Aggregation and Breakup of Colloidal Particle Aggregates in Shear Flow, Studied with Video Microscopy

V.A. Tolpekin, M.H.G. Duits, D. van den Ende, J. Mellema

Research output: Contribution to journalArticleAcademic

80 Citations (Scopus)


We used video microscopy to study the behavior of aggregating suspensions in shear flow. Suspensions consisted of 920 nm diameter silica spheres, dispersed in a methanol/bromoform solvent, to which poly(ethylene glycol) (M = 35.000 g) was added to effect weak particle aggregation. With our solvent mixture, the refractive index of the particles could be closely matched, to allow microscopic observations up to 80 m deep into the suspension. Also the mass density is nearly equal to that of the particles, thus allowing long observation times without problems due to aggregate sedimentation. Particles were visualized via fluorescent molecules incorporated into their cores. Using a fast confocal scanning laser microscope made it possible to characterize the (flowing) aggregates via their contour-area distributions as observed in the focal plane. The aggregation process was monitored from the initial state (just after adding the polymer), until a steady state was reached. The particle volume fraction was chosen at 0.001, to obtain a characteristic aggregation time of a few hundred seconds. On varyiation of polymer concentration, cP (2.2-12.0 g/L), and shear rate, Y (3-6/s), it was observed that the volume-averaged size, Dv, in the steady state became larger with polymer concentration and smaller with shear rate. This demonstrates that the aggregate size is set by a competition between cohesive forces caused by the polymer and rupture forces caused by the flow. Also aggregate size distributions were be measured (semiquantitatively). Together with a description for the internal aggregate structure they allowed a modeling of the complete aggregation curve, from t = 0 up to the steady state. A satisfactory description could be obtained by describing the aggregates as fractal objects, with Df = 2.0, as measured from CSLM images after stopping the flow. In this modeling, the fitted characteristic breakup time was found to increase with cP.
Original languageEnglish
Pages (from-to)2614-2627
Issue number7
Publication statusPublished - 2004


  • IR-59370

Cite this