GENERAL DISCUSSION Dr. B. A. Pethica (Unilever Res. Port Sunlight) said Frens Mysels and Vijayendran suggest that their derived (surface-pressure apparent-molecular-area) curves for a collapsing film stabilized by sodium dodecyl sulphate indicate a rapid increase in intermolecular attractions as the surface becomes very crowded. The form of the curves is however typical of those for collapsing insoluble monolayers and it is probable that a combination of solution and stacking effects in the collapsing film will explain the results. Dr. K. J. Mysels (R. J. Reynolds Winston Salem) said Pethica has raised the question whether the unexpected reversal of curvature in the ( l l A ) curves at low areas per molecule might not be due to a partial collapse of the surface monolayer. There is no way to answer that question with complete certainty on the basis of our observations.However the surfactant layer is highly ionized and not compact in our experiments (ca. 30A2/ion). This and the fact that variations in the rate of concentration-due to differences in thickness of the bursting film-by a factor of 10 do not produce significant deviations from the measured (a,@) curve are arguments against monolayer collapse as an explanation for the inversion of curvature. The reversal of curvature is also not to be explained as a time effect due to desorption. If the reverse curvature were not intrinsic for the film surface there could not be a beginning to the propagation of the aureole and therefore no such phenomenon to produce relaxation effects. In that case only a single shock-wave at the rim of the hole would propagate into an undisturbed film.In our opinion the reversal of curvature might indicate something like the incipient formation of a closely-packed vertically-oriented monolayer of long-chain ions. In such a layer important cohesive forces between chains would come into effect and act in the direction of inversion of the curvature. Dr. G. Frens (Philip Res. Lab. Eindhoven) said One would like to relate the surface tensions of super-saturated surfaces as obtained from the interpretation of the aureoles in bursting soap films to data for surface tension as a function of surface coverage below the c.m.c. We have pointed out that a problem here is the lack of agreement between the data of the different authors who have determined (17,A) curves for NaLS below the c.m.c.Mrs. Lucassen-Reynders of Unilever Research Laboratories has tackled this problem and some of her results were made available to me. She had shown earlier that two equations and relate surface pressure (ao -a) to adsorption (r) and r to the surfactant concentra- tion c for ionized surfactant.l These equations have three constants (a rW and Hs) E. H. Lucassen-Reynders J. Phys. Chem. 1966,70 1777. 38 GENERAL DISCUSSION whereas (ao - a) and c are the experimentally measured variables. Mrs. Lucassen computed which values of the constants a r" and Hs gave the best fit between theory and experimental data. The best set of constants is that for which the parameter 60 r( I E 40- ch Ei a 20- has its minimum. She analyzed some twenty sets of experimental data from the literature.Of these the surface tensions of NaLS as measured by Elworthy and MyFels gave satisfactory results. The other data led to unreasonable values for a I?" and Hs probably because there was too large a scatter in the experimental points. - - 1 I 1 1 0 20 LO 60 80 100 A (A') FIG. l.-(IT A ) curve for NaLS ; I and I1 calculated by Lucassen-Reynders ; U experimental points (ref. (2)); V " anchoring " point at c.m.c. (ref. (2)). Mrs. Lucassen's procedure will not only sort out the best set against experimental data it also produces a (l3,A) curve (or go - a against I?) through eqn (1). In fig. 1 we have combined this (l3,A) curve below the c.m.c. with the curve at high surface coverages which we obtained from OUT bursting experiments. The values ll = 30.3 mN m-l and A = 0.415 nm2 at the c.m.c.(where the soap film data are anchcred in the (&A) plane) were taken from the literature.'. The value of ll agrees with our own measurements. One would expect eqn (1) to break down near the c.m.c. as it predicts that the surface pressure would become infinite on saturation (I? = P). It is seen however that only a short interpolation is necessary in order to connect smoothly the two parts of the curve in fig. 1 into one (l3,A) plot which might then be meaningfully interpreted in a more complete equation of state than eqn (I). Dr. A. T. Florence and Dr. K. J. Mysels (Strathclyde University and R. J. Reynolds Winston SaZem) said In reply to Goodman we have been concerned with the question whether the existence of the aureole may not be due to the presence of impurities or hydrolysis in the sodium-dodecylsulphate used.As mentioned in our paper we used P. H. Elworthy