ed to remain the same all along the series, the only .... ed, as the quantity y~lp/r(13C-31P) 6 isl~. 0"4O/o .... Herrmann KW (1966)J Colloid Interface Sci 22:352. 8.
Colloid & Polymer Science
Colloid Polym Sci 266:441-448 (1988)
Micellar properties of zwitterionic phosphobetaine amphiphiles in aqueous solution: Influence of the intercharge distance Y. Chevalier, L. Germanaud1) and P. Le Perchec Laboratoire des Mat&iaux Organiques, CNRS, BP 24, Vernaison, France 1) Centre de Recherche Elf Sotaise, BP 22, Saint Symphorien d'Ozon, France
Abstract: Zwitterionic amphiphiles of the general formula H(CH2)y +N(CH3)2(CH2)n PO2C6H~-, where the number of intercharge methylenes n is varied, were studied in dilute aqueous solution. Their critical micellar concentrations show a peculiar variation with n, first increasing as n varies from 1 to 4 and then slowly decreasing as methylenes are added up to 10. This behavior is interpreted as being the consequence of two opposite contributions. The first is the classical CMC lowering due to the increase of hydrophobic character with the total number of methylene groups in the surfactant molecule. The second contribution is the increase in the dipole moment of the zwitterionic headgroup with n, leading to stronger dipole-dipole repulsions between headgroups at the micellar surface. Experimental results suggest that the dipole moment does not increase linearly with n because of the polymethylene chain flexibility. This is supported by 13CNMR relaxation experiments.
Key words: Zwitterionic amphiphiles, phenylphosphinatobetaines, micelles, 13C__NMR relaxation.
Introduction Zwitterionic amphiphiles [1] present remarkable stability against changes in external conditions. In contrast to ionic surfactants, they are rather insensitive to ionic strength. Even with multivalent ions, they do not precipitate in hard water media [2-5]. This is, of course, due to their overall electrical neutrality. They differ from polyoxyethylene nonionic amphiphiles in that they are not so sensitive to temperature and they generally do not show cloud point phenomena upon heating. In spite of these obvious advantages, few studies of zwitterionic amphiphiles exist. Most work has dealt with betaines (ammonioacetates) which are most frequently chosen at random [6-11]. Since no systematic studies have been carried out, it was not evident that any correlations between the chemical structure and physical properties existed. Some series of homologous compounds have been worked out, but this work was confined to small zwitterionic headgroups, namely compounds having from 1 to 3 methylenes 371
separating the charged functions. The CMC decrease upon lengthening the hydrophobic chain of alkyldimethylammonioacetates has been described [12-16]. The free energy of micellization per added methylene was found to be - 2.7 to - 2.9 kJ/mol, which is close to that of other nonionic amphiphiles. There is, however, no reason for that result to be general. The most extensive studies are those of Laughlin [17-20] who related phase diagrams to the hydrophilicity (hydration) of the anionic part of the zwitterion. Again, this work is only concerned with short intercharge distances. The present work addresses quantitatively the influence of the intercharge distance on the micellar properties of zwitterionic amphiphiles. The intercharge bridging unit is a polymethylene chain - (CH2)n- , containing from n = 1 to 10 methylenes. For that purpose, a complete series of homologous amphiphilic compounds have been prepared and their tensioactive properties have been studied as aqueous solutions. The zwitterionic amphiphiles are new phosphobetaines,
Colloid and Polymer Science, VoL 266. No. 5 (1988)
442
namely (alkyldimethylammonioalkyl)-phenylphosphinates of the chemical formula O-
CH 3
I
I
H ( C H 2 ) y +~'~(CH2)n ~3= O CH3
C6H5
n was varied from I to 10 and y was set at 12 or 18. They are abbreviated in the following to CyNnPr
ured using benzene as a reference (Rg0o= 3.165 10-5 cm -1) and are averages over 20 measurements: No angular dependence was observed. 3. NMR relaxation: 13C NMR measurements were carried out on a Bruker WP 80 spectrometer under broadband IH decoupling conditions. Temperature was regulated at 30 ~ _+ 1 ~ The way in which the spectral lines were assigned is described in the Appendix. The spin-lattice relaxation times T1 were measured by the inversion-recovery method. T 1 w a s calculated through a three-parameters fit of 14 values of I(t) to the relationship [24-26]
I(t) = I(oo)(1 + ke-t/rl).
