Colloid Stability in Aqueous and Won-Aqueous MediaINTRODUCTORY PAPERBY J. TH. G. OVERBEERvan’t Hoff Laboratory, University of Utrecht, The NetherlandsReceiued 28th July, 1966Colloid stability is determined by the interaction between particles during a (Brownian) en-counter. The forces playing a role in encounters (van der Waals forces, double-layer interaction,interaction between adsorbed molecules, such as surfactants, polymers and small molecules) arediscussed with emphasis on unsolved problems. It is pointed out in which of the papers of thisDiscussion the different problenls are treated.GENERAL ASPECT OF COLLOID STABILITYSince colloidal particles dispersed in a liquid are always subject to Brownianmotion, encounters between the particles occur frequently and the fate of the colloidaldispersion is determined by the interaction between particles during such an en-counter.When attraction predominates the particles will adhere more or lesspermanently and the dispersion will coagulate ; in the case of predominant repulsionthe system will remain in the dispersed state.Van der Waals forces are always present, whatever the composition of thesystem, and it has been proved that such forces always result in attraction betweenparticles of the same material. Therefore, a colloidal suspension can only be stableif a sufficiently strong repulsion counteracts the van der Waals attraction. Thisrepulsion may be based on the interaction of electrical double layers-mainly butnot exclusively important in aqueous systems-or on the interaction of layers ofadsorbed uncharged molecules, including the solvent molecules themselves, amechanism which may act in any solvent.The interaction between particles and adsorbed molecules does not necessarilylead to repulsion.Especially, if the adsorbed molecules are polymeric, one mole-cule may be attached to two particles and thus form a bridge between the particles,enhancing the attraction rather than the repulsion, and leading to sensitizedflocculation.Qualitatively our picture of colloid stability is well established and for severalof the factors entering the picture quantitative or semi-quantitative theories exist.At present we are at a stage of refinement of our understanding and control ofcolloid stability, but much refinement has still to be done before the situation issatisfactory.In this introductory paper, I shall treat the main interaction forcesone by one and try to indicate in each case what critical experiments or improvementsin the theory should be made.TeZak’s contribution to this Discussion, in which he gives a systematic descriptionof all states between the homogeneous solution and the macrocrystal, with a strongemphasis on the variability of the phenomena, has a wider scope than most of theother contributions. It includes and rather stresses such aspects as complex forma-tion, nucleation and crystal growth and reserves only a modest place for colloidalstability in the more restricted sense.8 INTRODUCTORY PAPBRTHE ELECTRICAL DOUBLE LAYERThe structure of the electrical double layer is relatively well established, notin the sense that the value of the surface charge or the surface potential can be pre-dicted from first principles, but that the relation between charge and potential andthe spatial extension of the double layer and the repulsion between overlappingdouble layers are fairly well known-fairly well, but possibly not well enough.Colloid stability is governed by the difference between repulsion and attractionand a relatively small error in one of the two may make our estimate of the stabilitygrossly wrong.The relation between charge and potential difference between the two phasesis dealt with by Levine and Bell, who will discuss a number of refinements to thePoisson-Boltzmann equation, such as the influence of ion-size, dielectric saturation,discreteness of charge, cavity potential, etc.They will also show the influence ofthese refinements on the interaction of double layers and on colloid stability.TeEak lays great stress ori the interaction between individual ions, as opposed tothe collective behaviour of ions in double layers. Lyklema has obtained data oncharge and potential difference at the AgP/water interface at elevated temperatures.Since he finds a more simple, more " Gouy-like " structure of the double layer anda correspondingly simpler behaviour in coagulation, high temperature may wellbecome an important tool in double-layer research.Instead of the potential difference between the two phases the zeta-potential isoften used.It has the advantages that it can be determined in practically all casesof interest, at practically any electrolyte concentration, and that it gives informationon