J. Chem. Soc., Faraday Trans. I, 1984, 80, 1525-1537 Adsorption of Ions at the Cellulose/Aqueous Electrolyte Interface Part 1 .--Charge/pH Isotherms BY THELMA M. HERRINGTON* AND BRIAN R. MIDMORE Department of Chemistry, University of Reading, Reading RG6 2AD Received 14th September, 1983 Charge/pH isotherms are reported for various cellulose fibres : cotton linters, bleached sulphate pulp and unbleached sulphate pulp. This charge was determined as a function of pH for 1.0, 0.1, 0.01 and 0.001 mol dm-3 NaC1. The effect of various cations on the charge on bleached sulphate pulp was also investigated. The results for cotton linters and bleached sulphate pulp are analysed in terms of polyelectrolyte theory. The coagulation of cellulose fibres at the mesh of a paper machine is a complex process with many variables.Many workersly have determined the zeta potential of cellulose fines as a function of pH and added electrolyte. This study was undertaken to find out the charge at the cellulose surface rather than at the plane of shear. Previous workers investigating the surface charge of cellulose have treated the system as a polyacid and have investigated the ion-exchange reaction between acidic protons and metal cations cellulose-H + M+ + cellulose-M+ + H+. This reaction in aqueous solution is equivalent to the adsorption of hydroxide ions, as the exclusion of water from the cellulose molecule is not analytically detectable: OH- / cellulose-H +OH- --+ cellulose, H The acidity of cellulose has generally been attributed within the cellulose structure. -H,O --+ cellulose-.to occasional carboxy groups Neale3 titrated oxycellulose against alkali ; only the ‘end-point ’ of the titration was determined, which gave the carboxy content. He then dispersed the oxycellulose in sodium chloride solution and measured the pH. This gave him a value for the degree of dissociation, a, and he estimated a pK of 4.2 for the acid group in oxycellulose. Neale in effect determined a single point on a pH/alkali adsorption isotherm at ca. pH 7. The work of Heymann4 which determined the quantity of acid released after dispersing cotton linters in neutral electrolyte solution also gave a single point on this isotherm. The first to find more points on this isotherm was David~on.~ He worked on oxycellulose prepared by the action of alkaline hypobromite on cotton using the adsorption of methylene blue, which has a high affinity for the carboxy group.However, the ionic strength was not held constant. None of these workers attempted 15251526 THE CELLULOSE/ AQUEOUS ELECTROLYTE INTERFACE to determine the surface area. The first work attempted on cellulose using the titration technique with comparison to a blank to determine the absorption of hydroxide ions was performed by Edelson.6 He titrated microcrystalline cellulose both in water and in 0.1 mol dm-3 KCI, varying the pH between 2 and 11. Only in the latter case was the ionic strength kept effectively constant. Two acid groups were detected, having pKvalues of 4.0 and 9.2, respectively. The latter was concluded to be ammonia present on the surface as an impurity, while the former was confirmed to be a c6 carboxy group.This was concluded as oxycellulose prepared by reaction with nitrogen dioxide gave a pK of 4.1. Using geometrical approximations, a charge density of 0.0 16 C m-2 can be estimated for the crystallites at the neutralization point. From the work of Davidson5 and Edelson6 a pH value of CQ. 2 for the P.Z.C. of cellulose seems reasonable. As far as we can ascertain this is the first work to determine the charge/pH isotherms on various types of cellulose fibres and also to determine the surface area of these fibres by different techniques. EXPERIMENTAL MATERIALS All salts were AnalaR grade and were used without further purification. The standard sodium hydroxide solution prepared from stock? was checked to be free of carbonate by titration; a concentration of only 0.1 % of carbonate gave two end-points.All water used was double distilled and had a conductivity of < 5 x The cotton linters were Hercules Chemical Cotton (Holden Vale Manufacturing Co., Haslingden, Lancashire), which is prepared from raw cotton linters by wet and dry mechanical cleaning, pressure digestion in caustic soda, bleaching, washing and drying. The degree of polymerization was 1040 and the carboxy content of the sample (determined by Hercules) was 1.5 x lop2 mol kg-l. The bleached sulphate pulp (STORA32) and unbleached sulphate pulp were Scots Pine supplied by Pira (Leatherhead, Surrey) and had never been dried. Both pulps were used in their wet form (16%, w/w, solids) and were stored at 4-6 "C.The cotton linters came in dry sheet form; a blender was used to disperse the fibres in water allowing a standard time of 90 s to produce the slurry, which was filtered on a sintered glass funnel to give a wet pulp. The fibres were treated with acid to remove unwanted cations and to produce the acid form of acidic groups present: 20 g of the fibres were soaked in 2 dm3 of 0.1 mol dmp3 HC1 for 20 h, washed and then soaked in conductivity water for a further 20 h to remove the last traces of acid; the supernatant was checked for the absence of chloride. After further washing the fibres were sucked as dry as possible. 0-l cm-l. DETERMINATION OF CHARGE/PH ISOTHERMS The basic technique consists of the potentiometric titration of a fibre suspension in aqueous electrolyte solution of varying ionic strength with H+ or OH- ions using a glass plus a silver, silver chloride electrode for pH determination.A sample of electrolyte solution of the same volume and concentration as the fibre suspension is then titrated with the same acid or alkali. The difference between the amounts of H+ or OH- ions that produce a given pH in the fibre suspension and the corresponding pH in the blank sample of the electrolyte gives the amount of H+ or OH- ions adsorbed by the surface. The surface charge is defined as where rH+ and r o H - are the surface excess of H+ and OH- ions. Silver, silver chloride electrodes were prepared ele~trolytically.~ The bias potentials of such electrodes did not exceed 0.2 mV. The potentials were measured at 25k0.05 "C using a Metrohm E636 titroprocessor.The electrodes were standardized using 6 standard buffers.' The assumption that ycl- = YNsCl was made; this is reasonable as for most pH values (4-10), mHCl + mNaCl. Y~~~~ was obtained from tables.8 Thus pH is defined as pH = ( E - p ) eIRT+ log (YNaclmC1-)*T. M. HERRINGTON AND B. R. MIDMORE 1527 A polypropylene guard was fitted to prevent tangling of the fires on the electrodes. The titration was carried out in a gas-tight vessel under a current of pure nitrogen to avoid contamination with CO, The greatest problem in extending this titration technique, which has been used for silica gel9 and metal oxides,1° to a suspension of cellulose fibres is the attainment of equilibrium. Because cellulose fibres are highly porous, there is the problem of obtaining uniform distribution of the electrolyte solution within the fibre.Coupled with this is the problem of stirring and ensuring adequate mixing in the bulk solution. To complete the titration in a reasonable period of time, i.e. ca. 4 h, a maximum of 1.8 g of fibre could be added to 450 cm3 of solution. This weight limitation proved to be a problem in the case of the cotton linters because of their very low charge density per gram. More rapid titrations of ca. 2 h in duration led to erroneous peaks and troughs in the differential plot of e.m.f. against volume, indicating non-attainment of equilibrium. There are two dilution effects, one caused by the water in the wet fibre and the other by the addition of acid or alkali.These two effects were corrected for by determining the expected e.m.f. of the solution if the fibres were dry and the acid or alkali of infinite molarity. Allowance was also made for the fact that certain areas of the fibre which are accessible to water are inaccessible to electrolyte. Thus a certain amount of water is effectively removed from the system, increasing the concentration of electrolyte present. As the amount removed i s constant, the absolute increase in electrolyte concentration is proportional to the total electrolyte concentration. Therefore, as we are only interested in variations in hydrogen-ion concentration, this effect only assumes significance below pH 2. The volume of water so absorbed by the fibre was determined in connexion with the negative adsorption experiments (Part 2 of this series).The continuous titration method indirectly determines the hydrogen-ion concentration by measurement of e.m.f. The precision of the titration technique between the pH values of 4 and 10 is governed by errors in the concentration of titrant, volume added and weight of fibre, which amount to 0.2%. The reproducibility of the results between batches of fibre was very high, the error being k 1 %. Even relatively large errors in the e.m.f., e.g. 0.3 pV, at pH 4 generate only small errors of kO.3 pmol. However, below pH 3 the