Composition and Energy Relationships for Some Thin Lipid Films and the Chain Conformation in Monolayers at Liquid- Liquid Interfaces BY D. M. ANDREWS,* E. D. MANEV t AND D. A. HAYDON$ Laboratory of Biophysical Chemistry and Colloid Science University of Cambridge Free School Lane Cambridge CB2 3RT England Received 2nd April 1970 Optically black films have been formed in aqueous media from solutions of glyceryl mono-oleate in aliphatic hydrocarbons. The thicknesscs of the hydrocarbon cores of the films were estimated from electrical capacitance measurements and the compositions from interfacial tension data. The thicknesses and the compositions were found to be interrelated in a simple way and were markedly dependent on the chain length of the hydrocarbon solvent. An electrical potential applied across a liquid film subjects it to a large compressive force under which most types of film became significantly thinner.From thickness measurenients in applied fields the strengths of the steric interactions which stabilize the films were calculated. From a knowledge of the steric interaction together with an estimate of the London-van der Waals forces from contact angle measurements the curve of potential energy against film thickness has been cal- culated for one system. The magnitude of the steric interaction at a given film thickness varies considerably for films of different solvent content. As a consequence a general picture of the time- average conformations of the hydrocarbon chains of glyceryl mono-oleate in the black films and at different hydrocarbon/water interfaces may be deduced.Detailed investigations of optically black lipid films in aqueous media have so far been inhibited by the difficulty of making simple stable films of well-characterized substances. However solutions of glyceryl mono-oleate in aliphatic hydrocarbons and other nonpolar solvents form relatively stable films in aqueous so1utions.l The thickness of these films may be measured by optical and electrical 3 9 methods and the adsorption of the oleate in the film may be estimated from interfacial tension measurements. From these data the composition of the films may be calculated. The free energy of formation of the films may be found from the contact angles between the thin film and bulk interface and may be interpreted to give the magnitude of the London-van der Waals force^.^ A uniform electric field normal to the surfaces of the black film exerts a compres- sional force which may be considerably larger than the London-van der Waals forces and which may produce appreciable thinning of the film.6 This thinning is opposed by the steric interaction of the chains of the adsorbed oleate and from the dependence of film thickness on applied field strength it is possible to deduce the magnitude of the steric interactions and to infer qualitatively the conformation of the oleate chains both in the black film and in the adjacent oil-water interfaces.The free energy against thickness reIationship has been deduced for films of n-decane in saturated sodium chloride stabilized by glyceryl mono-oleate. Films * present address Research Dept. Unilever Limited Port Sunlight Cheshire.j- present address Institute of Physical Chemistry University of Sofia Sofia Bulgaria. 2 present address Physiological Laboratory Univ. of cambridge Downing Street Cambridge. 46 D . M . ANDREWS E D. MANEV AND D . A . HAYDON 47 of glyceryl mono-oleate in other non-polar solvents have also been examined and although these studics have been less detailed they have revealed some important respects in which the solveiit may influence the oleate chain conformation and the composition of the films. EXPERIMENTAL METHODS ‘The films were fornied in a 1 mm hole in a Fluon vessel as described previ~usly.~ For experiments with volatile solvents the top of the cell was sealed by means of a glass lid. Capacitances were measured as described previo~sly.~ When a d.c.bias was applied across the film (from a potentiometer) a 1 pF capacitor was used to isolate the capacitance bridge. Interfacial tensions were determined by the drop-volume technique.’ The activity coefficients of the glyceryl mono-oleate in the more volatile solvents were obtained by vapour pressure osmometry using a Hewlett Packard type 302 Vapour Pressure Osmometer. The sensitivity of this instrument was increased by a factor of ca. 3 beyond the makers’ specifica- tion. In order to obtain good reproducibility for both tension and osmometer results it was necessary to equilibrate the apolar solutions with the appropriate aqueous phase for at least 24 h prior to the experiment. MATERIALS The glyceryl mono-oleate was obtained from Sigma and was found by thin layer chromato- graphy (kindly carried out by Dr.H van Zutphen) to be >99 % pure 1-isomer. No signifi- cant ageing of the interfacial tensions was observed except at low concentrations. The solvents were all of puriss grade and were 2 9 9 % by g.1.c. Before use they were passed through an alumina column to remove trace surface active impurities. A.R. NaCl was roasted at 700°C to remove organic impurities. The water was twice distilled first from a commercial still and then from a Pyrex still fitted with a quartz column condenser and receiver. All experiments were carried out at 20°C. RESULTS The two effects with which this paper is primarily concerned are illustrated in fig. 1 and 2. First the specific capacitance of the film increases considerably when number of carbon atoms in alkane FIG. I .-The specific capacitance of black films formed from solutions of glyceryl mono-oleate (ca.0 aqueous phase 0.1 M NaCl; El aqueous phase saturated NaCl. 12 mM) in normal alkanes. 48 THIN LIPID FILMS applied potential (V) FIG. 2.-The specific capacitance (C 0) and thickness of the hydrocarbon region (8,) of black f i l m under applied potentials. The films were formed from glyceryl mono-oleate in n-decane and the aqueous phase was saturated NaCl. the films are formed from hydrocarbon solvents longer than n-nonane (fig. 1). Secondly for a film formed from n-decane the specific capacitance increases with increasing potential across the film (fig. 2). Further results for other solvents are given in table 1. TABLE 1 - system V(V) C (nF mi-*) & 8 (nm) rp 0.1 M NaCl n-decane satd. NaCl cyclohexane n-hep t ane 2,2,4-trimethyl-pentane n-decane n-tetradecane n-hexadecane CCI 0.0141 0.106 0.0141 0.141 0.0141 0.141 0.0141 0.141 0.0141 0.1061 0.0141 0.141 0.0141 0.141 0.0141 0.141 3.83 4.14 4.42 4.94 3.98 4.47 4.1 1 4.59 4.19 4.40 5.73 5.92 6.51 6.38 5.63 5.63 2.08 2.10 2.11 2.12 2.07 2.09 2.08 2.10 2.10 2.12 2.14 2.14 2.14 2.14 2.15 2.15 4.8 4.5 4.2 3.8 4.6 4.1 4.5 4.05 4.4 3.9 3.3 3.2 2.9 2.95 3.4 3.4 0.53 0.72 0.73 0.70 0.73 1.03 1.18 0.85 D.M. ANDREWS E . D. MANEV A N D D . A . HAYDON 49 All the systems had capacitances and conductances which were frequency dependent. The dispersions were simple in form and corresponded accurately to those which would be expected for two parallel-sided isotropic layers in series. The two layers concerned are the hydrocarbon core of the film and the aqueous phase respe~tively.~~ * * Above and below the dispersion region the capacitance does not vary with frequency and the data here reported were obtained in the low frequency constant-capacitance region (at ca.1 kHz). It was confirmed that a given r.m.s. a.c. potential had a similar influence on the capacitance to the equivalent d.c. potential. When the electric field across a film was changed the re-establishment of equili- brium often took of the order of minutes or hours. This was especially so for films formed from solvents of higher molecular weight and arose evidently from the difficulty of escape of such solvents from the films. Thus when the potential was suddenly increased bright spots of surplus liquid appeared in the film after a period of seconds or minutes presumably from a squeezing-out or disproportionation process.The specific capacitance of the film was constant once the spots were visible but the diffusion of the spots to the edge of the film where they coalesced with the meniscus took much longer.1° In many of the systems the films became markedly less stable under potential differences of more than ca. 150 mV (r.m.s.) and precise capacitance measurements could not be made. The use of saturated NaCl as the aqueous phase greatly helped to minimize inaccuracies