In shear-induced coagulation processes of colloidal nanoparticles, which are customarily used in most industrial applications, dense clusters of particles are usually formed. The characterization of their size, structure, and distribution is a key factor for the quality of the final product. Light scattering or laser diffraction techniques are still the most common techniques to perform this characterization, because of their speed, excellent statistics and great simplicity. However, a major obstacle in the effective utilization of light scattering is the lack of realistic structural models for dense clusters of spheres which allow one to extract reliable information from scattering data. In addition, the commonly used Rayleigh-Debye-Gans (RDG) theory fails in the case of dense structures and particles with a size comparable to that of the wavelength of the incident radiation. These conditions are the rule rather than the exception in many coagulation processes.

In this work, a methodology is presented that aims at filling this gap. First of all, a tunable fractal dimension Monte-Carlo algorithm is used to generate dense clusters with a desired fractal dimension. Since it is well known that the tunable fractal dimension algorithm cannot generate clusters with fractal dimension larger than 2.5 (Thouy and Jullien, 1994), a new procedure has been developed to create cluster with fractal dimension up to 3. This procedure starts from clusters with a fractal dimension equal to 2.5 and uses a Voronoi tessellation of the space occupied by the cluster to make it progressively denser by moving particles initially located on its surface to its interior, in order to fill its empty space. In this manner, the entire range of cluster fractal dimension encountered in typical aggregation processes, ranging from 1.8 to 3, is covered. The cluster structure is than characterized by means of its pair correlation function. The pair correlation function is finally used to compute scattering properties of cluster applying a mean-field version of the T-Matrix theory, proposed by Botet et al.(1997), which can provide reliable scattering behavior of dense clusters with arbitrary primary particle size. This methodology is applied to the analysis of shear-induced coagulation experiments of polymer colloids, and it is shown that only by using the mean field T-matrix theory the experimental results can be correctly interpreted.