precision rapidly deteriorates and small errors in e.m.f. generate large errors in ArOH-. Below pH 3 the continuous titration technique was not employed and instead the single-point method was used in which the acid concentration in the supernatant was determined by direct titration against alkali.HYSTERESIS EXPERIMENT A hysteresis experiment was performed on the bleached sulphate pulp to determine the reversibility of the system. It was performed in the same way as the normal continuous titration, using 0.1 mol dmP3 NaCl as the electrolyte. After 4 h of titration with alkali to pH 10.5, the titrant was changed to hydrochloric acid and the pH lowered to its initial value over a further period of 4 h. From fig. 1 it can be seen that the degree of hysteresis is small, amounting to 14% of the maximum charge at its greatest at pH 9.5. The amount of hysteresis continued to drop to a value of 8 % of the maximum charge, which suggests that some at least of this hysteresis is not permanent and is time-dependent.In view of the very high pH, 10.5, to which the suspension was raised, this degree of hysteresis seems very small, SURFACE AREA Extensive determinations of the surface area of cellulose fibres were made both by B.E.T. and negative-adsorption techniques. The effect of beating and of oxidation on the surface area were investigated. These measurements are discussed in detail in Part 2 of this series. In this paper, the surface charge for the adsorption isotherms is recorded in C g-' of fibre, as this is the quantity which is directly determined. SODIUM-ION ADSORPTION MEASUREMENTS The adsorption of sodium ion was determined at a few points on the isotherm to check the charge balance. The sodium adsorbed together with the negative adsorption of chloride (this is in fact negligible: see Part 2 of this series) should exactly balance the surface charge.The adsorption of sodium was determined using the principle of the single-point technique but1528 THE CELLULOSE/AQUEOUS ELECTROLYTE INTERFACE 1c 9 8 PH 7 6 5 4 I I I amount of OH- added/pmol amount of H" added/pmol I I I I I I i 200 I I 1 I I00 200 0 Fig. 1. Hysteresis curve for bleached sulphate pulp in 0.1 mol dm-3 NaCl: ., forwards; V, backwards. Table 1. Sodium adsorption results for bleached sulphate pulp PH st7+/c g-1 - So& g-' 0.001 rnol dm-3 NaCl 10.89 3.50 9.39 2.84 4.91 1.41 3.10 2.90 1.35 0.01 mol dm-3 NaCl 10.82 2.92 3.10 analysing for sodium using a flame photometer. The experiment was performed in and mol dm-3 NaC1.In table 1 it can be seen that the sodium adsorption agrees well with the hydroxide adsorption. This indicates that it is the hydroxide ions that are responsible for the charge on the cellulose surface, which is balanced by sodium-ion adsorption in the liquid side of the electrical double layer. RESULTS Charge/pH isotherms at 25 "C were obtained for various cellulose fibres using the continuous titration and single-point techniques. The fibres chosen were cotton linters, a bleached sulphate commercial pulp and an unbleached sulphate commercial pulp. The isotherms were constructed over a pH range 1.3-10.3 and sodium chloride was used as the indifferent electrolyte. Four concentrations of indifferent electrolyte were employed, namely 1.0, 0.1, 0.01 and 0.001 mol dm-3.The isotherms are given in fig.T. M. HERRINGTON AND B. R. MIDMORE 1529 Fig. 2. Charge/pH isotherm at 25 "C for cotton linters. [NaCl]/mol dmP3 as follows: 0, 10-3; 8, 10-2; 0, 10-1; 0, 100. 3 2 - I 00 u . 0 cs: I 1 C Fig !. Charge/pH isotherm at 25 "C for bleached sulphate pulp. [NaCl]/mol dm-3 as foiiows: 0; 10-3; 0, 10-2; a, 10-1; 0, 100.1530 THE CELLULOSE/AQUEOUS ELECTROLYTE INTERFACE Fig. 4. Charge/pH isotherm at 25 "C for unbleached sulphate pulp. [NaCl]/mol dmP3 as foiiows: 0, 10-3; 8, 10-2; 0, 10-1; 0, 100. Fig. 5. Effect of various cations on the charge/pH isotherm at 25 "C for bleached sulphate pulp. A, 0.05 mol dm-3 BaC1,; 0, 0.05 mol dm-3 CaC1,; 0, 0.1 mol dm-3 NaCl; A, 0.1 mol dm-3 KC1; ., 0.1 mol dm-3 (C,H,),NCl.T.M. HERRINGTON AND B. R. MIDMORE 1531 2-4. The reader is referred to the original thesis’ for the experimental data from which these plots are drawn. For the bleached sulphate pulp, the effects on the isotherm of various cations were investigated ; solutions of calcium, barium, potassium and tetraethylammonium chloride were used at 0.1 mol dm-3 with respect to the chloride ion. The resulting charge/pH isotherms are shown in fig. 5 . DETERMINATION OF a AND CALCULATION OF pK In the case of a simple weak acid, such as