摘要:

Faraday Discuss. Chem. SOC., 1983, 76, 305-315 Concentrated Dispersions of Aqueous Polyelectrolyte-like Microphases in Non-polar Hydrocarbons BY HANS-FRIEDRICH EICKE AND RUDOLF KUBIK Physikalisch-Chemisches Institut der Universitat Basel, Klingelbergstrasse 80, CH4056 Basel, Switzerland Received 12th May, 1983 Properties of individual aqueous microphases and of their dispersions (microemulsions) in non-polar media are discussed, with particular emphasis on the implication of the Poisson-Boltzmann equation and the role of the semi-diffuse ionic double layer in such sys- tems in general, and calorimetric and 3Na-n.m.r experiments are presented. Ensemble pro- perties are demonstrated by means of electro-optic Kerr-effect measurements which elucidate, in particular, temperature-dependent reversible coagulation of microphases, solvent effects on their interaction and electric-field effects on the critical (one-phase demixing) coacervation state.Finally, H 2 0 self-diffusion studies have been made to examine the percolation region. An Arrhenius plot of the latter process displays two distinct branches which can be reason- ably well correlated with the expected underlying molecular processes. At first sight the long-lasting interest in ensembles of microphases (so-called microemulsions) might appear strange. These systems of aqueous or oily (lipid) microphases within the continuous antagonistic liquid phase are characterized by their thermodynamic stability as inferred from their spontaneous, reversible, form- ation and long-lasting stability, which is not due to a kinetic barrier.In the simplest case the microemulsion consists of three components, one which is to be dispersed in another (the continuous phase) and a third component, the amphiphile (surfactant), which makes the stable dispersion possible. The attention which these systems still find has been, apart from their many practical applications, clearly directed in recent years towards the elucidation of scientific concepts regarding interactions in colloidal systems. The apparent reason is the suitable model character of these systems, which in some instances form almost isodispersed suspensions of spherical entities.2 Moreover, these are susceptible to physico-chemical modifications producing more or less strongly interacting micro- phases that are frequently distinguished, according to model considerations, as ‘hard’ and ‘soft’ spheres.It appears straightforward to consider these ensembles of microphases as globu- lar liquids and hence to apply a variety of liquid model theories to these systems. In addition, a comparison with Monte Carlo or molecular-dynamics calculations seems appropriate owing to the particularly simple structure of these ‘liquids’. However, not only the structure but also the size of these ‘molecules’ which form the liquid are remarkable: thus physical phenomena may be observed easily which are otherwise to be detected only with considerable effort. The subject under discussion leads in a natural way to a subdivision of the paper in two main sections, one concerning the properties of the individual microphases which are essential for an appreciation of the ensemble properties, and the second describing phenomena originating from the interaction of the microphases.306 POLYELECTROLY TE-LIKE MICROPHASES IN NON-POLAR SOLVENTS PROPERTIES OF INDIVIDUAL MICROPHASES The dispersions to be discussed here consist of three components, water, Aerosol OT (sodium di-2-ethylhexylsulphosuccinate) and aliphatic or aromatic hydrocar- bons as the continuous oil phase.The thermodynamic model of aqueous micro- phases to be considered contains five main energy contributions to the total free energy of the microemulsions system, G,,,2,4 the electrostatic field energy of the semi-diffuse ionic double layer, Gf, the corresponding free energy due to the (in- homogeneous) distribution of the counterions in the diffuse double layer, G::, the surface free energy of the hypothetically uncharged water/oil interface (comprising short-range effects which are expected to be independent of curvature), yUnA (where A is the interface of a single microphase), and finally an entropic term arising from the mixing of the dispersed particles within the continuous oil phase, G,,,, i.e.Gme = G" + Gf + G E + yUnA + Gmix. (1) G" is an arbitrary reference state and is independent of the water content of the microphases. By solving the Poisson-Boltzmann (PB) equation with appropriate boundary conditions it is possible to calculate Gf [for details concerning GEx and the boundary conditions see ref. (4)]: n Vmic where E~ is the relative dielectric constant, E, the permittivity of a vacuum, - V CD the electric field and ViC the polar core volume of a single microphase.Within the approximation of the PB equation the equilibrium condition for the microphase ensemble is obtained by forming the first derivative with respect to A, i.e. (dGme/dA)p,T = 0 = yun - 2Gf/A + 3kT</A. (3) Eqn (3) shows that under equilibrium conditions the surface tension y (defined as yun - 2Gf/A) must be compensated by the entropy of mixing of the microphases in the continuous oil phase. This third term may be d e r i ~ e d , ~ according to the Percus-- Yevick and Carnahan-Starling theories for hard-sphere suspensions, by substituting 4: = lntp - 1 + ( ~ ( 4 - 340)/(1 - cpI2 + ln(Vodvmic) where cp is the volume fraction of the dispersed phase, V,,,and V,,, are the molar volumes of the apolar solvent and the microphases, respectively, and kT is the molecular thermal energy.A theoretical estimate shows that the third term is ca. -0.2 mJ m-2 for cp = 0.5. Hence y must be a small (positive) quantity. Experimen- tally one find^^?^ y = 10-4-10-3 mN m-*, the so-called 'ultra-low' interfacial tensions by which the thermodynamic stability of the system is manifested. Neglecting the entropy of mixing, the equilibrium condition demands yun = 2Gf/A. Gf can be calculated numerically via the PB equation: Gf/A = Y(Ro,-&im)k7'/f,m* (2') YIRo,hm(Ro)] is the result of functionalizing a large number of numerical solutions of the PB equation. It depends on the polar core radius of the microphase, R,, and the interfacial area, Sam, covered by one amphiphilic molecule.Y is a geometric parameter which is defined between lim Y = 0 (Ro -+ 0) and lim Y = 1 (R, --+ co, i.e. a plane interface).H-F. EICKE AND R. KUBIK 307 Since ylln contains only short-range interaction energies on a molecular scale, Gf/A should thus be independent of the interfacial curvature. Eqn (2') makes it possible to test this assertion for systems for which Ro andfa, are known. The inset of fig. 1 shows the constancy of Gf/A for two similar systems, i.e. H 2 0 + AOT + i-C,H1, and H,O + AOT + n-heptane at 298 K . The value yun = 12 mN m-' must then be used to calculate the interfacial area covered by one AOT molecule as a r t p t l , , , , , I , ) 2 .O L .O 6.0 8.0 I I, 0 2 4 6 8 10 &/nm Fig.1. Experimental values (0) and a theoretical plot (--) of the interfacial area of an AOT molecule as a function of the water core radius of the microphase (the calculation of the theoretical plot takes yun = 12.0 mJ Inset: the electrostatic-field energy density of the semi-diffuse double layer is plotted against the water core radius of the microphase for the systems H 2 0 + AOT + isoctane (A) and H 2 0 + AOT + n-heptane (0) at 298 K: R, andS,, are determined as in fig. 5. function of the core radius. The coincidence between experimental points and theoretical plot is satisfactory. Similar calculations of Gf/A for other mesophase structures are p~ssible.~ Another experiment which demonstrates clearly the role of the semi-diffuse double layer of the microphases is the determination of the enthalpy of a micro- emulsion, H,,, which can be obtained by mixing two microemulsions with different volume fractions of water.Thus Him = Nrnix(wo) - Hmx,ref (e-g. wg = 0 ) (44 where Hmlx,ref is the enthalpy of a micellar solution and w0 = [H,Q]/[AOT]. The mixing process is spontaneous and endothermic; hence, it must be driven entro- pically. However, simple geometric considerations show that the total number of microphases decreases during this process. Accordingly, the entropy source cannot be the dispersion of water but has to be attributed to the internal structure of the microphase, i.e. to the development of the semi-diffuse double layer. The enthalpy of a microemulsion, H m e , can be described by applying thrt Gibbs-Helmholtz relation to eqn (1).This yields am, = H " -I- H,", -I- HF (4h) where H (which corresponds to the same reference state as G") comprises the lattice and cohesion energies of the surfactant within the micellar state, the heats of hydration of the surfactant head groups and the difference in solvation energy be- tween the microphase and micellar reference state. H;,, corresponds to yunA and is calculated from308 POLYELECTROLYTE-LIKE MICROPHASES IN NON-POLAR SOLVENTS Htn = [yun - T(dYun/d n p , T]A"A (44 where N is the number of moles of microphases per system and NA is Avogadro's number. H f (corresponding to Gf) is essentially determined l o by If the equilibrium condition [eqn (l), i.e. neglecting the entropy-of-mixing term] is considered, then H m , = H e + [GfNNA ( 5 ) where [ = 3 + (dlne,/dlnT),,T - 2(dlnyun/dlnT),,T.Dividing eqn ( 5 ) by the total interface of the system, where cam is the weighed-in concentration of surfactant and V the volume of the system, eqn ( 5 ) can be rewritten, i.e. (6) (Hrne - H )/(camNA v) = Affme/(cam VNA) = [Yunfam/2. Eqn (6) has the advantage that two experimentally accessible parameters can be used to test the theoretical predictions. A theoretical estimate of the constant [ is difficult and requires a careful consideration of a variety of physical phen0mena.l' Essen- tially it is determined by the temperature dependence of E, and yun. A rough estimate yields 4 < [ < 11. The experimental results are plotted in fig. 2 with yun = 12.0 mN m-l, T = 298 K, E , = 78.3 (H,O) and reference state H"(w, = 2.8) rO.[ is ac- cessible, according to eqn (6), from the slope of the linear plot; its value is 5.1 f 0.1 [correlation coefficient 0.9992, number of points (n) 91 which fits into the above estimated range of [. 23Na n.m.r. spectroscopy offers the most direct approach to studying the semi- diffuse double layer in the present system. This technique is particularly suited to investigate the counterion distribution within the microphase and hence to ob- tain information about the binding parameter p, which is defined as the number of bound counterions divided by the total number of surfactant ions. As an operational t 2 .o 1.0 &,/10-20 m2 Fig. 2. Relative enthalpy of one surfactant molecule in kT units plotted against the interfacial area of an AOT molecule.H-F.EICKE AND R. KUBIK 309 1500 N 3 1000 c - v 500 definition, p corresponds to the counterions which are in direct contact with the interface. An appropriate description exists in the Stern-Helmholtz-Gouy-Chapman model, which formally separates the immobilized counterion layer (also affected by non-electrostatic forces) from the mobile, semi-diffuse part of the double layer, which is solely influenced by Coulombic forces. The solubilization of water in these non-polar micellar solutions causes the gradual formation of the Gouy-Chapman region. This process simultaneously influences the number of bound counter- ions. Under the assumption that the non-electrostatic forces are not affected during the solubilization process, the change in the number of 'bound' counterions is direct- ly proportional to the change in the counterion concentration, c,, within a spherical shell of radius R , - 6.ci = ci(R, - 6) may be calculated directly by solving the PB equation. 6 denotes the smallest possible distance between counterion and surfactant ion, i.e. ca 0.1-0.5 nm. Hence c, is to a first approximation proportional to p /?= c c j (7) where C is a constant; p can be related to the transverse relaxation time, T2.13 The sodium ions exist at different intervals from the water/oil interface and accordingly experience different molecular environments. If the exchange between different positions of the ions is fast compared with the relaxation time (which is the case for aqueous microphases 14) then A - - - where xi denotes the mole fraction of counterions in the ith state.Since 1/T, de- pends on the distribution function of the counterions, on the time constant of the relaxing (fluctuating) field gradients and on their absolute values, a rough approxim- ation is used. This is the two-state model in which the counterions can exist in a free and a bound state. Since the experimental relaxation rates l/T2 are much greater than 1/T2,f [the relaxation rate of 23Na is (in an infinitely dilute solution) 15.8-17.5 Hz 16], one can write (within the framework of the two-state model) where 1/T2,b and 1/T2,f are constants, i.e. independent of w O . Replacing /3 by eqn (7) yields \ 'A --A 1 I I I I c 0 10 20 30 40 "0 Fig. 3. 23Na relaxation rate (1/T2) plotted against the amount of solubilized water (wo = [D,O]/[AOT]) for the system D20 + 0.15 mol dm-3 AOT + c-C6HI2, 298 K.Parameters: &,(D2O)= 77.9, Gf/A = 6.0 mJ mP2; experimental value (A) and theoretical plot (-).3 10 POLYELECTROLYTE-LIKE MICROPHASES IN NON-POLAR SOLVENTS 1/T2 z k C, (Ro - 6). (7') Fig. 3 confirms the quality of the approximation derived in eqn (7'). In plotting this diagram the values of c, (Ro - d), which were calculated within the framework of the PB equation as a function of wo, had to be multiplied by an arbitrary constant k in order to fit the experimental plot. The function l/T2(wo) is rather insensitive to a variation of 6 within the range 0 < 6/nm < 0.5. The relatively large value of l/T,(ca. 500 Hz) at large wo compared with that of free Na+ indicates that even at complete development of the semi-diffuse double layer a considerable portion of the counterions are located close to the interface.ENSEMBLE PROPERTIES OF MICROPHASES (MICROEMULSIONS) Apart from phenomena based on the properties of individual micropkases, such as those discussed in the first part of this paper, there exists a set of features which I -I b i d 290 300 31 0 320 T K Fig. 4. Experimental relaxation times of the electric birefringence plotted against temperature for the system H 2 0 + AOT + organic solvent. [AOT] = 0.18 mol dm-3. 0, 2,2,4-trimethyl- hexane, w0 = 48; A, 2,2,4-trimethylpentane, wo = 48; A , n-hexane, w0 = 48; ., c-C,H,,, kv0 = 60.H-F. EICKE AND R. KUBIK 31 1 are brought about by interacting microphases and hence are typical ensemble pro- perties.One such physical phenomenon is the electro-optical Kerr effect. This effect decreases with decreasing size (which is approximately proportional to wo) of the microphase or concentration of the particles. The phenomenon is also sensitive to the nature of the continuous hydrocarbon phase. The latter observation is a strong indication that the Kerr effect is substantially due to dispersion interactions between the microphases. Fig. 4 demonstrates that organic solvents with small molar vol- umes* have a significant effect on the relaxation times (z) derived from electrical birefringence measurements. This observation is in agreement with recent findings from light-scattering experiments, where the close relation between the partial molar volumes and the mutual interaction potential of the microphases is elucidated, time-domain spectroscopic experiments and less direct estimates from the electronic polarizability of solubilized water * in aqueous microphases which are sufficiently small to penetrate the surfactant film.Closely related to this discussion is the effect of a fourth component (which may be considered as a cosurfactant), which has been studied experimentally: in fig. 5 Gf/A is plotted against nitrobenzene concentration P G I I , I , I , I ] 0 20 40 60 80 c(C,H ,NO,)/mmol dm- Fig. 5. Electrostatic-field energy density plotted against nitrobenzene concentration for the systems H 2 0 + AOT + iso-octane and nitrobenzene (as additive); [H,O]/[AOT] = 50; geometric parameters (&,&,) were determined by small-angle light scattering.and shows a linear decrease in the energy density of the electrostatic field. Under the explicit assumption of no mutual interactions between surfactant and cosurfactant (i.e. small perturbation of the surfactant film) the equilibrium condition 2Gf/A z yun should hold and hence Yun should decrease with c(C,H,NO,). From a molecular point of view the penetration of the cosurfactant reduces the interfacial charge den- sity, thus decreasing 2Gf/A. The reduction in yun is described by i.e. by the Gibbs adsorption isotherm, where Ti is the interfacial concentration, pi the chemical potential of the uncharged component i (or correspondingly the neutral equivalent of an ionic surfactant) and ywlo the interfacial tension of a water/oil interface.An estimate of the nitrobenzene interfacial concentration yields a [C,H,NO,]/[AOT] ratio of ca.0.25. Qualitatively the plot of z against T for the H,O + AOT + c-C6HI2 system differs appreciably from the other plots in fig. 4 regarding the slope and shape of the curves. The slopes are believed to be determined essentially by the tendency of the microphases to segregate from the continuous oil phase. This decreasing mutual solu- bility of the non-polar phase and the microphases with increasing temperature can be attributed to two molecular processes: (i) possible desolvation of the microphases and, probably even more important, (ii) decomposition of the outer hydration * More correctly, partial molar volumes should be considered.312 POLYELECTROLYTE-LIKE MICROPHASES IN NON-POLAR SOLVENTS spheres of the ionic surfactant molecules.This conclusion is inferred from 2H n.m.r. experiments, which indicate that above a certain amount of solubilized water (D20) the mole fraction of bound water molecules stays constant; i.e. on further addition of water the chemical shift (v)? is predominantly determined by ‘free’ water molecules. Hence, if n b denotes the number of bound water molecules per surfactant molecule, and if pure D20 is taken as reference state, vf = 0, the overall experimental value is v=(nb/WO)Vb, i.e. directly proportional to the mole fraction of bound water molecules. A plot of v against l/w, should be a straight line, which is true for wo values above the range 23 < wo < 27, where the spheres of hydration are fully devel- oped.A similar limiting value is obtained from light-scattering experiments and H-n.m.r. measurements.2o The slope, nbvb, cannot be split into individual factors without additional information regarding the temperature dependence of vb or n b . The temperature dependence of v b is assumed to have the same order of magnitude as vf. In the range 288-323 K the value of nbvb decreases by ca. 50% (83.3-42.8 Hz) whereas vf is almost unaffected. This then can only be explained by a significant decrease in n b , i.e. a dehydration of the outer hydration spheres. Fig. 4 also shows that with increasing temperature most of the plots pass through a maximum where the solution is still transparent with no indication of increased scattering.These maxima probably represent critical states of a so-called one-phase disintegration in an apparent single colloid system,21 which is also called a coacervation. The relaxation-time plots are terminated by the rather sudden onset of turbidity. The region between maximum and turbid end-point of the plot of z against T is only precisely reproducible by repeating the measurements from low to high tempera- tures, indicating a slow phase separation beyond TCit. At T < Tcrlt the system res- ponds reversibly to all variations of physical parameters. Two oil phases separate out, one of which also displays field-free birefringence and contains almost all the water and surfactant. At all concentrations above wo z 40 (where the sensitivity of the equipment is sufficient) the H 2 0 + AOT + i-C8HIs system has two relaxation times, as indicated by analysing the rise and decay curves of the birefringence signal.At concentrations above ca. 0.2 mol dm-3 both relaxation processes separate into single consecutive exponential processes in the neighbourhood of the critical point, z~~~ (fig. 6). A relevant finding in this connection is the different response of the two processes to an electric field (fig. 7). The slower process is strongly field dependent, as indicated by a shift of the critical temperature to lower temperatures, while the t I F t Fig. 6. Rise and decay relaxation times of electric birefringence (An). H,O + AOT + iso- octane: (a) [AOT] = 0.3 mol dm-3, [H,O]/[AOT] = 65, 294.6 K; (b) [AOT] = 0,144 mol drnF3, [H,O]/[AOT] = 65, 297 K.t Measured in Hz from a resonance frequency of 13.81 MHz.H-F. EICKE AND R. KUBIK 313 A 10 - 8 - 0 288 293 29 8 TIK Fig. 7. Electric-field-induced shift in zmaX =crit of the slower birefringence relaxation process towards lower temperatures. E = a, 4.15 x lo5 and A , 0 V m - l , [AOT] = 0.252 mol dm-3, [H,O]/[AOT] = 65. faster process is apparently independent of the field. This observation points to two different polarizable entities. At temperatures sufficiently below the critical tempera- ture the ratio of z,/z, for a given system is nearly constant and varies between 5 and 6. Approaching the critical region this value reduces to ca. 1, as is also the case for extrapolation to infinite dilution. Semi-quantitative considerations regarding the field dependence of the two processes lead one to expect a sudden increase in the dielectric constant at a temperature close to the critical point.22 If the critical tem- perature with respect to the slower process is lowered, then the curvature of the dielectric constant as a function of the critical volume fraction of water is positive and increases as the critical temperature is shifted to lower temperatures.For a given temperature shift the degree of increase in the dielectric constant at the criti- cal temperature depends essentially on the ratio of the electrical energy density and the interaction energy density per component particle, i.e. W , , / u ~ + W2,/u: - 2W12/ulv2 where the subscript 1 refers to the oil phase and 2 to the polarizable par- ticle, vi are the respective molecular volumes of the components and Wij has the dimension of energy times volume.Also at the critical temperature the electrical conductivity experiences a sudden increase by several orders of magnitude, frequently denoted as percolation. The same tendency is observed with the self- diffusion coefficient of the water molecules 2 3 (fig. 8). The increase in the latter oc- curs less abruptly. The inset shows that in a semi-logarithmic diagram the plot splits314 POLYELECTROLYTE-LIKE MICROPHASES IN NON-POLAR SOLVENTS 8.00 7 6.00 3 I 2 Q 4.00 1 2.0 0 10.0 t / 283 293 303 313 T K Fig. 8. Temperature dependence of the H 2 0 self-diffusion coefficient for the system H 2 0 + 0.17 mol dm-3 AOT + [2Hl,]n-hexane, [H,O]/[AOT] = 50.5.Inset: Arrhenius plot of the H 2 0 self-diffusion coefficient. into two branches. The activation energy corresponding to the steeper branch can be satisfactorily attributed to a three-step process: (i) the energy needed to create a hole in the apolar environment, (ii) the polarization energy of the apolar environment and (iii) the energy needed to abstract a water molecule from the bulk of the solubi- lized water (evaporation energy). This part of the curve corresponds to the steep increase in the electrical conductivity which, according to these results, does not occur in bicontinuous structures as is occasionally stated. The branch with the lower activation energy has probably to be attributed to microphases which are diffusing more independently.However the experimentally obtained self-diffusion coefficient is too large to be described exclusively by diffusion of the microphases. Hence ano- ther contribution has to be considered; this is probably found in the exchange pro- cesses that occur during the frequent collisions between the rnicropha~es.~~ A more detailed analysis regarding the percolation process, in particular charge production, will be discussed in a forthcoming paper,25 together with the observation of non- Newtonian viscosity in the systems as they approach critical regions. In conclusion, it appears that all the experiments described here indicate the increased formation of a network-like structure with increasing temperature. This work is part of the Swiss National Science Foundation, Project No.2.527- 0.82. H-F. E. acknowledges discussions with A. M. Cazabat and D. Langevin. 1. D. Robb, Microemuisions (Plenum Press, New York, 1982). J.Th.G. Overbeek, Faraday Discuss. Chem. SOC., 1978, 65, 7.H-F. EICKE AND R. KUBIK 315 H. Christen, H. F. Eicke and P. J. Xia, Colloid Pslym. Sci., 1983, in press. B. Jonsson and H. Wennerstrom, J . Colloid Interface Sri., 1980, 80, 482. A Vrij, E. A. Nieuwenhuis, H. M. Fijnaut and W. G. M. Agterof, Faraday Discus&. Chem. Soc., 1978, 65, 101. J. K. Percus and G. J. Yevick, Phyy. Rev., 1958, 110, 1. N. F. Carnahan and K. E. Starling, J . Chem. Phys., 1964, 51, 635. A. M. Cazabat, D. Langevin, J. Meunier and A. Pouchelon, in Surface Phenomena in Enhanced Oil Recovery, ed. D. 0. Shah (Plenum Press, New York, 1981), p. 161. K. Shinoda and S. Friberg, Adv. Colloid Interface Sci., 1975, 4, 281. 1980, 76, 1287. 10 G. Gunnarsson, H. Wennerstrom, G. Olofsson and A. Zacharov, J . Chem. Soc., Faraday Trans. 1, l 1 R. Kubik, Ph.D. Thesis (University of Basel, 1983). l 2 B. Lindman and H. Wennerstrom, Topics Curr. Chem., 1980, 87. l 4 E. Sjoblom and U. Henriksson, J. Phys. Chem., 1983, 86, 4451. l 6 R. K. Harris and J. E. Mann, NMR and the Periodic Table (Academic Press, London, 1978). l 7 B. Lemaire, P. Bothore and D. Roux, J . Phys. Chem., 1983, 87, 1023; S. Brunetti, D. Roux, A. M. Bellocq, G. Fourche and P. Bothorel, J . Phys. Chem., 1983, 87. 1028. R. Kubik, Dip/. Thesis (University of Basel, 1979); R. Hasse, Ph.D. Thesis (University of Basel, 1983). l 9 H. F. Eicke and R. Kubik, Ber. Bunsengens. Phys. Chem., 1980, 84, 36. 2 o H. F. Eicke, Phys. Bull., 1982, 38, 31 1. 2 1 H. G. Bungenberg de Jong, in Colloid Science, ed. H. R. Kruyt (Elsevier, Amsterdam, 1949),chap. 8. 2 2 P. Debye and K. Kleboth, J. Chem. Phys., 1965, 42, 3155. 2 3 H. F. Eicke and M. Holz, Ber. Bunsenges. Phys. Chem., 1983, in prep. 2 4 H. F. Eicke, J. C. W. Shepherd and A. Steinemann, J . Colloid Interface Sci., 1976, 56, 168. 2 s H. F. Eicke and R. Hilfiker, unpublished work. M. Wong, J. K. Thomas and T. Novak, J. Am. Chem. Soc., 1977, 99,4730. A. Abragam, The Principles of Nuclear Magnetism (Clarendon Press, Oxford, 1961).