the diffuse part of the double layer. The disadvantage is that the theoreticalrelation between zeta-potential and electrokinetics is far from simple, and as yetthere is no good understanding of where the slipping plane (if it is a plane) is situatedwith respect to the phase boundary. Nevertheless in at least eight of the paperspresented at this discussion (Watillon and Mrs. Joseph-Petit, Hall, Ottewill andShaw, Matijevic, Kratohvil and Stryker, McGown and Parfitt, Romo, Tehk andMicale, Lui and Zettlemoyer) the zeta-potential is used for information on the doublelayer, three being about suspensions in non-aqueous media.In the papers byRomo and by Micale, Lui and Zettlemoyer the influence of trace amounts of wateron the zeta-potential is stressed.The above information deals mainly with single double layers. Direct experi-mental information on the interaction of double layers is much scarcer. I wouldlike to cite the classical work of Bergmann, Low-Beer and Zocher 1 on Schillerlayers, and mention the modern work on black soap films (see Overbeek 2) of whicha good example obtained with a new technique is presented here by Mysels andJones, who find a quite acceptable agreement with the theory for the interaction ofdiffuse double layers.To close my remarks about the double layer I should say that accurate data ofcharge and potential difference for a variety of interfaces especially at low electro-lyte concentrations are still greatly needed and that experiments, in which the doublelayer repulsion is determined as free as possible from other interactions, such asthose mentioned above on Schiller layers or soap films, are extremely valuable.VAN DER WAALS FORCESSince Kallmann and Willstaetter 3 suggested that van der WaaIs forces are re-sponsible for the attraction between colloid particles, much work, both theoreticaland experimental, has been done on these forces.The earlier theories (Tomlinson,J . TH. G . OVERBBEK 9Rradley,s de BoeQ Hamaker 7) treated the forces between atom or molecules asstrictly additive and did not take into account that the forces might be modified bypassage through a dense medium.Lifshitz,s basing his treatment on electromagneticfluctuations in a dense medium, derived an exact expression for the attraction be-tween macroscopic objects (consisting of many atoms or molecules) in terms of themacroscopic dielectric constant and dielectric loss factor, which had to be knownover the complete frequency range of dispersion. This treatment has been extended(Dzyaloshinskii, Lifshitz and Pitaevskii 9) to include even the interaction betweentwo fiat plates of different composition separated by a dense medium. At lowdensities, Lifshitz’ treatment is equivalent to the earlier treatment based on addi-tivity of the forces between pairs of molecules.Unfortunately, the necessary dielectric data are not yet known with sufficientat;curacy to be able to apply Lifshitz’ theory directly.Only for the retarded attrac-tion forces, effective at “ large ” distances (e.g., > 1000 A) between the objects, thesituation is more favourable, since in this case only the “ static ” dielectric con-stant (square of the refractive index extrapolated to long wavelengths) has to beknown. But, although the retarded force may play a role in some colloidal phe-nomena, the non-retarded force is much more important. Independent measure-ments of the van der Waals forces of some accuracy exist only in the retarded region,whereas measurements in the non-retarded range are still highly inaccurate (for areview sce Overbeek and Van Silfhout,lO Van Silfhout 11).In the present situationthe best source of information on van der Waals forces are flocculation experi-ments and experiments on thin liquid films, but to a certain extent this is beggingthe question.In this Discussion the papers by Watillon and Mrs. Joseph-Petit and by Ottewilland Shaw show how the stability of polystyrene latex dispersions can be used to derivevalues for the van der Waals attraction. Use is made of the variation of the stabilitywith particle size, particle charge and electrolyte content, but the resulting valuesfor the van der Waals attraction are not yet internally consistent. The followingfactors should be considered in the search for an explanation of this discrepancy.(i) The electrostatic repulsion may have been incorrectly estimated.The in-fluence of this uncertainty can be minimized by comparing sols with different particksizes, but with identical surface charge or potential (at flocculation).