(1)
Equilibrium recovery curves were always observed as monoexponential decays. Materials and methods
Materials The general step-by-step procedure of synthesis of CyNnP# compounds is described elsewhere [21,22]. Starting from a,c0dibromoalkanes, dimethyldodecyl (or octadecyl) amine and diethylphenylphosphonite, this procedure affords very pure compounds free of inorganic salts with different interionic distances (n = 1 to 10).
Methods 1. The Krafft boundaries were determined by direct observation of the dissolution temperature of a series of mixtures of water and crystals as they were slowly heated up in steps of 0.5 ~ 2. Light scattering: The light scattering device is composed of a Spectra Physics Argon laser working at the 5145 A green line and a Malvern goniometer, as described in Ref. [23]. The laser power was 300 mW. The photomultiplier was set perpendicular to the incident light beam. The temperature was regulated at 25 ~ _+1~ for all the compounds except C12N1Pr and C12N2Pr which were heated at 30 ~ (above their Krafft temperature). Raleigh ratios were meas-
~" J
C12N 1 P r
k
C 12N Z P(~
Results and discussion
Krafft boundaries All the compounds are miscible with water at room temperature except those having intercharge bridges of n = 1 or 2 methylenes. The phase diagrams of C12N1Pr and C12N2Pr were obtained and are shown in Fig. 1. The Krafft points of the n = 1 and n = 2 compounds are 27.5 ~ and 28 ~ respectively. The uncertainty is 0.5 ~ because of the step by step procedure of heating. Homologues with an octadecyl chain were not studied since they are thought to have higher Krafft temperatures than the dodecyl compounds. The low Krafft temperatures of large intercharge bridged compounds (n _> 3), even with an octadecyl chain, are rare and promising for solubilization applications. As a comparison, the Krafft temperature of octadecyltri-
f
E
t --cmc
-3'
rafft -4,
9 " " Tn~,~,d
I'~ " ~
20
"
"
" T~mrature
pc)"
Fig. 1. Phase diagrams of C12NIP# and C12N2P~5 showing the Krafft boundaries (the concentration scale is logarithmic for better visibility in the low concentration region)
Chevalier et aL, Micelles of zwitterionic phosphinatobetaine arnphiphiles Table 1. Values of critical micellar concentrations of studied phenylphosphinatobetaines C y N n P r in water as measured by light scattering.
443
C~ (mu)
C12NnPQ
1.5 n
CMC (raM) for y = 12
CMC (raM) for y = 18
1 2 3 4 5 6 8 10
0.5 1.0 1.4 1.56 1.4 1.1 0.60 0.47
0.28 0.60 0.60 0.57
1.0,
0.5,
methylammonium bromide was measured as 39 ~ (measured in the concentration domain where the Krafft boundary is at constant temperature). The high Krafft point of C12N2Pq5 is in good agreement with observations of a gel phase (L/~) in concentrated solutions [27]. Critical micellar concentrations
The critical micellar concentration (CMC) values were measured by means of light scattering and are collected in Table 1. They are shown in Fig. 2 as a function of the number of intercharge methylenes n (for y = 12 only). The CMC values are all around 10-3 M for y = 12 and 10 -4 M for y = 18. The CMC range of dodecyl compounds is the same as many other zwitterionic amphiphiles reported in the literature [5, 716]. These values are closer to those of ionic surfactants with similar alkyl chains (CMC = 2.10 -2 M for dodecyltrimethylammonium chloride [28]) than with those of polyoxyethylene nonionic ones. In that way, although the net charge of the zwitterionic molecule is zero, it acts more like an ionic than a nonionic amphiphile. Indeed, a strong condensation of counterions occurs at the surface of ionic micelles, the order of magnitude of the fraction of bound counterions is generaUy 0.7 to 0.8 [28]. When this fraction approaches one with aromatic organic counterions, as with dodecylpyridinium salicylate, the CMC is reduced to some 10-3 M [29], which is highly consistent with our CMC values of zwitterionic amphiphiles, their bound fraction being exactly 1 (the "counterions" are covalently bound at the micellar surface). The difference of one order of magnitude between the CMC of ionic and zwitterionic amphiphiles can be attributed to the incomplete charge neutralization of ionic micelles by
II Fig. 2. Critical micellar concentrations of C12NnPg5 as a function of n