ISSN:0301-7249    

DOI:10.1039/DC9837600305

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

Faraday Discuss. Chem. Soc., 1983, 76, 317-329 Fourier Transform Carbon- 13 Relaxation and Self-diffusion Studies of Microemulsions BY BJORN LINDMAN, THOMAS AHLNAS, OLLE SODERMAN AND HARALD WALDERHAUG Physical Chemistry 1, Chemical Center, University of Lund, P.O.B. 740, S-220 07 Lund, Sweden AND KRZYSZTOF RAPACKI AND PETER STILBS Institute of Physical Chemistry, University of Uppsala, P.O.B. 532, S-751 21 Uppsala, Sweden Received 23rd May, 1983 Multi-field 13C n.m.r. relaxation (spin-lattice relaxation and nuclear Overhauser enhance- ment) and multi-component Fourier-transform n.m.r. self-diffusion measurements have been applied to provide information on different aspects of the dynamic structure of micro- emulsions. Studies were performed over extensive concentration regions for four-component systems composed of sodium p-octylbenzene sulphonate, butan-1-01, water and n-decane or toluene while limited investigations were made for other systems.The I3C relaxation data were analysed on the basis of a two-step model of relaxation to provide information on short- and long-range molecular dynamics and on the degree of order of the alkyl chains with respect to hydrophobic-hydrophilic interfaces. The local motion of the alkyl chains is gener- ally about as rapid (10- s) as that of liquid hydrocarbons. The effect of the organization in the system and the creation of hydrophobic-hydrophilic interfaces is a slight anisotropy in the molecular motion. This anisotropy is similar in magnitude to that of liquid-crystalline phases, but tor the microemulsions the lifetime of the anisotropy is very short, one or two nanosec- onds.According to the self-diffusion results, a closed water droplet structural model may apply for alcohols with longer alkyl chains. For the above-mentioned systems with butanol as cosurfactant the self-diffusion coefficients of both water and hydrocarbon are quite high and inconsistent with structurally segregated domains, unless these are very short-lived. Much interest in the field of microemulsions is presently focused on structural and dynamic aspects, as these are important for both a fundamental understanding of the systems and the design of technical applications. While there is a range of systems which may be termed microemulsions, most work is concerned with four- component systems consisting of a surfactant, a medium-chain alcohol (butanol, pentanol or hexanol) as ‘co-surfactant’, hydrocarbon and water.- 3 The structural problem can be expected to be more intriguing the shorter the alcohol. A change from, for example, decanol to butanol is expected to lead to dramatic structural changes in surfactant systems. Therefore, it was considered to be of particular rele- vance t o study microemulsions with a cosurfactant like butanol. On the initiative of M. Kahlweit it was suggested, during the International Symposium on Surfact- ants in Solution (Lund, 1982), that coordinated studies were started between different groups on a particular model microemulsion system. This led to the formation of the ‘European Microemulsion Group’.The system chosen for collaborative work within this ‘Lund Project’ was sodium p-octylbenzene sulphonate, butan- 1-01, water and decane or toluene. For a system of this type, where some of the components show318 13C N.M.R. STUDIES OF MICROEMULSIONS important pairwise miscibility, the relation to simple solutions is apparent. This miscibility is counteracted by the tendency of the surfactant to induce organized structures in aqueous systems. It seems, therefore, to be of particular interest to establish to what extent organized structures are formed as well as to characterize them dynamically, i.e. molecular conformation and order and structural lifetime. In this paper we consider two methods which are applicable to these problems. First, magnetic-field-dependent 3C n.m.r.relaxation of the different carbons pro- vides information on the ordering of the chains with respect to internal interfaces, on chain dynamics and on slow motions extending over distances of the order of ag- gregate size. Order parameters are related to packing constraints given by interface curvature and are thus related to solution structure. Secondly, the translational motions over large distances of the different components provide information on aggregate geometry, the possible confinement to closed domains etc. There are a number of methods which are capable of providing self-diffusion coefficients but the Fourier-transform n.m.r. technique 4-6 has several advantages: it is reliable and ac- curate and provides simultaneously the self-diffusion coefficients of all (normally ") components in a single rapid (normally 10 min or less) experiment without isotope labelling; difficulties arise if transverse spin relaxation is rapid but such problems are rarely prohibitive for lH and 3C nuclei in microemulsions. The present paper briefly describes the principles of the two approaches with reference to studies of simple surfactant systems.In particular, however, measure- ments are presented for the above-mentioned model system chosen for collaborative work. To cover in any detail a complex microemulsion system like this is a con- siderable undertaking; note that the I3C relaxation study for any reliable analysis requires 4-6 sets of relaxation data (different fields etc.) for each of the 15-20 car- bons for each sample composition.Our data are still far from complete, meaning that full analysis is not yet possible. Further studies are underway to extend the data obtained from these microemulsion systems. EXPERIMENTAL SAMPLE PREPARATION AND CHEMICALS The compositions of the different samples investigated in the self-diffusion studies are shown in fig. 1. The detailed phase diagrams of these two four-component systems are under study by P. Stenius' group in Stockholm and, of course, a correlation of the present findings with the phase equilibria will be an essential part of this research. The samples were prepared by weighing into glass ampoules (ca. 1 g total amount in a 3 cm3 ampoule for the diffusion work and ca. 10 g total amount for the relaxation work). After temporary sealing with rubber stoppers and cooling, the ampoules were flame-sealed, equilibrated at 25 "C and then inspec- ted through crossed polarizers, if necessary, to determine the phase state.Isotropic solutions were then transferred to n.m.r. tubes for the measurements. All samples were prepared from chemicals of puriss or equivalent grade chemicals. The sodium octylbenzene sulphonate was from the special batch prepared for the 'Lund Project' by P. Bothorel's group in Bordeaux to ensure sample uniformity between the various labora- tories. Heavy water (Ciba-Geigy or Norsk Hydro, Rjukan, Norway, 99.8% 2H enrichment) was needed to provide an internal field-frequency lock. 3C N.M.R. RELAXATION AND 13C-( 'H) NUCLEAR OVERHAUSER EFFECT (n.0.e.) MEASUREMENTS The 3C T , relaxation measurements were performed at magnetic field strengths of 8.5, 2.1 and 1.4 T on a Nicolet NM 360 spectrometer, a Jeol FX-90Q spectrometer and a Jeol FX-60B.LINDMAN et al. 319 butanol/SOBS = 2 water toluene butanol/SOBS = 2 water decane Fig. 1. (a) Sample composition notation and the result of a coarse trial and error phase-region investigation at 25 "C along the indicated dilution lines in the 'Lund project' system sodium octylbenzene sulphonate + D20 + butan-1-01 + toluene (weight fractions). (b) The corres- ponding diagram for the analogous decane system at 25 "C.320 3C N.M.R. STUDIES OF MICROEMULSIONS spectrometer, respectively. The T I were measured using the fast inversion recovery technique. In addition, n.0.e. experiments were performed at 8.5 T on the Nicolet spectrometer.The detailed procedures for these measurements as well as for extracting the T , and the n.0.e. may be found e l ~ e w h e r e . ~ ~ ~ The temperature was 27 “C for all relaxation measurements. FOURIER-TRANSFORM N.M.R. SELF-DIFFUSION MEASUREMENTS The measurement procedure has previously been outlined in a series of paper^.^-^ The latest developments of these procedures make possible convenient and accurate multi- component self-diffusion studies through a single measurement series. Individual self- diffusion coefficients of each component (typically 4) in the concentrated microemulsion sy- stems discussed in the present paper were measured on protons at 99.6 MHz and can be performed in < 10 min per sample. Water self-diffusion coefficients are evaluated from the averaged ROH-water proton signal according to a previously described proced~re.~ All diffu- sion measurements were performed at 25 “C.SELF-DIFFUSION PRINCIPLES The pulsed-gradient spin-echo experiment is a convenient method for the detection and measurement of the lateral displacement of matter. In the present implementation and application of the technique the self-diffusion in isotropic multi-component solutions is detected through a Fourier-transform procedure. It is important to note that individual molecular self-diffusion coefficients are mea- sured separately and that the monitoring time is of the order of 100 ms. During this time span all measured (average) self-diffusion coefficients correspond to displace- ments over macroscopic distances several orders of magnitude larger than any struc- tural element in the solution.knowledge of individual-component self- diffusion behaviour in multi-component solutions provides a unique source of in- formation with regard to the solution structure. In micellar solution the distinct separation of the hydrocarbon droplet-like micelles and the continuous aqueous pseudophase is reflected in very slow surfactant and solubilized hydrocarbon diffusion. Water diffusion is 1-2 orders of magnitude more rapid. l4 Similarly, the distinct separation of water into droplets in a continuous hydro- carbon solution in inverted micellar solutions manifests itself in rapid hydrocarbon diffusion and very slow water diffusion. The term ‘microemulsion’ implies some rather distinct separation into hydro- philic and hydrophobic domains as well.According to our findings this is not nor- mally the 9-1 The prevalent structural model of ‘microemulsions’, up to very recently, has had an unfortunate origin in normal macroscopic emulsions. It has been argued that the simultaneous incorporation of large amounts of hydrocarbon and water into solution is accomplished through formation of emulsion-like aggre- gates, only of smaller size. Suggestions of bicontinuous structures of various kinds have also been rnade.l5-l9 Since isotropic solutions can be obtained over rather wide concentration ranges in ‘cosurfactant microemulsion systems’ also without surfactant, and since numer- ous examples of simultaneously water- and hydrocarbon-soluble compounds exist [e.g.alkylamides, substituted aromatics and short-chain alcohols (which exhibit quite large one-phase areas in three-component water + hydrocarbon systems)] it is more reasonable to use normal solutions as a starting point for any structural As discussed in earlierB. LINDMAN et al. 32 1 characterization of microemulsions. Only a few microemulsion systems exhibit self- diffusion behaviour consistent with the ‘closed-droplet type’, and our results indicate that different ‘microemulsions’ can structurally span the whole range between the two extremes. Their structure is sensitive to constituent as well as to the relative composition of the microemulsion. Only in one case, the AOT (sodium di-2-ethyl- hexyl sulphosuccinate) + water + hydrocarbon systems, have we found evidence for a distinct water-in-oil structural model within the whole, extensive isotropic solution region, regardless of the relative proportions of the constituent^.^^ RESULTS AND DISCUSSION Fig.2 illustrates the very pronounced effect of the alcohol chain length on the component self-diffusion behaviour in a typical cosurfactant microemulsion. It is seen that a maximum ‘structure-breaking’ effect occurs for C4-C5 alcohols, which parallels the behaviour of the composition extent of the isotropic phase region. Beginning with the C9 and C10 alcohols, the results do become consistent with a closed-water-droplet structural model, otherwise the degree of structural segregation between predominantly hydrophilic and hydrophobic constituents is evidently low.Note that although the alcohol self-diffusion coefficients decrease strongly with al- cohol chain length, there is actually an increase in alcohol self-diffusion throughout the same series, when compared with the neat alcohols. With reference to the sample compositions indicated in fig. 1 ( a ) and (b), the results of a systematic survey of the self-diffusion behaviour in the ‘Lund project’ systems is summarized in fig. 3(a)-(c). Except for the water-rich regions with es- sentially normal micelles, the diffusion coefficients are high and inconsistent with structurally segregated domains. For most surfactants rapid spin relaxation makes self-diffusion studies on the surfactant constituent impossible or difficult by our tech- nique.The aromatic proton part of sodium octylbenzene sulphonate, however, has isolated bands with sufficiently long transverse relaxation times for a determination of the surfactant self-diffusion coefficients. As shown by the results in fig. 3(a)-(c), the surfactant diffusion is very rapid (typically 5 times more rapid than in micellar systems). This information clearly suggests that any local order and short-term ag- gregate size are less than those of micelles and that this particular isotropic solution phase cannot be much more organized than a partly associated liquid throughout most of its composition region. The most evident structural model would therefore be one of disrupted or distorted micelles for this and most other cosurfxtant microemulsions.I3C RELAXATION THEORY AND PRINCIPLES A rather general important observation in n.m.r. relaxation studies of associated surfactant systems is that of rather long relaxation times combined with a marked magnetic-field dependence.*OP2 This is incompatible with a single-exponential decay of the correlation functions describing the time dependence of the interaction of the nuclear spins. Instead it demonstrates that motions which span widely different time-scales make significant contributions to relaxation. An important feature of the current picture of surfactant systems such as micelles and liquid crystals is that of liquid character on the molecular level combined with marked organization and order over longer distance^.^^.^^ As examples we cite the findings of diffuse high- angle X-ray diffraction in combination with characteristic low-angle X-ray diffrac-322 sx10-'2 - I I I I I 1 I \ - L 3C N.M.R.STUDIES OF MICROEMULSIONSB. LINDMAN et al. 323 1 . 5 e I v1 % 1.0 6 h I 0 - 0.5 20 31 43 54 65 76 sample 1.5 ,- I v1 N E 1.0 E: 5 a I 0 . 5 70 71 72 73 74 75 76 77 sample 9 7 2t x L I I I 8 t 1 70 71 72 73 74 75 76 sample Fig. 3. Some self-diffusion results [with reference to the notation in fig. l(a) and (b)] for the 'Lund project' systems. (a) and (b) A , water; 0, toluene; 0 , butanol; x , surfactant; (c) A , water; 0, butanol; 0, decane; x , surfactant. of relaxation for microheterogeneous systems of the type considered here. The model assumes that the nuclear interaction is modulated in two steps over widely different time-scales.There are rapid local motions within the structure which do not give complete isotropy but reduce the interactions to a level which can be character- ized by an order parameter. This is described, as for anisotropic liquid-crystalline systems, as the degree of orientation of a molecular vector with respect to the normal to a hydrophobic-hydrophilic interface. The remainder of the interaction must be modulated by motions which (isotropically) change the direction of the hydro- phobic-hydrophilic interface. These motions fall in the range 10- 7-10-9 s for solutions of spherical micelles or cubic liquid crystals. For rod micelles and anisotropic liquid crystals they are much longer. In fact, for the latter case they are generally too long (if the microcrystallites are not very small) to affect the n.m.r. parameters.Detailed discussions on the interpretation of 3C n.m.r. relaxation in micro- heterogeneous systems may be found elsewhere 7 , *, 2 3 where investigations of two- and three-component surfactant systems are also reported. The expressions for T I and n.0.e. of 13C are (under wide-band proton decoupling conditions) given by and324 3~ N.M.R. STUDIES OF MICROEMULSIONS where N is the number of protons directly bonded to the carbon atom, rCH is the carbon-hydrogen distance (taken to be 1.09 A) and the other symbols have their usual meaning. The various reduced spectral densities, J((o), are for the applied two- step model of relaxation given by where 7: and zs are the correlation times of the fast local and slow overall motions, respectively (exponential correlation functions are assumed), S is the (local) order parameter defined by 1 2 S?-(3 c0s20 - l>f (4) where 0 is the angle between the C-H bond vector and the local director, which is taken to be perpendicular to the hydrophobic-hydrophilic interface, and the symbol ( > f indicates an ensemble average over the fast local motion.The two-step model of relaxation, as briefly sketched above, has been rather extensively applied for simple micellar systems.8 These studies have involved the determination of a larger number of experimental relaxation data than minimally required in the analysis, thus providing a critical examination of the model used. Significant support for the model is also provided by the close agreement between the deduced order parameters and those of liquid- crystalline phases with similar composition.8.29 RESULTS AND DISCUSSION I3C relaxation studies were performed for 10 different sample compositions (see table 1) in different parts of the extensive isotropic-solution phase of the sodium octylbenzene sulphonate + butanol + toluene + water system. Measurements were performed of T I at 1.4 and 8.5 T and of the n.0.e. at 8.5 T while complemen- tary measurements (mainly of T , at 2.1 T) were performed for some sample compo- sitions and are under way for others. The experimental results are shown in fig. 4 and table 1. Note that in addition to a concentration dependence and a variation along the chain the main characteristics of the T , are that they are rather long at the same time as they are markedly field dependent; furthermore, the n.0.e.is below the theo- retical upper limit (1.99). This demonstrates that there is an important contribution to relaxation from slow motions but also that these average only a small part of the total interaction [cf. eqn (1)-(3)]. As judged from the line width, T2 is found to be rather long as well. Since T, are difficult to quantify for narrow lines they are not reported. However, the T, observations give support for the general interpreta- tion. In particular, note that since T2 is very strongly dependent on zf, the long T2 show that zf cannot be very long. Using the two-step model of relaxation as outlined above, the relaxation data ( T I and n.0.e.) were analysed to provide the three parameters z:, zf and S.S and z: values along the chain are presented in fig. 5 and 6 for four sample compositions while zt data are presented in table 2. While 7: gives quantitative information on local alkyl chain dynamics, S and z: contain information on amplitude and time-scale of more long-range motions and are thus related to solution structure. The packing of surfactant alkyl chains varies with the shape of surfactant aggregate, sphere, reversed sphere, rod, lamellar e t ~ . , ~ O 3 3 and it has in particular been observed that the order-parameter profileB. LINDMAN et al. 325 varies with structure for liquid-crystalline phase^.^^-^ If aggregate rotation and mole- cular lateral diffusion are assumed to determine zs, then a small 7:. value implies high ‘effective’ curvature and is inconsistent with large ordered domains.I I I 2 0 9 1 .o 1 2 3 L \ - I I 1 I I I 1 I I P h l P h 2 1 2 3 4 5 6 7 8 2.0 1.6 1.2 0.8 0.4 9 Fig. 4. Relaxation rate and Overhauser enhancement (y) for sample 3 (see table 1 for sample composition) as a function of carbon position. The main figure refers to sodium octylbenzene sulphonate and the insert to butanol. 17 and 0 refer to (T,)- measured at 1.4 and 8.5 T, respectively, while x refer to r,~ measured at 8.5 T. The error bars correspond to an approximately 80% level of confidence. For points without error bars, the errors are smaller than the size of the symbols. Numbering of alkyl-chain carbons starts from the polar head- group. Phl and Ph2 refer to the meta and para carbons in the phenyl ring.From the data in fig. 5 it is inferred that the order parameters are about as large in the microemulsions as in solutions of simple micelles and in liquid crystalline phases. This demonstrates a marked ordering of the alkyl chains over periods of the order of at least s. For such periods there is no marked difference between the microemulsion solutions and liquid crystals (where, of course, ordering is on a long time-scale). However, if one considers T: a very clear difference is apparent. Thus it is striking that 7: lies in the range 1-2 ns over a very wide composition range, demon- strating that the effective curvature of the hydrophobic-hydrophilic interfaces of the microemulsions seems to agree throughout with an effective radius corresponding to less than the length of the extended surfactant molecule.These findings are consist- ent with a predominance of small aggregates. If there are anisometric aggregates to an important extent, or if there are spheres (of oil or water) with radii markedly larger than the surfactant molecule length, then these findings indicate that the aggregates must be very flexible. In fact our results suggest that the hydrophobic- hydrophilic interfaces are very flexible and change direction extensively within a few nanoseconds. The rather even distribution of cosurfactant between different domains 3 8 reduces the interfacial energy and makes possible such high interface flexi- bility. The remark that with the considerable pairwise miscibility of components one should not be far from the rather structureless limit of simple solutions also points to such a situation. As regards the z: values, they are of the order of 10 ps, which is close to what is observed for simple alkanes; 8 , 3 9 an analogous observation has been made for lipid326 3C N.M.R.STUDIES OF MICROEMULSIONS M -. G v, I V m I V 4 I u v, v,mm 8 y g +I +I +I +I cr, e(u9 ? ?"\9 9999099000 meInnmemC\ICI 0 0 0 0 0 0 0 0 99999999 +I +I +I +I +I +I +I +I -?-?\9'?\9?\9'9 m * o o t - m v , C \ I 9900000080 +I +I +I +I +I +I +I +I woc't-00qqCY?oq 9~00000000 mFmr-lnr-Paw +I +I +I +I +I +I +I +I +I +I c?c?Y-"C?"c?c?"c'! + * e 0 0 0 0 m o r - m 0 0 0 0 0 0 0 0 0 0B. LINDMAN et al. 327 0.5 s 0.25 0 H 1 2 3 4 5 6 7 8 1 2 3 , 4 5 6 7 8 I I I I I I I 1 1 2 3 4 5 6 7 8 I 1 1 1 1 1 1 1 1 2 3 L 5 6 7 8 carbon no.Fig. 5. The order parameter (see under Theory and Principles) for the alkyl chain of sodium octylbenzene sulphonate in four samples. (a), (b), ( c ) and ( d ) correspond to samples I, 5,9 and 11, respectively (see table 1 for sample compositions). Numbering of carbons starts at the carbon closest to the phenyl ring. The error bars correspond to an approximately 80% level of confidence. P 3 E si 1 I 1 , I t L 1 2 3 L 5 6 7 8 3 0 VJ - u ,a 20 10 0 carbon no. 0 P 4 P 46 1 2 3 L 5 6 7 8 Fig. 6. The fast correlation time (see under Theory and Principles) for the alkyl chain of sodium octylbenzene sulphonate in four samples. (a), (b), (c) and ( d ) correspond to samples 1, 5 , 9 and 1 1, respectively (see table 1 for sample compositions).Numbering of carbons starts at the carbon closest to the phenyl ring. The error bars correspond to an approximately 80% level of confidence.328 3C N.M.R. STUDIES OF MICROEMULSIONS Table 2. Slow-motion correlation time, T:, for different sample compo- sitions (given in table 1) sample T 3 0 - 9 s 3 4 5 6 7 9 10 11 1.8 & 0.4 1.3 k 0.4 0.8 & 0.2 1.7 & 0.5 1.0 & 0.8 1.1 0.3 1.7 4 0.4 1.7 & 0.5 bilayers in liquid crystals.40 7: decreases on moving away from the polar head. For alkanes and non-polar derivatives thereof the correlation time varies little between different met h y lenes. CONCLUSIONS There is at present no theoretical model available against which we can quan- titatively test our results. Whilst awaiting the development of such models one may loosely talk about the degree of structure or organization in the isotropic solution phases.In one limit one has the structureless case of simple solutions such as alcohol + water mixtures and in the other (for surfactant systems with liquid chains) the well organized liquid crystals. The microemulsions clearly fall in between these two situations. Both the present approaches, i.e. Fourier-transform n.m.r. self-diffusion and multifield 13C n.m.r. relaxation, are sensitive to the degree of structure in the system but in very different ways. The self-diffusion results emphasize that over wide con- centration ranges there are no marked long-lived structural barriers to translation over macroscopic distances for either hydrophilic or hydrophobic compounds.The self-diffusion work also emphasizes that this situation is extremely dependent on cosurfactant chain length. The 13C relaxation results show that the motion of the alkyl chains is about as rapid as that of liquid hydrocarbons and emphasize also that a substantial fraction of the surfactant molecules are oriented and have marked motional anisotropy. Thus there is organization in the system and the creation of hydrophobic-hydrophilic interfaces which induces slight anisotropy in the motion. This anisotropy is similar in magnitude for the microemulsions as for liquid- crystalline phases and micelles but for the microemulsions the lifetime of the anisotropy is very short, i.e. one or two nanoseconds. Microemulsions, ed. L.M. Prince (Academic Press, New York, 1977). Microemulsions, ed. 1. Robb (Plenum Press, New York, 1982). P. G. De Gennes and C. Taupin, J . Phys. Chem., 1982, 82, 2294. P. Stilbs and M. E. Moseley, Chem. Scr., 1980, 15, 176. P. Stilbs and M. E. Moseley, Chem. Scr., 1980, 15, 215. P. Stilbs, J . Colloid Interface Sci., 1980, 87, 385. T. Ahlniis, 0. Soderman, C. Hjelm and B. Lindman, J . Phys. Chem., 1983, 87, 822. B. Lindman, P. Stilbs and M. E. Moseley, J . Colloid Interface Sci., 1981, 83, 569. * H. Walderhaug, 0. Soderman and P. Stilbs, to be published.B. LINDMAN et al. 329 l o B. Lindman, N. Kamenka, T-M. Kathopoulis, B. Brun and P-G. Nilsson, J. Phys. Chem., 1980,84, l 1 P. Stilbs, J . Colloid Interfuce Sci., 1982, 89, 547. l 2 P. Stilbs, K. Rapacki and €3.Lindman, J . Colloid Inlegace Sci., in press. l 3 P. Stilbs and B. Lindman, J . Colloid lnterjuce Sci., in press. l 4 B. Lindman and H. Wennerstrom, Top. Curr. Chem., 1980, 87, 1. 2485. K. Shinoda, J . Colloid Interfuce Sci., 1968, 26, 70. S. Friberg, I. Lapczynska and G. Gillberg, J . Colloid Interface Sci., 1976, 56, 19. l 7 M. Lagues, R. Ober and C. Taupin, J . Phys. Lett., 1978, 39, 487. l 8 Y. Talmon and S . Prager, J . Chem. Phys., 1978, 69, 517. l 9 L. E. Scriven, in Micellization, Solubilization and Microemulsions, ed. K. L. Mittal (Plenum Press, 2o H. Wennerstrom, G. Lindblom and B. Lindman, Chem. Scr., 1974, 6, 97. 2 1 U. Henriksson, L. Odberg, J. C. Eriksson and L. Westman, J . Phys. Chem., 1977, 81, 76. 2 2 B. Halle and H. Wennerstrom, J. Chem. Phys., 1981, 75, 1928. 2 3 H. Wennerstrom, B. Lindman, 0. Soderman, T. Drakenberg and J . B. Rosenholm, J . Am. Chem. 24 H. Wennerstrom and B. Lindman, Phys. Rep., 1979, 52, 1. 2 5 B. Lindman and H. Wennerstrom, in Solution Behaviour of Surfactants, ed. K. L. Mittal and E. J. Fendler (Plenum Press, New York. 1982), vol. 1, pp. 3-25. 2 h G. J. T. Tiddy, Phys. Rep., 1980, 57, 1. 2 7 V. Luzzati, in Biological Membranes, ed. D. Chapman (Academic Press, New York, 1968), chap. 3 , ** K. Fontell, Prog. Chem. Fats Other Lipids, 1978, 16, 145. 2 9 T. Ahlnas, 0. Soderman, H. Walderhaug and B. Lindman, in Surfactants in Solution, ed. K. L. 3o J. N. Israelachvili, S. Marcelja and R. G. Horn, Q. Rev. Biophys., 1980, 13, 121. 3 1 D. J. Mitchell and B. W. Ninham, J . Chem. Soc., Furaday Trans. 2, 1981, 77, 601. 3 2 K. A. Dill and P. J. Flory, Proc. Natl Acad. Sci. USA, 1981, 78, 67. 33 D. W. R. Gruen and E. H. B. de Lacey, in Surfhctants in Solution, ed. K. L. Mittal and B. Lindman 34 J. Seelig, Q. Reu. Biophys., 1977, 10, 353. 3 5 J. Charvolin, P. Manneville and B. Deloche, Chem. Phys. Lett., 1973, 23, 345. 3 6 T. Klason and U. Henriksson, in Solution Behaviour of Surfactants, ed. K. L. Mittal and E. J. 37 T. Klason and U. Henriksson, to be published. 3 8 J. Biais, P. Bothorel, B. Clin and P. Lalanne, J . Disp. Sci. Technol., 1981, 2, 67. 39 R. R. Lyerla Jr, H. M. McIntyre and D. A. Torchia, Macromolecules, 1974, 7. 11. 40 M. F. Brown. J. Chem. Phys., 1982, 77, 1576. New York, 1977), vol. 2, p. 877. Soc., 1979, 101, 6860. p. 71. Mittal and B. Lindman (Plenum Press, New York, 1983), in press. (Plenum Press, New York, 1983), in press. Fendler (Plenum Press, New York, 1982), vol. 1, p. 417.