(ii) An incorrect estimate of the distance between the phase boundary (wherethe van der Waals constant changes its value) and the plane of the surface charge(from where the distance of repulsion is calculated). Quite often these planes aresupposed to coincide, but this is almost certainly incorrect. It might be advisableto introduce the distance between these planes as a parameter in the theory.(iii) Neglecting the influence of transmission of the van der Wads force througha medium may be a more serious error than is usually assumed. A first estimatecould be obtained by using Lifshitz’ theory with relatively simple models for thedispersion curves.In this connection it is also important to develop Lifshitz’method for more complicated geornetrics (spheres, spheres surrounded by layersof different composition).Further work on the direct measurement of non-retarded van der Waals forces,if possible between objects in a dense medium, is highly desirable. Optical datashould be obtained especially in the far ultra-violet to be used in Lifshitz’ equations.Continued work on the stability of isodispersed sols with careful control of the surfacecharge and wide variation of conditions is desirable. Work on thin films (equi-librium thickness under pressure, light scattering, rate of drainage, etc.) ought tolead to important information10 INTRODUCTORY PAPERPROTECTION AGAINST COAGULATION BY ADSORPTION OF NEUTRALMOLECULESProtective action of hydrophilic colloids on hydrophobic ones has been recog-nized early in the history of colloid science.It was interpreted as the envelopmentof hydrophobic particles by a layer of the hydrophilic colloid making the particlesas stable against coagulation as the hydrophilic colloid. The particles of a goldsol protected by gelatin behave as ‘‘ gelatin particles with a golden heart ”.Later, it was found that the phenomenon is not restricted to aqueous systemsand that protective agents need not necessarily be polymeric. The work of Vander Waarden, Mackor and van der Waals 12-14 on the stabilization of carbon blacksuspensions in a liquid hydrocarbon by adsorbed mixed aromatic-aliphatic mole-cules of modest mass is a good example.In non-aqueous and in particular lion-polar systems repulsion between particlesby electric charge is usually of minor importance.Non-aqueous suspensions,therefore, have to be stabilized by some variant of protective action. Given thetremendous technical importance of non-aqueous suspensions (paints, inks, pig-ments for synthetic polymers, etc.) it is no wonder that in this Discussion muchattention is given to protected colloid systems. Crow1 and Malati and Walbridgeand Waters give some pure examples of stabilization of suspensions in hydrocarbonsby polymeric surfactants. Clayfield and Lumb treat the case of separating carbonfrom metal by a polymeric surfactant.In McGown and Parfitt’s work on thedispersion of rutile in p-xylene by aerosol OT obviously electric charge plays a role.There is no indication that the simple presence of an adsorbed layer of unchargedsurfactant molecules contributes to the protection. Glazman stresses the fact ofprotection in aqueous media by non-ionic surfactants of low and intermediate molarmass. Mrs. Taylor and Haydon point out the parallel between the stability of thinhydrocarbon films in water and colloid stability in non-aqueous media. They con-firm the essential correctness of Mackor and van der Waals ideas, which predicta very steep repulsion for a film thickness equal to twice the length of the stabilizingchains.It is striking that in so many of these cases polymers are used as protectiveagents.They are obviously favourable in several respects. Their standard freeenergy of adsorption is proportionally larger than that of a small molecule of similarchemical nature. Therefore they are more easily adsorbed. Simply on account oftheir size they are expected to form thicker layers, keeping the particles more widelyseparated. Nevertheless, they should not be adsorbed too strongly, ix., at toomany points, because then the layer formed will be thin and the protection inadequate.The detailed chemistry of the adsorption process becomes a point of majorinterest in this field as pointed out by Slater and Kitchener in their paper on theclosely related field of flocculation caused by small amounts of polymers.The theory of protective action is still in a rather primitive state.Mackor andvan der Waals 14 introduced the notion “ entropic repulsion ”, because the loss oftranslafional freedom of the “ wriggling tails ” of the adsorbed molecules when thelayers interpenetrate, leads to a loss of entropy and thus to repulsion. This theoryhas only been worked out for very simple cases.We should try to reach a more complete force-distance relation, using morecomplete statistics for the interpenetrating chains, taking into account that polymermolecules may have several