their counterions. It could then be postulated that zwitterionic micelles closely resemble ionic ones in salt solutions. The CMC values of octadecyl compounds are rather high, as compared to what could be calculated from the CMC of dodecyl compounds and the conventional decrease of the CMC with the alkyl chain length [12-16]. CMC values of some 10-6 M are indeed expected, which is a factor of 100 less than the measured ones. This point demands further clarifications as regards the validity of light scattering for very low CMC measurements. Preliminary surface tension experiments show a good agreement with light scattering measurements for dodecyl compounds while CMC values of octadecyl ones are in error and the expected CMC range (10 -6 M ) is actually obtained. Such large discrepancies between surface tension and light scattering CMC values (a factor of 10) have been pointed out with nonionic amphiphiles [30], and the formation of small enough micelles for light scattering to be weak was put forward as the origin of this. Within the conventional picture of micelle formaton, light scattering should detect the correct CMC as measurements are performed, down to the observation of solvent thermal scattering and such discrepancies, if not due to experimental artifacts (during filtration), reveal some physical phenomena. The results obtained from light scattering are of significance, then. As two very different transitions are observed, both cannot be called "CMC". The results of surface tension experiments seem to be closer to the standard definition of a CMC,
444
Colloid and Polymer Science, VoL 266. No. 5 (1988)
Q
b
|
|
|174
|
|
|174
Parallel
~iperallel
REPULSION
ATTRACTION
WATE R
YDROCARB~N
i I(miceller~ core J (
while the results of light scattering would indicate a morphology transition in the aggregate size or shape. The remarkable feature shown in Fig. 2 is that the CMC increases as n varies from 1 to 4, and decreases from this maximum for larger values of n. The CMC is related to the free energy of micellizationAGmicby the approximate relation [28] AGmic = RT In (CMC).
(2)
Because of the negative free energy of transfer of a methylene from water to a micelle, addition of a methylene to a surfactant molecule generally causes a CMC decrease. This is actually not the case: the CMC increases with n for the low n compounds. In other words, the CMC increases in spite of an increased overall hydrophobicity (as estimated from the chemical formula) of the molecules. This may seem peculiar and is actually in total disagreement with the hydrophile-lipophile balance (HLB) concept, as calculated from increments per group [31]. For the CMC to increase with n, another contribution to the free energy of micellization, other than just the free energy of transfer of methylenes from water to alcane AGh, has to be considered. This additional contribution has to be of the opposite sign to AGh and its absolute value has to be larger than that ofA Gh for I < n < 4 and lower for n > 4. Solvation of molecules in water is expected to remain the same all along the series, the only strongly hydrated group being the phosphinate one. It is suggested that the repulsive interaction between amphiphilic molecules in the micellar aggregate comes from electrostatic dipole-dipole repulsions between headgroups.
Fig. 3. Schematic representation of repulsive and attractive dipole-dipole interactions showing that they are repulsive at a micellar interface
Indeed, dipole-dipole interactions are repulsive or attractive according to their parallel or antiparallel mutual orientation. At the micellar surface, the alkyl chain dives into the micellar hydrophobic core, compelling the dipoles to a parallel average mutual orientation. The interhead group interactions are thus repulsive (Fig. 3). The minimization of this interaction is possible by making the zwitterionic moiety lay parallel to the interface. This situation occurs with lecithins in lamellar structures (vesicles or liposomes) [32, 33]. It is possible in this case because of the large area occupied by the two hydrocarbon chains of lecithins. It is not believed that such a situation may occur with single chain compounds, especially when the zwitterionic moiety is very bulky (n is varied up to 10 methylenes). Moreover, the conformational entropy has to be as large as possible. Disorder is the general rule for liquid solutions of flexible molecules, so that many conformations will be present unless a huge enthalpic gap exists for some preferred conformations. If a more precise theory is to be drawn, the significant quantity that enters the dipolar contribution to AGmic is the projection of the average dipole moment onto the normal to the interface (neglecting curvature effects). The average orientation, namely the order parameter S = (3 cosaO - 1)/2, and the average dipole strength have to be known for that purpose. The total free energy of micellization AGmicis the sum of a hydrophobic term AGh, a dipolar term AGaip and other terms that are expected to be independent of H:
AGmir = AGh + AGaip + constant terms.
