionic surfactants in water by temperature uersus concentration diagrams.l The phase diagram of C,E, is then particularly simple7. and comprises three boundary curves distinguishing (i) the monomeric state below the c.m.c. (0.15 7; ), (ii) the hexagonal HI phase extending from 40 to 607; at temperatures below 5 "C and (iii) the two-phase area above ca. 60 "C. In between these boundaries is the ordinary micellar phase. The large area in which the micellar state exists is a typical property of octyloligo-oxyethylenes. Cloud-point temperatures for the surfactants C,E, and C,E, are given in table I .The values for C,E, are shown graphically in fig. I (aj. At concentrations below 0.5:.1,, clouding becomes progressively difficult to observe. We have verified by light scattering that well above the 1.c.b. the aqueous phase obtained after complete phase separation contains micelles at very low concentration. We therefore conclude thatK . WECKSTR~M AND M. ZULAUF 2949 Table 1. Cloud-point temperatures for the non-ionic surfactants C,E, and C,E, in water 0.5 1 .o i .5 2.0 3.0 5.0 10.0 15.0 20.0 25.0 30.0 40.0 50.0 55.0 60.0 65.0 43.4 41.6 40.9 40.3 40.0 39.7 39.7 40.4 41 - 5 43.4 45.6 51.1 57.9 61.1 65.3 69.9 62.6 61.1 60.6 60.2 59.8 59.5 59.4 59.6 60.0 61.1 62.8 66.9 72.8 60 46 0.5 1 5 10 50 0.5 1 5 10 50 Fig.1. Lower consolute boundaries of C,E, in (a) water, (b) 0.5 mol dm-3 NaCl, (c) 0.9 mol dmP3 NaCl and ( d ) 1.3 mol dmP3 NaC1. Points corresponds to the 11 lowest concen- trations in table 1. I 4 2 I922950 POLY (OXYETHYLENE) SURFACTANTS '42 7 0.5 1 5 10 50 ' 4 2 0.5 1 5 10 50 ' 4 2 0.5 1 5 10 50 '9 2 Fig. 2. Lower consolute boundaries of C,E, in (a) 0.2 mol dm-3 NaI, (b) 0.5 mol dm-3 NaI, (c) 0.9 moi dm-3 NaI and ( d ) 1.3 mol dm-3 NaI. the 1.c.b. will rise below 0.5% and merge towards the c.m.c. curve. However, direct observation of clouding by recording optical absorbance (see Experimental section) failed in this concentration region. The neat break in absorbance as a function of temperature found above 0.5% is blurred out below, and we have no direct experimental data on the 1.c.b.there. Note that for surfactants with a short hydrocarbon chain (such as C,E,) the c.m.c. is also less neatly defined than for the longer-chain homologues. We have determined the lower consolute boundary of C,E, in the presence of NaCl, KC1, CsC1, LiC1, NaI, NaBr and NaF. Fig. 1 (b)-(d) shows the 1.c.b. for three NaCl concentrations. The 1.c.b. is shifted to lower temperatures with a small change in curvature of the boundary. Within 0.2-0.4 "C similar 1.c.b. curves are obtained with corresponding molarities of KCl and CsC1. On the other hand, NaI shifts the 1.c.b. to higher temperatures (fig. 2). The salts NaBr and LiCl (fig. 3) are less effective in lowering the boundary position. With NaI, LiCl and NaBr the 1.c.b.is shifted to higher or lower temperatures with practically no changes in the shape of the boundary. Critical concentrations ('4;) for various salt conditions, as obtained from the rectilinear diameter, are given in table 2. The diameter does not show marked deviations from linearity for the salt condition studied, when the surfactant volume fraction is plotted on a logarithmic scale. Shifts in temperature position of the 1.c.b. from the binary value can be interpreted in terms of micellar solubility. As first observed with protein solubility, the effectiveness of salts in reducing solubility followsK. WECKSTROM AND M. ZULAUF 295 1 57 V 2 55 52 u c 50 ‘42 J I I I I 0.5 1 5 10 50 0.5 1 5 10 50 Fig. 3. Lower consolute boundaries of C,E, in (a) 0.5 mol dm-3 NaBr, (b) 1.3 mol dm-3 NaBr, (c) 0.5 mol dm-3 LiCl and (6) 1.3 mol dm-3 LiC1. I 4 2 I41 Table 2.Values of the interaction parameters Hi and Si, their percentage changes, the interaction energy, w, and the