ISSN:0301-7249    

DOI:10.1039/DC9837600317

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

GENERAL DISCUSSION Dr. B. Vincent (Bristol University) said: Dr. Croucher’s estimation of the thick- ness (L) of the steric barrier is based on the dimensions of an extended, single chain at the interface. However, I doubt whether this is realistic considering the structure of the graft-copolymer stabilizer used. This presumably consists of a poly(dimethylsi1oxane) (PDMS) backbone with poly(methy1 methacrylate) (PMMA) side chains. The latter will become incorporated into the matrix of PMMA chains during the latex polymerization, leaving only relatively short PDMS loops and tails extending into the dispersion medium. Thus the value of L is probably significantly less than the value of 100 nm estimated. Considering this in combination with the relatively large PMMA particle diameter (0.7-0.9 pm) could mean that some (weak) floccs are indeed present in the system, particularly at the higher volume fraction used in the rheological studies.Could Dr. Croucher justify the use of eqn (25) to relate L ( T ) to x ( T ) , and also the separate splitting of eP into two terms ( E , + E ~ ) ? Surely, it would be better to consider the net pair potential in evaluating this term? Dr. M. D. Croucher (Xerox, Ontario) said: Since we were unable to experimen- tally measure the thickness of the steric barrier, our estimation was indeed based upon the dimensions of a single chain at the interface. I would disagree however with Dr. Vincent’s argument that only relatively short PDMS loops and tails will extend into the n-hexadecane dispersion medium.Recent work on the morphological features’ of these particles has indicated that the PDMS stabilizer is nut anchored via the PMMA to the surface of the particle but that the PDMS is incorporated throughout the core of the particle. This indicates that long tails could conceivably be stabilizing the particles. Therefore, our estimate of the barrier thickness could be quite realistic. These data also cast some doubt on the traditionally held view of the loop-and-tail conformation of the stabilizer at the surface of the particle. The splitting of E~ into E, terms follows the traditional theories of steric stabiliz- ation.2 It would indeed be more appropriate (and more complicated) to consider the net pair potential as indicated by Dr. Vincent. Eqn (25) was used to obtain a value for the barrier thickness as a function of temperature because of its simplicity. We could have estimated the values for L(T) from the well known equation3 for the expansion, due to excluded-volume effects, of a polymer in dilute solution. Calcula- tions of this type gave results which were similar to those obtained from eqn (25).If anything, eqn (25) underestimates the expansion of the chain at the interface caused by increasing the solvency of the dispersion medium for the steric stabiliser. 0. Pekcan, M. Winnik and M. D. Croucher, J. Polym. Sci. Lett., 1983, 21, 1011. D. H. Napper, J. Colloid Interface Sci., 1977, 58, 390. P. J. Flory, Principles ofpolymer Chemistry, (Cornell University Press, Ithaca, 19531, chap. 12. Dr. J. S. Higgins (Imperial College, London) said: Following Dr.Vincent’s query of the extent of the polymeric layer (100 nm) proposed in the paper, I would point out that the agreement with Dawkins and Taylor [ref. (1 1) quoted in Dr. Croucher’s paper] is not strictly relevant. In their system the stabilizer is a block copolymer of332 GENERAL DISCUSSION polystyrene with poly(dimethylsi1oxane). The PS is incorporated in the core and the PDMS is thus end-linked to the particle surface in contrast to the samples of Dr. Croucher. In his case the stabilizing PDMS is grafted at arbitrary points and its extent from the surface will be severely restricted. Indeed large portions of the molecule may well be included within the core particle. The value of 100 nm is based on the end to end distance of a very expanded PDMS molecule of M, = 5 x los.The arguments about extension of chains restricted at a surface is probably irrele- vant with this graft system unless there is of order one graft per molecule (and this value is not given). It would seem that the radius of gyration of the PDMS molecule in solution might give a more realistic idea of its extent away from the surface and this is probably the upper limit. Dr. Th. F. Tadros (ICZ, Jealott 's Hill) said: In his paper Dr. Croucher accounted for increase of viscosity with increase of temperature by increase in hydrodynamic volume fraction as the solvent becomes better for the chains on increase of temper- ature. Surely this effect cannot be large enough to account for the observed increase. Even if one accepts a barrier thickness of 100 nm (which in my judgement is an overestimate) an increase of say 10% in the barrier thickness will only lead to a small increase in qh(qh = qS[l + (6/a)I3) since S/a = 1/8 is fairly small.An alternative explanation of the results could be given in terms of the possible flocculation of the suspension with increase of temperature. As the temperature is increased, the solvent becomes better for the chains leading to some desorption. This will lead to some flocculation and hence the viscosity increases. Could Dr. Croucher comment on this effect and state what evidence he has that flocculation indeed did not occur? Dr. M. D. Croucher (Xerox, Ontario) replied: There are two factors which we feel are important in controlling the rheological properties of sterically stabilized disper- sions.The first is the hydrodynamic volume of the particles (which is temperature dependent) and the second is the thickness of the steric barrier and the effect this will have on the microstructural properties of the dispersion in shear flow. If the steric barrier is thin then it is essentially undeformable, however, as it becomes larger it will become capable of being deformed. The effect this plays in controlling the rheological behaviour remains to be explored. Although the barrier thickness is only thought to increase by l0-15% over the temperature range studied, OH will increase by ca. 7-10%, which is not insignificant when one compares the increase in the viscosity as a function of the volume fraction as shown in fig.2 of our paper. The temperature dependence of the viscosity is therefore not inconsistent with this explanation. Although it is difficult to rule out weak flocculation completely in these concentrated dispersions, the Hamaker con- stants of the core and the dispersion medium were chosen to give VA sz 0. Secondly, after the rheological measurements had been made the dispersion was diluted and its particle size measured. There was no change in the particle size after shearing the dispersion. With a high-molecular-weight stabilizer it is often difficult to redisperse the particles after they have flocculated, but evidence for this was not apparent in the dispersion that was used in these experiments. Dr. J. H. R. Clarke (UMZST, Manchester) said: Although all the systems discus- sed in the paper by Croucher and Milkie show shear-thinning behaviour, they men- tion that some similar colloidal systems apparently exhibit shear-thickening as the strain rate is increased from zero.Are these cases well characterized? Would they comment on the possible mechanism of shear-thickening? On the basis of com-GENERAL DISCUSSION 333 parisons with simple liquids, shear-thinning is normally to be excepted except at very large strain rates. Could thickening arise from shear-induced flocculation or from some complex hydrodynamic effects? in the literature which indicate that sterically stabilized dispersions exhibit shear- thickening as a function of increasing shear rate. This has been found to occur at a critical shear rate and critical volume fraction.Two explanations for discontinuous dilatancy have appeared in the literature: The treats this effect as an experi- mental artifact while the second3v4 treats dilatancy as an intrinsic property of the dispersion in steady shear flow. Experimentally, it has been observed4 that rapid acceleration of the rheometer is not a necessary condition for discontinuous dilatancy as had been suggested.’v2 This casts some doubt on the explanation that the phenomenon is an experimental ar- tifact. The second explanation is that abrupt shear-thickening is caused by a tran- sition from an ordered to a disordered flow pattern. Hoffman3 has monitored these shear-induced changes optically and observed the loss of two-dimensional hexagonal symmetry at the point of melt fracture in the sample which is found to occur over a narrow range of shear rates.There is, at the present time, a paucity of data on discontinuous dilatancy which makes it difficult to draw any definitive conclusions as to the fundamental mechanism of shear-thickening. Dr. M. D. Croucher (Xerox, Canada) said: There are numerous T. A. Strivens, J. Colloid Interface Sci., 1976, 57, 476. C. E. Chaffey and I. Wagstaff, J. Colloid Interface Sci., 1977, 59, 63. R. L. Hoffman, Trans. SOC. Rheol., 1972, 16, 155. S. J. Willey and C. W. Macosko, J . Rheol., 1978, 22, 525. Prof. L. V. Woodcock (Uniuersity of Bradford) said: My colleague Prof. M. F. Edwards and I have turned our attention to the mechanism of this shear-thickening in concentrated colloidal dispersions.When colloidal dispersions are sheared at high concentrations the ratio of stress (oXy) to the strain rate (9) (i.e. the ‘apparent viscosity’) is sometimes seen to increase with strain rate. This well known, but largely unexplained, transport phenomenon, termed ‘shear-thickening’, is encountered in many, if not all, colloidal dispersions at sufficiently high concentrations and shear rates. It is rare, however, in single- component liquids or simple fluid mixtures, which in the non-Newtonian region generally exhibit pseudoplasticity or shear-thinning behaviour. We have recently been investigating the physical molecular basis of shear-thickening rheology with a view to using the predictive powers of computer simulation techniques to aid the engineer in the design of the many industrial processes in which these shear- thickening materials are involved.The sudden onset of shear-thickening gives rise to flow difficulties and attendant heat-transfer problems in such diverse operations as pipe flow, roll coating and mixing1 Only limited progress has been made towards alleviating the uncertain conse- quences of shear-thickening at the engineering design stage for two basic reasons.2 First, unlike the linear viscosity, the non-linear stress-strain rate ratio lacks a precise definitive value which is independent of the boundary conditions, system size and shear mechanism. Shear-thickening of colloidal dispersions is particularly difficult to characterize in a quantitative manner3 owing to the strong dependence on experi- mental artifacts.For example, both surface effects and bulk viscoelastic effects are inextricably and inevitably involved, since superficially induced changes in both shape and size of the stress tensor occur for all finite-gradient irreversible deform- ations in real processes.334 GENERAL DISCUSSION Secondly, although there exists already some literature on theories of shear- thickening based on molecular models,2 these tend to be rather speculative conse- quences of colloidal forces or structures, and lack the generality that seems to be required from the body of experimental evidence. Thus, quantitative predictions from the molecular data are not available to the engineer. Whilst considering the possible ways in which computer simulation at the mole- cular level might be helpful in providing the 'experimental' framework for theor- etical predictions, and reviewing the previous non-equilibrium molecular-dynamics (NEMD) studies on simple liquids, we believe we have the general explanation for the phenomenon of shear-thickening already.This stems from previous NEMD studies on dense simple liquids of the effects of both compressive4 and shear ~ t r a i n s . ~ The reason evidently rests in a bifurcation of the relative relaxation times for momentum dissipation by the normal transport process and by structural re- arrangement. Generally speaking, it seems to have little to do with the peculiar facets of the intermolecular potentials, colloid particle shapes or effective pair potentials. A long-standing experimental observation is that shear-thickening in concen- trated colloidal dispersions is often associated with dilatancy, i.e.an attendant expansion of the sheared fluid (indeed, so much so that the terms are unfortunately often used interchangably). Whereas shear-thickening refers to the behaviour of the rate of momentum-transfer equation-of-(sheared) state axu/j(j), dilatancy refers to the thermodynamic mechanical equation-of-state as a function of strain rate p , V(j)T. One can deduce from the thermodynamic equilibrium conditions at constant pressure (C?G/C?~)~ = V or constant volume (aA/aV), = p (G and A being the Gibbs and Helmholtz free energies, respectively) that since for equilibrium C?G 2 0 or 8A 2 0 and p and I/ are necessarily positive, then at least small perturbations from equilibrium for all systems lead to dilatancy, which include shear-induced pressure increases at constant volume also.Hanley and Evans6 have further shown the pres- sure to be a non-analytic function of strain rate, notably: lim+o p = PO + p ~ j " where rn < 2 and found from simulation studies to be ca. 3/2. Thus, in the non-linear region compressive stresses and hence bulk viscoelastic relaxation are essential parts of the momentum dissipation processes. Intuitively one expects these to become increasingly important with strain rate. Consequently, con- trary to the widely used assumption in non-linear hydrodynamic calculations that the thermodynamic functions of statefb, V, T ) are independent of the strain rate under steady-state shearing conditions, there is a significant, and possibly dominant, change in the thermodynamic equation-of-state p , V, T ( j ) as a function of shear rate.It seems to be this change that provides the thermodynamic driving force for the structural rearrangement leading to shear-thinning behaviour. Even liquid argon, when sheared at sufficient rates (ca. 1 0 I 2 s-l) can be shown to be a viscoelastic, dilatant, shear-thinning material. That even the simplest liquids may be both dila- tant and shear-thinning is noteworthy because it begs the question not so much 'why do colloidal dispersions shear-thicken?' (in the absence of structural response this would be expected for all liquids), but 'why do they not continue to shear-thin at the higher shear rates?'. The molecular mechanisms of shear-thinning in simple Lennard-Jones liquids have been studied using NEMD by Heyes et al.5*7 Their calculations show that onGENERAL DISCUSSION 335 the application of an applied strain, not only is there the usual increase in stress, which levels off after the initial viscoelastic response, but also an almost identical and simultaneous increase in the normal pressure (at constant volume) of the same magnitude. The system exhibits a slight shear-stress overshoot but does not reach a shear-thickened steady state associated with the rate of stress creation being equal to its rate of dissipation, as might be anticipated.What happens instead is that in a certain thermodynamic state p , V (f)= the relaxation time for structural rearrange- ments or layering of molecules along the plane of shear becomes less than that for the dissipation of momentum.The increasingly slower (with j ) momentum-transfer process is ‘short circuited’, and both the shear stress and the normal pressure build- up are relieved by internal structural rearrangement, i.e. an alignment of layers of molecules along the plane of the shear. We already have substantial experience from computer studies of analogous behaviour for compressive stresses in the very same simple Lennard-Jones type liquids at high densities, which lead to either glass formation (compressive ‘thicken- ing’?) vis-a-vis homogeneous nucleation (compressive ‘thinning’?). In these isotropic relaxation processes we have just the same situation; in a certain thermodynamic (non-equilibrium) state the slow relaxation of the compressive stress is ‘short cir- cuited’ by local structural rearrangements when the relaxation time for nucleation or local ordering becomes shorter than that for the configurational re-equilibration to the new metastable compressed state, the relaxation time for which is ever- lengthening with increase in density or pressure.Evans and Watts8 have reported NEMD computations which used an effective colloidal pair potential of the screened Coulomb type, and which were specifically designed to observe shear-thickening. Their models, however, were found to behave like simple liquids and were dilatant and shear-thinning. The reasons may be partly due to NEMD boundary conditions being especially favourable to shear-thinning but may also be due to the fact that medium-dependent effective pair potentials are tailored for specific static properties and are generally both state- and property-dependent.Thus, whilst an effective pair potential may always be defined to produce an equilibrium property such as the static two-body structure factor, there can be no expectation that such an effective pair potential should simulate time-dependent non-linear transport phenomena, particularly where changes in the structural state are involved. This would be especially so where the phenomenon in question is not directly determined by the equilibrium manifesta- tions of an effective pair potential, but by the relaxation times. Returning to the results of Heyes et al.,’ their observations can be used to explain why many colloidal dispersions could be expected to shear-thicken when subjected to strain rates of the order of laboratory or engineering time-scales.The shear-thin- ning relaxation time for simple ‘Lennard-Jones’ liquids, 10- s for argon, lengthens to 10-1 (by corresponding states) for a colloid-size particle lo4 times greater in dia- meter and 10l2 times more massive, and by a further factor of at least lo3 when they are suspended in a hydrodynamic medium, for example the original Lennard- Jones fluid at the same original temperature and density. These simple arguments predict an increase in diffusional relaxation time on going from a simple liquid like argon to a concentrated colloidal dispersion of approximately 10’ 5 . If the structural relaxation times leading to shear-thinning are of the same magnitude, then the reason that many colloidal dispersions do not shear-thin, or do not continue to shear- thin at higher shear rates, is simply that the relaxation time for that process is too long, or much longer than the strain rate which leads simultaneously to the dilat- ancy.The structural changes in the dilatancy effect itself remain to be investigated.336 GENERAL DISCUSSION A simplified concept of two divergent characteristic relaxation times, z, for momentum transfer, which lengthens with dilatancy on shearing, and z, for struc- tural reordering for shear-thinning, which shortens with the dilatant effect, seems to offer a general molecular-level explanation for both shear-thinning and shear- thickening in concentrated colloidal dispersions.It further suggests that all colloidal dispersions, where the medium is essentially a perfect isothermal heat bath, should first shear-thin at sufficiently low strain rates in the initial deviation from linearity and ultimately shear- t hicken. Approaching zero strain rates (i, + O)z, > z,, and as the shear rate increases the system begins to exhibit normal shear-thinning. The relaxation time z, decreases whilst zm for the momentum transfer increases with the thermodynamic change in the sheared state until z, > z, - and the ‘short circuit’ of the normal or Newtonian momentum-transfer process begins. There could follow a range of shear rates where the lower, shear-thinned ‘viscosity’ remains constant (the second Newtonian region!), until eventually the strain rate i, will be faster than z, so that the structural relaxation time for the system to shear-thin is too long for it to respond.Conse- quently, the apparent viscosity continues to increase as both z, and z, become too slow for either dissipation mechanism to occur. Consequently, the stress will con- tinue to increase faster than the shear rate, leading ultimately to solid-like behaviour, reminiscent of glass formation due to excessive compressive stresses and slow relax- ation times, or material failure. This is precisely what is seen to happen in laboratory viscometric studies of highly concentrated dispersions. These measurements, however, often show the ‘ap- parent viscosity’, in the transition to shear-thickening, to be dependent only on the steady-state shearing rate and not to exhibit hysteresis or thixotropy.If the spectrum of structural relaxation times for shear-thinning is broad we would expect to see time-dependent thixotropy (from the shear-thickened state), but if it is narrow and much greater than the shear rate then we would have effectively a metastable steady (sheared) state with time-independent, reproducible rheological behaviour on that time-scale. M. F. Edwards and W. C. MacSporran, Proc. tnst. Chem. Eng. Meeting, Chester, 1983. * Dispersion Rheology 1980 (A survey of industrial problems and academic progress), report by H. A. Barnes to Process Technology Group, RSC Industrial Division. D.C-H. Cheng, Further Observations on the Rheological Behaviour of Dense Suspensions, paper presented at a symposium in Eindhoven, The Netherlands, 1983 (preprint).C. A. Angell, J. H. R. Clarke and L. V. Woodcock, Adv. Chem. Phys., 1981, XLVIII, 397. D. M. Heyes, J. J. Kin, C. J. Montrose and T. A. Litovitz, J. Chem. Phys., 1980, 73, 3981. D. J. Evans and H. J. M. Hanley, Physica, 1981, 108A, 567. D. M. Heyes, C. J. Montrose and T. A. Litovitz, J . Chem. SOC., Faraday Trans. 