(3)
Chevalier et al., Micelles of zwitterionic phosphinatobetaine amphiphiles
AGmic
O.
445
shown this important feature [35, 36]. The electrostatic attraction between the two opposite charges of the flexible dipole may also make folded conformations more probable. One must be aware, too, that steric constraints at the interface do not allow any conformational possibility. An experimental measurement of the intercharge chain flexibility is thus the only reliable way for discussion to be continued. Such an experiment has been performed by 13C N M R relaxation.
13C N M R relaxation
The 13C N M R spin-lattice relaxation rates (1/T~) of a methylene are dominated by the dipolar interaction Fig. 4. The two n-dependent contributionstO the free energy of of the i3C spin with 1H nuclear spins of its two directly bound protons. The static geometry of the magnetic micellization interaction is thus well established (r13C - H --=1.09 .~). The contribution of the 1/2 spin of the 31p nucleus to the relaxation rate of the a-methylene can be neglectAGh is well documented [34]. It is a linear function ed, as the quantity y~lp/r(13C-31P)6 isl~ 0"4O/o of 2y~H/r(lBC---1H)6, taking r(~3C-Blp) = . [37,38]. of the number of methylenes n: The 13C relaxation is thus only sensitive to molecular AGh = A + B n . (4) motions. Slow and fast motions with very different time scales are observed in micellar solutions and a B is estimated from the free energy of transfer of complete interpretation of these motions requires the methylenes from water to alcanes as - 3 kJ/mol [34]. measurement of relaxation rates at several different AGaip can thus be obtained from the experimental data Larmor frequencies [39-44]. It is intended here to and this known contribution (Fig. 4). ZlGdip is a observe the qualitative features of molecular motions strongly increasing function of n at low n values and of the intercharge polymethylene moiety using 13C reaches a plateau as n becomes large. This effect can be N M R relaxation data obtained at one Larmor frequenattributed either to a change in the orientation of cy only. It is made possible because of the extensive dipoles, or to a non linear increase of the dipole data already known about molecular motions in micelmoment with n because of the intercharge chain flexi- lar systems [39-44]. It should be firstly noticed that the bility. Orientation should change from a direction per- relaxation rates of the 13C nuclei of the methyl and pendicular to the interface (or close to it) for low n methylene groups, directly bound to the nitrogen values, to a direction parallel to the interface at large n atom, are of comparable magnitude all along the series values. The origin of such an effect could be the of dodecyl chained surfactants (Fig. 5). As a consequincreased hydrophobicity of the intercharge chain ence, it may be assumed that the slow motion is the with n, which makes the zwitterion lay on the interface same for all these compounds. This absence of variain order to avoid water contact. Although a physical tion of relaxation rate with n could be a cancellation of origin can be ascribed to this orientation effect, steric contradictory effects, but it would be a great coincidconstraints act against it. The second origin of variation ence. This slow motion is a combination of the rates of of AGdip seems more likely. The first one cannot be rotation of the micelles and of diffusion of the surfaccompletely rejected nevertheless, and the most reason- tant molecules at the curved interface [45]. This result able statement is to say that both are operative. As the suggests that the micellar size does not depend much number of possible conformations of a polymethylene on the intercharge distance n. On this basis, the relaxachain undergoing gauche-trans isomerizations grow as tion rate variations along the intercharge polymethyl3n, a linear growth of the dipole moment with n is com- ene moiety are attributed to differences in the anisopletely unrealistic. Dipole moment measurements on tropic fast motion rates and average orientation of non amphiphilic zwitterionic molecules in water have 13C-1H bonds relative to the micellar radius. Figure 6 ~)
~
4.