critical concentrations, ’4; (in volx), for C,E, c s -Hi - si w ‘4; salt /mol dm-3 /kJ mol-’ AHi PA) /J K-I mo1-1 ASi (%) /J mol-l (~01%) ~~~ NaI 0.5 1.3 NaBr 0.5 1.3 NaCl 0.5 1.3 NaF 0.2 0.5 LiCl 0.5 1.3 12.14 12.51 11.42 11.13 11.06 10.39 11.11 10.48 11.28 10.93 - 4.4 - 7.6 1.8 4.3 4.9 10.7 4.5 9.9 3.0 6.0 40.08 40.61 39.02 38.58 38.47 37.45 38.55 37.58 38.81 38.28 - 1.9 - 192.2 -3.3 -404.2 0.8 21 1.9 1.9 370.7 2.2 407.9 4.8 773.9 2.0 381.8 4.5 722.6 1.3 289.3 2.7 481.3 6.8 8.0 7.0 6.6 8.6 10.2 11.0 15.5 8.2 6.8 - 11.63 - 39.33 - 94.3 8.6 H2O2952 POLY (OXYETHYLENE) SURFACTANTS a certain order, known as the Hofmeister or lyotropic series.16 With CEE5, NaF, LiC1, NaCl, KCl, CsCl and NaBr t,he salt micelles out, whereas salting-in is found with NaI.The salt-induced changes in 1.c.b. shape consist of a progressive deviation from the binary 1.c.b. shape with increasing surfactant concentration. In other words, if the binary 1.c.b. is shifted in temperature by an amount such that it is superimposed on the observed 1.c.b. in the presence of salts at low surfactant concentrations (e.g. 1 vol % ), deviations occur progressively with increasing surfactant concentration (fig. 6). This deviation is - 5.8 "C for C,E, at 50 vol;/i in 1.3 rnol dm-3 NaCl (data not shown). With NaF the deviations are even bigger: - 8.9 "C, at 45 vol % CEE, in a salt solution of 0.5 mol dm-3.The boundaries thus have a flatter shape. Cloud-point temperatures for a 2 vol % CEE, solution as a function of salt molarity are shown in fig. 4. The curves are convex when C,E, is salted out and concave when salting in occurs. Similar salt trends at low surfactant concentration have been reported in several previous studies.,, 1 5 9 17-20 DISCUSSION Recent n.m.r.2y 22 and neutron-scattering m e a s u r e m e n t ~ ~ ~ - ~ ~ provide strong evidence that the consolution phenomenon observed in aqueous solutions of several poly(oxy- ethylene) surfactants (CEE5,C12E,,Cl,~E,Cl,~E) arises from attractive forces between micelles having a fairly constant size. The static and dynamic neutron-scattering results can be accounted for if we assume that micelles remain small, but that they interact according to a direct interparticle potential of the hard-sphere type with an attractive short-range Data analysis on this basis revealed that an increase in temperature corresponds to an increase in the attractive potential depth, with other parameters (micelle size or shape) unchanged.The concentration dependence is thereby reproduced without additional parameter variation. Clouding occurs at a finite value of the attractive potential. The increasing attraction was interpreted as being due to dehydration of the oligo-oxyethylene head groups, allowing closer contacts between micelles to occur upon collisions.26 In fact, both neutr~n-scattering~, and n.m.r. self-diffusion experiments21 indicate a decrease in hydration of the polar layer with increasing temperature.The salts probably change the interparticle interactions, or micelle-solvent interactions, so that the temperature at which phase separation occurs changes. This implies that the salt effects are, to a first approximation, independent of the salt/surfactant molar ratio (fig. 1-3), and that strong ion-micelle interactions are absent. Evidence has been presented before in favour of complex formation between polyvalent ions and lithium with the oligo-oxyethylene parts of non-ionic surfactant~,~~ the ether linkages probably acting as polydentate ligands. Fig. 4 shows that the anions effectively determine the cloud-point temperatures. Cations are known to be smaller and to bind more hydration water than anions.28 Therefore the 1.c.b.shifts are not caused primarily by competition between the ions for free water, but by a more specific anion effect. This