in the capacitances for films formed from low-molecular- weight solvents. Unless the apolar and aqueous phases were in perfect equilibrium the capacitances became anomalously high. In saturated NaCl this effect was considerably reduced owing presumably to the lower solubility of the solvents and a wider range of films could be examined.A film thickness was obtained from the specific capacitance C by means of the equation c = &,&/6 where E~ is the permittivity of free space E is the dielectric constant and 6 is the thickness of the hydrocarbon core of the film. The validity of this procedure and in particular the justification for ignoring contributions to the capacitance from the polar group region and the electrical double layers in the aqueous phases has been examined in previous papers.3* * * 9* The hydrocarbon core is regarded as bulk liquid hydrocarbon composed of a mixture of oleyl chains and the appropriate so1vent.12 The proportions of the two components have been estimated from the adsorption data as described below.The dielectric constants of the two components have been assumed additive on a volume fraction basis. Anisotropy in the hydro- carbon has been ignored. Even in hydrocarbon crystals this effect is small and the film structure more closely resembles a liquid than a crystal. The resulting dielectric constants are all close to 2.1 and the calculated thicknesses are shown in fig. 2 and table 1. From measurements of molecular models the thickness of the glyceride polar groups layer was estimated to be 0.45 nm. The total film thickness 2jL was therefore taken to be (6 + 0.9)nm. Such optical measurements as have been reported (e.g. films of glyceryl distearate and n-hexane 13) are consistent with the present estimated thickness. In order to estimate the composition of the black films the colligative properties of glyceryl mono-oleate and its adsorption at the various oil-water interfaces were studied.Glyceryl mono-oleate aggregates strongly in apolar media at concentra- tions above ca. moll-l. When the apolar medium is equilibrated with 0.1 M 50 THIN LIPID FILMS loglo [activity concentration (mol PI)] CC14 - 3.0 -2.5 loglo [activity concentration (mol 1-91 - 3.0 - 2.5 Iog o[activity concentration (mol 1-91 FIG. 3.-hiterfacial tensions of glyceryl mono-oleate in various solvents as a function of log, (activity) (Closed points) and or log, (concentration) (open points). (a) 0 n-decane+0.1 M NaCl ; A n-decane+ saturated NaCl ; 0 0,2,2,4trimethylpentane + saturated NaCl ; v v carbon tetrachloride+saturated NaCl (top and right hand axes). (b) All aqueous phases saturated NaCl.V n-hexadecane ; A n-tetradecane ; El m n-heptane ; 0 0 cyclohexane. NaCl the onset of the aggregation is sufficiently sharp to be described as a critical micelle concentration. When equilibrated with saturated NaC1 however the aggregation is less sharp and there is no clear c.m.c. This point is illustrated by the interfacial tension curves for decane in fig. 3(a). Black films tended to be stable D. M. ANDREWS E. D. MANEV AND D. A . HAYDON 51 only at concentrations above the c.m.c. or aggregation region although films were occasionally sufficiently stable just below this concentration range for some types of measurement to be made. All the films of table 1 were formed above the aggregation region. The adsorption of the glyceryl mono-oleate at the oil/water interface was found by means of the Gibbs equation which as the two solvents were effectively insoluble in each other and the oleate was very strongly adsorbed reduces to l4 where y is the interfacial tension T2 is the surface concentration of the glyceryl mono-oleate and a2 is the activity of the oleate in the oil phase.Curves of y against log, (concentration) are shown in fig. 3 (a) and (b) for each of the systems of table 1. For only four systems however was it possible to determine the activity co- effiicients fi. The latter were obtained from vapour pressure osmometer measure- ments via the osmotic coefficients g by means of the equation -dy = T2RTd In a2 (2) J'd lnfi = J'dg+ l(g - l)d In x2. (3) The results may be seen in fig. 3 (a) and (b). For the remaining systems the vapour pressure osmometer