K. J. Mysels J. Colloid Interface Sci. 1966 21 331. I. Weil J. Phys. Chem. 1966 71 738. GENERAL DISCUSSION 39 solutions well above the c.ni.c. so as to solubilize any surfactant impurities and thus reduce any such effects. These effects particularly those of dodecyl alcohol cannot be very significant. This is shown by some yet unpublished experiments of Dr. A. T. Florence at the R. J. Reynolds Lab. He combined the purification by foaming of a solution below the c.m.c. with photographic studies of bursting. Enough dodecyl alcohol was initially added to the solution to give rigid films. These show a peculiar bursting behaviour.2 Foaming rapidly reduced the alcohol concentration to the point where the films become mobile and bursting produces the usual aureole.This aureole persisted in the experiments without qualitative change through several days of further purification by foaming. used a similar purification method. They obtained indistinguishable results in solutions of sodium dodecyl sulphate and sulphonate. This indicates that hydroylsis of the sulphate is not a limiting factor under these conditions. Razouk and Mysels Prof. A. Vrij (Utrecht Netherlands) said I would ask Prins the following questions (i) in fig. 3 of his paper he has plotted a calculated equilibrium thickness (graph 2) for an electrolyte concentration of 3 x mol ~ m - ~ which is the total concentration of sodium lauryl sulphate. In my opinion however he should have used the con- centration of free soap ions (i.e.the c.m.c.) which is about a factor of 4 smaller. (ii) He proposes that the pressure in the film could be higher than in the surrounding atmosphere. Does this mean that in that case the van der Waals’ attraction forces exceed the double-layer repulsion forces ? (iii) The plots of calculated electrolyte concentrations as a function of height in fig. 5 are linear. Does this have any physical significance ? Dr. A. Prins (Unilever Res. Lab. Netherlands) said In reply to Vrij the activity of surfactant molecules is indeed affected by micelle formation. However since it is not known how the counter ion activity changes with the concentration in the presence of micelles we have assumed that for the calculation of the film thickness the nominal Concentration of the counter ions has to be used.This is confirmed by the agreement between the measured and the calculated film thickness at the bottom plateau border. In addition as appears from fig. 4 of our paper exchange of most of the surfactant by salt results in exactly the same profile of the film. Moreover a calculation based on a salt concentration equal to the c.m.c. would result in an even thicker film making the discrepancy between theory and experiment even bigger. The suggested explanation of the water vapour pressure equilibrium is indeed that the pressure in the film exceeds atmospheric pressure in order to compensate for the decrease in water vapour pressure caused by the higher salt concentration. The linear relation between the salt concentration and the height in the film suggests that it is perhaps possible to give a simple explanation for the observed phenomena.The physical significance however is not known at present. Dr. M. N. Jones (University of Manchester) said With regard to the paper by Prins and van den Tempel Reed and I have attempted to measure the thickness profile of films up to 10 cm in height drawn from sodium n-dodecylsulphate solutions at low concentrations ( N 0.009 M) and found that while a given film had a thickness which remained constant within -0.5 nm over a period of several hours there was considerable variation (k 10 nm) in thickness between different films drawn from K. J. Mysels and A. T. Florence in CZean Surfaces G. Goldfinger ed. (New York 1970). W. R. McEntee and K. J. Mysels J. Phys. Chem. 1969,73 3018. Razouk and K.J. Mysels J. Arner. Oil Chem. SOC. 1968,45 381. 40 GENERAL DISCUSSION the same solution. At such low ionic strength the film thicknesses were very depen- dent on the method used to humidify the atmosphere in the film chamber and it was possible to observe even with the naked eye that different elements of the film had different thicknesses. The addition of electrolyte however had a marked effect on the thickness profiles and at higher ionic strengths (-0.15 M 1 1 electrolyte) we found that for a given film although the thickness decreased with time this change was only small -0.5 nm and when the system was left to thermostat for a long period (- 70 h) the decrease was even smaller which implied that it was due to evaporation. What is perhaps more significant is that under these conditions the initial black film thickness of freshly drawn films was always constant within k0.05 nm and we concluded that these are true equilibrium thicknesses.At higher salt concentrations (-0.2 M 1 1 electrolyte) we