becomes evident when the cloud-point temperatures, Td, are plotted as a function of the ionic surface charge density, l/r2 (fig. 5 ) . With chloride as the common anion, Td values do not depend in a marked way on the surface charge density of the cations (crosses in fig. 5). However, we find a regular decrease in Td with decreasing size of the anions when sodium is the common cation. These solubility changes can be further correlated with the structure-breaking ability of the anions. It is well established that structure breaking increases with increasing anionic radius.29 Thus large ions like iodide perturb the tetrahedral hydrogen-bonding pattern of water more than smaller monovalent anions.A comparison between such ionic structure-breaking properties and theK. WECKSTROM AND M. ZULAUF 2953 2 . 3 5 6 I 1 I I I I 0 0.2 0.5 0.9 1.3 C,/mol dm-3 Fig. 4. Cloud-point temperatures of aqueous C,E, solutions (2 vol"/) as a function of salt molarity for the following salts: (1) NaI, (2) NaBr, (3) LiCl, (4) NaC1, ( 5 ) KCL, (6) CsCl and (7) NaF. 7 0 6G LJ iz- 50 40 - 1'- I \ F- 0 .I I 5 r-2 / R -2 Fig. 5. Cloud-point temperatures of aqueous C,E, solutions (2 vol x , 0.5 mol kg-' salt) as a function of r-2, proportional to the surface charge density Z2e2/r2 of either cations (crosses) or anions (circles); 2 is the charge number, e is the electronic charge and Y the crystallographic radius.412954 POLY(OXYETHYLENE) SURFACTANTS observed 1.c.b.shifts shows that structural changes in the solvent cannot solely explain the observations; for example, F- is only a slight structure maker, whereas C1-, Br- and I- are structure breakers.29 This does not correlate with the observed salting-out and salting-in trends. It is also unlikely that the hydration of C,E, increases owing to a partial disruption of the water network, since an increase in temperature lowers the hydration numbers. Which ion actually causes salting-out and which salting-in depends in general on the dielectric and acid-base properties of the With the surfactant C,E, salting-in was observed only with NaI. We will now apply the lattice expression devised independently by Flory and Hugginsg-13 to study how the position with regard to temperature of the lower consolute boundary is related to the free energy of the oligo-oxyethylene-water interaction, and how salts change this interaction.The theory of interacting micelles outlined above predicts (principally correctly) the 1.c.b. shape but with a critical concentration that is too high in comparison with experiment.26 The low critical concentration of the 1.c.b. is most certainly due to the size difference between the micellar aggregates and the solvent water molecules. The Flory-Huggins (FH) theory, originally applied to polymers, takes such size differences into account as a starting principle. Examples of systems where the FH theory has been applied include polymer and non-ionic micellar solution^.^^^ 34 The highly directional interactions occurring when water is the solvent are not explicitly accounted for in the FH theory, which makes its application to aqueous solutions non-rigorous. Alternative theories are in the process of development, but they are not yet easily applicable to the present ~ y s t e m .