was insufficiently sensitive and there seemed no obvious alter- native method of obtaining the activity coefficients.The adsorption in these instances was therefore estimated by assuming that the activity correction was identical to that for the heptane system. Both determined and estimated surface concentrations are shown in table 2. TABLE 2 - system 0.1 M NaCl n-decane satd. NaCl cyclohexane n-hep t ane 2,2,4-triniethyl-pentane n-decane n-tetradecane n-hexadecane cc14 (darnax 21.8" 28.1 31 .O 27.3* 22.9" 24.8" 24.5" 26.6 r x 10-'8 (molecules m-2) 2.5t 3 .O 3.3 2.9 3.2t 3.4t 3.4.f 2.9 *-(dy/d 1ogloc)max ; t based on activity correction for n-heptane system The volume fraction (p of the oleate chains in the film was calculated as follows. The thickness 6 of a film was estimated from the specific capacitance using an arbitrarily chosen dielectric constant (say 2.1).Assuming that the partial molar volume of the oleate chain in the film was equal to the molar volume of 1-heptadecene in bulk the first approximation to was deduced from T2 the surface concentration of oleate (table 2). From this value of 40 and on the assumption that the dielectric constants of the hydrocarbons in the film were equal to their bulk values a new thickness was calculated and the procedure repeated. The second approximation to (p differed only slightly from the first. The assumption of bulk density for the hydrocarbon in the film was made by analogy with the interior of surfactant micelles in aqueous s o l u t i o ~ i ' ~ - ~ ~ and is consistent with the electrical conductances of the films.12 The assumption that the adsorption in the film was for present purposes identical to that at the bulk interface with which the film was in equilibrium has been justified by theoretical considerations,14 and has also been substantiated experi- mentally for one system by contact-angle measurements.Thus from the contact 52 THIN LIPID FILMS angles for systems below the c.m.c. the difference between the film tension and twice the bulk tension may be found and it may be shown that the variation of this difference with glyceryl mono-oleate activity is extremely small compared to the variation of either of the individual tensions. It then follows l4 that the adsorption at the film and bulk interfaces differs only very slightly (< 1 x). DISCUSSION FREE ENERGY AS A FUNCTION OF THICKNESS In the absence of an electrical potential difference across a film the forces acting normally to the film surface are assumed to be the London-van der Waals compression and the steric repulsion which originates from the interaction of the oleate chains.All other forces such as those from electrical double-layer overlap and from dipole- dipole interactions are assumed to be negligible. Experimental evidence so far available is entirely consistent with this assumption. * For unit area of film the London-van der Waals forces FL are given by F L = -A/6~c6? (4) where A is the Hamaker constant. In the present systems the retardation correction should be less than ca. 15 % and will be disregarded. At equilibrium therefore FL+Fs = 0 (5) where Fs is the steric repulsion per unit area.When a potential Vexists across the film there is an additional force Fe of compression given by Fe = -CV2/26,. (6) At equilibrium F'+&+Fe = 0 and therefore (7) Fs = (A/6n6i)+(CV2/26,). (8) The Hamaker constant for the present systems may be found from contact-angle measurements. At potentials of more than about 50 mV across the film the electrical forces exceed the London-van der Waals forces and for 370 mV applied potential the former are some 50 times the latter. From eqn (8) and a knowledge of the capacitance over a range of applied potential the variation of the steric repulsion force with film thickness may be found. This has been done for all the systems but only for the n-decane + saturated NaCl system have the measurements been taken to relatively high potentials.A plot of Fs against 6 for this system is shown in fig. 4 (inset). The change AAs in free energy of the film due to the steric interaction may be calculated by means of the relationship AAs = - [dLF,d6,. (9) Jo3 The total free energy change AA of the system as the film thins may therefore be written D. M. ANDREWS E. D. MANEV AND D. A . HAYDON 53 AA and its components are shown in fig. 4. The onset of the steric repulsion is so sharp as not to affect appreciably the depth of the minimum. This incidentally justifies the assumption which was made in the calculation of the Hamaker c~nstant.