5
Im
(~
"r
8
9
10
Colloid and Polymer Science, Vol. 266. No. 5 (1988)
446
~
i/T,~ ) lS
+++
+
16
.7
N~
O~
5u
n
n=3
~n=6 n:4
0.,~
'
"1" ~ " .
n:5
.~~.n:8
Fig. 5. Spin-lattice relaxation rates (1/T1) of 13C nuclei adjacent to the nitrogen atom for all the dodecyl compounds Carbon Number X
shows the variation of relaxation rates of 13C nuclei of the intercharge arm. It is represented in such a way that only variations are displayed: relaxation rates of the 13C are relative to that of 13C1(which is directly bound to the nitrogen atom). The assignment of the spectra is reported in the appendix. It may be firstly observed that the intercharge arm is rigid for n < 3 since all the 13Chave the same relaxation rate. The increased flexibility with the number of the carbon atom is observed as decreasingvalues of relaxation rate, which reveals decreasing order parameters or correlation times of fast motions. Chain flexibility is thus observed for n > 3, which is in agreement with the maximum value of the CMC at n = 4. The effective dipole moment (that which is operative for interhead group interactions) grows linearly with n from n - 1to n = 3 and departures from this behavior are only allowed for n > 3. The flexibility is nevertheless somewhat hindered, as shown by the plateau reached for large n values. The relaxation rate decrease is less than that of a free chain, like the alkyl chain of surfactants in micelles [39-44]. This reduced mobility can be attributed to the bulkiness and strong hydration of the terminal phenylphosphinate group, but may also be caused by the electrostatic attraction of the anionic phenylphosphinate group by the positive charge of the nitrogen atoms. A more quantitative interpretation would require additional experimental data which will be obtained later from relaxation rate measurements at different Larmor frequencies. Conclusion
A previously unknown behavior of zwitterionic amphiphiles has been observed in a new series of
Fig. 6. Relative spin-lattice relaxation rates (see text) as a function of the number of the carbon atom for the different compounds studied. The numbering of the carbon atoms starts from the nitrogen side
phenylphosphinatobetaines. As the number of intercharge methylenes n is increased, the CMC increases, reaches a maximum at n -- 4 and decreases. This peculiar behavior has been explained by the intervention of two opposite effects: the continuous increase ofhydrophobicity with n, and the increase of the dipole moment of the zwitterionic part which is modulated by chain flexibility. This flexibility has been evidenced as occuring for n values larger than 3, in good agreement with the CMC variations. 4-9
CHiCHi-----CHiN
1211
phenyl group
115
1
CH=) C2H~--- -- CH~PO2(~-
1'
6'
1[
l~stppm) from "rMS ~
11
~s . . . .
5
Fig. 7. 13C NMR spectrum of a 0.5 M solution of C12N6P# in CD3OD
Chevalier et al., Micelles of zwitterionic phozphinatobetaine amphiphiles
447
Table 2. Assignment of the 13C NMR spectra run in CD3OD. Chemical shifts 6 are given in ppm relative to TMS. Carbon atoms are numbered from the nitrogen atom Dodecyl group (for all the compounds): 13C number NCH3 1
2
3
4-9
10
11
12
6 (ppm)
23.18
27.13
30.24 30.49
32.93
23.18
14.39
51.0
64.99
Intercharge polymethylene moiety 13C number
1
2
3
5
4
6
C12N1Pr
6r 64.47 6~1~ 67.80
C12N2P#
6~xp 61.16 6caI~ 58.90
27.30 26.30
C12N3Pr BzN3Pr
6oxp 65.50 6exp 65.74 6r 66.46
17.60 17.93 16.84
29.46 29.68 30.06
C12N4Pr
6exe 64.74 6c~1r 64.74
24.26 24.59
21.01 20.79
32.40 32.87
C12N5P#
6~xp 64.98 6r 64.99
22.65 22.93
28.33 28.54
23.50 23.90
33.00 33.36
C12N6Pr BzN6Pr
6exp 65.26 6exp 65.54 6r162 64.99
23.01 23.03 23.18
26.64 26.64 27.13
31.05 31.25 31.65
23.70 23.69 24.15
33.34 33.32 33.36
C12N8P#
6exp 65.21 6talc 64.99
23.29 23.18
27.23 27.13
-
31.95 31.90
6exp 64.92 6c~1c 64.99
23.14 23.18
27.03 27.13
m
D
m
D
m
6exp 14.34 6c~1c 14.39
23.41 23.18
32.78 32.93
C12N10Pr
8Pr ~)
m
30.00 30.24
30.00 30.24
32.16 31.90
7
8
24.09 24.15
33.50 33.36 32.07 31.90
24.00 34.15
9
10
23.93 24.15
33.65 33.36
33.51 33.36
~) numbered from the methyl group
Appendix For the purpose of assignment, 13C NMR spectra were run in deuterated methanol. In this solvent, no aggregation occurs and lines are narrow (Fig. 7). Internaltetramethylsilane(TMS) was used as a chemical shift reference. Assignment was carried out considering all the studied compounds in such a way that the influence of the substitutents is the same all along the series, depending only on their relative position from the 13C (only ~e,/~, y and 6 influences have been taken into account). This method is supported by the work of Grant and Paul [46] which shows that the shift of a 13C line can be written as
6(13Ck) = Constant (k) + •
nikAi.