acetic, the ionization constant is given by pK, = pH +log [( 1 - a)/a]. (3) For a polymer carrying a large number of ionizable groups, the apparent ionization constant of an average ionizable group is defined by pK = pH + log [( 1 - a)/al (4) where K will vary with the degree of ionization, since the charged polymer will interact with the hydrogen ions.If the required electrostatic Gibbs energy for the removal of an equivalent of protons at a given degree of ionization is AGel(a), then pK = pK, + AGel (a) log,, e/RT where KO is characteristic of the ionizing group when electrostatic interactions with other ionizing groups are absent. Thus pH = pK, -log [( 1 - a)/a] + AGel (a) log,, e/RT. (6) pK, can be determined from a plot of pK against a by extrapolating a to zero where AGel = 0. If conformational changes occur as the titration proceeds, then there is a non-electric contribution to the Gibbs energy, AGno,.el, and then The presence of AGnon.el is revealed either by the presence of a maximum or minimum in the plot of pK against a or by the construction of a Henderson-Hasselbalch plot.It was found empiricallyll that a plot of pH against log [( 1 - a)/a] yielded a straight line for most polyelectrolytes. Thus (8) pH = pK,, - n log [( 1 - a)/a] where K,, is the average ionization constant and n is a constant depending on the nature and concentration of the polyacid and the ionic strength of the solution. A conformational transition is indicated by a break in the slope.12 The degree of neutralization, a, was determined from the charge/pH isotherms by making the initial assumption that only one acid group was involved in producing the change between the P.Z.C. and the horizontal region of the curve. This assumption is shown to be a good one in the case of cotton linters and bleached sulphate pulp. The horizontal plateau region was taken as the neutralization point and therefore the degree of dissociation at a pH Y was given by (9) a t Y ) = ( Y ) 0 0 /OF where ahY) is the charge at pH Y and 0: is the charge at the point of neutralization. The charge at pH 7 and 0.1 mol dmP3 sodium chloride was taken as an estimate for1532 THE CELLULOSE/AQUEOUS ELECTROLYTE INTERFACE Fig.6. Variation of pK with a for cotton linters. [NaCl]/mol dm-3 as follows: 0, 0, 10-2; *, 10-1; 0, loo. 0 7 6 E 5 4 3 Fig. 7. Henderson-Hasselbalch plot for cotton linters. {NaCl]/mol dm-3 as follows 0, 10-3; 0, 10-2; 0, 10-1; 0, 100.T. M. HERRINGTON AND B. R. MIDMORE 1533 Fig. 8. Variation of pK with a for bleached sulphate pulp.[NaCl]/mol dm-3 as follows: 0, 10-3; 0, 10-2; 0, 10-1; a, 100. 9 8 7 PH 6 5 a Fig. 9. Henderson-Hasselbalch plot for bleached sulphate pulp. [NaCl]/mol dm-3 as foiiows: 0, 10-3; 0, 10-2; 0, 10-1; 0, 100.1534 THE CELLULOSE/AQUEOUS ELECTROLYTE INTERFACE Saf. Thus for bleached sulphate pulp at pH 6.03 and 0.01 mol dm-3 NaCl: Sor.03) = - 2-59 C 8-1 Saf = -2.90 C 8-l a(6.03) = 0.893 and pK = pH +log [( 1 - a)/a] = 6.03 - 0.92 = 5.1 1. In this way a plot of pK against a was made and extrapolated back to a = 0 to find an estimate of pKo, the pK of the acid at zero dissociation. Using these calculated values for a Henderson-Hasselbalch plots were constructed. The variation of pK with a for cotton linters is shown in fig. 6 and the Henderson-Hasselbalch plot in fig.7. The corresponding plots for bleached sulphate pulp are shown in fig. 8 and 9. DISCUSSION COTTON LINTERS CHARGE/PH ISOTHERMS The form of the charge/pH isotherm in the pH range 2.5-7.0 is characteristic of a monofunctional polyacid. The charge increases as expected with increasing pH as the degree of dissociation of the acid increases. The charge also increases with increasing electrolyte concentration. This can be understood either in terms of the increasing capacitance of the double layer predicted by the differentiation of the Gouy-Chapman equation13 or by the decreasing apparent pK of the acid group. The precise location of the P.Z.C. is not clear, although a value between pH 1.5 and 2.5 is estimated from the isotherms. The 0.1 and 1 .O mol dmP3 isotherms appear to give a slightly higher P.Z.C.than the lower concentrations of electrolyte, and this is probably explained by specific adsorption of chloride at the higher concentrations of sodium chloride. The increase in charge after pH 7.0 may be explained in two ways. It may be that the swelling of the fibre, which occurs in alkali solutions, opens up new acid groups with which the alkali reacts. Alternatively there may be a second weak acid group on the cellulose surface. The