always obtained uniform films of constant thicknesses. It would seem from these observations that the diffusion of non-volatile components along the film is dependent on ionic strength and I would like to know whether they have made any observations on thinner films. Eqn (6) of this paper predicts that the rate of thinning should be much slower for thin films and t b s is contrary to our observations. Dr. A. Prins (Unilever Res. Lab. Netherlands) said In reply to Jones a measurable thickness profile as described in our paper appears only when films axe used which contain not too much salt.By increasing the salt concentration the film profile becomes more and more parallel and a Perrin film has a uniform thickness which however is determined by other forces than are discussed in our paper. Under the usual conditions of observation the rate of thinning of large films is always deter- mined by evaporation and not by the process described by eqn (6). In reply to Lyklema Bruil carried out his film thickness measurements at 0.5 cm above the bottom plateau border. It might well be that his results were affected by the phenomena discussed in our paper which should result in slightly thinner films than predicted by the DVLO theory at that level. Dr. H. Sonntag (Deutsche Akademie der Wissenschaften Berlin) said Clunie et al. expect that films stabilized with non-ionic surface-active agents only might form second black films.I cannot agree with this opinion because (i) we only found first black films in polar systems with various non-polar surface-active agents and never obtained an alteration of film tension; (ii) aqueous films between oil droplets stabilized with nonylphenol ethylene oxide (20) formed second black films after adding K2S04 or MgS04 as electrolyte. With Ca(NO,) or BaCl e.g. we obtained only first black films at all concentrations. I think the experimental results are still insufficient to give a theoretical explanation of the formation of second black films. Moreover I would ask whether they are quite sure that there is no building-up of adsorbed multilayers of surface-active agents above the critical micelle concentration ? Dr.J. M. Corkill (Procter & Gamble Ltd. Newcastle-upon-Tyne) said In reply to Sonntag it is generally accepted that the stability of first black films is determined by a balance between van der Waals’ attraction and electrostatic double-layer repul- sion. For aqueous foam films stabilized by a non-ionizable surface-active agent the existence of any diffue double layers can only result from ion adsorption at the film-core/surfactant-monolayer interface. In the absence of ion adsorption one would therefore not expect the formation of first black films. Any stable films that GENERAL DISCUSSION 41 are formed in such systems are likely to be of the second black type. Recent radio- tracer experiments using 35S-labelled decyl methyl sulphoxide have shown the equilibrium black film radioactivity to be virtually independent of electrolyte con- centration in bulk solution (up to 4 kmol m-3).The measured activity corresponds to an area of 0.34+0.02nm2 per surfactant molecule in the film surface. This value is identical to the area per surfactant molecule at the bulk-solution/air interface obtained from surface tension measurements on the same system. Dr. S. Levine (Manchester University) and Prof. G. M. Bell (Chelsea College) said In the paper by Clunie Corkill Goodman and Ingram eqn (1) for the double- layer repulsion whch is based on the DLVO theory is applied to electrolyte concentrations as high as 5 mol/l. However this is derived from the Poisson-Boltz- mann (Gouy-Chapman) equation which in fact should not be used for concentrations of 1 1 electrolytes greater than say 0.3 mol/l.at the rather low value of the Stern potential (20 mv) quoted by the authors. The corrections to the Poisson-Boltzmann equation have been considered by many authors.2 Dr. J. M. Corkill (Procter & Gamble Ltd. Newcastle-upon-Tyne) said In reply to Levine we agree with his comment concerning the limitation of the Gouy-Chapman theory to electrolyte concentrations below -300 mol m-3. However we have found good agreement between experimental and theoretical A* values both below and above the strict limits of applicability of double-layer theory. In these film calculations the contribution to the potential energy from electrical double-layer repulsion is much smaller than that due to van der Waals’ attractive forces (maximum of 40 % for 7.2 nm films).The influence on A* of errors in calculating the double- layer repulsion term for high ionic strength films will probably be small. This may explain the consistency between calculated A* values for films at high and low ionic strengths and their agreement with the A* value calculated for the electrolyte-free second black film. Prof. A. Scheludko (Sofia) (communicated) Exerowa and Kolarov have measured the thickness of microscopic horizontal films of C1 SOCH3 aqueous solution (the compound was supplied by Dr. Goodman) the concentration of the surfactant being the same as in the paper of Clunie et al. ( 2 . 