2, 1983, 79, 611. D. J. Evans and R. 0. Watts, Chem. Phys. 1980, 48, 321. Prof. B. J. Ackerson (Oklahoma State University, U.S.A.) said: Prof. Clark and I have been involved in determining the shear-induced microstructure of very dilute (0.1 wt%) aqueous suspensions of submicron particles. The most recent review is given in ref. (1). We find that at zero shear our samples form b.c.c.polycrystals. At low rates of shear the polycrystals are replaced by a flowing twinned b.c.c. crystal structure. Further increase in the rate of shear gives freely slipping distorted h.c.p. layers, followed by a transition to strings of particles aligned along the velocity direction, and finally a transition to an amorphous structure. The amorphous structure exhibits distortion of the local order consistent with the fluid being distor- ted by the applied stress. While these systems may be too dilute to measure viscosityGENERAL DISCUSSION 337 with any accuracy, the viscosity is expected to decrease with increasing rate of shear through the transitions we have observed, until the amorphous structure appears. At this point, a shear-thickening is expected. Recently we have modelled the flowing crystal structure using two-body screened-Coulomb interactions and find that a shear-independent potential gives a good quantitative representation of the mea- sured data.B. J. Ackerson and N. A. Clark, Physica, 1983, 118A, 221. Mr. C. W. J. Beenakker (Instituut-Lorentz, Leiden, The Netherlands) said: De Gennesl conjectured a few years ago that in a sheared suspension of hard spheres (with negligible Brownian motion) an infinite cluster of spheres appears when the concentration exceeds a certain critical value. The transition is similar to percolation or sol-gel transitions. This model predicts at the critical concentration an anomaly in the plot of viscosity against concentration of the type observed by Buscall and McGowan.P. G. de Gennes, J . Phys. (Paris), 1979, 40, 783. Prof. W. B. Russel (Princeton University, U.S.A.) said: The idea of de Gennes concerning the formation of an infinite cluster of spheres rests on the effect of hydrodynamic interactions. In a shearing flow the high viscous resistance to motion along lines of centres at small separations increases the probability of finding spheres near contact.' The argument implicitly assumes a large value of the Peclet number (convection/diffusion). The values of the low shear viscosity and shear modulus reported here were obtained at small Peclet numbers where Brownian motion dominates, Hence the anomaly observed is more likely a true thermodynamic phase transition due to weak attractive potentials as discussed in the paper.G. K. Batchelor and J. T. Green, J. Fluid Mech., 1972, 56, 401. Dr. R. Buscall (ZCI, Runcorn) replied: I do not think that the clustering predicted by de Gennes is responsible for the anomalous viscosity behaviour we observe. If this were so, then I would also expect to see a similar anomaly in viscosity curves for stable latices which approximate more closely to idealized hard spheres. The data of Krieger and coworkers, ourselves and others do not, however, show any such anomaly. Viscosity<oncentration data for two such latices are shown in fig. 1. The data lie on a smooth curve which is described well by the Dougherty-Krieger equa- tion. In the case of our weakly flocculated system we believe the transition to be thermodynamic rather than hydrodynamic in origin.Dr. M. J. Gamey (Unilever Research, Port Sunlight) said: Would Dr Buscall care to comment on the mechanism by which NaCMC causes weak flocculation of the latex. I would not have expected NaCMC, an anionic, hydrophilic macromolecule, to be adsorbed by the polystyrene latex and suspect that floculation arises from a combination of the total ionic strength of the medium and macromolecular exclu- sion (depletion flocculation). Dr. R. Buscall (ICI, Runcorn) said: We have not investigated the mechanism of flocculation in detail, but we do not believe it to be depletion flocculation per se, although depletion flocculation may well intrude as the volume fraction is raised.338 GENERAL DISCUSSION 15 s 10 5 1 1 I I 0 0.1 0.2 0.3 0.4 0.5 cp Fig.1. Plot of reduced viscosity against volume fraction for polystyrene latices dispersed in mol dm-3 sodium chloride solution: e, 2.0 pm diameter; 0, 3.2 pm diameter; Solid line, Dougherty-Krieger equation with k , = 2.6; p = 0.585. The reason for this view is that flocculation occurred irrespective of the concen- tration of added polymer and under the conditions of our experiments there was very little, <0.05%, polymer in the supernatant. The polymer used was a commercial sample and no further purification was carried out on the materials actually used in the rheological experiments. Polycarboxylates tend to bind divalent cations strongly and it would be surprising if our material had not picked up some divalent cations (e.g. Ca2 + or Mg2 +).It was found that if a small sample was converted to the pure sodium form by repeatedly saturating a solution with sodium chloride and dialysing out the excess then this purified material acted as a steric stabilizer rather than a flocculant. It is thus suspected that divalent ions were involved in the flocculation. Possibly, divalent ions form weak bridges between the adsorbed layers. This grade of NaCMC appears to adsorb on polystyrene but this is not to say that all grades do: the degree of carboxylation may be important. Prof. A. Vrij (University of Utrecht, the Netherlands) said: I would like to ask how you obtained your sedimention results for g = 2. It seems to me that a cen- trifugal field of g = 2 must be strongly perturbed by the gravitation field (g = 1).Dr. R. Buscall (ICI, Runcorn) said: The measurements were done using a swing- out rotor with tube holders whose centre of gravity was close to the centre of the measuring tube. Under such conditions the tubes adopt an orientation such that there is a resultant force due to the combined effects of gravity and the centrifugal field which acts down the axis of the tube. Thus by running at suitably low speeds a small net acceleration can be achieved. Obtaining accelerations of the order of 2g admittedly requires very low rotational speeds to be maintained and as a conse-GENERAL DISCUSSION 339 quence the results are not as accurate as those obtained at higher g. A more serious criticism of our technique would be that with our present centrifuge there is a vari- ation in acceleration in the sediment about its midpoint of ca.+lo%. However, small accelerations would be even harder to achieve with a rotor of large enough diameter to reduce this to a few percent. Unfortunately our experience is that floc- culated sediments show highly non-linear, and thus interesting, behaviour at very low g and so we are forced to do the best we can in this region. Prof. H. F. Eicke (Uniuersity of Basel, Switzerland) said: I would like to address two questions to Dr. Langevin. (1) According to extensive self-diffusion measure- ments of all components constituting the multicomponent system and applying a variety of alcohols with different chain lengths,’ C4- and C,-alcohols increase re- markably the self-diffusion coefficients of those components which are confined to the interior or at the interface of the proposed droplets. Alcohols of these chain lengths were used in Dr.Langevin’s study. To what extent do the above findings invalidate the droplet picture (hard-sphere mode!), in particular the theoretical pre- dictions using generalized scaling laws and calculations of correlation lengths? (2) Could Dr. Langevin elucidate her arguments as to why conductivity measure- ments are unsuitable for the observation of percolative phenomena? The contact at a ‘single point’ between the emulsion droplets (considered as hard spheres) certainly can not be considered too literally. Also, how is a collision between droplets to be understood within a network structure (percolation regime)? P. Stilbs, K.Rapacki and B. Lindman, J. Colloid Interface Sci., in press. Dr. D. Langevin (ENS, Paris) replied: (1) We have used the droplets picture for systems along the demixing line closest to the oil corner in the phase diagrams. The systems studied by Stilbs, Rapacki and Lindman, and in particular those of fig. 2 of the paper by Lindman et al. in this discussion, contain a much larger amount of alcohol than ours: more than twice that for the butanol and pentanol systems. It is well known that alcohol fluidizes the interfaces, so that their systems might well be molecular solutions rather than microstructured fluids. Moreover, the corresponding more structured systems along the demixing lines (containing less alcohol but equivalent amounts of water, oil and surfactant) are not hard-sphere systems: the butanol systems would be overcritical and the pentanol ones close to critical. Fin- ally, according to their composition, they are in the structural inversion zone.Thus the droplet pictur? is unlikely to be valid at all. Finally, and in relation to Prof. Eicke’s last point, the theoretical predictions for behaviour close to critical points given by scaling laws are universal and should be valid whatever the structure: droplets, bicontinuous or a molecular solution. (2) Electrical-conductivity measurements are a suitable technique to observe per- colative phenomena only when, during their collision, two droplets can exchange water contents. Of course, the ‘single-point’ contact should not be taken too literally here, and would apply to perfectly spherical, rigid and permanent conducting spheres dispersed in a non-conducting solvent.In such a case, two spheres require an infinite time to expel the solvent between them.l When the average network structure is reached, the relative number of isolated droplets is certainly reduced, but the structure probably evolves by the same mech- anism: exchanges of surfactant molecules between the oil, water and interfacial domains. ’ P. G. de Gennes, PhysicochemicaZ Hydrodynamics, 198 1, 2, 3 1 .340 GENERAL DISCUSSION Dr. W. van Megen (Royal Melbourne Institute of Technology, Australia) said: Could Dr. Langevin tell me how A in eqn (2) is obtained and what information relating to the interparticle forces can be derived from it? Dr. D. Langevin (ENS, Paris) answered: The parameter A in eqn (2) characterizes the strength of the perturbation potential added to the hard-sphere potential of interaction between two droplets [ref.(2) of our paper]. The expressions shown in eqn (2) have been used to fit the light-scattering data. Several experimental values for p are thus given in table 1 of our paper. From these p values one can deduce a first estimate for A . A second estimate can be deduced by fitting the variation in scattered intensity I against q with the form of n given in eqn (2) [I cc q/(8n/i?q)]. The two experimental determinations of A are generally slightly different [ref. (6) of our paper]. This is not surprising since three or more particle interactions are neg- lected in the .n formula of eqn (2), whereas they probably contribute to the virial coefficient p.Dr. Th. F. Tadros (ICI, Jealott’s Hill) said: In her paper Dr. Langevin differen- tiates between two systems: attractive system and hard sphere. The only difference between the two systems is in the chain length of the cosurfactant. I cannot see how a small increase in cosurfactant chain length (say from C, to C,) could lead to a change from an attractive to hard-sphere system. If the surfactant is kept the same how could one visualize that, say, a reduction in alcohol chain length from C6 to C5 would lead to penetration of surfactant chain and hence to attraction. It is hard to believe that this small effect will lead to such an increase in attraction, unless such a change is accompanied by considerable reorientation (or disruption) of the surfact- ant layer, allowing penetration and overlap of the chains to take place.Could Dr. Langevin please comment on this effect? Dr. D. Langevin (ENS, Paris) said: The surfactants which we have used in our measurements were mainly SDS and butanol or pentanol. Although very close to hard-sphere systems, the pentanol systems were slightly attractive (p < 7 instead of PHS = 8). The large increase of attractive forces when pentanol is replaced by butanol (p < - 10) can be entirely explained on a theoretical basis from the differ- ent interpenetration volumes of droplets containing pentanol and butanol, assuming that RHS = Rw + lalcohol, where lalcohol is the alcohol chain length [see ref. (28) of our paper]. This can be understood qualitatively using the following Arguments: (i) The radius R is less than Rw + lsDs [see ref.(6) of our paper], probably because the SDS chains are folded, unlike the case in simple micelles. Then R - RHS < l s ~ s - lalcohol. (ii) The virial coefficient p is proportional to the volume of interpenetration and then varies very rapidly with the difference R - RHS. Finally, it must be recalled that much more progressive changes in interactions can be observed by varying the water-to-soap ratio (i.e. the droplet size), the salt concentration, the nature of the oil etc. Prof. V. Degiorgio (Uniuersity of Pauia, Italy) said: I direct my first question to Dr. Langevin and my second to Dr. Langevin and Prof. Lindman jointly. (1) The description of a microemulsion in terms of a critical system implies the existence of a critical point above (or below) which one observes phase separation.It is not clear to me how percolation theory may be used for such a system, since the former is usually developed for a single-phase system.GENERAL DISCUSSION 34 1 (2) In the cases in which there is evidence of bicontinuous structure, does this appear above or below or both above and below the phase-separation point? The term critical point should indicate a continuous transition (the same structure below and above the phase-separation point). Dr. D. Langevin (ENS, Paris) replied: (1) The site percolation theory is used in single-phase systems, as Prof. Degiorgio states. However, the site-bond percolation theory can be applied to systems that undergo phase separation like gels [ref.(33) of our paper]. (2) It is not clear (at least from the existing experiments) if, when a bicontinuous structure is likely to exist, this structure appears above or below the phase- separation point. The systems close to this point are always in an intermediate structural range: the electrical conductivity a is much larger than that corresponding to a disconnected structure, but smaller than the one corresponding to a completely bicontinuous structure [a linear as a function of cpw; see ref. (20) of our paper]. In other words, the transition towards bicontinuous structure is very progressive and the critical points in water-oil microemulsion systems are usually located in this transition region. This is not in contradiction to the structural-continuity require- ments at the critical point, since percolation is also a second-order transition [see ref.(33) of our paper]. Prof. B. Lindman (Chemical Center, Lund, Sweden) remarked: Our self-diffusion work on these ionic microemulsion systems has not yet been concerned with their behaviour as a function of the composition or temperature distance from critical points. However, this would be an interesting aspect of the present type of work. For simpler systems, in particular two-component non-ionic surfactant-water systems, we have investigated changes in surfactant aggregates on approaching the critical point.’ As Prof. Degiorgio suggests, there appears in these systems to be a con- tinuous transition in aggregation on passing the phase-separation point.P. G. Nilsson, H. Wennerstrom and B. Lindman, J. Phys. Chem., 1983, 87, 1377. Prof. H. Hoffmann (University of Bayreuth, West Germany) said: It is generally assumed that the thickness of an interface scales with the interfacial tension; the thickness becomes larger with decreasing interfacial tension. This means that under very low interfacial tensions the interface is no longer a sharp boundary but a rather diffuse layer. Would it not be possible to apply this concept also to the microemul- sion phase? In that case the interface between the droplets and the continuous phase would become diffuse and it would no longer be possible to speak of well defined structures. I admit that the situation is complicated by the presence of detergent molecules, but it is likely that the interfacial tension of droplets of a hydro- carbon + alcohol mixture in an alcohol/water continuous phase is already low, and so the above-mentioned principle could probably be applied.I wonder whether Dr. Langevin could comment on this problem. Dr. D. Langevin (ENS, Paris) replied: Prof. Hoffmann’s question raises several points. First, it is true that the thickness of an interface scales with the interfacial tension y ; however, two distinct contributions have to be distinguished:’ (i) an ‘intrinsic’ density profile whose width l scales with the correlation length of the concentration fluctuations in the coexisting bulk phases, 5, and (ii) a roughness contribution 5, due to thermal motion at the liquid interface, which is generally of a similar order of342 GENERAL DISCUSSION magnitude to I; [ is proportional to y - 3 .For instance, if E is the distance to a critical point c - E - " and y - E" with p x v/2. So both 5 and [ diverge as E - " . For microemulsion systems the situation is slightly different. When a microemul- sion is in equilibrium with oil and/or water, in many cases the tension is low not because a critical point is approached but because a surfactant layer, probably monomolecular, is present at the interface.2 Then 1 << 5 and the large thickness of the interface is mainly due to thermal fluctuations. Let us now consider a small microemulsion droplet. The thickness of the inter- face between this droplet and the continuous phase will similarly be of the order of the surfactant chain length plus a roughness contribution.The roughness term is expected to be much smaller than that for a flat interface since the contribution of the deformation modes of wavelengths larger than the droplet radius (which is domi- nant there) will be absent. The second point is relative to the problem of the surface tension of the droplet. This quantity does not have a real physical meaning due to the microscopic charac- ter of the droplet. However, one could try to use an average picture of the interac- tion energies between surfactant molecules at the interface. As discussed in ref. (3), the attractive contribution can be represented by an interfacial free energy per unit area of aggregate yo, where yo is close to 50ergcm-2, characteristic of the liquid hydrocarbon-water interface.This is balanced by repulsive terms which are gener- ally more difficult to handle: in ionic surfactants the electrostatic repulsion of polar heads is one such term. One then arrives at an equivalent surface tension for the droplet of the same order of magnitude as the surface tension of the flat interface. One must, however, recall that shape fluctuations in microemulsions are prob- ably induced by fast exchanges of surfactant molecules that can be driven by other forces. Finally, all the X-ray and neutron scattering experiments on microemulsions seem to prove that the oil-water interfaces in the case of droplet (or other) structures are relatively well defined. The scattered intensities at large q follow Porod's law, and the interface thickness remains of the order of the surfactant chain length.4 J.Meunier and D. Langevin, J. Phys. Lett., 1982, 43, L185. See ref. (20) of our paper. J. N. Israelachvili, D. J. Mitchell and B. W. Ninham, J. Chem. Soc., Faraday Trans. 2, 1976, 72, 1525. L. Auvray, J. P. Cotton, R. Ober and C. Taupin, J. Phys. (Paris), to be published. Prof. B. U. Felderhof (RWTH, Aachen, West Germany) said: The dependence of the collective diffusion coefficient on concentration as given in Dr. Langevin's fig. 2 shows a disagreement between theory and experiment. Has she attempted to explain the discrepancy? If these spheres can interpenetrate or coalesce then surely it is incorrect to use the expressions for the hydrodynamic interactions between hard spheres.In addition, could it be that the finite lifetime of the spheres she has alluded to has an influence on the spectrum of scattered light? Dr. D. Langevin (ENS, Paris) replied: Fig. 2 of our paper shows that even when attractive interactions are included in the hydrodynamic motion there is a discrep- ancy between calculated and experimental slopes of the variation in diffusion coeffi- cient with cp at low cp. We believe that these discrepancies arise from the transient character of the aggregates that the droplets can form. This is slightly different from the case of micelles in water, where the surfactant molecules can disassemble completely andGENERAL DISCUSSION 343 where the micelle itself has a transient character. This could lead to modifications of the diffusion process that become important close to the c.m.c.l The lifetime of water-oil microemulsion droplets is apparently much longer because the exchanges involve only small fractions of their constituents. The time-scales of these exchanges are s for alcohol molecules.2 These pro- cesses are probably involved in the droplet-aggregation mechanism.A reversible-aggregation model is likely to account for the deviations from single-particle Brownian motion which we observe. However, to our knowledge a corresponding calculation is not available in the literature. s for surfactant molecules and G. D. Phillies, J. Colloid Interface Sci., 1982, 86, 226. J. Lang, A. Djavanbakht and R. Zana, J. Phys. Chem., 1980, 84, 1544. Prof. V. Degiorgio (University of Pavia, Italy) said: I am interested in the electric birefringence results shown in fig.4 of Prof. Eicke's paper. What is the explanation for the appearance of the peak at a temperature ca. 10°C below T,? Does the amplitude of the electric birefringence follow a pattern similar to that shown by the relaxation time? Prof. H. F. Eicke (University of Basel, Switzerland) replied: The maxima in the z(T) curves are due to a counterbalance of growth and sedimentation of clusters of aqueous microphases. The distances between these maxima and their corresponding T, values (lower consolute points as indicated by the onset of turbidity, fig. 4 of my paper) increase with decreasing partial molar volumes of the hydrocarbons forming the continuous phase (keeping the other components constant).This observation reveals that the attractive potentials between the aqueous microphases increase in the same direction. Amplitudes of the electric birefringence and the relaxation times of the underly- ing processes have very similar patterns if plotted against temperature. B. Lamaire, P. Bothorel and D. Roux, J. Phys. Chem., 1983, 87, 1023. Prof. H. Hoffmann (Bayreuth University, West Germany) said: It was not clear to me from the discussion paper what Prof. Eicke assumes is the cause and the mechan- anism for the appearance of electric birefringence in his systems. I wonder whether he could be more specific as to the physical nature of the phenomena? I would also like to know whether the electric birefringence decay curves from which he evaluates two time constants could not be fitted as well on the basis of a single distorted decay process in which the distortion might possibly be caused by interactions or dispersity in the system? Prof.H. F. Eicke (University