The effect of the concentration of dispersed particles on the mechanisms of low-frequency dielectric dispersion (LFDD) in colloidal suspensions

There are two mechanisms which are currently used to explain the low-frequency (kHz range) dispersion of the dielectric permittivity of suspensions in electrolyte solutions (LFDD). The first, the surface diffusion mechanism (SDM), associates the LFDD with the diffusion of bound ions along the particle surface caused by the applied electric field. The second, the volume diffusion mechanism (VDM), follows from the generalization for alternating fields of the classical theory of the relaxation effect in electrophoresis and associates the LFDD with the diffusion of free ions in the diffuse double layer. It has been found that VDM is much more strongly dependent on particle concentration than SDM, opening new possibilities for the investigation of each of these two mechanisms separately. The reason is that when the concentration of particles in suspension increases, the characteristic length for the propagation of volume diffusion processes decreases together with the decrease of the free electrolyte volume, whereas the characteristic length for the surface diffusion remains constant. Correspondingly, when particle concentration is raised, the relaxation time of the VDM effect must decrease, whereas it must remain constant for the SDM mechanism. Thus, by varying the concentration of particles in suspension, one can separate the dispersion curves of SDM and VDM. A simple model is elaborated which can be useful to predict the volume fraction dependence of the parameters of LFDD; in particular, its amplitude and critical frequency. The results are compared with experimental data obtained with polymer latex dispersions of volume fractions ranging from 3 to 16%. It is found that the dielectric behaviour (the volume fraction dependence of both the amplitude and critical frequency of LFDD) of the dispersions is reasonably well explained with our model, thus demonstrating that VDM prevails in the systems studied. Experimental data previously found by other authors are also discussed in the light of the model presented. (C) 1998 Elsevier Science B.V. All rights reserved.