2 ~ M) and electrolyte NaCl M. White films have been obtained with thickness of 460-520A instead of 48 8 measured by Clunie et al. at low electrolyte concentration.The outer capillary pressure was 290 dyn cm-2 which yields for the potential of the diffuse electric layer 4o = 17 mV as calculated according to the DLVO theory and with a van der Waals- Hamaker constant K = A/4n = 4 x 10-14.2 The calculation has been carried out with the complete formulae and has account of all the necessary conditions as described in ref. (2). The certainty of the value obtained for 4, is diminished because of the considerable effect of the van der Waals component of the disjoining pressure with such a low value of the potential and thicknesses; the van der Waals constant is known only with limited accuracy.l It is certain however that the potential is considerably lower than that of water (30 mv) which is evident from the fact that the equilibrium thickness is smaller with the same electrolyte concentration (700 without the surfactant against - 500 A here).The significant difference in thickness corn- pared to 48 A of Clunie et al. is completely explained by the higher outer pressure of the cell in our work. J. S. Clunie J. M. Corkill and B. T. Ingram unpublished results. see e.g. S. Levine and G. M. Bell Disc. Faraday SOC. 1966 42 69. 42 GENERAL DISCUSSION In fig. 1 the curve disjoining pressure II against film thickness h at M NaCl is presented TI being calculated from the approximate expression with y = 0.164 corresponding to & = 17 mV. I1 = 64cRTy2 exp (-xh) - K/h3 I I 1 1 1 I 1 400 6 0 0 8 00 h fi FIG. 1 As is seen the force barrier corresponds to some 500 dyn cm-2 so that with a highex outer pressure this barrier would be overcome and the film would either collapse or form a more stable film (Perrin’s film) the latter not being described by the DLVO theory.The outer pressure in the vertical frame of Clunie et al. also includes the hydrostatic pressure which at the upper end of the frame is - 5000 dyn cm-2 C10HZIS03CH3 NaLS m X t 2 0 0- I 0.2 0.3 0.040 0.316 CNaCl (M) FIG. 2 so that the force barrier is definitely exceeded and a thermodynamically more stable Perrin film is formed which then covers the whole frame. Up to this point therefore the fiIms described behave normally. Exerowa and Kolarov have also studied the transition first/second (Perrin’s) black film with the same surfactant by means of the new method of foam destruction by a stream of a-particles. As shown in ref. (3) the rate of foam column reduction (dZ/dt) subjected to cr-particle bombardment possesses a sharply outlined maximum near the electrolyte concentration corresponding to the first /second black film tran- sition with microscopic films.GENERAL DISCUSSION 43 M) and for compari- son with NaLS (5 x M) are shown. As an a-ray source Pu239 with intensity 558 a-particlesls was used. In the second case the maximum is at a NaCl concentra- tion of 0.316 M very close to the value obtained using the contact angle method by us and by My~els.~ For the surfactant used by Clunie et al. a maximum at 5 x M NaCl has been observed. Therefore the first/second black film transition does not show at first sight any peculiarity compared to the above results. The striking and very interesting moment in the paper of Clunie et al.is the thicken- ing of the film from 40 to 70 A at a NaCl concentration of 7 x M. They interpret this effect as a result of the increase in 4,-potential. This explanation is based mainly we think on the fact that the 4,-potential with this surfactant has as we have shown an especially low value and therefore can easily increase on changing the conditions e.g. the electrolyte concentration. Quantitative interpretation on the basis of the DLVO theory however seems to be groundless in this case. As is well known the DLVO theory gives a maximum and a minimum for the energy and dis- joining pressure isotherms when the film thickness decreases. The maximum explains the first black films but is not able to explain Perrin films. In order to describe the latter another minimum with a consequent rise of the disjoining pressure above the outer pressure should exist at very small thicknesses.Therefore at thicknesses close to those of the Perrin films significant deviations from the DLVO theory are expected in the direction of positive disjoining pressure. Such a large deviation has been observed in the measurements of the contact angle black film/bulk liquid in the region of first/second black film tran~ition.