of Basel, Switzerland) replied: The electric birefrin- gence phenomenon in three-component (water, Aerosol OT, oil) systems, so-called microemulsions, is observed experimentally to increase (i) with concentration of the microphases (i.e. spherically shaped water droplets covered by a surfactant mono- layer) at constmt temperature, (ii) with increasing size of the microphases (i.e. pro- portional to [H,O]/[surfactant]) at constant concentration and temperature, and (iii) with increasing temperature at constant concentration and size of microphases. (See fig.4 of our paper, where the relaxation time could be replaced by the amplitude of the birefringence, since both plots exhibit almost the same dependence.) Phase diagrams reveal that the above experimental procedures change the com- position of the systems towards the respective lower consolute points. Studies of344 GENERAL DISCUSSION permittivity and, in particular, small-angle light scattering, clearly demonstrate’ the pronounced formation of microphase aggregates as the critical regime is ap- proached. The simultaneously incipient birefringence signals indicate an apparent relation between the aggregational state of the microphases and the electric bire- fringence. The latter statement is nicely illustrated by fig. 4 of our paper, which shows the solvent dependence of the birefringence due to the sensitivity of microphase aggregation (attractive potential) towards the partial molar volumes of the hydro- carbons forming the continuous medium.2 The dependence of the birefringence on the size of the microphases is caused by their increasing polarizability, which leads to increasing interactions between them (and probably their aggregates) in the presence of an electric field.This conclusion may be inferred from a study of the variation of relaxation time with pulse length of the applied electric field: the relaxation times are observed to decrease with decreasing pulse length. Orientation of such non-spherical aggregates, and hence artificial birefringence, can then be visualized easily. A ‘single distorted decay’ would be incompatible with the exponential decays observed exclusively in this study: a distorted decay would necessarily require the assumption of more than one exponential relaxation process.We observe two relax- ation processes, which are assumed to be independent since both respond rather differently to variations in the electric field strength and concentration (see fig. 6 of my paper). G. Furler, Ph.D. Thesis (University of Basel, 1984). B. Lamaire, P. Bothorel and D. ROUX, 5. Phys. Chem., 1983, 87, 1023. Dr. Th. F. Tadros (ICZ, Jealott ’s Hill) said: I have a question on the sign of the free-energy term Gf in Prof. Eicke’s paper. Eqn (2) shows that Gf is positive, which means that on differentiation Gf/A becomes negative [eqn (3)]. The conclusion reached in the analysis is that Gf is large and negative, whereas yun is large and positive such that (yu, - 2Gf/A) is small and positive.This is at variance with Overbeek’s analysis. The conclusion reached by Overbeek is that the electrical term Sod$ is large and positive, and hence Yun is large and negative. Could Prof. Eicke clarify this difference? J. Th. G. Overbeek, Faraday Discuss. Chem. SOC., 1978, 65, 7. Prof. H. F. Eicke (University of Basel, Switzerland) answered: The original oil/water interfacial free energy (without surfactant) is decreased to yun by the ac- cumulation of surfactant in the interface. This residual interfacial tension (yu,) is further decreased by 2Gf/A (if ionic surfactants are used) owing to the formation of a semi-diffuse double layer (entropically driven) which simultaneously produces a molecular capacitor with a surface energy density Sod+ = Gf/A, according to Over- beek.’ This is the physical interpretation of eqn (3)’ which describes the equilibrium condition of a microemulsion under the above-stated conditions.In Overbeek’s ap- proach yun (Overbeek) comprises yun (this paper) and dGF!dA. The second term is the ‘driving force’ to charge Overbeek’s molecular capacitor. J. Th. G. Overbeek, Faraday Discuss. Chem. SOC., 1978, 65, 7. Prof. H. N. W. Lekkerkerker (Free University of Brussels) said: Prof. Eicke men- tioned that microemulsions are more rare than has been assumed so far. This is an interesting point to consider. One wonders whether he would be prepared to go as far as to say that only AOT/water/oil systems are true microemulsions.GENERAL DISCUSSION 345 Dr.G. J. T. Tiddy (Unilever Research, Port Sunlight) said: I address a general remark to Dr. Langevin, Prof. Eicke and Prof. Lindman. In considering whether microemulsions are bicontinuous or form closed droplets it may be significant to examine the relevant surfactant/oil/water phase diagram. While the overall micro- emulsion composition is given by point A on fig. 2, during the close approach of two 0 W S Fig. 2. Schematic illustration of local composition changes on close approach of microemul- sion droplets for a ternary oil (O)/surfactant (S)/water (W) system. Initial local composition A moves to region B/C on the close approach of droplets. droplets the local composition will move to higher surfactant levels represented by B/C.If region B/C is occupied by a lamellar phase then the droplets are unlikely to fuse because the contact zone between the droplets will resemble the stable surfac- tant bilayers of the lamellar phase. However, if this composition region is occupied by L, solution then a different possibility arises. It is likely that the L2 structure is one of ‘smallish’ surfactant aggregates. On close approach of two droplets the sur- factant films will become disrupted to form small aggregates, and hence the micro- emulsion droplets will fuse. Of course the small aggregates will rapidly diffuse to readsorb at a nearby oil/water interface, giving rapid fluctuations in droplet size and shape (as expected from the model of Friberg).The local composition fluctuations should be consistent with the activities of the various components, so that one expects that surfactant activities will not vary very strongly with composition in this region (i.e. critical-composition fluctuations will occur). If measured activities were available one might compute (from a Boltzmann-type equation) the concentration of small aggregates that is likely. This would give some idea of how much of the oil/water interface is disrupted in a bicontinuous system. ’ See fig. 8 in S. Friberg, L. Rydhag and T. Doi, Ado. Chem. Ser., 1976, 152, 28. Prof. B. Lindman (Chemical Center, Lund, Sweden) said: This is a very significant remark and we agree completely that phase-diagram studies are very important for any microemulsion work.In particular, as Dr. Tiddy stresses, much can be learnt on microemulsion structure from the nature of adjacent phases in combination with information on component chemical potentials. In this connection I should like to mention that in a collaborative study between the Swedish Institute of Surface Chemistry in Stockholm and our group in Lund the component activities are deter-346 GENERAL DISCUSSION mined experimentally (Eva Sjoblom) and analysed (Bengt Jonsson) on the basis of a theoretical model put forward by HAkan Wennerstrom and Bengt Jonsson. Dr. D. Langevin (ENS, Paris) said: Dr. Tiddy’s suggestion is interesting and is corroborated by several facts: (1) There are no critical points (or, if one prefers, attractive type of systems) along the demixing line for the water/oil microemulsion region (the line farthest from the surfactant corner) if this region is disconnected from the oil/water region by a sequence of liquid-crystalline phases.In these systems the interactions are of the hard-sphere type and electrical percolation is never observed. (2) The surfactant layer has to be fluid for two droplets to interpenetrate and to fuse, as evidenced from our measurements [see ref. (6) of our paper]. On the con- trary, if organized phases are formed, the surfactant layer is expected to be rather rigid [see ref. (36)]. The fusion mechanism need not necessarily be driven by the formation of smaller aggregates. Indeed, you can imagine distorted aggregate shapes where the area per polar head is not changed.Of course, more experimental data are needed to under- stand the fusion mechanism. Prof. H. F. Eicke (University of Basel, Switzerland) said: Small-angle and quasi- elastic light-scattering measurements as well as small-angle neutron-scattering studies suggest’ a certain polydispersity of the droplets (as a function of the water content). Whether the mechanism suggested in your comment is operative in the present case cannot be inferred from the above investigations. However, regarding the conclusions drawn from the proposed mechanism, recent determinations of the temperature-dependent self-diffusion coefficients of water and stationary and time- dependent electric conductivities (after field jumps)2 strongly indicate considerable potential barriers for water and charge-carrier transport processes in the percolation regime where a network structure of droplets is assumed.M. Kotlarchyk, S. H. Chen and J. S . Huang, J. Phys. Chem., 1982,86, 3273. H. F. Eicke, R. Hilfiker and M. Holz, Helu. Chim. Acta, 1984, 67, in press. Dr. R. Kubik (University of Basel, Switzerland) said: I address my remarks to The order parameter Sii is defined by Prof. Lindman. -- where cos26i denotes the time average of the angular fluctuations of the ith co- ordinate axis of a cartesian-coordinate system x,y,z which is attached to an am- phiphile molecule, with respect to the local director axis 2’. The director 2’ character- izes the average orientation of an ensemble of molecules grouped together in one microdomain. In a hydrophilic-lipophilic interface, for example, the director z’ is at any point of the interface identical with the normal to the interface.l In the theory2 the correlation time 7s: is a measure of the reorientation rate of the microdomain (which can be a micelle or any other local liquid-crystalline structure) or generally the change of direction of the normal to a hydrophobic-hydrophilic interface.7: is related to the ‘radius’ r of a microdomain, to the microviscosity of the environment q and to the thermal energy kT byGENERAL DISCUSSION 347 As 7s: lies in the range of 1-2 ns, one obtains as a rough estimate r = 6-7 A, which is, as mentioned in ref. (3), approximately half the length of the extended octylbenzene sulphonate molecule. As ~ t a t e d , ~ this is consistent with the predominance of small aggregates (which consist of very few molecules) or even, in the limit of simple solutions, with no aggregation at all.So now I have a question. Is it possible to extend the concept of the order parameter S, which is used in the case of structured lyotropic mesophases, to cases where one is not sure whether the surfactant molecules are anchored within a hydrophobic-hydrophilic interphace or in microdomains, which is a necessary con- dition to define the local director z' and therefore the angle 8? * J. Seelig, Q. Rea. Biophys., 1977, 10, 353 H. Wennerstrom, B. Lindmaii, 0. Soderman, T. Drakenberg and J. B. Rosenholm, J. Am. Chem. SOC., 1979, 101, 6860 B. Lindman, T. Ahlnas, 0. Soderman, H. Walderhaug, K. Rapacki and P. Stilbs, Faraday Discuss.Chem. SOC., 1983, 76, 317. Prof. B. Lindman (Chemical Center, Lund, Sweden) replied: First a comment: 7; may be influenced by other dynamic processes, mainly lateral diffusion and a very rapid chemical exchange of surfactant molecules. Secondly, the basic assumption of the two-step model of relaxation devised by Wennerstrijm is that local and overall motions occur on widely different time-scales and thus give independent contri- butions to relaxation. The theory has been tested extensively for surfactant systems with well defined aggregates, such as liquid crystals and micellar solutions, and found to give a good description of the relaxation behaviour (2H, I3C, 1 7 0 , 14N). As the solutions become less and less organized and 7: decreases, the basic assumption will progressively become questionable.For solutions with a low degree of organiz- ation and only small aggregates it is, of course, not applicable. As indicated in the present study, microemulsions with butanol or pentanol as cosurfactant are inter- mediate between simple non-associated solutions and well organized surfactant phases. However, even if one encounters cases where the two-step model would be a bad approximation the multi-field relaxation studies should, by providing limits of structural lifetimes, give pertinent structural information. Dr. Th. F. Tadros (ZCI, Jealott 's Hill) said: In his paper Prof. Lindman focused attention on the effect of the chain length of the cosurfactant in which he clearly showed that for a given surfactant (sodium dodecyl sulphate), D for D20 decreases rapidly with increase in alcohol chain length.The analysis shows that only above C , or C , alcohols are the results consistent with the separation into hydrophobic and hydrophilic domains that are indicative of the presence of definite water cores. Have similar measurements been made whereby the cosurfactant chain length was kept constant while the surfactant chain length (and structure) was changed? Are vari- ations in surfactant structure and chains as important as cosurfactant? Could Prof. Prof. Lindman also explain the implications of such changes. Prof. B. Lindman (Chemical Center, Lund, Sweden) answered: This is an import- ant point. We have indeed studied different surfactants (C,-C, 6) but not systemati- cally varied the alkyl chain length at fixed compositions as for the cosurfactant.From these experiments it appears that the separation into hydrophobic and hydro-348 GENERAL DISCUSSION philic domains is, as expected, favoured by a lengthening of the alkyl chain length of single-chain surfactants. However, the surfactant length seems to be less important than that of the cosurfactant. Definite conclusions on these matters of importance for the formulation of microemulsions in various contexts must await the self- diffusion studies presently under way. Dr. J. H. R. Clarke (UMIST, Manchester) said: I should like to present some results which are relevant to several of the points raised by the previous three speakers (Langevin, Eicke and Lindman). These are the results of dynamic light- scattering measurements on the system H,O/AOT/p-xylene at a water-to-surfactant molar ratio of 10.At high volume fractions the g(')(t) correlation function fits extremely well to the sum of two exponentials, which we interpret, following the theory summarized by Prof. Vrij, as arising from collective diffusion (concentration fluctuations) and self-diffusion (polydispersity fluctuations). Fig. 3 compares the 2 .o 1.6 1.2 I v) N E 2 5 0.8 3 I 0 .L 0 I I I 1 y+ 1 I 0.2 0.L 0.6 0.8 volume fraction, cp Fig. 3. Diffusion data for the system H,O/AOT/p-xylene Dl(.) and D z ( V ) are coefficients from the long and short time components of g(')(t). D, refers to best-fit single exponen- tials. N.m.r. data (Lindmann et al.) for H 2 0 ( x ) and AOT (0) diffusion are also shown.The broken line is a prediction of the Stokes-Einstein equation. self-diffusion data with tracer diffusion coefficients of H 2 0 and AOT determined by Fourier-transform n.m.r. spectroscopy on the same system.2 The agreement is satis- factory, showing that the constituents of the microemulsion particles diffuse at the same rate as the particles themselves (the water diffusion is faster than expected and this point has been discussed2). This agrees with the conclusion of Prof. LindmannGENERAL DISCUSSION 349 that the microemulsion droplets have some real integrity in this system. In this case one might then attempt to predict the dependence on viscosity of the diffusion coefficient of a test particle with the infinite-dilution hydrodynamic radius of the microemulsion droplets.The simplest prediction is that of the Stokes-Einstein law, which is shown on fig. 3. At low volume fractions it works well. At high volume fractions the observed rate of diffusion is an order of magnitude faster than pre- dicted. This discrepancy may be related to some ‘percolation’ mechanism for mass transfer in these microemulsions, as suggested by Dr. Langevin. This work will be discussed in more detail in a future a r t i ~ l e . ~ A. Vrij, J. W. Jansen, J. K. G. Dhont, C. Pathmamanoharan, M. M. Kops-Werkhoven and H. M. Fijnaut, Faraday Discuss. Chem. SOC., 1983, 76, 192. B. Lindmann, P. Stilbs and M. E. Moseley, J. Colloid Interface Sci., 1981, 83, 569. J. H. R. Clarke, K. Regan and D. Nicholson, to be published. Dr.D. Langevin (ENS, Paris) said: Several inelastic light-scattering experiments on polydisperse oil/water microemulsion systems have already been interpreted in terms of mutual and self-diffusion processes. Self-diffusion coefficients have been compared here with those measured with forced Rayleigh-scattering experiments and the agreement between spontaneous and forced-scattering data was found to be satisfactory. A. M. Cazabat, D. Chatenay, D. Langevin, J. Meunier and L. LCger, in Surfactants in Solution, ed. K. L. Mittal and B. Lindman (Plenum Press, New York, in press). Dr. B. H. Robinson, Dr. P. D. I. Fletcher and Mr. A. M. Howe (University of Kent) said: An alternative approach which can give information on interactions in micro- emulsion systems is provided by kinetic studies.We have carried out kinetic mea- surements on relatively simple three-component microemulsion systems, e.g. Aerosol- OT (AOT)/water/alkane, in a concentration region (the oil-rich domain) where the existence of discrete water droplets is not in doubt. We have used a variety of techniques to characterize these systems, including small-angle neutron scattering, photon-correlation spectroscopy, time-resolved fluorescence studies and ultracentri- fugation . By means of an indicator method1y2 we have measured rate constants for the transfer of a solubilized ionic species between water droplets as in reaction (l), where A is a hydrophilic solute, e.g. an aquo-ion: Ions of varying size and charge, e.g. H ‘(as), Zn2 +(as) and Fe(CN):-, are all trans- ferred at similar rates.Hence the exchange process is controlled by interactions between the droplets themselves, rather than the transport properties of the indi- vidual ions within the dispersion. Typically, for AOT-stabilized dispersions we find k- values in the region of lo7 dm3 mol- s- l . If every encounter between the droplets resulted in exchange, then k (from the Smoluchowski equation) would be of the order of 1 O l o dm3 mol- s - l . Therefore, ca. 1 in lo3 encounters results in exchange of A. The only mechanism consistent with the data is shown in reaction (2):3 50 GENERAL DISCUSSION encounter transient pair dimer It would appear that in the more energetic droplet collisions (ca. 1 in lo3), the water pools ‘fuse’ to form a transient droplet dimer.The fusion process is associated with a large enthalpy of activation (ca. 100 kJ mol-*) and a large positive entropy of activation. The pool contents are then randomly distributed between the two drop- lets that are formed on the subsequent breakdown of the transient dimer. Hence the rate of water exchange is identified with that for solute exchange. This mechanism is thought to operate since all solubilizates studied exchange at the same rate. Apart from the obvious relevance of these results to reaction kinetics in these systems, the results reveal the dynamic nature of the droplets and also have impli- cations for the interactions between droplets. Clearly, if the droplets were ‘hard spheres’ then no exchange by this mechanism would be possible. Mechanism (2) implies ‘sticky’ collisions between the droplets; i.e. the presence of an attractive interaction. The question which then arises is: how do these kinetic measurements relate to the more commonly determined interaction parameters such as the osmotic compressibility term /??3 These kinetic measurements are useful in that they reveal a significant attractive interaction in a largely elastic system, i.e. only ca. 0.1% of the droplet collisions result in fusion. Another aspect of this study should be empha- sized. We measure the forward rate constant for exchange [equated in mechanism (2) with the rate of formation of transient droplet dimers]. We have no direct measure- ment of the lifetime of the transient droplet dimer. [This lifetime must, however, be short (of the order of ps), since no appreciable equilibrium concentration of the droplet dimers is observed experimentally]. Since commonly determined interaction parameters are equilibrium properties, there may be no direct correlation between the two types of measurement. The following experiments, however, demonstrate that there is a correlation be- tween the exchange rate and the stability of the microemulsion phase with respect to temperature. Changing the alkane solvent or adding a fourth component (e.g. toluene, cholesterol or benzyl alcohol) can decrease or increase the exchange rate. In all cases studied there is found to be a corresponding shift in the temperature range over which the microemulsion phase is stable. For example, addition of 0.1 mol dm-3 benzyl alcohol typically increases k by a factor of ca. 10, whereas addition of toluene reduces the rate constant by a similar amount. There are cor- responding opposite shifts in the microemulsion-phase temperature region in the presence of these additives. In conclusion, it has been clearly demonstrated that: (1) The droplets are deformable on contact. (2) Fusion and ‘breakdown’ of droplets occurs on a fast time- scale (ms-ps). (3) Information from the kinetic studies is obtained directly on the important interactions. In particular, k is strongly correlated to the temperature stability of the droplet system. (4) Such dynamic measurements may be more sen- sitive to the behaviour of the system than p or S(Q),“ since in microemulsions the attractive (short-range) component appears to be of particular importance. (5) Related kinetic measurements on the rate of size equilibration on mixing small and large droplets confirm the general validity of the analysis. (6) The conclusions of this work suggest that there is a similarity in trends of behaviour with allGENERAL DISCUSSION 35 1 microemulsion systems, including four-component ones, and that there is a link between ‘fusion’ and percolation. P. D. I. Fletcher and B. H. Robinson, Ber. Bunsenges. Phys. Chem., 1981, 85, 863. P. D. I. Fletcher, A. M. Howe and B. H. Robinson, J. Chem. SOC., Faruday Trans. I , to be submitted. A. M. Cazabat, D. Chatenay, D. Langevin and J. Mennier, Furaduy Discuss. Chem. SOC., 1983,76, 291. J. B. Hayter, Faruday Discuss. Chem. SOC., 1983, 76, 7 .