~ , ? ~ ~ The FH expression for the solvent chemical potential is solutions of where indices 1 and 2 refer to two molecular components, 4 is the volume fraction (b1 + 42 = l), N = V2/ & is the molecular volume ratio of the components, R is the gas constant and w is an interaction energy parameter. The parameter w is a measure of the oligo-oxyethylene-water free energy of intera~ti0n.l~ We use the following (2) expression for co : 1 3 3 33 with the two temperature independent parameters Hi and Si.When applying eqn (1) to the experimental demixing temperatures we make the following assumptions. (1) Previous experimental studies and the results presented here suggest that the phase separation is mediated by oligo-oxyethylene-water interactions. We analyse the salt effects in terms ofchanges in oligo-oxyethylene-solvent (water plus salt) interaction. The two components in the FH equation are thus the oxyethylene of individual micelles and the aqueous solvent. The hydrocarbon regions are not included in the solute volume fraction. Note that the favourable combinatory entropy of mixing arises from the increased volume over which the components are distributed.12 The micellar hydrocarbon forms the aggregate interior and does not mix with the aqueous solvent.Let 0, be the volume fraction of micellized surfactant. Then the volume fraction of rlicellar oligo-oxyethylene is given by CD = Hi - TS, where a is the ratio of the oxyethylene volume to the total surfactant monomer volume. With the specific volumes given in ref. (7) we have a = 0.584 for C,E,. (2) We assumeK. WECKSTROM AND M. ZULAUF 2955 7c 60 V L 1 5c 40 Fig. 6. Fit of the theoretical lower consolute boundaries (see the text) to experimental cloud points of the surfactant C,E, in (a) water, (b) 0.5 rnol dm-3 NaCl and (c) 1.3 mol dm-3 NaCl. that the combinatory entropy of mixing of the micellar oligo-oxyethylene domain can be calculated as for a linear polymer of the same molecular weight, an approximation also used in recent extensions of the FH theory to highly branched (a geometry similar to the micellar oligo-oxyethylene).(3) When phase separation occurs, one of the phases contains surfactant at low concentration. We do not consider the effect of this surfactant on the water chemical potential in the aqueous phase. (4) In the presence of salts the electrolyte contribution to the water chemical potential should be considered. However, we assume that the salt/surfactant molar ratio is the same in both phases, which cancels the salt effect on the chemical potential. Combining eqn ( 1 ) and (2) gives (4) -Hi 4: R[ln (1 - 4 2 ) + (1 - 1/N) 4 2 1 -Si 4: Td = The critical concentration 4; corresponding to the minimum of Td is related to N by13 We then proceed as follows: first the cloud-point temperatures are expressed as a function of headgroup polymer volume fraction using eqn (3).Then eqn (4) is fitted to the experimental data by adjusting the three parameters N, Hi and Si with a least-squares program. The theoretical 1.c.b. for C,E, in water is shown in fig. 6(a). The theoretical boundary deviates slightly, both in form and in the position of the critical point, from the experimental values. This is most probably due to the neglect of terms dependent on solute volume fraction in eqn (3). A more comprehensive analysis of the exact 1.c.b. shape is not pursued here. The following parameter values define the binary 1.c.b.: N = 4520, Hi = - 11.63 kJ mol-1 and Si = -39.33 J K-l mol-l. The error in the Hi and Si values due to the deviation of2956 POI,Y(OXYETHY LENE) SURFACTANTS the theoretical boundary from the experimental values is 15"/<;.The agreement is sufficiently good to study shifts of the boundary with temperature. Eqn ( 5 ) gives a critical concentration of 1.5 vol for the oxyethylene polymer, which is satisfied (a 1 ~ 0 1 % ) by the theoretical boundary (fig. 6). Note that small deviations of the theoretical boundary minima from the condition given by eqn ( 5 ) have been observed before.13 It is not possible in our case to deduce a micellar size from the value of AT, partly because of the approximations mentioned above, but also because the minimum ('4:) in the theoretical 1.c.b. deviates from the experimental minimum [eqn (4)].For C,E, we have7 VEo = 328 A3 and Vw = 30 A3. The definition of N. would give an aggregation number ( u ) of 413, which is roughly five times the aggregation number obtained from neutron-scattering studies' (80 10). For the salt 1.c.b. we fix N = 4520, and thus assume that the micellar size is not affected by added salts. This is supported by the neutron-scattering results for C,E,. which show that the micellar size remains rather constant with increasing temperature, even if the oligo- oxyethylene chains are dehydrated and thus change conformation. It was seen in fig. 1-3 that salts shift the position of the binary 1.c.b. in temperature with only small changes in the boundary shape. We thus retain the binary 1.c.b. shape and explain the observed deviations from this shape by a change in the local salt concentration due to salt-deficient or salt-rich regions around the oxyethylene chains.This implies that the 1.c.b. would be essentially shifted in a parallel manner if the bulk salt concentration was the same as a function of surfactant concentration. The binary 1.c.b. is thus simply shifted in temperature by amounts given by the respective phase separation temperature at low (1 vol ) surfactant concentration. Table 2 summarizes values obtained for Hi and Si for two salt concentrations, together with the respective percentage changes. Also included in the table are values for the interaction parameter cr) at 25 "C. The calculations show that moderate changes in the interaction between the micelles and the aqueous solvent can explain the observed 1.c.b.shifts. Even if the hydrophobic regions do not determine the 1.c.b. shifts they affect the magnitude of Hi and Si. It is thus necessary to study salt effects on particles with hydrocarbon domains of different sizes before a comparison can be made between the absolute values of Hi and Si and experimental quantities. We propose in the following a simple mechanism to account for the observed 1.c.b. shifts. Cations are known to have a strong hydration with a relatively slow exchange of hydration water They are also known to be strongly repelled from a dielectric discontinuity, e.g. from a water/air or water/oil interface, by dielectric image-charge f o r ~ e s . ~ ~ - ~ O Anions show a larger variation, so that the repulsion of the ion decreases in the order F-, C1-, Br- and I-. Iodide can even show a net attraction.This provides an explanation for the strong variation of the salt effect with anion; for example, Aveyard and Heselden have devised a semi-empirical approach for the calculation of salting coefficients which is based on the effect of the electrolyte on the surface tension of the water.42 We propose to discuss the solubility shifts in terms of two separate steps. First is the formation of regions around the oligo-oxyethyleae chains with either lower or higher salt concentration as compared with the bulk. The existence of such regions is entropically unfavourable. This has recently been proposed as the reason for salt-induced miscibility shifts by Garvey and R ~ b b ~ ~ and by Florin et aZ.44 Secondly, the existence of, for example, a salt-deficient region leads to aK.WECKSTROM AND M. ZULAUF 2957 reduction in the oxyethylene-water interaction with increasing salt concentration, in order to satisfy the chemical-potential equilibrium of the water. Fluoride and chloride ions induce the formation of salt-deficient regions around the oligo-oxyethylene chains. The salts NaC1, KCl, CsCl and NaF thus salt-out the surfactant strongly by (mainly) a dehydration mechanism, with characteristic changes in the enthalpic interaction between solute and solvent. These salts also show a marked deviation in 1.c.b. shape from the binary 1.c.b. shape (fig. 6). With NaBr and LiCl the changes in 1.c.b. shape are small, even at 1.3 mol dm-3 salt, which suggests the absence of sai t-deficient regions and renders