(5)
The influence of methyl and methylene groups was taken to be that described by Lindeman and Adams [47, 48]. Some additional
compounds were synthesized to ensure a correct assignment: two ammoniophosphinates bearing a benzyl group instead of the dodecyl one (BzNnPr dodecyltrimethylammonium bromide and sodium octylphenylphosphinate (8PgS). The 13C NMR shift of the dodecyl moiety was the same all along the series and was consistent with that of dodecyltrimethylammonium bromide, the Lindeman and Adams empirical parameters and the ordering pyramid [49]. The influence of the replacement of a proton by a dimethylammonium group was then deduced as a = + 50.6 ppm,/~ = 0.0 ppm andy = - 5 . 8 ppm. Sodium octylphenylphosphinate gives, in the same way, ce = +19.5 ppm,/~ = +1.5 ppm and y = - 0 . 5 ppm for the replacement of a proton by the phenylphosphinate moiety. The experimental and calculated 13C chemical shifts of all the compounds are listed in Table 2. The spectra were run with 8000 points, providing an uncertainty of 0.05 ppm. The actual uncertainty is, however, somewhat larger because of concentration dependence of chemical shifts. The agreement between experimental and calculated chemi-
448 cal shifts is thus good, the accuracy of the empirical parameters not being excellent by themselves. A simple conformational change from a trans to a gauche conformation indeed leads to a y effect of the order of 5 ppm [50-52]. 13C nuclei are coupled with the phosphorus atom up to 3 bonds. 1Iv_ c = 85 Hz for C12N1Pr ll~_ c = 87 Hz for C12N2Pr and llv- c = 95-100 Hz for the other compounds. For all compounds, 2lp-c = 2-3 Hz and 3Iv c = 13-16 Hz. References
1. Bluestein BR, Hilton CL (1982) Editors of Amphoteric Surfactants, Surfactant Science Series 12, Dekker, New York 2. Parris N, Weil JK, Linfield WM (1973) J Am Oil Chem Soc 50:509 3. Noble WR, Linfield WM (1980) J Am Oil Chem Soc 57:368 4. Fernley GW (1978)J Am Oil Chem Soc 55:98 5. Toil K, Nakagawa T (1963) Kolloid Z Z Polym 189:50 6. Toil K, Nakagawa T (1963) Kolloid Z Z Polym 188:47 7. Herrmann KW (1966)J Colloid Interface Sci 22:352 8. Dahanayake M, Rosen MJ (1984) ACS Symp Set 253:49 9. Rosen MJ, Zhu BY (1984) ACS Symp Ser 253:61 10. Zhao F, Rosen MJ (1984)J Phys Chem 88:6041 11. Kamenka N, Haouche G, Faucompr6 B, Brun B, Lindman B (1985) J Colloid Interface Sci 108:451 12. Toil K, Kuriyama K, Nakagawa T (1963) Kolloid-Z Z Polym 191:48 13. Beckett AH, Woodward RJ (1963) J Pharm Pharmaco115:422 14. Molyneux P, Rhodes CT, SwarbrickJ (1965) Trans Faraday Soc 61:1043 15. Swarbrick J, Daruwala J (1969)J Phys Chem 73:2627 16. Swarbrick J, Daruwala J (1970)J Phys Chem 74:1293 