points of attachment to the particles and that thc ad-sorption density itself decreases when the layers interpenetrate. Such a theorywould also be very helpful for a more complete understanding of sensitizationJ .TH. G . OVERBEEK 11SENSITIZATION OF LYOPHOBIC SUSPENSIONS BY SMALL AMOUNTS OFPOLYMERSA polymer molecule, having many points of possible attachment to a surface,can in principle as easily form loops between adsorption sites on one particle as formbridges between sites on different particles. In the last mentioned case, agglomer-ates are formed and flocculation occurs. A low concentration of polymer and a highconcentration of particles will promote this sensitization.Sensitized flocculation cannot simply be treated as an interplay between at-tractive and repulsive force between the particles. The elementary step in thiscase is not the approach of two particles (which, in a stable sol, may be prohibitedby an activation energy of 30 kT or more), but the formation of a single adsorptioncontact (which, even when the adsorbed group is charged, would only have anactivation energy of a few times kT).Once the polymer bridge is made betweentwo particles with single adsorption contacts on each of them, it can be strengthenedand made virtually irreversible by occupying more adsorption sites on each particle.This lack of reversibility does not permit the particles to find the closest packingby sliding along each other and explains the openness of the flocculate.It is evident that the quantitative theory of sensitized flocculation should havemany points in common with the theory of protective action, including consider-ations on the specificity and on the thermodynamics of adsorption of polymer mole-cules.However, as indicated above, for sensitization also kinetic factors andenergies of activation, both for adsorption and desorption, are essential.Several papers in this Discussion deal with sensitized flocculation. In the firstplace we should mention the paper by La Mer, who stresses the different natureof flocculation by polyelectrolytes and that by small ions, and offers the rate ofrefiltration through the flocculate as a tool to distinguish the two types of aggrega-tion. La Mer also proposes to keep the terms coagulation and flocculation sharplyseparated, the former for precipitation by electrolytes (forming dense flocs), thelatter for precipitation by polyelectrolytes (open flocs).It might be useful to spendsome time of the Discussion on this point of nomenclature.Slater and Kitchener also discuss flocculation of aqueous suspensions by poly-mers. Since their theoretical treatment of the phenomenon differs from that byLa Mer in essential points, we may expect an interesting discussion. La Mer andSlater and Kitchener both stress the importance of the specific chemical aspectsin the interactions leading to sensitization.In La Mer’s paper, in that of Hall on the interaction of clay with hydrolyzedaluminium solutions and, although perhaps somewhat less clearly, in the paper byMatijevic, Kratohvil and Stryker, we see the possibility that hydrolyzed polyvalentions form polymers and that charge reversal and flocculation by ‘‘ polyvalent ions ”,as already pointed out by Troelstra and Kruyt,ls, 16 is more akin to sensitizationthan to flocculation by simple compression of the double layer.THE ROLE OF THE SOLVENT AS A PROTECTIVE AGENTUndoubtedly, the structure of a liquid near an interface deviates from its struc-ture in bulk.There is no doubt either that the special structure of water near theinterface with particles dispersed in it will affect their stability by influencing thestructure of the electrical double layer and possibly even by modifying the van derWaals forces. But it is further conceivable that this layer of modified solventaffects the stability in a more direct way, e.g., by increasing the viscosity in theaeighbourhood of the particles or by acting as a protective, impenetrabfe layer12 INTRODUCTORY PAPERAn example of such an influence of small neutral molecules can be found in ahitherto unpublished observation by Mackor.During his work on the influenceof acetone on the electrical double layer at the interface AgI/water 17 he found thatacetone changed the coagulation of AgI by electrolyte in a qualitative way. IfAgI is coagulated in water by the addition of simple salts it has a rather voluminousopen structure and is not easily peptized by washing away the precipitant electro-lyte. In the presence of 5-98 % (vol/vol) of acetone the coagulate is more compactand sandy aiid can be easily and completely repeptized by dilution, which suggeststhat acetone is firmly bound to the surface of the Agl particles and prevents an actualAgl-AgT