ISSN:0301-7249    

DOI:10.1039/DC9837600331

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

Faraday Discuss. Chem. SOC., 1983, 76, 353-362 Electric Birefringence Studies on Concentrated Aqueous Polyoxyethylene Surfactant Systems BY PAUL G. NEESON AND BARRY R. JENNINGS Electro-optics Group, Physics Department, Brunel University, Uxbridge AND GORDON J. T. TIDDY Unilever Research, Port Sunlight Laboratory, Quarry Road East, Bebington, Wirral, Merseyside L63 3JW Received 13th May, 1983 Pulsed-electric-field birefringence measurements (Kerr constant, B, and relaxation time, t) have been made on polyoxyethylene surfactant solutions in the concentration region just prior to the formation of liquid-crystal phases. For solutions in the pre-hexagonal region both B and t initially increase with temperature and then decrease as the hexagonal phase melts. This is consistent with the presence of interacting rod micelles.For the pre-lamellar region the behaviour is more complex; at high concentration B increases with temperature and con- centration, while t increases with concentration and decreases with temperature. At lower concentrations B changes sign while z exhibits a maximum and becomes very short. Tentative explanations based on changes in the nature of micellar interactions and a rod/disc shape transition are proposed. It is well known that alkyl polyoxyethylene surfactants [C,H2, + 1(OCH2CH2)m, C,EO,] form micelles in dilute solutions, giving way to lyotropic mesophases at higher surfactant concentrations. l , However, the sizes and shapes of the micelles are not well established. It has recently been suggested3 that in the concentration region prior to the formation of liquid crystals the micelles have a shape similar to that in the mesophase.Where the mesophase is anisotropic (lamellar or hexagonal) the micelles may become large. Furthermore, at low concentrations the structure of the micelles in the region of the lower consolute temperature (the ‘cloud point’) is a matter of controversy. While it is generally accepted that large aggregates are pre- sent at the cloud point, some argue that the aggregates consist of long rod rni~elles,~, while others - maintain that they are composed of small, almost spherical, micelles. Both could undergo the critical concentration fluctuations that occur close to a lower consolute temperature. Indeed, it has been proposed3 that two distinct types of behaviour may be observed; critical concentration fluctuations of small micelles could occur for surfactants with short alkyl chains and/or large head groups, while secondary aggregation of disc micelles is the suggested mechan- ism for surfactants where the micellar solution is completely replaced by a lamellar phase at higher temperatures.In this study we have investigated the micellar region precursive to the formation of mesophases with the non-ionic surfactants tetra- oxyethylene dodecyl ether (C, 2E0,), hexaoxyethylene dodecyl ether (C, zEO,) and octaoxyethylene dodecyl ether (C , ,EO,) using the pulsed electric birefringence tech- nique. Employing these surfactants, we can examine the transformation of micellar solutions to cubic (Il), hexagonal (H,) or lamellar (La) phases.3 54 ELECTRIC: BIREFRINGENCE OF NON-IONIC SURFACTANTS The application of short-duration pulsed electric fields to solutions containing anisotropic aggregates initiates reorientation of the particles, inducing the medium to become temporarily birefringent.The time for the decay of this effect after cess- ation of the pulse is characterised by a relaxation time (z) given l 2 by the equation (1) where An is the amplitude of the birefringence at time t after the cessation of the pulse for which t = 0 and An = An,. In dilute solutions z provides an extremely sensitive measure of particle size,' whilst for the concentrated media particle inter- actions and the angular correlations between orienting elements must also be con- sidered.The amplitude of the electrically induced transient can be characterised l4 by the Kerr constant B An = Ano exp( - t / z ) where A is the vacuum wavelength of the analysing beam and E is the amplitude of the electric field. The range of surfactants studied was chosen so that the transformation of micel- lar solutions (L,) to cubic (I1), hexagonal (H,) or lamellar (L,) phases could all be examined. In addition, we have initiated studies on the dilute regime adjacent to the micellar lower consolute boundary. The results for solutions adjacent to I 1 and HI phases are given first, followed by a description of data for the L1/Lb transition. Finally we report some initial measurements of the complex behaviour observed at lower concentrations for one surfactant (C 2E04).EXPERIMENTAL The surfactants were obtained from Nikko Chemicals (Japan). They were used as received, having a stated purity of > 97%. [See ref. (3) for further details.] Water was deionised and doubly distilled. All sample concentrations are given as wtO/S surfactant. Electric birefringence measurements were made with a conventional apparatus. * Light of 633 nm wavelength (A) from a 1 mW helium-neon laser entered the sample cell after being linearly polarised at 45" azimuth to the electric-field direction. After leaving the cell, the light traversed a quarter-wave plate set in parallel azimuth to the initial polarisation state and fell on an analysing polariser which was slightly offset from the 'crossed' azimuth relative to the initial polariser. This arrangement gave optimum detection sensitivity and enabled the sign of the induced birefringence to be evaluated.Light transmitted through the analyser fell on the photocathode of a photomultiplier whose output was displayed on an oscilloscope, and fed to a transient digitiser-microcomputer assembly for direct analyses of birefringence amplitudes and decay rates. The cell consisted of a glass trough 5 cm in length, holding a pair of stainless- steel electrodes spaced 2 mm apart along its length. Some 1 cm3 of sample filled the inter- space, across which pulsed electric potential differences of up to 2.2 kV were applied for durations between 30 and 5 0 0 p . The duration was limited in each experiment to that just sufficient to allow the birefringence to attain a steady value in order to minimise heating and electrophoretic effects.The cell temperature was maintained to a thermal stability of & 0.1 "C. RESULTS AND DISCUSSION Measurements were made for a range of applied field strengths and in all cases the birefringence responses were of regular exponential form (fig. 1). The steady birefringence amplitude (An) always had a quadratic dependence on the applied field amplitude (E), while z was independent of E. All transients were symmetrical, withP. G. NEESON, B. R. JENNINGS AND G. J. T. TIDDY 355 Fig. 1. Electric-field-induced transients. Lower curve, electric-field pulse applied across a 2 mrn electrode gap. Upper curve, experimental transient. ( a ) 10.0% CI2EO4 at 2.5 "C, positive birefringence; (b) 19.87/, C, ,E04 at 2.5 "C, negative birefringence.the same value of z characterising the build-up and decay of the birefringence. Thus eqn (1) and (2) could be used to characterise the observed behaviour. At high con- centrations [sections (i) and (ii) below] the sign of B was negative, while at low concentrations [section (iii)] B was positive. (i) MICELLAR SOLUTION ADJACENT TO CUBIC (I A partial phase diagram for C,,EO, is shown in fig. 2(a). The L, region is bounded at high surfactant concentration by a cubic ( I 1 ) phase up to 16 'C, and above this by an hexagonal phase below 59 "C, there being no other mesophases at higher temperature^.^ The cubic phase is thought to consist of close-packed spheri- OR HEXAGONAL ( H I ) PHASES (C12EOB AND C12EO6)3 56 ELECTRIC BIREFRINGENCE OF NON-IONIC SURFACTANTS 60 30 30 2.1 c, 10 N ; i ’ E 6- 0.5 N - 0 - I 0 I I I 1 - - L - - I I 1 I I 30 T f C 60 Fig. 2. ( a ) Partial phase diagram of CI2EO8 [redrawn from ref. (3)]; 0 , data points. (h) Variation of z with temperature at 34.9% C,2E08. (c) Variation of B with temperature at 34.9% C , 2EOs. cal micelle~,~~ hence here L1 should also contain spherical micelles, and no induced birefringence is expected for this region. The H1 phase is thought to occur because spherical micelles are transformed to rods under the influence of intermicellar repul- sions and/or a decreased head-group hydration with increasing temperature. Both of these lead to a decrease in head-group area at the micelle surface (designated as a), resulting in a micelle shape change due to alkyl-chain packing constraint^.^ Thus in this part of L, we expect rod micelles to be present, these growing larger with increasing temperature.For CI2EO8, the upper temperature limit of the H1 phase [i.e. T > Tk, fig. 2(a)] is thought to correspond to a critical decrease in hydration ofP. G. NEESON, B. R. JENNINGS AND G. J. T. TIDDY 357 the ethylene oxide groups, resulting in the formation of small micelles and eventually the occurrence of the lower consolute b ~ u n d a r y . ~ Thus we expect no birefringence for the L1/I1 region, increasing birefringence and z with increasing temperature for the L,/H, region, and a decrease in both B and z at about Tk. The results obtained on a 34.9% C,,EO, sample are in good agreement with this [fig. 2(b) and (c)].(The H, phase occurs above ca. 40%.) No birefringence was detected between 15 and 30 "C, but above 30 "C B increases up to ca. 45 "C and then decreases. The values of z are coo 300 v1 \ x 200 100 14 N I + 10 E g 6 N - I 2 TIT Fig. 3. ( a ) Partial phase diagram of C12E06 [redrawn from ref. ( 3 ) and (16)]; a, data points at 34.4%; x , data points at 24.9% (a bicontinuous cubic phase2s3 exists in region V,). (6) Variation of z with temperature at 34.4% (a) and 24.9% ( x ) CIZEO6. (c) Variation of B with temperature at 34.4% (0) and 24.9% ( x ) C1 2E06.358 ELECTRIC BIREFRINGENCE OF NON-IONIC SURFACTANTS relatively low and are less temperature dependent but qualitatively show the same behaviour. There is a slow increase up to 45 "C followed by a sharper decrease.Further measurements were made in a second surfactant (C12E06) to support this interpretation. Because of its smaller head group (lower a value) no I, phase is observed and the H1 phase forms only below 37 "C. At higher temperatures and/or at high concentrations a lamellar phase occurs [fig. 3(a)].3p16 As is the case for the I1/HI transition of CI2EO8, the HI/La transition is due to a reduction in a; here it falls below that necessary for the alkyl chains to pack into rods. This results in a rod/disc micelle shape transition at the H1 melting temperature. Thus we expect an induced birefringence of the same form as with C12E08, but having a maximum value for both B and z at ca. 22 "C lower temperature. Data for two concentrations of C12E06 are shown in fig. 3(b) and (c).As reported in a preliminary communi- cation," at 34.4% C12E06 maxima in both z and B are observed at 22 "C. As with C12E08, the rate of increase of z is less steep on the low-temperature side than the rate of decrease at high temperatures. Also shown are similar data for 24.9%, where both B and z are generally lower, and the maxima occur at ca. 24.5 "C. The shift of temperature for the maxima compared with C1 2E08 is that expected from the phase diagrams. Also the larger birefringence and z values at higher concentrations are consistent with an increase in micelle size and/or increasing orientational alignment of the micelles on approaching the hexagonal phase. The slight increase in the tem- perature of maximum z and B values at 24.9% is also consistent with the require- ment of a less hydrated head group for a to reach the critical value for the rod/disc transition at the lower concentration, as the role of intermicellar repulsions is re- duced.Note that at the highest temperatures, 35-40 "C, the z values for 24.9% are longer than those for 34.4%. Similar results are observed for this region of the phase diagram with C12EO4. These data are described in more detail below. (ii) MICELLAR SOLUTIONS ADJACENT TO THE LAMELLAR PHASE (La) (C12E04) At low surfactant concentration and temperature, C1 2E04 exhibits the phases shown in fig. 4(a).3118 The micellar solution (L,) is adjacent to a lamellar phase (L& and a lower consolute boundary (W + L,) also occurs. Dilute surfactant solution (W) is found in equilibrium with the lamellar phase in the area designated (W + LJ.In addition, an HI phase (not shown) occurs at ca. 35-45% surfactant below -2 "C. This phase is metastable, the equilibrium structure being ice plus L,. In the L, region at temperatures below the La phase, the micelles are thought to be discs which increase in size with temperature, until they are large enough for a disorderlorder transition with formation of L, phase to occur.3 Thus we expect both B and z to increase towards the phase transition. Birefringence responses were recorded as a function of temperature at 34.9 and 41.3% C12E04 [fig. 4(b) and (c)]. The results given in fig. 4(c) show that B does increase with temperature, but z decreases. However, both z and B increase with concentration, as expected.In order to verify this dependence we made further measurements at constant temperature (2.5 and 1 1 . 1 "C) as a function of concen- tration (fig. 5). (The data for 34.9% and 1 1 , l "C have been reported previou~ly.~~) In this case both B and z increase towards the La phase, as anticipated for an increase of orientational correlations and size of disc micelles with increased concentration. However, at 2.5 "C reducing the concentration below ca. 30% leads to an increase in z followed by a dramatic decrease, together with a sign change of B. This is discussed further below. The increase in B with temperature is consistent with growth of theP. G . NEESON, B. R. JENNINGS AND G. J. T. TIDDY 359 I I I I 0 4 8 12 16 TI" C Fig. 4. (a) Partial phase diagram of C12E04 [redrawn from ref.(3) and (18)]; (- * -) experi- mental scan lines; @, data points at 34.9% C12E04; x , data points at 41.3% C12EO4. (b) Variation of z with temperature at 34.9% (@) and 41.3% ( x ) C12E04. (c) Variation of B with temperature at 34.9% (@) and 41.3% ( x ) C12E04. micelles. The value of z is determined both by the size of the micelles and the interactions between them. At a given concentration, as their size increases with temperature, the number of micelles decreases, hence their average separation also increases. Since the interactions are a function of micelle separation these may be expected to decrease. If the positional correlations due to interactions decrease with increasing temperature faster than increased order occurs due to micellar growth, then a decrease in z can be accounted for.That a decrease in the range of order with increasing temperature does occur in these solutions is confirmed by the occur- rence20 of streaming birefringence only below ca. 10 "C. Also, n.m.r. linewidths of concentrated C1 ZE04 solutions show a marked decrease with increasing tem-360 (6) - 1 I - I / I - - - - - L A I I I 1 ELECTRIC BIREFRINGENCE OF NON-IONIC SURFACTANTS 400 300 v1 3 . k d 200 100 7 N 6 > 5 , E L - 2 3 B 1 2 1 0 - 1 wt% C,,EO, Fig. 5. ( a ) Variation of z with C,,EO, concentration at 2.5 "C ( A ) and 1 1 . 1 "C (0). (b) Variation of B with CI2EO4 concentration at 2.5 "C ( A ) and 11.1 "C (0). perature.20 Note that we expect a rapid increase in z to occur very close to the L, transition.At 2.5 "C the values of z increase with decreasing surfactant concentration in the range 20-30% while this is not observed at 11.1 "C. The 2.5 "C region corresponds to the CI2EO6 solutions at ca. 40 "C, since the hexagonal phase occurs with CI2EO4 below -2 "C. Thus rod micelles are expected to be present at lower concentrations and below the cloud point. We attribute the longer relaxation times to the presence of large rod micelles. The cloud point of CI2EO4 is at ca. 4 "C. Thus the micelle shape change from rods to discs is expected to occur between -2 "C and 4 "C, with the transition temperature being higher, the lower the concentration. The continuous decrease in B over this region indicates that a gradual change in micelle properties is occurring.Further measurements on this part of the phase diagram for CI2EO5 and C1 2E06 are required to fully substantiate this explanation. (iii) DILUTE SURFACTANT SOLUTION Only a few measurements of z and B were made for dilute CI2EO4 solutions (< 20%). Note that z decreases sharply and there is a sign change in B (fig. 5 ) . AtP. G. NEESON, B. R. JENNINGS AND G. J. T. TIDDY 36 1 very low concentrations z increases again. At 0.10% the value of z at 4.0 "C (17 p s ) is in good agreement with that reported by Hoffman et al. (20 ps). The value at low concentration we attribute to the presence of loose aggregates of micelles formed via long-range attractive interactions. At higher concentrations, where the interactions are weakly repulsive and there is little orientational correlation, the values may be more representative of individual micelles.We suggest that the dramatic increase in z and change of sign for B above 15% reflects the onset of strong repulsions giving strong nearest-neighbour positional and orientational correlations. The concen- tration at which B changes sign (17-19%) appears to be roughly invariant for a range of temperatures (2-10 "C). The sign change is also observed at 25 "C for lower concentrations (10%) of the other two surfactants. Much detailed experimental work is required before definitive conclusions can be drawn about the detailed mechanisms. GENERAL COMMENTS The pulsed electric birefringence measurements on solution composition adjacent to the hexagonal phase are entirely consistent with the presence of interacting rod micelles which initially increase in size with increasing temperature, giving larger z values before undergoing a transition to more rapidly reorienting aggregates as the hexagonal phase melts.By contrast, in the pre-lamellar region micellar growth at constant concentration is suggested to give shorter z values. The difference may be explained by the growth of rods in one dimension while discs grow in two dimen- sions. Thus the centre-centre separation of disc micelles increases more rapidly than that of rod micelles, hence the effect on micellar interactions could be different in the two cases. The present study represents an initial survey of the type of information that can be obtained on concentrated micellar systems by this technique.Further experi- mental work is required to fully substantiate the hypotheses proposed to account for the data. However, this study does illustrate the potential of birefringence measure- ments for obtaining structural information on micellar solutions. The S.E.R.C. is thanked for a CASE studentship to P. G. N. during the course of which this work was undertaken. H. Wennerstrom and B. Lindman, Phys. Rep., 1979, 52, 1. G. J. T. Tiddy, Phys. Rep., 1980, 57, 1. D. J. Mitchell, G. J. T. Tiddy, L. Waring, T. Bostock and M. P. McDonald, J . Chem. Soc., Faraday Trans. 1, 1983, 79, 975. R. Kjellander, J . Chem. SOC., Faraday Trans. 2, 1982, 78, 2025. D. J. Cebula and R. H. Ottewill, Colloid Polym. Sci., 1982, 260, 118. C. Tanford, Y. Nozaki and M. F. Rohde, J . Phys. Chem., 1977,81, 1555. E. J. Staples and G. J. T. Tiddy, J . Chem. Soc., Faraday Trans. I , 1978, 74, 2530. M. Corti and V. Degiorgio, J . Phys. Chem., 1981, 85, 1442. R. Triolo, L. J. Magid, J. S. Johnson Jr and H. R. Child, J . Phys. Chem., 1982, 86, 3689. l o J. B. Hayter and M. Zulauf, Colloid Polym. Sci., 1982, 260, 1023. l 1 M. Zulauf and J. P. RosenbuSch, J . Phys. Chem., 1983, 87, 856. l 2 H. Benoit, Ann. Phys., 1951, 6, 561. l 3 B. R. Jennings, in Molecular Electro-Optics, ed. S . Krause (Plenum Press, New York, 1981), chap. l4 E. Fredericq and C. Houssier, Electric Diochroism and Electric Birefringence (Clarendon Press, l 5 J. Badoz, J . Phys. Radium, 1934, 5, 497. l 6 J. S . Clunie, J. F. Goodman and P. C. Symons, Trans. Faraday Soc., 1969, 65, 287. l 7 P. G. Neeson, B. R. Jennings and G. J. T. Tiddy, Chem. Phys. Lett., 1983, 95, 533. 2, pp. 27-60. Oxford, 1973).362 ELECTRIC BIREFRINGENCE OF NON-IONIC SURFACTANTS T. A. Bostock, M. P. McDonald, G. J. T. Tiddy and L. Waring, Surface Active Agents (SOC. of Chem. Ind., London, 1979), vol. 181. l9 P. G. Neeson, €3. R. Jennings and G. J. T. Tiddy, Croat. Chern. Acta., in press. 2 o T. A. Bostock, Ph.D. Thesis, (Sheffield City Polytechnic, 1981). *' H. Hoffmann, H. S. Kielman, D. Pavlovic, G. Platz and W. Ulbricht, J . Colloid Interface Sci., 1981, 80, 237.