17. Laughlin RG (1978) Adv Liquid Crystals 3:41 18. Laughlin RG (1978) Adv Liquid Crystals 3:99 19. Laughlin RG (1981) J Soc Cosmet Chem 32:371 20. Nilsson PG, Lindman B, Laughlin RG (1984) J Phys Chem 88:6357 21. Germanaud L (1984) Thesis, University of Lyon I 22. Germanaud L, Brunel S, Chevalier Y, LePerchec P, Bull Soc Chim Fr, in press 23. Arnaud B, Legr6 J, Drifford M (1974) J Chim Phys 71:591 24. Sass M, Ziessow D (1977)J Magn Reson 25:263 25. Canet D, Levy GC, Peat IR (1975)J Magn Reson 18:199 26. Kowalewski J, Levy GC, Johnson LeRF, Palmer L (1977) J Magn Reson 26:533 27. Gallot B, Germanaud L, Chevalier Y, LePerchec P (1988)J Colloid Interface Sci 121:522 28. Lindman B, Wennerstr/Sm H (1980) Top Curr Chem 87:1
Colloid and Polymer Science, Vol. 266. No. 5 (1988)
29. Angel M, Hoffmann H, LSbl M, Reizlein K, Thurn H, Wunderlich I (1984) Progr Colloid Polym Sci 69:12 30. Balmbra RR, Clunie JS, CorkillJM, GoodmanJF (1962) Trans Faraday Soc 58:1661 31. Davies JT (1957) Proc Int Congress Surface Activity 1:426 32. Biildt G, Gaily HU, Seelig A, SeeligJ, Zaccai G (1978) Nature 271:182 33. Biildt G, Gally HU, SeeligJ, Zaccai G (1979)J Mol Bio1134:673 34. Tanford C (1973) The Hydrophobic Effect: Formation of Micelles and Biological Membranes, Wiley 35. Kirchnerova J, Farrell PG, Edward JT (1976) J Phys Chem 80:1974 36. Kaatze U, Bieler H, Pottel R (1985) J Mol Liquids 30:101 37. Kulpe S, Seidel I (1979) Krist Tech 14:1089 38. Kulpe S, Seidel I (1980) Krist Tech 15:149 39. Wennerstr6m H, Lindman B, S6derman O, Drakenberg T, Rosenholm JB (1979)J Am Chem Soc 101:6860 40. Ahln~is T, S6derman O, Hjelm C, Lindman B (1983) J Phys Chem, 87:822 41. Ahln~is T, S6derman O, Walderhaug H, Lindman B (1984) In: Mittal KL, Lindman B (eds) Surfactants in Solution, Henum, New York 1:107 42. Lindman B, Ahln~is T, S6derman O, Walderhaug H, Rapacki K, Stilbs P (1983) Faraday Discuss Chem Soc 76:317 43. Walderhaug H, S6derman O, Stilbs P (1984) J Phys Chem 88:1655 44. S/Sderman O, Walderhang H, Lindman B (1985) J Phys Chem 89:1795 45. Alexandre M, Fouchet C, Rigny P (1973) J Chim Phys 70:1073 46. Grant DM, Paul EG (1964) J Am Chem Soc 86:2984 47. Lindeman LP, Adams JQ (1971) Anal Chem 43:1245 48. Wherli FW, Wirthlin T (1976) Interpretation of Carbon-13 NMR Spectra, Heyden 49. Bengsch E, Perly B, Deleuze C, Valero A (1986) J Magn Reson 68:1 50. Grant DM, Cheney BV (1967) J Am Chem Soc 89:5315 51. Cheney BV, Grant DM (1967)J Am Chem Soc 89:5319 52. Tonelli AE (1978) Macromolecules 11:565 Received September 3, 1987; accepted October 5, 1987 Authors' address: Y. Chevalier Laboratoire des Mat~riaux Organiques CNRS BP 24 69390 Vernaison, France
