The purpose of this study was to gain insight in the factors determining the stability of hydrophobic sols in the presence of polymers, with the emphasis on the destabilisation of sols by polymers and the role played by salts therein.
In chapter 1. the practical importance of polymer stabilisation and destabilisation is shown by several examples, a.o. in industrial applications, in water purification and soil structure improvement. Thereafter the choice of the PVA-AgI system as a model for this study was explained. PVA has a simple structure and is uncharged and its concentration in solution may be readily determined. This is important for adsorption measurements. AgI provides a good model for the dispersed phase: the properties of the electrical double layer on AgI in the presence of salts and low molecular weight organic substances have been investigated extensively and the specific surface area can be determined easily. Moreover, with a combination of PVA and AgI one has the advantage of being able to acquire information on the properties of the first layer on the surface by a comparison with known data on the butanol-AgI and ethylene glycol-AgI systems.
The characterisation of the materials used is described in chapter 2. The specific surface area of AgI was determined by three independent methods. The results of these methods agreed well with each other. The average radius of the AgI particles turned out to be about 500 Å. From viscosimetric measurements on PVA solutions the molecular weights and configurational parameters of PVA, such as the radii of gyration, the length of a statistical chain element and the linear expansion factors were determined. In addition, it was shown that the PVA used is essentially uncharged.
In chapter 3. the measurement of the amount of PVA adsorbed per ml is treated. The adsorption isotherm shows a pronounced high affinity character. The maximum amount adsorbed is 1-1.5 mg/m 2
, depending on the molecular weight and the degree of hydrolysis of the PVA. The maximum adsorption increases somewhat with increasing molecular weight; for PVA with 12% of acetate groups it is distinctly higher than for PVA which is nearly completely hydrolysed. At maximum adsorption one fourth of the segments at most can be in contact with the surface; the remaining parts of the molecule protrude into the solution in the form of loops and tails. From measurements of the adsorption as a function of time and from 'two- step' adsorption experiments it could be deduced that the adsorption of segments is reversible. However, desorption of whole polymer molecules is not measurable.
In chapter 4. measurements are described to obtain the layer thickness and the coverage in the first layer on the surface by PVA. From protection measurements qualitative information was obtained about the layer thickness. The protective power appeared to be slightly dependent on the molecular weight and to depend somewhat more strongly on the degree of hydrolysis of the PVA. The thickness of the adsorbed layer was viscosimetrically determined as a function of the amount adsorbed. The maximum layer thickness is about 100 Å. By measuring the electrophoretic mobility of polymer covered particles the layer thickness was likewise estimated. These results are in good agreement with the viscosimetric results.
The coverage in the first layer on the surface was estimated from the shift of the point of zero charge and from the change in the surface charge on adsorption of polymer, in comparison with the same properties of AgI in the presence of butanol and ethylene glycol. A reasonable estimation for the percentage of the surface which is occupied by PVA turned out to be 70% for amounts adsorbed of more than half the maximum.
With the help of these data the distribution of segments in the adsorbed layer could be obtained. For amounts adsorbed between 0.5 and 1.0 mg/m 2
a HOEVE distribution applies. In the first layer on the surface the polymer volume fraction is about 70%. At a distance equal to the thickness of the first layer a discontinuity occurs, the volume fraction dropping to 56 %, and in the remaining part of the adsorbed layer the segment distribution is exponential. If more than 1 mg/m 2
is adsorbed possible end effects occur: due to the presence of long tails at the ends of a polymer molecule the thickness increases more strongly with the amount adsorbed than predicted from the HOEVE distribution.
The model for the segment distribution is somewhat oversimplified: it appeared to be impossible to account for the differences between different molecular weights at a given amount adsorbed.
Results with respect to the flocculation of AgI by PVA have been given in chapter 5. Flocculation was found to be optimal if a special method is used for the mixing of PVA and Agl. Most efficient flocculation is obtained if a given volume of sol with uncovered particles is added to an equal volume of a sol with nearly completely covered particles. This phenomenon could be easily explained on the bridging model: flocculation occurs because loops of the adsorbed layer of one particle attach to the other. In this way a network of AgI particles interconnected by polymer bridges is formed. For the explanation of the efficiency of the way of mixing irreversibility of the adsorption of the polymer molecules is essential.
Another important condition for efficient flocculation is the presence of a small amount of electrolyte. On these grounds the flocculation should be referred to as sensitisation. The minimum salt concentrations which are needed for flocculation are in the ratio of about 100:10:1 for salts with univalent, bivalent and trivalent counterions, respectively. Critical flocculation concentrations measured after a fixed time of flocculation were found to depend on the sol concentration. From measurements of the initial rate of flocculation, and from experiments in which the flocculation time was adjusted to the sol concentration, it was shown that this dependence on the salt concentration has a kinetic origin. The flocculation by bridging was found to be a bimolecular process.
The critical flocculation concentrations were found to depend only slightly on the molecular weight of the PVA. For a PVA with a higher acetate content the amount of electrolyte needed was found to be significantly lower.
In chapter 6. an attempt has been made to interpret the flocculation theoretically. To that order the free energy of interaction between a covered and an uncovered particle has been calculated. On account of the complicated nature of the problem only an approach for flat surfaces has been considered.
In addition to the VAN DER WAALS attraction and the double layer repulsion the contribution to the free energy of interaction due to the adsorbed polymer has to be calculated. This contribution was formally split up in two terms, the first being the adsorption attraction due to the gain in free energy on account of the adsorption of segments on the second particle. The second term is the configurational repulsion which is caused by the entropy loss if a loop becomes two bridges by the adsorption of the middle segments of the loop. The fundamental assumption used to evaluate these two terms is that the number of segments which, at a given interparticle separation H
, adsorbs on the second particle equals the number of segments which, in the absence of the second particle, would lie beyond a distance H
from the surface. Using the theories of HOEVE and HESSELINK and the distribution of segments derived in chapter 4. these two polymer contributions to the free energy of interaction could be obtained.
It was found that the VAN DER WAALS attraction is negligibly small in comparison to the other terms. The total free energy of interaction has the following characteristics. At small salt concentrations a maximum occurs at large distances due to the double layer repulsion, whilst at distances of some tens of Ångströms a minimum is present, sufficiently deep for irreversible flocculation. The function of salt is to suppress the maximum at large distances by partial compression of the double layer, so that the particles can approach each other to a distance corresponding to the minimum in the free energy. The system will then flocculate.
Although the magnitude of the polymer contribution, especially at small distances, is somewhat doubtful on account of various approximations made, the theory does give a good explanation for the amount of salt, with ions of different valencies, which is needed for flocculation. The theoretical predictions with respect to the effect of the amount of adsorbed polymer agree also with the experimental observations. From this it follows that the theory is essentially correct. Indications were obtained that a theory which is applicable to spherical particles would agree even better with the experiments. The development of such a theory would be a promising next step.
In conclusion, this study firmly establishes the bridging model for flocculation by polymer. It appeared possible to interpret several aspects quantitatively. Especially the function of indifferent electrolytes emerged clearly.