contact.Whether such an influence of the solvent itself, e.g., of water, occurs more or lessgenerally is still an open question.The arguments for such a role of the solventderive on the one hand from the lack of agreement between experiments and atheory based exclusively on van der Waals forces and double layer interaction(" there must be some additional factor '7, and on the other hand on determinationsof properties of the solvent near an interface such as increased viscosity, nuclearmagnetic resonance, certain interpretations of electrokinetics, etc. Unfortunately,the whole situation is still somewhat ambiguous. The theories of van der Waals+double layer interactions may in themselves be inaccurate. The " additionalfactor " may be just a necessary refinement of these theories.Stigter 18 has calculated from the viscosity of micellar solutions that there isnot a layer of increased viscosity around each micelle.Lyklema, Scholten andivysels 19 concluded from the rates of drainage of soap films that the whole innercore of a soap film has the same viscosity as bulk water. On the other hand,Derjaguin 20 has given examples of changes of viscosity of liquids near interfacesto a very great depth and he pointed out that the thickness of black soap films 21frequently remains thicker than can be explained by double layer repulsion alone.In this Discussion the paper by Derjaguin on the effect of lyophile surfaces onthe properties of boundary liquid films and that by Johnson, Eecchini, Smith,Clifford and Pethica on the stability of polyvinylacetate sols and on the nuclearmagnetic resonance of water in these sols relate to this problem and may hopefullyinitiate an interesting exchange of ideas.The most desirable advance in this field would be a theory which explains theprotective action of the solvent in a mechanistic way and on the other hand moreaccurate experiments connecting stability with " structure " or " adsorption " ofthe solvent.THIN FILMSIt will be clear from some of the above remarks that thin detergent films are animportant source of information for several of the finer points of colloid theories.The combination of a small thickness determined by the interactions, in which weare interested, with a relatively largz area, which makes optical and other observa-tions easy and the applicability of straightforward thermodynamics, are advantages.It is fortunate that several papers in this Discussion on colloid stability are devotedto thin film work.The papers by Mrs. Taylor and Haydon, by Mysels and Jones,and by Derjaguin have been cited earlier. I should also mention here Vrij's paperon fluctuations in the thickness of soap films which may lead to rupture (the equivalentof coalescence in emulsions) and the paper by Corkill, Clunie and Goodman whoobtain consistent values for the thickness of the films using a variety of techniquesand find from their measurements that a water layer of about 20A thickness istenaciously held between the two surfactant layersJ .TH. G. OVERBEEK 13CONCLUSIONThe general framework in which colloid stability can be interpreted appearsto be well established. There are sufficient uncertainties in the quantitative aspectsof the theory and in the detailed interpretation of certain observations, to expecta lively and fruitful discussion.1 Bergmann, Low-Beer and Zocher, 2. physik. Chem. A , 1938, 181, 301.2 Overbeek, J. Physic. Chem., 1960, 64, 11 78.3 Kallmann and Willstaetter, Naturwiss., 1932, 20, 952.4Tomlhson, Phil. Mag., 1928, 6,695.5 Bradley, Phil. Mag., 1932, 13, 853.6 de Boer, Trans. Furaduy SOC., 1936, 32, 10.7 Hamaker, Physica, 1937, 4, 1058.SLifshitz, Doklady Akad. Nauk. S.S.S.R., 1954, 97, 643; 1955, 100, 879; Zhw. Eks. Teor.9 Dzyaloshinskii, Lifshitz and Pitaevskii, Adu. Physics, 1961, 10, 165.10 Overbeek and Van Silf hout, Proc. Symp. InterinoZeculur Forces (Pontifical Academy of Sciences,11 Van Silfhout, Proc. Kon. Neder. Akud. Wetens. B, 1966, 69, 501, 516, 532.12 Van der Waarden, J. Colloid Sci., 1950, 5, 317.13 Mackor, J. Colloid Sci., 1951, 6,492.14 Mackor and van der Waals, J. Colloid Sci., 1952,7,535.15 Troelstra, Thesis (Utrecht, 1941), p. 81.l6 Troelstra and Kruyt, Kulloid Beilzefte, 1943, 54, 251.l7 Mackor, Rec. trav. chini., 1951, 70, 663, 747, 763, 841.l8 Stigter, ZVth Znt. Congr. Surface-Active Substances (Brussels, 1964), no. B IV-2.l9 Lyklema, Scholten and Mysels, J. Physic. Chem., 1965, 69, 116.20Derjagui1-1, Trans. ConJ Colloid Chem. (ed. Acad. Sci. Ukrain, S.S.S.R., Kiev, 1952), p. 26.Karasev and Derjaguin, Colloid J. (Russ.), 1953, 15, 365.21 Derjaguin and Titijevskaja, Disc. Furuday SOC., 1954, 18, 27.Fiz., 1955, 29, 94 ; Soviet Physics JETP, 1956, 2, 73.Rome, April, 18-24, 1966)