~ Therefore the application of the DLVO theory when calculating the constants at thicknesses exceeding Perrin’s film thickness less than 30A is quite uncertain. It is appropriate to recall the measurements on films of concentrated aqueous solutions of fat acids.6 With 2 M butyric acid solution very thick equilibrium films of some lW0A have been obtained.The thickening at such high electrolyte concentration is not due to the diffuse electric layers. The high positive values of the disjoining pressure in aqueous solution of butyric acid have been confirmed by Voropaeva using the method of crossed platinum fibre^,^ i.e. with different phase surfaces. Maybe in the work of Clunie et al. a similar situation appears. Other possible reasons for the deviations from the DLVO theory are discussed in ref. (8). As far as there is no theory to explain Perrin films we suggested that all these deviations from the DLVO theory should be combined in a third component of the disjoining pressure rl[bl. This component must rise sharply at small thicknesses in order to explain the Perrin films. The object should be a study of this I&,l and in this sense we find the results in the paper of Clunie Corkill Goodman and Ingram of considerable interest.In fig. 2 the results obtained with C10H21SOCH3 (2.2 x This suggests a volume origin for the effect. Dr. J . M . Corkill (Procter & Gamble Ltd Newcastle-upon-Tyne) said In reply to Scheludko we agree with the explanation advanced by Scheludko for the Occurrence of thick (-50 nm) microscopic films below the transition concentration found by us l A. Scheludko and D. Exerowa KolloidZ. 1960 168 24. D. Exerowa Kolloid-Z. 1969,232,703 ; Proc. 5th Int. Congr. Surface Activity 1968,2 153. D. Exerowa and D. Ivanov Compt. Rend. Acad. Bulg. Sci. 1970,23 547. T. Kolarov A. Scheludko and D. Exerowa Trans. Furaday Soc. 1968,64,2864. F. Huisman and K. Mysels J . Phys. Chem. 1969,73,489.A. Scheludko and D. Exerowa Ann. Univ. Sofia Fac. Chim. 1959/60,54 205. ’ T. Voropaeva B. Deryaguin and B. Kabanov Kolloid-Z. 1962,24,398. * A. Scheludko Adv. Colloid Interface Sci. 1967 1 391. A. Scheludko Ann. Univ. Sofia Fac. Chim. 1967/68 62,47. 44 GENERAL DISCUSSION for large vertical films. For second black (Perrin) films i.e. films with thicknesses independent of electrolyte concentration we agree that the dominant repulsive force llbl does not behave simply as a cut-off potential. However we find it difficult to invoke a J&, which is not a double-layer repulsion to explain the occurrence in the decyl methyl sulphoxide system of first black films i.e. films where the thickness decreases with increasing electrolyte concentration. This difficulty becomes greater when we consider the lower concentrations and higher thicknesses involved in the transition from second to first black films with different added electrolytes.For example with NaOH a first black film of thickness 47 nm is formed at a concentration of 0.1 mol m-3. The thickness of this first black film then decreases progressively with increasing NaOH concentration. Dr. B. A. Pethica (Unilever Res. Port Sunlight) said Clunie et al. suggest that the transition from second to first black type in their films stabilized by a non-ionic surfactant is due to ion adsorption in the head-group plane. It is highly improbable that an adsorption of ions governed by a Stern adsorption potential will give such a sharp transition as is shown on their fig. 2. The adsorption of ions can be tested by electrophoretic methods using e.g.oil drops stabilized by the same non-ionic or by surface potential measurements. The film transitions are probably associated with salt effects on smectic phase transitions in solution. Bearing in mind that the film ionic composition is quite possibly different from that of the bulk associated solution it would be useful to have related data on the solution properties of the stabilizer. The determination of the apparent Hamaker constants in their paper is open to objection. The set of equations (1)-(5) rest on the improbable assumption that the term B (which contains a Stern potential) does not vary with film thickness. Assuming that the electrical repulsion is a simple exponential function of thickness the electrical term disappears from the argument leaving ACT a direct function of the Hamaker constant A thickness and the Debye-Hiickel reciprocal length.The resulting values of A vary by a factor of two over a 10 K temperature interval which leaves grave doubts as to the physical significance of the calculations. Dr. J. M. Corkill (Procter & Gamble Ltd. Newcastle-upon-Tyne) said In answer to Pethica’s comment concerning the abruptness of the film transition we refer to our reply to Lyklema; it is not necessary to postulate a corresponding lamellar mesomorphic phase transition in the bulk solution. The assumption that the Stern potential