behaviour of the 1.0 mol dm-3 NaCl isotherm is anomalous. Below pH 5.0 it lies in the expected order, above the 0.1 mol dm-3 isotherm. However, above pH 5.0 it drops to where it is almost coincident with the 0.01 mol dm-3 isotherm, and at pH 8.5 it levels to a plateau. This behaviour seems to suggest that both swelling and a second weak acid group may be involved.The swelling of the fibre is suppressed by the high electrolyte concentration, which explains the drop in charge of the 1 mol dm-3 isotherm, the effective number of acid groups having been reduced. This would also explain the plateau region above pH 8.5, if the fibre in the lower concentrations of NaCl continued to swell and therefore react after this group had been neutralized. Thus the 1 mol dm-3 isotherm may be considered to be the system in which any complicating effects of swelling are removed. POLYELECTROLYTE PLOTS The straight lines of the Henderson-Hasselbalch plots are indicative of a poly- electrolyte in which no detectable conformational transitions occur. The slopes of unity for the 0.1 and 1.0 mol dm-3 plots indicate that at these NaCl concentrations the system is behaving like a simple monobasic acid, such as acetic, for which n is always unity.The values of n and pK,, at other electrolyte concentrations are given in table 2.T. M. HERRINGTON AND B. R. MIDMORE 1535 Table 2. Values of n and pK,, obtained from the Henderson-Hasselbalch plots for cotton linters and bleached sulphate pulp [NaCl]/mol dmP3 n 0.00 1 0.01 0.1 1 .o cotton linters 1.97 1.44 1 .o 1 .o 5.15 4.20 3.95 3.65 bleached sulphate pulp 0.001 2.05 5.10 0.01 1.92 4.35 0.1 1.58 3.85 1 .o 1.26 3.75 The increase in n with decreasing electrolyte concentration is as expected and reflects the increasing surface potential with decreasing electrolyte concentration. The increase in pK,, is also a reflection of this phenomenon.The form of the pK against a plots is again indicative of a polyelectrolyte in which there is no detectable conformational transition. Such transitions generally show up as local minima in the pK against a plots and it is evident that no such minima exist. The monotonically increasing pK is a reflection of the increasing surface potential with increasing a. In Part 3 of this series it is shown how surface potential can be calculated from these plots. The 1.0 and 0.1 mol dmP3 NaCl systems both show a constant pK, consistent with the very low surface potential caused by the high ionic strength. The plots are extrapolated back to zero dissociation where pK = pKo. From the plots it is clear that the pKo values cluster around a value of ca.4.0; the values are lower, as expected, with increasing activity of NaCl. This is in good agreement with previous estimates obtained by other workers5$ and seems to be consistent with the hypothesis that the alkali adsorption is due to reaction with a carboxylic acid group which has replaced the primary alcohol group in the cellulose structure. BLEACHED SULPHATE PULP CHARGE/~H ISOTHERMS Like the isotherms for cotton linters, the bleached sulphate isotherms are also characteristic of a monofunctional polyacid. However, in this case it is clear that a much higher charge density is involved. This is confirmed by the fact that the neutralization point is achieved at a much higher pH. The precise location of the P.Z.C. is clearer because of the higher charge density and it is estimated to be at pH 1.75f.0.25.There is also clear evidence of charge reversal below pH 1.5. A mechanism for this reversal may be the protonation of the large numbers of oxygen atoms within the cellulose structure. As with the cotton linters, there is a degree of reaction above pH 8 after neutralization has occurred; however, here it is less marked because of the overall high charge density. Again, it is interesting to observe that the 1 mol dm-3 isotherm drops below the 0.1 mol dmP3 isotherm and becomes coincident with that for 0.01 mol dm-3 NaCl. This again is good evidence of the high electrolyte concentration suppressing swelling and thereby reducing the surface charge. The1536 THE CELLULOSE/AQUEOUS ELECTROLYTE INTERFACE e L 0 0 5 1 0 Q Fig.10. Variation of pK with a for various cations on bleached sulphate pulp. A, 0.05 rnol dm-3 BaCI,; 0, 0.05 mol dm-3 CaCI,; 0, 0.1 mol dm-3 NaCl; A, 0.1 mol dm-3 KCl. higher crossover point of pH 6.1, compared with pH 5.0 for cotton linters, also indicates a higher charge density, electronic effects dominating swelling effects until a higher pH. When fig. 5 is studied it