~ For the systems listed in table 1 the steric repulsion against film thickness is shown in fig. 5 ; the rise of the steric repulsion is as steep or more so than in the n-decanef saturated NaCl system.8L (m> FIG. 4.-Free energy changes as a function of film thickness for glyceryl mono-oleate + n-decane films in saturated NaCI. The dashed curves represent the separate London-van der Waals and steric interaction contributions. The inset shows the steric repulsion force Fs also as a function of film thickness. Hamaker constant = 3.48 x J. FILM COMPRESSIBILITY COMPOSITION AND CHAIN CONFORMATION It can be seen from table 1 that the thicker films contain a greater volume fraction of solvent (1 -@) than do the thinner films. In fact as the adsorption of the glyceryl mono-oleate is almost the same in each system (table 2) the film thickness is directly proportional to the amount of solvent in the film. (The finding that F > 1 for the n-tetradecane and n-hexadecane systems is attributed to the inaccuracy in the estimation of the activity correction.) The slopes of the curves in fig.5 are inversely proportional to the compressibility of the films. The thicker films where F is low are thus much more compressible than the thinner films where (Pz 1. There is therefore a direct relationship between oleate chain density and repulsive force. The nature of this force or of the related interaction free energy has been discussed by a number of authors.20 The Helmholtz free energy change of the system as the film thins is for unit area of film where 0 is the film tension y is the interfacial tension of the interfaces between the equilibrium bulk phases ni is the number of moles of i in the system and pim and p i are the chemical potentials of i in the system before and after the film has thinned.In most instances and certainly for the present systems the second term on the right-hand side is negligibly For a system in which adsorption equilibrium AA = A - A = ( ~ ~ - 2 y ) + C n i ( ~ i - ~ i > (1 1) 54 THIN LIPID FILMS with the bulk phases is maintained the first term is calculable in principle from the adsorption isotherms for the single interfaces and the thin films respectively.21 A crude attempt to do this for a thin lipid film was made previou~ly.~ An alternative approach which is theoretically less satisfactory but which more readily yields an answer is to assume the adsorption to be independent of film thickness and to estimate the osmotic pressure changes in the film produced by the overlap of the chains of the stabilizer 23 This has also been attempted for the present systems.l The important conclusion is that whichever approach is used it is found that a repulsive force sufficient to stabilize the film is generated by a very small overlap of the oleate chains of the two monolayers.Furthermore in order to explain the relatively high compressibility of the thick films the volume fraction of the 0 SL (nm) FIG. 5.-The steric repulsion force Fs as a function of film thickness for the systems of table 1. Saturated NaCl (Hamaker constant assumed to be 3.48 x J Is) ; (l) n-hexadecane ; (2) n-tetra- decane ; (3) carbon tetrachloride ; (4) cyclohexane ; (9 n-decane) ; (6) 2,2,6trimethylpentane ; (7) n-heptane. 