other mechanisms possible, e.g.these involving the ethylene parts of the molecule. The salt-induced partial-miscibility shifts thus reflect changes in the specific accommodation of the oxyethylene chains in the solvent structure. K.W. acknowledges financial support from the Academy of Finland and thanks Prof. I . Danielsson and Dr D. Worcester for interesting discussions. We also thank iP referee for constructive suggestions. D. J. Mitchell, G. J. T. Tiddy, L. Waring, T. Bostock and M. P. McDonald, J . Chem. Soc., Faraday Trans. I , 1983, 79, 975. C. Tanford, ‘The Hydrophobic EJgect (Wiley, New York, 1980). M. Corti and V. Degiorgio, Opt. Commun., 1975, 14, 358. M. Corti and V. Degiorgso, J. Phys. Chem., 1981, 85, 1442.M. Corti, V. Degiorgio and M. Zulauf, Phys. Rev. Lett., 1982, 48, 1617. T. Nakagawa, in Non-ionic Surfactants, ed. M. J. Schick (Marcel Dekker, New York, 1967), chap. 17. M. L. Huggins, J . Chem. Phys., 1941, 9, 440. lo M. L. Huggins. J . Phys. Chem., 1942, 46, 151. l 1 P. J. Flory, J. Chem. Phys., 1941, 9, 660; 1942, 10, 51. l2 P. J. Flory, 1)iscuss. Furaday Soc., 1970, 49, 7. l 3 P. J. Flory, Principles of Polymer Chemistry (Cornell University Press, Ithaca, New York, 1953). :’ V. Luzzati, W. Mustacchi and A. Skoulios. Discuss. Faraday Soc., 1958, 25, 43. ’ M. Zulauf and J. P. Rosenbusch, J . Phys. Chem., 1983, 87, 856. M. Grabo, Ph.D. Thesis (University of Basel, 1983). L.. Marszall, Tenside Deterg., 1981, 18, 25. l 6 F. Hofmeister, Arch. Exp. Puthol.Pharmakol., 1888, 24, 247. I7 F. Tokiwa and T. Matsumoto, Bull. Chem. Soc. Jpn, 1975,48, 1645. IH M. J. Schick, J . Colloid Sci., 1962, 17. 801. H. Schoit and S . K. Han, J. Pharm. Sci., 1977, 66, 165. ‘Lo 13. Schott, A. E. Royce and S. K. Han, J . Colloid Interface Sci., 1984, 98, 196. 1’. G. Nilsson and B. Lindman, J . Phys. Chem., 1983, 87, 4756. 22 E. J. Staples and G. J. T. Tiddy, J . Chem. Soc., Faraday Trans. I , 1978, 74, 2530. 23 M. Corti, V. Degiorgio, J. B. Hayter and M. Zulauf, Chem. Phys. Lett., 1984, 109, 579. 24 M. Zulauf, K. Weckstrom, J. B. Hayter, V. Degiorgio and M. Corti, in Proc. 5th Int. Symp. Surfactants in Solution, Bordeaux 1984 (Plenum Press, New York, in press). 25 M. Zu!auf, K. Weckstrom, J. B. Hayter, V. Degiorgio and M. Corti, J. Phys. Chem., in press. 26 .I. B. Hayter and M. Zulauf, Colloid Polym. Sci., 1982, 260, 1923. 27 H. Schott. J . Colloid Interface Sci., 1973, 43, 150. p y 13. E. Verrali, in Wuter, A Comprehensive Treatise, ed. F. Franks (Plenum Press, New York, 1973), 3o F. A. Long and W. F. McDevit, Chem. Rett., 1952, 52, 119. 31 R. Koningsveld and L. A. Kleintjens, J. Polym. Sci., Polym. Symp., 1977, 61, 221. 32 R. Kjellander and E. Florin, J . Chem. Soc., Faraday Trans. 1, 1981, 77, 2053. :x3 R. Kjellander, J . Chem. Soc., Faraday Trans. 2, 1982, 78, 2025. 31 J. Goldfarb and L. Sepulveda, J. Colloid Interface Sci., 1969, 31, 454. J. S. Walker and C. A. Vause, J . Chem. Phys., 1983, 79, 2660. ‘j6 c‘. A. Vause and J. S. Walker, Phys. Lett. A , 1982, 90, 419. j 7 L. A. Kleintjens, R. Koningsveld and M. Gordon, Macromolecules, 1980, 13, 303. R8 R. Aveyard and S. M. Saleem, J. Chem. Soc., Faraday Trans. I , 1976, 72, 1609. 0. Ya. Samoilov, Discuss. Faraday Soc., 1957, 24, 141. vol. 3, p. 211.2958 POLY(OXYETHYLENE) SURFACTANTS 39 R. Aveyard, S. M. Saleem and R. Heselden, J . Chem. SOC., Faraday Trans. 1, 1977,73, 84. 40 B. E. Conway, Adv. Colloid Interface Sci., 1977, 8, 91. 41 L. Pauling, The Nature of the Chemical Bond (Cornell University Press, Ithaca, 3rd edn, 1960). 42 R. Aveyard and R. Heselden, J. Chem. SOC., Faraday Trans. 1, 1975,71, 312. 43 M. J. Garvey and I. D. Robb, J . Chem. SOC., Faraday Trans. I , 1979, 75, 993. 44 E. Florin, R. Kjellander and J. C. Ericksson, J. Chem. SOC., Faraday Trans. I , 1984, 80, 2889. (PAPER 5/055)