is clear that the position concerning the effect on the isotherm of changing the cation is not very straightforward. Perhaps the most striking feature is the calcium isotherm, which after a value of - 2.66 C g-l at pH 5.46 remains constant to within k0.004 C g-l until pH 9.22. This is also the lowest value for So, above pH 6.5 in spite of the high affinity of calcium for carboxy groups.It is suggested that this behaviour is caused by the calcium ions suppressing swelling. This suppression cannot be due simply to the bivalency of the calcium ion, as no such effect is seen with barium. The effect is specific to calcium and probably indicates substantial adsorption of calcium ions into the Stern layer, thereby considerably reducing interlamellar repulsion by lowering ~ 6 . Here the size variations of the hydrated cation are not the dominant factors in determining So,, unlike the silica system studied by Tadros and L~k1ema.l~ They found that at high pH and high salt content, the OH- adsorption increased beyond the surface density of silanol groups. Most of the charge is neutralized by cations penetrating into the pores of the silica-gel structure.The extent of penetration depends on the size of the cation as OH- adsorption increased in the order (C,H,),N+ < Li+ < Na+ < K+ < Cs+. For the cellulose adsorption isotherms it is seen that the OH- adsorption for 0.1 mol dm-3 (C,H,),NCl rises above the potassium isotherm above pH 6.5. Again for silica the charge, for example, at pH 9 for 0.01 mol dm-3 (C,H,), NCl is about one third that of the equivalent sodium chloride concentration, while for bleached sulphate pulp it is approximately the same. This must reflect the relatively large pores in the cellulose fibre compared with those in the silica sol. POLYELECTROLYTE PLOTS In a similar way to the cotton linters, the straight lines of the Henderson-Hasselbalch plots are indicative that again we have a system in which there are no detectableT. M.HERRINGTON AND B. R. MIDMORE 1537 conformational transitions. The larger values for n which lie in the range 1.26-2.05 also indicate greater surface charge density. The pK against a plots for bleached sulphate show a similarity to those for the cotton linters, the pK, values being close to those of the latter. Overall it is clear that the main contributor to the charge is the same as that for cotton linters, namely a carboxy group at the C, position of the glucose molecule. The pK against a plots for the various cations on bleached sulphate pulp are shown in fig. 10. They offer useful ways for comparing the differing affinity of the various cations for the carboxy group. The effect of swelling is eliminated as a,N ( i e .the amount of carboxy groups available) is determined individually for each cation. From the plots the affinity lies in the order Na+ < K+ < Ba2+ < Ca2+. The barium and calcium plots show similarities to the 1 mol dm-3 NaCl plots as they flatten off at a = 0.9. UNBLEACHED SULPHATE PULP The charge/pH isotherm for unbleached sulphate pulp is shown in fig. 4. The charge per gram at pH 7 is four times greater than for bleached sulphate pulp and some 25 times greater than that of cotton linters. There is no tendency to form a horizontal plateau with increasing pH, but on the other hand the charge does not show the rapid increase without limit of a silica The behaviour suggests that several acid groups are contributing to the surface charge. Unbleached sulphate pulp contains a considerable quantity of lignin, so that phenolic and other acid groups may contribute to the surface charge as well as carboxy groups. M. J. Jaycock and J. L. Pearson, J. Appl. Biotechnol., 1975,25, 827. M. B. Donnan, T. W. Healy and P. F. Nelson, Colloids SurJ, 1981, 2, 133. S.M. Neale and W. A. Stnngfellow, Trans. Faraday SOC., 1937, 33, 881. E. Heymann, Trans. Faruhy Soc., 1942,38, 209. G. F. Davidson, J . Text. Inst., 1948, 39, 87. B. R. Midmore, Thesis (University of Reading, 1983). ti M. R. Edelson and J. Hermans, J. Polym. Sci., Part C, 1963, 2, 145. * R. A. Robinson and R. H. Stokes, Electrolyte Solutions (Butterworths, London, 2nd edn, 1959), p. 433. G. H. Bolt, J. Phys. Chem., 1957, 61, 1166. lo G. H. Parkes and P. L. deBruyn, J. Phys. Chem., 1962,66, 967. l1 A. Katchalsky and P. Spitnik, J. Polym. Sci., 1947, 2, 432. l2 A. M. Kotliar and H. Morawetz, J . Am. Chem. Soc., 1955, 77, 3692. l3 J. Lyklema and J. Th. G. Overbeek, J. Colloid Sci., 1961, 16, 595. l4 Th. F. Tadros and J. Lyklema, J. Electroanal. Chem., 1968, 17, 267. (PAPER 3/ 1620)