0.1 M NaCl (Hamaker constant assumed to be 6.76 x J ’) ; (S) n-decane.oleate chain segments in the overlapping parts of the monolayers must be very small compared to the average volume fraction @ in the film. As in these particular systems the thickness of the hydrocarbon core of the film is closely similar to twice the extended chain length of the oleate it is inferred that at any given time only a small fraction of the oleate chains are fully extended. In the thinnest films (6 <3.2 nm@= 1 and the oleate chains of each monolayer are packed into a thickness of 1.5-1.6 nm. Films thicker than this must be stabilized by the greater tendency of the oleate chains to extend themselves beyond ca. 1.5 nm from the interfaces although as noted above only a small proportion of them are required to do so. The tendency to extend more fully is apparently related to the nature of the solvent.From the foregoing arguments it is possible to construct a qualitative picture of the time-average volume fraction q of oleate chain segments as a function of distance from the film interfaces (fig. 6). The curve for the thin (e.g. hexadecane) films cannot be appreciably in error as the amount of solvent present is undetectable by the present methods. For the thicker films the form of the curve is less certain, D. M . A N D R E W S E . D . MANEV AND D. A . HAYDON 55 although as any segments located outside the miniinurn volume for the chains (i.e. the region 1.45 nm thick between z = & 2.3 nm and z = & 0.85 nm) necessarily leave a similar sized hole within this region the shaded areas must be equal. -2.3 - 0.05 0 0 * 8 5 2.3 z (m) FIG.6.-A schematic representation of the time-average volume fraction 'p of oleate chain segments as a function of distance across the hydrocarbon region of a film (a) when n-hexadecane and (6) when n-heptane is used as the solvent. While the above remarks apply to the thin film it is probable that if the curves of fig. 6 are correct for the film they are also good indications of the situation in the soliated monolayers. Thus for the thick films where the individual monolayers are not compressed to significantly less than their maximum thickness this must be so. For the thin films it is difficult to give a satisfactory argument in the absence of quantitative relationships. However only a very small overlap of the monolayers at very small segment concentration is necessary to stabilize the film.If therefore the oleate chains in the monolayers were fully extended prior to film formation it is almost certain that more work would have been required to compress them into the close-packed state than to stabilize the thick film and a thick film would haveresulted. As a thick film was not formed it is concluded that the chains in the separate mono- layers must have been already contracted into the nearly close-packed state. D. M. Andrews was in receipt of a maintenance grant from Unilever Limited Port Sunlight during the course of this work. E. D. Manev thanks the British Council for the award of a Scholarship. The authors thank Mr. A. R. Taylor for considerable experimental assistance in the determination of the activity coefficients and Dr. J. L. Taylor for the determination of the interfacial tension against concentration curve for the system hexadecane + saturated NaCl.56 THIN LIPID FILMS D. M. Andrews Ph.D. Diss. (Cambridge 1970). R. J. Cherry and D. Chapman J. Mol. Biol. 1969 40 19. T. Hanai D. A. Haydon and J. L. Taylor Proc. Roy. SOC. A 1964 281 377. H. Sonntag and H. Klare Kolloid-Z. 1964 195 35. D. A. Haydon and J. L. Taylor Nature 1968,217,739. D. A. Haydon and J. Th. G. Overbeek Disc. Furuday SOC. 1966,42 76. ’ R. Aveyard and D. A. Haydon Trans. Furaday SOC. 1965 61,2255. T. Hanai D. A. Haydon and J. L. Taylor J. Theor. Biol. 1965 9 278. J. L. Taylor and D. A. Haydon Disc. Furaday SOC. 1966,42 51. lo D. M. Andrews and D. A. Haydon J. Mol. Biol. 1968 32 149. l1 C. T. Everitt and D. A. Haydon J . Theor. Biol. 1968 18 371. l 2 T. Hanai D. A. Haydon and J. L. Taylor J. Thzor. Biol. 1965 9 433. l3 H. T. Tien and E. A. Dawidowicz J. Colloidlnterfuce Sci. 1966 22,438. l4 G. M. W. Cook W. R. Redwood J. L. Taylor and D. A. Haydon Kolloid-Z. 1968,227 28. l5 G. S . Hartley Ann. Reports. 1948 45 50. l6 A. B. Scott and H. V. Tartar J. A m r . Chem. SOC. 1943,65,692. l7 J. M. Corkill J. F. Goodman and T. Walker Truns. Furuduy SOC. 1967 63 768. l8 D. F. Billett and D. A. Haydon to be published. l9 E. J. W. Verwey and J. Th. G. Overbeek Theory of the Stability of Lyophobic Colloids (Elsevier 2o R. H. Ottewill in Nonionic Surfuctunts ed. M. J. Schick (Marcel Dekker New York 1967) 21 E. L. Mackor and J. H. van der Waals J. Colloid Sci. 1952 7 535. 22 E. W. Fischer Kolloid-Z. 1958 160 120. 23 D. H. Napper Trans. Furatlay SOC. 1968 64 1701. Amsterdam 1948). vol. 1 p. 649.