ISSN:0301-7249    

DOI:10.1039/DC9837600353

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

Faraday Discuss. Chem. SOC., 1983, 76, 363-373 Viscoelastic Detergent Solutions BY HEINZ REHAGE AND HEINZ HOFFMANN Institut fur Physikalische Chemie der Universitat Bayreuth, UniversitatsstraDe 30, D-8580 Bayreuth, West Germany Received 13th May, 1983 Aqueous solutions of some cationic detergents show viscoelastic behaviour at very low concentrations. Rheological measurements indicate that the strong viscoelasticity of these solutions can be traced back to the existence of a three-dimensional structure built up from rod-like micelles. As long as the lengths L of the rods are smaller than their mean distance of separation D, the unsheared solutions are not elastic. The L values increase linearly in this concentration range with the detergent concentration and the axes of the rods seem to be randomly oriented.In the concentration range of overlapping rods some of the experimental data can be explained on the basis of the Doi and Edwards theory for stiff rods. In order to explain the elasticity of the solution it must be assumed that a three-dimensional network is formed by the rods as soon as L > D. The network is of a temporary nature. The dynamic aspect comes about because of the continuous formation and breaking of contacts between individual free rods. The viscoelastic properties of the network solutions have been evaluated by dynamic and static rheological measurements. The effects of temperature and the chain length on the viscoelastic properties were determined and are discussed. Dilute solutions of ionic and non-ionic detergents usually behave as Newtonian liquids.Only a few surfactant systems are known to have more complicated rheolog- ical behaviour. Solutions of these systems are elastic. This phenomenon is relatively rare and is not completely understood, in spite of the effort that has recently been undertaken in order to explain it. - For a few surfactant systems the phenomenon has been observed in very dilute solutions with volume fractions of the detergent of the order of 10-3.1-3,10-12 Most of the known viscoelastic detergent systems are built up of cationic surfactants with pyridinium or trimethylammonium headgroups. The elastic effects of these solutions are spectacular, especially in view of the rather small detergent concentrations. The aggregates that are present in these solu- tions have been characterized by static and dynamic light-scattering, electric- birefringence, n.m.r., rheological, kinetic and neutron-scattering experiments. - The results of these investigations indicate that non-spherical rod-like micelles are present in these solutions.The axes of these rods are randomly oriented when the solutions are at rest. From light-scattering, electric-birefringence, rheological and neutron-scattering experiments the size parameters of the aggregates can be cal- culated. The results of these different experimental techniques are in fairly good agreement. The lengths of the rod-like micelles increase linearly with the total deter- gent concentration. At very low shear rates the solution behaves as a Newtonian liquid.The viscosity is very low compared with the solvent viscosity. At some higher values of the shear rate, however, the solutions show the phenomenon of rheopexy. At a constant prolonged shear rate the viscous resistance increases as a function of the shear time. The results of these observations clearly show that the micellar ag-364 VISCOELASTIC DETERGENT SOLUTIONS gregates can be modified above a certain critical shear rate. The rheopectic behaviour of the detergent solutions can be explained by postulating the formation of a supermolecular structure during flow. This supermolecular shear-induced structure has network properties. The state of this structure is evidently complex. It is continuously broken by the shearing forces but is also reforming all the time.It is characterized by a balance between the structure-building and -breaking forces of the shear field. When the stirring action is stopped, the shear-induced structure decays with very long time constants which are of the order of a few hundred seconds. After a lapse of sufficiently long time no discernible changes occur within the solution. The physical changes which occur during flow are reproducible. At concentration ranges where the dimensions of the rod-like micelles are equal to the spacings between them, the rheological properties change dramatically. The viscous and elastic properties of the solutions increase by several orders of mag- nitude within a very small concentration range and the solutions show complicated rheological behaviour. The viscoelastic properties of the detergent solutions depend on various conditions.We shall here report further measurements on viscoelastic surfactant systems which have been carried out in order to obtain more information on the structure of these solutions. In particular, we shall study the influence of temperature and chain length on the viscoelastic properties of the detergent solu- tions. The measurements were carried out on alkylpyridinium salicylates and alkyl- trimethylammonium salicylates, which have been well characterized in previous investigations. - EXPERIMENTAL PREPARATION OF THE SOLUTIONS The alkylpyridinium salicylate and the alkyltrimethylammonium salicylate solutions were prepared as previously described by ion-exchange procedure from the corresponding chlorides or by dissolving the salicylates which had been previously synthesized.Both methods gave identical results. The solutions were left standing for two days in order to reach equilibrium.' METHOD Two types of viscometer were used for measuring the rheological properties of the deter- gent solutions. The viscosities at low detergent concentrations were determined with a modified Zimm-Crothers viscometer. From these values the size parameters of the aggregates could be calculated. According to Doi and Edwards the viscosity of a dilute solution of rod-like molecules is given by (1) where qo is the viscosity of the solution at zero shear rate, qs is the viscosity of the pure solvent, P is the number of rods per unit volume and L is the length of the rods.As long as L is shorter than the mean spacing between the aggregates, the viscosity of the detergent solution is very low and small compared with the solvent viscosity. As a consequence, very accurate measurements are necessary in order to obtain information on the dimensions of the rod-like micelles. The Zimm-Crothers viscometer is a very sensitive instrument and it was possible to measure the viscosity of the dust- and air-free solutions with an accuracy of 0.2% at very low shear rates. The lengths of the rod-like aggregates can be calculated using eqn (1) when the con- centration P is known. The value for P can be obtained either from the electric birefringence measurements or it can be calculated using the simple equation q o = qs(l + PL3) ? = cM M/xr2pL (2)H.REHAGE AND H. HOFFMANN 365 where r is the short radius of the rod-like micelles, cM is the concentration of the aggregated detergent, M is the molecular weight of the surfactant monomers and p is the density of the rodlike aggregates. Inserting eqn (2) into eqn (1) leads to an equation where the only un- known variable is L Eqn (3) is only correct for the dilute-solution region. Interference arises when the rotational volumes of the rods actually touch one another or when the rods begin to overlap. The condition for overlapping rods is i. >> 1/L3. Because the normal viscosities change with time, dynamic experiments were carried out with a rotary viscometer in the oscillating mode (Contraves Low Shear 30 sinus). In this way it was possible to obtain information without disturbing the internal supermolecular structures. In the general case a sinusoidal deformation or strain is applied.The response of the liquid to the periodic change consists of a sinusoidal shear stress p21, which is out-of-phase with the strain by the phase angle 6. The shear stress is made up of two components. The first component is in-phase with the deformation and the second one is out-of-phase with the strain. From these quantities the storage modulus G’ and the loss modulus G” can be calculated according to the equations P2 1 G’ = ~ C O S 6 Y P 2 1 G = Tsin 6 Y (4) ( 5 ) where Pzl is the amplitude of the shear stress, .i, is the amplitude of the deformation and 6 is the phase angle between the stress and the strain. It is convenient to express the periodically varying function as a complex quantity which is termed the complex viscosity 1q”l.This quant- ity may be calculated using the equation It can be shown that for most dilute solutions there exists a simple correlation between dynamic and steady-state flow characteristics.’ At most detergent solutions the magnitude of the complex viscosity Iqxl at a certain angular frequency w coincides with the steady-state viscosity qm at the corresponding shear rate j . 1 6 The simplest mechanical model which can describe a viscoelastic detergent solution is called the Maxwell element. It consists of a spring and a viscous element (dashpot) connected in series. The spring corresponds to a shear modulus Go and the dashpot to a viscosity q. The behaviour of the Maxwell material under harmonic oscillations can be obtained from the following equations w2z2 1 + w2T2 G’(co) = Go (7) From these equations we see that for wz >> 1, G’ approaches a limiting value which is iden- tical with the shear modulus Go.Under such experimental conditions the solution behaves as an elastic body. At low frequencies wz << 1 and G’ becomes proportional to w2. This region is called the terminal zone and the solution behaves as a liquid. Viscoelastic solutions can be sub-divided into energy and entropy elastic systems. The free energy per volume G at a deformation y consists of an energetic and an entropic term366 4 10‘ - 5 - 3 - 2 - g loo - G 5- 1 6 ’ - b 3- 2 - 5- 3 - 2 - VISCOELASTIC DETERGENT SOLUTIONS (f) = (z) - T ( f ) * (9) In most real systems energy and entropy changes can occur.The elasticity of an ideal network is entropy controlled. In this picture stresses are caused by the chain orientation. From the theory of rubber-like elasticity it can be shown that the shear modulus of an ideal network depends on the number of elastically effective chains between the cross-links * Go = vkT (10) where v is the number of elastically effective chains in unit volume. RESULTS Results concerning the rheological properties of the studied systems are given in fig. 1-5. Fig. 1 shows the storage modulus G’ and the loss modulus G” as a function rubbery plateau f 1 0-* .2 3 5 2 3 5 2 3 5 2 3 5 1 o - ~ 1 o-2 lo-’ 1 oo 10’ w/s - ’ Fig. 1. Storage modulus G ’ ( 0 ) and loss modulus G” ( W ) as a function of the angular frequ- ency (CPySal, T = 20 “C, c = 50 mmol dm-j).of the angular frequency o for a 50 mmol dm-3 solution of cetylpyridinium salicyl- ate (CPySal). The dynamic characteristics of this solution exhibit certain typical features. At high frequencies the storage modulus approaches a constant value. In polymer systems this region of the curve is called the rubber plateau. At high frequ- encies the solution behaves as an elastic body. At low frequencies G’ becomes pro- portional to m2. In this region the detergent solution behaves as a liquid. The func- tion G” passes through a maximum and a minimum. The position of the maximum approximately corresponds to the onset of the rubber plateau. The angular frequ- ency at the maximum is denoted wM.The value of the loss modulus at the maximum is found to be 0.5 of the value of the storage modulus on the rubber plateau. The maximum in G” is attained at a frequency which is lower than the intersection point of the two curves. The ratio of the values of the loss modulus at the minimum and the maximum is 0.5. At low frequencies the loss modulus is proportional to w. The transition region from this proportionality to the position of the maximum value isH. REHAGE AND H. HOFFMANN 367 ca. 0.6 orders of frequency. The dynamic properties mentioned above are predicted from molecular theories of viscoelasticity of polymer melts and concentrated poly- mer solutions. The shape of the frequency dependence of the dynamic functions G'(w) and G"(w) point to the presence of a supermolecular structure with network properties.The experimental results can be interpreted in terms of the Graessley model or the theory of microviscoelasticity. These theories are based on the concept of entanglement. The theoretical predictions of these models indicate that the surfactant solutions have the same dynamic properties as temporarily entangled polymer solutions. The experimental results quoted above are typical for the pre- sence of a dynamic three-dimensional network. Other rheological properties of sur- factant solutions, e.g. from stress relaxation experiments, support this hypothesis.22 The storage moduli G' of the detergent solution for different temperatures are given in fig. 2, which shows the values for the storage modulus G' as a function of the - A lo3 - 5 - 3 - 2 - lo2 - a" E 5 - u 3 - 10' - ?- 2 - 5 - 3 - 2 - l o o ! ;;-; I " ' 2 3 5 ' " - I 2 3 5 ' 2 3 5 t o - ~ lo-' loo 10' w/s- Fig.2. Storage modulus G' as a function of the angular frequency for a 20 mmol dm-3 CPySal solution at different temperatures: (a) 12, (b) 20, (c) 25, ( d ) 30 and (e) 35 "C. angular frequency at different temperatures for a 20 mmol dm-3 solution of CPySal. The elastic properties of the surfactant solutions decrease with increasing temper- ature. The solution, equilibrated at 35 "C, shows little elasticity in the frequency range below 1 Hz and at temperatures of 50 "C the solutions behave as a Newtonian fluid. The supermolecular structures that are present in the solution and that are responsible for the viscoelastic properties seem to be completely destroyed under these experimental conditions.However, the curves suggest that the elastic pro- perties would be apparent again if the measurements were carried out at higher frequencies. Fig. 3 shows the magnitude of the complex viscosity as a function of the angular frequency for several alkyltrimethylammonium salicylates of different chain lengths, m. The curves show a sharp increase of the viscous resistance at a certain well defined characteristic concentration c *. Solutions with concentrations below c * were368 VISCOELASTIC DETERGENT SOLUTIONS clmmol dm - Fig. 3. Magnitude of the complex viscosity as a function of the detergent concentration for different alkyltrimethylammonium salicylates (7‘ = 20 “C, o = 0.01 s-l): A, Clb; e, C14; 0, c12; +, G o .not elastic while in the unperturbed state. For rn = 16 and m = 14, however, elasti- city could be induced by shearing of the solutions. This striking phenomenon can be explained by postulating the formation of a supermolecular structure during flow, as discussed above. Above the critical concentration c * the viscous resistance increases by several orders of magnitude within a very small concentration range. With de- creasing chain length the viscous resistance of the solutions falls rapidly. Similar behaviour is also found for the alkylpyridinium salicylates. Relevant results are represented in fig. 4 and 5. The storage modulus G’ and the loss modulus G” are plotted as a function of the angular frequency for different chain homo- logues.The elasticity and the viscous resistance of the detergent solutions increase with the number of CH2 groups. Note that G” for the shorter homologues is always considerably larger than G’. This means that under these experimental conditions the rheological properties are due mainly to viscous forces.H. REHAGE AND H. HOFFMANN to3- 5 - 3 - 2 - lo2 - 2 2”- 5 - - E 0 : 10’ - 5 - 3 - 2 - loo - 5 - 3 - 2 - lo-’ 369 A I l l I I l l I I l l I I l l I + 2 3 5 0 2 3 5 t Cojs- Fig. 4. Storage modulus G’ as a function of the angular frequency for different ridinium salicylates (RPyCl + Nasal, T = 30 “C, c = 25 mmol dm-3): a, c18; ., c14. alkylpy- c16; A,370 VISCOELASTIC DETERGENT SOLUTIONS DISCUSSION The fundamental feature underlying all the above observations is that the super- molecular structure which is responsible for the viscoelastic behaviour can be af- fected by the temperature and the chain length of the detergent system.It could be argued that this is due to different sizes of the rod-like micelles. At low detergent concentrations the dimensions of the rod-like particles can be evaluated from the viscosity measurements at zero shear rate, according to the theory of Doi and Edwards. Some results of these measurements are summarized in table 1. For com- Table 1. Lengths, L/& of the rod-like micelles as a function of temperature and concentration for solutions of CPySal clmmol dmW3 20°C 25°C 30°C 35°C 40°C 45°C 50°C 1 2 3 4 5 6 7 7.5 120 210 300 430 550 640 750 810 200 70 275 140 350 210 420 270 490 340 570 400 680 460 740 500 60 110 150 240 275 330 390 410 40 90 145 160 210 280 31 5 320 40 80 60 110 90 170 150 200 160 230 200 275 240 300 250 - parison purposes the dimensions of the aggregates have been calculated from light- scattering, electric- birefringence and small-angle neu tron-sca t tering experiments.Note that the different techniques give values which are in fairly good agree- ment. s* 6 , 8 , According to the data, the dimensions of the rod-like particles increase more or less linearly with the total detergent concentration. Detailed measurements carried out on the detergent system CPySal show that the rheological properties change dramatically in concentration ranges where the rod-like micelles start to o ~ e r l a p . ~ . 6, At the concentration where the dimensions of the rod-like particles are equal to the spacings between them, a supermolecular structure is formed.This structure has network properties. The dynamic rheological behaviour of such a de- tergent solution is shown in fig. 1 . In order to explain the experimental results we have to conclude that the rods form a network when they begin to overlap. It seems possible that the rod-like micelles acquire enough rotational energy to overcome the repulsive forces between them. When they touch they stick together. They may coalesce and form some sort of a supermolecular network. Judged from the electric-birefringence and the small-angle neutron-scattering data there exist still free, individual rods at high detergent concentrations. From all the information we have it seems that the rotation of these rods is highly restricted while the trans- lational diffusion of the rods seems to be little affected and is quite normal, as measured from the dynamic light-scattering experiments.For this reason we have to assume that the rod-like micelles in the system keep their identity and that they make contacts. As far as the contacts go we cannot be too specific at present. In a previous paper we proposed a model in which the rods are connected by physical contacts.16 It seems possible that the rod-like micelles can be held together by adhesion forces which are strong enough to overcome the electrostatic repulsion between the charged aggregates. Those attracting forces could be due to van der Waals forces or they could be due to the interfacial tension between the rods and the aqueous phase.H.REHAGE AND H. HOFFMANN 37 1 In this context it seems to be of significance that most of the known viscoelastic detergent systems are built up from cationic detergents with the pyridinium or tri- methylammonium headgroup. The interfacial tension between these groups and the surrounding water must be high. For this reason it seems possible that two interfaces whose charges have been largely neutralized by strongly binding counterions tend to stick together when they collide. In a more abstract or general sense, contact be- tween two rod-like aggregates could also be imagined as an energy minimum be- tween two rods in a particular position and orientation with respect to each other. Such an energy minimum can be explained on the basis of the DLVO-theory.The attractive forces between the rod-like aggregates would be due to van der Waals forces which, at short distances, are strong enough to overcome the repulsive electrostatic forces. However, the fact that the headgroup of the detergent seems to be an important factor in the formation of the supermolecular structure indicates that specific surface effects are involved. The experimental results show that the trimethylammonium headgroup is even more effective in building up supermolecular structures than the pyridinium group. Whatever the exact structure of the contacts may be, the exact nature of the binding affinities between the rod-like micelles is still an open question.The resulting supermolecular network structure which is formed from free indi- vidual rods can be visualized as very long chains which are connected by cross- linking. These flexible chains are probably strongly entangled and the result is a temporary network extending through the whole solution and providing complete molecular binding from one wall to the other. Such dynamic networks, which can be pictured as scaffolding, can be formed from many rod-like particles, e.g. vanadium pentoxide, betonide and carbon black suspension^.^^ - From some of these sys- tems microphotographs could be obtained. These photographs show that the rod- like particles are joined to others at their ends forming a continuous network We would therefore expect that such a structure can also be formed in the micellar systems.The supermolecular network structure built up of rod-like micelles cannot be permanent, as is clearly shown in fig. 1. At low frequencies there exists a terminal zone where the solution can flow. In this region the molecules are able to move with respect to each other. The dynamic aspect of the network must be emphasized. There is dynamic equilibrium between free, individual rods and between rods which are connected to each other. In the network solutions there will always be some fluctuations and the individual rods can undergo translational motion. A given rod might therefore be connected to another particle at one moment and be free at the following moment. Each contact can be characterized by a certain well defined lifetime tL.Furthermore, the size of the rods will vary as a consequence of kinetic processes. Some rods will dissociate completely and new ones will be formed. For this reason a certain shear stress can relax after a sufficiently long time. The solution behaves as a Newtonian liquid under these experimental conditions. Over time-scales < tL no relaxation process can occur and the solution behaves as a rub- ber. The intermediate time-scale is characterized by an equilibrium between relax- ation processes and the formation of new contacts. Table 1 reveals that the dimen- sions of the rod-like micelles ,decrease with increasing temperature. A three- dimensional dynamic network can be formed from short rods by increasing the number of junctions. From fig. 2 it is clear that the elastic properties of the detergent solutions decrease with increasing temperature.The different rheological behaviour is most directly expressed in the relaxation time of the network. The relaxation time z is a time constant describing the decay of the shear stress after the cessation of steady-state flow. Any shear would put a stress on the flexible chains of a network372 VISCOELASTIC DETERGENT SOLUTIONS structure. In the light of our model the relaxation time depends on the average lifetime of a contact point and on the motions of the flexible long-chain aggregates formed from rod-like micelles. The contact time tL is determined by the time cons- tant of free rotation and the adhesion energy between two contacts. Increasing the temperature results in a decrease in the contact time.If the average lifetime of a contact point is the dominating relaxation process we would therefore expect a decrease in the relaxation time constant with increasing temperature. The experi- mental data shown in fig. 2 validate these assumptions. In terms of this simple model the shear stress can only relax by the opening of contact points when the super- molecular structure becomes stretched. From this model we would expect that the stored energy decays with one single-relaxation time constant. A more detailed study of relaxation experiments after cessation of steady-state flow does not validate this assumption. For most of the solutions at least three time constants could be evaluated. Comparison of the relaxation processes of the anisotropy effect of the electrical conductivity and the stress relaxation after the cessation of steady-state flow shows that at least one of these time constants can be attributed to the form- ation of a supermolecular network structure.These experiments indicate that at least three different processes contribute to the decay of the stored energy. At high con- centrations of excess electrolyte, however, the dynamic behaviour of the solutions can be described in as that of a Maxwell material with only a single relaxation time constant. The experimental results of these solutions can be explained in terms of the proposed model. A detailed study of these solutions is still in progress.22 The influence of the chain length on the rheological properties of the detergent solutions can be evaluated from fig.3-5. Shortening of the chain leads to a dramatic reduction of the viscous and elastic properties. For a given circular frequency and the same detergent concentration there is a drop of some orders of magnitude for a decrease in the chain length of 2 CH, groups. At low detergent concentrations the lengths of the rod-like micelles can be cal- culated from viscosity measurements at zero shear rate according to the theory of Doi and Edwards. Some results are summarized in table 2. The dimensions of the rod-like aggregates increase with the number of CH2 groups for a homologous series. Small unisometric particles have rather high rotational diffusion constants and we predict a decrease of the contact time tL between two rods.In terms of these arguments we would expect an increasing relaxation time with decreasing hydrocar- bon chain length. The experimental data can be explained on the basis of the pro- posed model, as can be seen in fig. 4 and 5. The large differences in the flow behaviour of the solutions depend on the dynamic properties of the temporary net- work structure. At low shear rates or angular frequencies the solutions behave as Newtonian liquids. When the shear rate or the circular frequency co is comparable Table 2. Lengths, L/A, of the rod-like micelles as a function of the detergent concentration for alkyltrimethylammonium salicylate solutions ( T = 20 "C) concentration/mmol dm- detergent system 1 2 3 4 5 6 8 10 12 30 40 50 70 C1,TA Sal 385 440 - - - - - - - - - - - C14TA Sal 235 440 460 550 - - - - - -- - - - C, 2TA Sal - - 105 125 140 160 C,,TA Sal 170 - ,290 325 430 460 495 - - - - - - - - - - - - -H. REHAGE AND H.HOFFMANN 373 to the relaxation time constant z, the solution behaves more and more as an elastic body. Under these conditions G' attains a constant value. The characteristic fre- quencies of the plateau region are shifted to higher values by shortening the hydro- carbon chain of the detergent molecules. The relaxation time constants of these solutions and the average lifetime of a contact point between two micelles de- crease. The rheological measurements indicate that the dynamic properties of the net- work solutions depend strongly on the chain length of the detergent molecules. H. Hoffmann, G.Platz, H. Rehage, W. Schorr and W. Ulbricht, Ber. Bunsenges. Phys. Chem., 198 1, 85, 255. H. Hoffmann, G. Platz, H. Rehage, K. Reizlein and W. Ulbricht, Makromol. Chem., 1981, 182, 451. H. Hoffmann, G. Platz and W. Ulbricht, J . Phys. Chem., 1981, 85, 1418. W. Schorr and H. Hoffmann, J . Phys. Chem., 1981,85, 3160. H. Hoffmann, G. Platz, H. Rehage and W. Schorr, Ber. Bunsenges. Phys. Chem., 1981, 85, 877. J. Kalus, H. Hoffmann, K. Reizlein and W. Ulbricht, Ber. Bunsenges. Phys. Chem., 1982, 86, 37. H. Hoffmann, J. Kalus, K. Reizlein, W. Ulbricht and K. Ibel, Colloid Polym. Sci., 1982, 260, 435. H. Hoffmann, G. Platz, H. Rehage and W. Schorr, Adu. Colloid Interface Sci. 1982, 17, 275. H. Hoffmann, H. Rehage, G. Platz, W. Schorr, H. Thurn and W. Ulbricht, Colloid Polym. Sci., 1982, 260, 1042. l o H. Rehage and H. Hoffmann, Rheol. Acta, 1982, 21, 561. l 2 J. Ulmius, H. Wennerstrom, L. B. A. Johansson, G. Lindblom and S. Gravsholt, J. Phys. Chem., l 3 M. Doi and S. F. Edwards, J . Chem. Soc. Faruday Trans. 2, 1978, 74, 918. l 4 B. H. Zimm and D. M. Crothers, Proc. Nut1 Acad. Sci. USA, 1962, 48, 905. W. P. Cox and E. H. Merz, J. Polym. Sci., 1958, 28, 619. l 6 H. Hoffmann, H. Rehage, W. Schorr and H. Thurn, Proceedings of the International Symposium on Surfactants in Solution, University of Lund, Sweden, in press. l 7 K. Strenge and H. Sonntag, Colloid Polym. Sci., 1982, 260, 638. l 8 P. J. Flory, J. Chem. Phys., 1950, 18, 108. l 9 W. W. Graessley, J. Chem. Phys., 1971, 54, 5143. 2 o V. N. Pokrovsky and V. S. Volkov, Vysokomol. Soedin., Ser. A , 1978,20, 255 and 2700. 2 2 H. Hoffmann and H. Rehage, in preparation. 2 3 C. F. Goodeve, Trans. Faraday Soc., 1939, 35, 342. 24 E. A. Hauser and C. E. Reed, J . Phys. Chem., 1937, 41, 911. 2 5 C. M. McDowell and F. L. Usher, Proc. R. Soc. London, Ser. A, 1931, 131, 409 and 504. ' S. Gravsholt, J . Colloid Interface Sci., 1976, 57, 576. 1979, 83, 2232. G. V. Vinogradov and A. Ya. Malkin, Rheology of Polymers (Springer-Verlag, Berlin, 1980).