is independent of the separation between the charged surfaces is commonly employed in double-layer repulsion calculations although some authors have suggested that the surface charge is more likely to remain constant.For weakly overlapping double layers the difference between the two treatments leads to a very small difference in repulsion energy.l. The variations in A* with temperature that we have observed are hence unlikely to originate in a variation of the B term. We return to the point made in our paper that the temperature dependence of Aa for the first black films cannot be explained in simple terms such as a variation in Stern potential. Prof. J. Lyklema (Wageningen) said In fig. 2 of their paper Clunie et al. observe a steep rise in film thickness in the salt concentration range 0.07-0.1 M. An explana- tion is offered invoking ion absorption in the head-group plane of the surfactant. E. J. W. Verwey and J. Th. G. Overbeek Theory ofthe Stability ofLyophobic Colloids (Elsevier Amsterdam 1948).J. E. Jones and S. Levine J. Colloid Interface Sci. 1969,30 241. GENERAL DISCUSSION 45 Tf this were true one would also expect an analogous irregularity in the micelle formation behaviour of the used surfactant as a function of the salt concentration around the 0.07-0.1 M region. Was anything like that observed? Dr. J. M. Corkill (Procter & Gamble Ltd. Newcastle-upon-Tyne) said In reply to Lyklema the critical micelle concentration (surface tension data) of decyl methyl sulphoxide decreases with addition of salt but no abrupt discontinuities are observed. The dependence of critical micelle concentration on ionic strength is similar to that reported for other non-ionic surface-active agents. To account for the transition in film thickness it is not necessary to postulate an abrupt rise in ion adsorption.It is generally accepted that a charged film can exhibit two minima in its potential energy-thickness relationship. The secondary minimum at larger thicknesses results from superposing a double-layer repulsion term on the gravitational and van der Waals’ energy contributions. If the hydrostatic pressure is sufficiently large a secondary minimum does not occur and either a primary minimum (second black) film is formed or rupture takes place.2 In our system the hydrostatic pressure remains constant at the point of film thickness measurement (- 300 N m-2) as the solution composition is changed. The transition occurs when the surface charge of the film has risen sufficiently to create a secondary minimum rather than an inflection in the energy-thickness diagram.The observation that further increasing the ionic strength leads to a deepening of the secondary minimum in addition to moving it to lower thickness is consistent with the predictions of the DLVO theory. Prof. R. J. Good (Bristol University) said With regard to the paper by Clunie et al. it may be misleading to refer to film tension of for a second black film without distinguishing it physically from the film tension of a thick film and from surface tension. The state of stress of a second black film is a Hookean state characterized by an elastic modulus i.e. the tension in the film for small elongations is propor- tionate to the elongation. In this it is unlike the surface of a bulk liquid or of a thick film ; for these cases the static stress is independent of elongation.The dynamic stress in the latter two cases may be dependent on elongation through the Gibbs elasticity ; however this is a strongly time-dependent condition of stress. This difference between the second black film and the other surfaces exists because of the lack of a reservoir of fluid in the interior of the black film. So the black film acts mechanically simply as a linkage for transmitting the pull of the bulk-liquid surface to the Plateau border at the top of the frame and thence to the top of the frame itself. There is a contact angle formed by the surfaces of the plateau border and the film just as there is a contact angle at the bottom. I would ask whether Clunie et al. measured the contact angles directly and in particular the angle at the top of the frame? This set of measurements should enable them to tell whether the plateau border has a higher surface tension than the bulk liquid indicating appreciable depletion of surfactant. Dr. J. M. Corkill (Procter & Gumble Ltd. Newcastle-upon-Tyne) said In reply to Good we have directly measured the contact angle between the bottom of the film and the bulk solution at 298 I< using the refraction m e t h ~ d . ~ We found no significant difference between this measured equilibrium contact angle and that calculated from the film tension measurements. P. Mukerjee J . Phys. Chern. 1965 69 4038. J. Th. G. Overbeek J. Phys. Chem. 1960 64 1178. S. Frankel and H. M. Princen J. Phys. Chern. 1970,74 2580.