ISSN:0301-7249    

DOI:10.1039/DC9837600363

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

GENERAL DISCUSSION Dr. M. Holmes (Preston Polytechnic) said: I would like to put two questions to Dr. Tiddy: (1) The partial phase diagram of C, 2E04 (fig. 4 of the paper) is generally similar to other poly(oxyethy1ene) surfactant systems which may rather unusually exhibit an isotropic micellar phase bounded to high temperatures by what is normally believed to be a more highly ordered phase, the lamellar phase. In addition here the existence of a metastable hexagonal phase at low temperature suggests that not only the ordering but also the structure within the isotropic phase region may be unusual. The suggestion that the changes in z are due to a decrease in positional order coupled with an increase in orientational order of disc micelles as the temperature increases seems contradictory and appears to rest on the assertion that within the experimental region the micelles are disc shaped.How much evidence has been obtained from scattering or other techniques that the micelles are indeed disc micelles? (2) In the C12EOs and C12E06 the experimental measurements have been made along lines parallel to phase boundaries, whereas in C12E04 both when temperature and concentration are varied the La phase boundaries are being approached. It is possible that close to phase transitions small clusters of micelles form having short- range order of the type that exists in the La phase. Has Dr. Tiddy considered how the effects of such a clustering would manifest itself with this technique? Dr. G. J. T. Tiddy (Unilever Research, Port Sunlight) replied: (I) There is no unequivocal experimental evidence to demonstrate the existence of disc micelles in the L1 region of C12E04 above the cloud point.The La phase above L1 (at 24 "C) swells to large water separation (> 70 A) as indicated by low- angle X-ray diffraction and 2H n.m.r. measurements., This suggests that bilayers are extensive and relatively robust. Disc micelles in L1 are inferred by assuming that the surface area per surfactant molecule (a) of micelles in the L1 phase is similar to that measured for the La phase, and by considering the nature of the Ll/La transition, as described in ref. (2). It is not obvious how a model based on micelles having other shapes could account for all the observed behaviour. However, we regard the ex- planation given as tentative, forming the basis for further experimentation.(2) The formation of distinct micelle clusters implies the existence of long-lived concentration fluctuations in the solutions. We see no evidence of critical opales- cence (as occurs just at below the cloud point), indicating that any clusters are small. Also, the repulsive forces between micelles due to steric and hydration forces would prevent clusters forming. However, we feel that the results for C12E04 do reflect pre transi tional effects. I. G. Lyle, C. D. Adam, K. Rendall and G. J. T. Tiddy, unpublished results. Trans. I, 1983, 79, 975. * D. J. Mitchell, G. J. T. Tiddy, L. Waring, T. Bostock and M. P. McDonald, J. Chern. Soc., Faraday Prof. H. F. Eicke (University of Basel, Switzerland) asked: Why are the alkyl- trimethylammonium or pyridinium salicylates viscoelastic? Is a particular inter- action necessary for the phenomenon of viscoelasticity in surfactant solutions?376 GENERAL DISCUSSION Prof.H. Hoffmann (University of Bayreuth West Germany) replied: Viscoelastic- ity is a general phenomenon in surfactant solutions. There are many more counter- ions than the salicylate which give rise to viscoelasticity in combination with cationic detergents. In general it can be said that viscoelastic systems can be obtained when the above-mentioned cationic surfactant ions are combined with counterions which bind strongly to the micellar interface. Viscoelastic systems are also obtained by a combination of betaine-type surfact- ants with anionic detergents or of cationic with anionic surfactants.Such systems have been studied in detail by Dr. Tiddy. We have also worked with perfluoro- systems which are viscoelastic at rather low concentrations. Viscoelasticity can also be induced by additives like decanol or naphthol to ionic surfactants. In all theses cases the additives or counterions induce growth of globular micelles to rod-like micelles and the solutions become viscoelastic in the overlap region. In some of the systems the elasticity is not obvious because the relaxation time for the system can be short, say in the ms and p s region. The solutions then appear to be normal, slightly viscous, Newtonian solutions. Dr. Th. F. Tadros (ICZJ Jealott’s Will) said: In Prof. Hoffmann’s paper, he ac- counted for the viscoelasticity in detergent solution in terms of the attraction be- tween the thin ‘rods’ formed, which leads to the formation of a three-dimensional network.An alternative interpretation of the viscoelasticity may be given in terms of the double-layer repulsion between the rods (which have some charges). This double-layer repulsion effect was also suggested to explain the viscoelasticity in sodium montmorillonite suspensions. Evidence of the role of double layers may be obtained from the effect of added electrolyte. Addition of electrolyte was shown to reduce viscoelasticity, and at sufficiently high electrolyte concentrations viscoelastic- ity disappeared altogether. Could Prof. Hoffmann comment on the role of double layers in these viscoelastic detergent solutions? Prof.H. Hoffmann (University of Bayreuth, West Germany) said: While the electrostatic repulsion between the rods seems to be of some importance for the occurence of elasticity it can not be the only significant parameter. This can clearly be concluded from the experimental observation that for most viscoelastic surfactant systems the storage modulus for a given oscillating frequency, in all the systems we have looked at, at first increases with increasing salt concentration. This clearly shows that the elasticity in the system is increasing with salt concentration usually over quite a large range of salt, for example between lop4 and mol dm-3 NaC1.l This would be difficult to understand if electrostatic repulsion were the main cause for the elasticity.For a 1 x mol dm-3 NaCl solution the Debye length is of the order of 30 A and it is clear that for rod-like micelles with a mean distance of ca. 1000 A the electrostatic interaction must be very small. Experimentally, we see in SANS measurements the correlation peak which is due to electrostatic repulsion disappearing with the addition of salt, while the elasticity is increasing. We must assume therefore that attractive terms between the rods are of importance for the formation of the contacts between the rods in the network model or for the slow-down of the rotation of the rods in a more general sense. The existence of attractive terms manifests itself also in the phase separation of viscoelas- tic systems for higher salt concentrations. This phase transition can be compared with the liquifaction of a gas.It is true that before the phase transition is reached the elasticity decreases with increasing salt concentration. This breakdown of the elasticity is not clear. It couldGENERAL DISCUSSION 377 possibly be related to the rods becoming more flexible with increasing salt concentration. H. Hoffmann, H. Rehage, K. Reizlein and H. Thurn, in Proc. Symp. on Microemulsions, Washing- ton D.C., August 1983, ed. D. Shah (Am. Chem. SOC., Washington D.C., in press). Dr. J. D. F. Ramsay (Harwell, Oxfordshire) said: The viscoelastic behaviour of the cationic detergent systems can be explained, as Prof. Hoffmann describes in his paper, by a network of interconnected rod-like micelles, although Dr. Tadros has suggested that similar behaviour might arise from strong electrostatic repulsion be- tween individual micelles, without invoking a network structure.To obtain further confirmation for the structure suggested it would be interesting to examine the effect of strain amplitude on the storage and loss moduli of these systems, and in parti- cular to determine the critical strain, yo for disruption of the network. It can then be shown1 that the work required for disruption, which is equivalent to the cohesive energy, E,, of the dispersion, is given by Jo where pc corresponds to a maximum peak stress in phase with the strain. Ec should then depend on the density of contacts in the network and the force of attraction between the rod-like micelles. J. D. F. Ramsay, S. R. Daish and C. J.Wright, Faraday Discuss. Chem. SOC., 1978, 65, 65. Prof. H. Hoffmann (Uniuersity of Bayreuth, West Germany) replied: In order to determine the linear viscoelastic deformation conditions we have investigated the effect of strain amplitude on the storage and loss moduli of viscoelastic detergent solutions. The results of these measurements show that the internal structures can be stretched by a factor 3 or 4 before the structures break. This result also points to the existence of a three-dimensional structure having similar properties to rubber networks. Dr. R. B. Jones (Queen Mary College, London) said: Prof. Hoffmann has con- cluded that the rheological behaviour of his solutions implies the existence of a network structure formed by rods sticking together. Is this network to be thought of as local clusters or as an infinite network in the sense of percolation theory? Is there an experimental number for the lifetime of such a network or for the lifetime tL of contact of two rods? Prof.W. B. Russel (Princeton University, U.S.A.) said: The paper by Dr. Rehage and Prof. Hoffmann mentions the use of the Doi-Edwards theory, for semi-dilute solutions of rigid rods, to interpret some of the data and states that the lengths extracted conform with those calculated from scattering experiments. What is not clear is why this does not provide a satisfactory interpretation for the bulk of the data. The trends described conform at least qualitatively with those to be expected for rigid rods in the semi-dilute regime: a dramatic increase in viscosity and storage modulus with increasing concentration, and a highly restricted rotational diffusion but free translational diffusion.378 GENERAL DISCUSSION Are there critical experiments which demonstrate quantitatively the inadequacy of this interpretation? Prof.H. Hoffmann (University of Bayreuth, West Germany) said: In several papers we have used the Doi-Edwards theory to calculate lengths for the rods in the overlap region from the viscosity. These values agree well with results from other techniques. Furthermore, we see the development of long tails in the transient electric-birefringence measurements when we pass into the semi-dilute region. All these results seem to be in qualitative if not in quantitative agreement with the Doi-Edwards theory, which was derived for stiff rods in the absence of attractive and repulsive forces between them.Recently the Doi-Edwards theory has been at- tacked by Pecora,l who claimed that for non-interacting systems the influence of restricted rotation should come in at considerably higher concentrations than at the overlap concentration, as proposed by Doi and Edwards. Based on these theoretical considerations it seems that in our systems attractive forces between the rods make the systems appear more crowded than they really are, and it is only by fortuitous compensation effects that we can use the Doi-Edwards theory. K. M. Zero and R. Pecora, Macromolecules, 1982, 15, 87. Prof. H. Hoffmann (University of Bayreuth, West Germany) said: The experi- ments show that the rods do not fuse together but continue to exist as independent rods even in the viscoelastic region.The main evidence for this is seen in the trans- lational diffusion coefficients from dynamic light-scattering data. These results clear- ly show that the mutual translation diffusion coefficients are little affected when we pass from the dilute to the semidilute region. If anything the translational diffusion becomes even faster. This could be due to the fact that in the dilute region the translational motion is determined by the two diffusion coefficients for diffusion both along and perpendicular to the axis of the rods. [D = (Dll + 20L)/3 with Dil > DJ. In the overlap region diffusion perpendicular to the axis is no longer possible and diffusion along the axis is unhindered.At the cross-over concentration the rotation of the rods becomes extremely restricted. This can best be observed from transient electric-birefringence decay curves (see Proceedings of the Symposium on Surfactants held in Lund, July 1982). On the other hand fusion of rods or even globular micelles to larger rods does seem to occur in solutions undergoing shear. It was shown by us that viscoelastic solutions can be induced by shearing of the solutions which are in the dilute region. These data can best be explained by postulating that small micelles which are forced to collide in the shear gradient fuse together and form larger rods which, when sheared long enough, will overlap and form a dynamic network. Furthermore, the equilibration time which is needed to reach equilibrium again after the solution has been heated for a short period and globular micelles have been formed during this process can be shortened by shearing.At present we do not know the lifetime of a contact. It is likely that this time constant is related to the network relaxation time, but unfortunately we do not know the quantitative correlation between these two parameters. We assume that percolation theories can be used for the description of the rheo- logical properties. Dr. J. H. R. Clarke (UMIST, Manchester) said: Prof. Hoffmann discusses the remarkable effects of surfactant chain length on the viscoelastic properties of deter-GENERAL DISCUSSION 379 gent solutions in terms of the ‘average lifetime of contact points’ between rod-like micelles.Have the authors considered the possibility that rod flexibility and its de- pendence on chain length also play a part in determining flow properties? Prof. H. Hoffmann (University of Bayreuth, West Germany) said: This seems to be a good possibility. So far, we only have detailed information on the flexibility of the rods for the CI6 system. For this system the rods seem to be fairly stiff with a persistence length of more than several thousand A. Intuitively one could imagine that the rods become more flexible with shorter chain length because the diameter of a rod is approximately equal to the extended length of the hydrocarbon chain, and a thinner rod is likely to have a higher intrinsic flexibility than an thicker one, other factors such as surface charge density being equal. Another possibility to explain the different 1 ~ * 1 values would be that some of the rods are sitting in energy minima which are determined by attractive and repulsive terms from the DLVO theory. It could be argued that the depth of the minimum becomes shallower with decreasing diameter of the rods. A third explanation would be that the Iq*l values are kinetically controlled by fluctuations in the length of the rods or even the lifetimes of the rods. It is known from kinetic relaxation measurements on micellar systems that both of these time constants depend very much on the chain length and decrease with decreasing chain length. It is conceivable that these processes control the rheological behaviour. It is of course also possible that all three factors mentioned contribute to the flow behaviour of these systems. More work is certainly necessary to achieve a better understanding. The experimental results clearly show that as far as the de- pendence of lq*l on chain length is concerned we are dealing with a general phenomenon.

ISSN:0301-7249    

DOI:10.1039/DC9837600375

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

LIST OF POSTERS Structural studies of water-in-oil microemulsions using CTAB as surfactant P. D. I. Fletcher, J. Mead, J. McDonald, C. Toprakcioglu, J. C. Dore and B. H. Robinson Chemistry and Physics Laboratories, University of Kent Critical scattering and phase stability in Aerosol-OT-stabilised microemulsions C. Toprakcioglu, J. C. Dore, B. H. Robinson and A, M. Howe Chemistry and Physics Labora- tories, University of Kent Rheological properties of ferrofluids S. Alker, R. W. Chantrell, P. R. Bissell and P. A. Bates Division of Physics and Astronomy, Preston Polytechnic Smectic-nematic transition in a lyotropic liquid crystal M. C. Holmes and J. Charvolin Division of Physics and Astronomy, Preston Polytechnic The structure and rheology of a very concentrated emulsion C.A. Mummk-Young, R. Buscall, J. McMahon and J. Cooper ICI Corporate Colloid Science Group, Runcorn, Cheshire Small-angle X-ray scattering of concentrated silica dispersions J. A. H. M. Moonen and A. Vrij Van 't HoffLaboratorium, Utrecht, The Netherlanh Dielectric constants of concentrated reversed micellar solutions J. Sjoblom, B. Jonsson and I. Lundestrom Institute for Surface Chemistry, Stockholm, Sweden Interactions of polio virus with sand and clay in electrolyte solutions V. L. Vilker Chemical Engineering Department, University of California, U.S.A. Light-induced viscosity changes of aqueous solutions containing cetyltrimethylammonium micelles and 9-substituted anthracenes T. Wolff, G. von Bunau and M. Muller Physikalische Chemie, Universitat Siegen, 05900 Siegen, West Germany The influence of shear on the coagulation of aqueous quartz dispersions H.N. Stein Laboratory for Colloid Chemistry, Eindhown University of Technology, The Netherlanh Preparation and stability of model, concentrated non-aqueous silica dispersions J. Edwards, S. Lenon and B. Vincent Department of Physical Chemistry, University of Bristol Morphology of concentrated flocculated dispersions J. McMahon, R. F. Stewart and D. Sutton ICI Corporate Colloid Science Group, PO Box 11, The Heath, Runcorn, Cheshire Investigation of various effects on the rheological properties of PVC plastisols H. Zecha Akademie der Wissenschaften der DDR, Institut fur Polymerenchemie, 153 Teltow- Seehof, Kantstrasse 55, East Germany Self-diffusion in a concentrated suspension: resummation of many-body hydrodynamic interactions C.W. J. Beenakker Instituut-Lorentz, Leiden, The Netherlands The mode of aggregation of montmorillonite platelets J. A. McShea Physical Sciences Branch, BP Research Centre, Sunbury-on-Thames Spherical agglomeration and its application to recovery of bitumen from tar sands S. Levine, B. D. Sparks and F. Weldon Meadus Division of Chemistry, National Research Council, Ottawa, CanadaRheology of foams and highly concentrated emulsions H. M. Princen Exxon Research and Engineering Company, Linden, New Jersey, U.S.A. Refractometry of concentrated dispersions G. F. Harding, H. G. Meeten, A. N. North and F. M. Willmouth Department ofphysics, Sir John Cass School of Science and Technology, 31 Jewry Street, London Steric effects in the coagulation of casein rnicelles by ethanol D. S. Horne Hannah Research Institute, Ayr, Scotland Structure of a concentrated microemulsion in the critical region J. Tabony, A. de Geyer and M. Drifford Centre d’Etudes Nuclkaires de Saclay, France

ISSN:0301-7249    

DOI:10.1039/DC9837600381

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

INDEX OF NAMES* Ackerson, B. J. 103, 219, 233, 235, 246, 247, Ahlnas, T. 317 Avery, R. G. 53 Bacon, J. 165 Beenakker, C. W. J. 230, 231, 240, 337 Benest, L. 53 Beresford-Smith, B. 65, 109, 111, 112, 115, Buscall, R. 256, 277, 337, 338 Cazabat, A-M. 291 Cebula, D. J. 37 Chan, D. Y. C. 65, 109, 111, 112, 115, 116 Chatenay, D. 291 Clark, N. A. 219 Clarke, J. H. R. 332, 348, 378 Corti, M. 116 Croucher, M. D. 261, 331, 332, 333 Cummins, P. G. 77, 11 6, 117, 1 19 Degiorgio, V. 99, 102, 258, 340, 343 Dhont, J. K. G. 19 Dickinson, E. 104, 112, 165, 232, 240, 241, Eicke, H-F. 305, 339, 343, 344, 346, 375 Felderhof, B. U. 99, 100, 106, 1 12, 179, 229, 241, 244, 342 Fijnaut, H. M. 19 Fletcher, P. D. I. 349 Garvey, M. J. 337 Gast, A. P. 189 Goodwin, J. W. 37, 243 Hall, C. K. 189 Hayter, J.B. 7, 97, 98, 99, 257 Heath, D. 203 Hess, W. 137 Higgins, J. S. 77, 331 Hoffman, H. 97, 103, 341,343,363,376,377, Holmes, M. 375 Howe, A. M. 349 Jansen, J. W. 19 Jeffrey, G. C. 37, 106 Jennings, B. R. 353 Jones, R. B. 97, 100, 179, 247, 377 Klein, R. 102, 115, 137, 233, 234, 235, 238, Kops-Werkhoven, M. M. 19 257, 258, 259, 336 116 242, 243, 244, 246 378, 379 239, 246 Kubik, R. 305, 346 Langevin, D. 291, 339, 340, 341, 342, 346, Lekkerkerker, H. N. W. 231, 239, 344 Levine, S. 258 Lindman, B. 317, 341, 345, 347 Lyle, I. G. 77 McGowan, I. J. 277 Meunier, J. 291 Milkie, T. H. 261 Moonen, J. 95 Muddle, A. G. 77 Mummt-Young, C. A. 255 Neeson, P. G. 353 Ottewill, R. H. 37, 103, 105, 106, 107, 243, Parentich, A. 37 Parker, R. 165 Pathmamanoharan, C. 19 Pusey, P. N. 93, 113, 117, 123, 229, 231, 232, 235, 259, 260 Ramsay, J. D. F. 53, 108, 110, 377 Rapacki, K. 317 Rarity, J. G. 243, 258 Rehage, H. 363 Richardson, R. A. 37 Robinson, B. H. 349 Russel, W. B. 110, 189, 247, 248, 250, 337, Snook, I. 106, 117, 151, 239, 241, 242, Soderman, 0. 317 Staples, E. J. 77 Stein, H. N. 251, 254 Stilbs, P. 317 Tadros, Th. F. 98, 116, 203, 231, 238, 250, 254, 255, 256, 260, 332, 340, 344, 347, 376 Tiddy, G. J. T. 97, 345,353, 375 Tough, R. J. A. 123, 229, 231, 232, 235, 260 van Megen, W. 102, 107, 110, 118, 151, 237, Vincent, B. 249, 331 Vrij, A. 19, 95, 100, 101, 102, 103, 106, 115 239, 247, 259, 338 Walderhaug, H. 317 Warner, M. 99 Woodcock, L. V. 239, 243, 333 349 260 377 247 240, 241, 247, 248, 340 * The page numbers in bold type indicate papers submitted for discussion.

ISSN:0301-7249    

DOI:10.1039/DC9837600383

出版商:RSC    

年代:1983

数据来源: RSC

摘要:

THE Date 1907 1907 1910 191 1 1912 1913 1913 1913 1914 1914 1915 1916 1916 1917 1917 1917 1918 1918 1918 1918 1919 1919 1920 1920 1920 1920 1921 1921 1921 1921 1922 1922 1923 1923 1923 1923 1923 1924 1924 1924 1924 1924 1925 1925 1926 1926 1927 1927 1927 1928 1929 1929 1929 1930 1931 1932 1932 GENERAL DISCUSSIONS OF FARADAY SOCIETY/FARADAY DISCUSSIONS OF THE CHEMICAL SOCIETY Subject Osmotic Pressure Hydrates in Solution The Constitution of Water High Temperature Work Magnetic Properties of Alloys Colloids and their Viscosity The Corrosion of Iron and Steel The Passivity of Metals Optical Rotatory Power The Hardening of Metals The Transformation of Pure Iron Methods and Appliances for the Attainment of High Temperatures in a Refractory Materials Training and Work of the Chemical Engineer Osmotic Pressure Pyrometers and Pyrometry The Setting of Cements and Plasters Electrical Furnaces Co-ordination of Scientific Publication The Occlusion of Gases by Metals The Present Position of the Theory of Ionization The Examination of Materials by X-Rays The Microscope: Its Design, Construction and Applications Basic Slags: Their Production and Utilization in Agriculture Physics and Chemistry of Colloids Electrodeposition and Electroplating Capillarity The Failure of Metals under Internal and Prolonged Stress Physico-Chemical Problems Relating to the Soil Catalysis with special reference to Newer Theories of Chemical Action Some Properties of Powders with special reference to Grading by The Generation and Utilization of Cold Alloys Resistant to Corrosion The Physical Chemistry of the Photographic Process The Electronic Theory of Valency Electrode Reactions and Equilibria Atmospheric Corrosion.First Report Investigation on Oppau Ammonium Sulphate-Nitrate Fluxes and Slags in Metal Melting and Working Physical and Physico-Chemical Problems relating to Textile Fibres The Physical Chemistry of Igneous Rock Formation Base Exchange in Soils The Physical Chemistry of Steel-Making Processes Photochemical Reactions in Liquids and Gases Explosive Reactions in Gaseous Media Physical Phenomena at Interfaces, with special reference to Molecular Atmospheric Corrosion. Second Report The Theory of Strong Electrolytes Cohesion and Related Problems Homogeneous Catalysis Crystal Structure and Chemical Constitution Atmospheric Corrosion of Metals.Molecular Spectra and Molecular Structure Colloid Science Applied to Biology Photochemical Processes The Adsorption of Gases by Solids The Colloid Aspect of Textile Materials Laboratory Elutriation Orientation Third Report Volume Trans. 3* 3* 6* 7* 8* 9* 9* 9* 10* 1 o* 11* 12* 12* 13* 13* 13* 14* 14* 14* 14* 15* 15* 16* 16* 16* 16* 17* 17* 17* 17* 18* 18 19* 19 19* 19 19* 20* 20* 20* 20* 20* 21* 21 22 22 23* 23* 24 24 25* 25* 26* 26 27 28 29Date 1933 1933 1934 1934 1935 1935 1936 1936 1937 1937 1938 1938 1939 1939 1940 1941 1941 1942 1943 1 944 1945 1945 1946 1946 I947 1947 1947 1947 1948 1948 1949 1949 1949 1950 1950 1950 1950 1951 1951 1951 1952 1952 1953 1953 1954 1954 1955 1955 1956 1956 1957 1958 1957 1958 1959 1959 I960 1960 1961 1961 1962 1962 1963 Subject Liquid Crystals and Anisotropic Melts Free Radicals Dipole Moments Colloidal Electrolytes The Structure of Metallic Coatings, Films and Surfaces The Phenomena of Polymerization and Condensation Disperse Systems in Gases: Dust, Smoke and Fog Structure and Molecular Forces in (a) Pure Liquids, and (b) Solutions The Properties and Functions of Membranes, Natural and Artificial Reaction Kinetics Chemical Reactions Involving Solids Luminescence Hydrocarbon Chemistry The Electrical Double Layer (owing to the outbreak of war the meeting was The Hydrogen Bond The Oil-Water Interface.The Mechanism and Chemical Kinetics of Organic Reactions in Liquid The Structure and Reactions of Rubber Modes of Drug Action Molecular Weight and Mokcular Weight Distribution in High Polymers (Joint Meeting with the Plastics Group, Society of Chemical Industry) The Application of Infra-red Spectra to Chemical Problems Oxidation Dielectrics Swelling and Shrinking Electrode Processes The Labile Molecule Surface Chemistry (Jointly with the Sociktk de Chimie Physique at Colloidal Electrolytes and Solutions The Interaction of Water and Porous Materials The Physical Chemistry of Process Metallurgy Crystal Growth Lipo-proteins Chromatographic Analysis Heterogeneous Catalysis Physico-chemical Properties and Behaviour of Nuclear Acids Spectroscopy and Molecular Structure and Optical Methods of Investi- gating Cell Structure Electrical Double Layer Hydrocarbons The Size and Shape Factor in Colloidal Systems Radiation Chemistry The Physical Chemistry of Proteins The Reactivity of Free Radicals The Equilibrium Properties of Solutions on Non-electrolytes The Physical Chemistry of Dyeing and Tanning The Study of Fast Reactions Coagulation and Flocculation Microwave and Radio-frequency Spectroscopy Physical Chemistry of Enzymes Membrane Phenomena Physical Chemistry of Processes at High Pressures Molecular Mechanism of Rate Processes in Solids Interactions in Ionic Solutions Configurations and Interactions of Macromolecules and Liquid Crystals Ions of the Transition Elements Energy Transfer with special reference to Biological Systems Crystal Imperfections and the Chemical Reactivity of Solids Oxidation-Reduction Reactions in Ionizing Solvents The Physical Chemistry of Aerosols Radiation Effects in Inorganic Solids The Structure and Properties of Ionic Melts Inelastic Collisions of Atoms and Simple Molecules High Resolution Nuclear Magnetic Resonance The Structure of Electronically Excited Species in the Gas Phase abandoned, but the papers were printed in the Transactions) Systems Bordeaux) Published by Butterworths Scientific Publications, Ltd Volume 29* 30 30 31* 31* 32* 32* 33* 33* 34* 34* 35* 35* 35* 36* 37* 37* 38 39 40* 41 * 42* 42 A 42 B Disc.1* 2 Trans. 43* Disc. 3 4* 5 6 7 8* Trans. 46* Disc. 9 Trans. 47 Disc. 10 11 12* 13 14 15 16* 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33* 34 35Date 1963 1964 1964 1965 1965 1966 1966 1967 1967 1968 1968 1969 1969 1970 1970 1971 1971 1972 1972 1973 1973 1974 1974 1975 1975 1976 1977 1977 1977 1978 1978 1979 1979 1980 1980 1981 1981 1982 1982 1983 1983 Subject Fundamental Processes in Radiation Chemistry Chemical Reactions in the Atmosphere Dislocations in Solids The Kinetics of Proton Transfer Processes Intermolecular Forces The Role of the Adsorbed State in Heterogeneous Catalysis Colloid Stability in Aqueous and Non-aqueous Media The Structure and Properties of Liquids Molecular Dynamics of the Chemical Reactions of Gases Electrode Reactions of Organic Compounds Homogeneous Catalysis with Special Reference to Hydrogenation and Bonding in Metallo-organic Compounds Motions in Molecular Crystals Polymer Solutions The Vitreous State Electrical Conduction in Organic Solids Surface Chemistry of Oxides Reactions of Small Molecules in Excited States The Photoelectron Spectroscopy of Molecules Molecular Beam Scattering Intermediates in Electrochemical Reactions Gels and Gelling Processes Photo-effects in Adsorbed Species Physical Adsorption in Condensed Phases Electron Spectroscopy of Solids and Surfaces Precipitation Potential Energy Surfaces Radiation Effects in Liquids and Solids Ion-Ion and Ion-Solvent Interactions Colloid Stability Structure and Motion in Molecular Liquids Kinetics of State Selected Species Organization of Macromolecules in the Condensed Phase Phase Transitions in Molecular Solids Photoelecttochemistry High Resolution Spectroscopy Selectivity in Heterogeneous Catalysis Van der Waals Molecules Electron and Proton Transfer Intramolecular Kinetics Concentrated Colloidal Dispersions Oxidation 387 Volume 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65* 66 67 68 69 70 71 72 73 74 75 76 * Not available; for current information on prices, etc., of available volumes, please contact the Marketing Of$cer, Royal Society of Chemistry, Burlington House, London Wl V OBN stating whether or not you are a member of the Society.

ISSN:0301-7249    

DOI:10.1039/DC9837600385

出版商:RSC    

年代:1983

数据来源: RSC