Influence of polymer charge on the shear yield stress of silica aggregated with adsorbed cationic polymers

Abstract

Flocs were produced by adding three cationic polymers (10% charge density, 3.0 × 10 5 g/mol molecular weight; 40% charge density, 1.1 × 10 5 g/mol molecular weight; and 100% charge density, 1.2 × 10 5 g/mol molecular weight) to 90 nm diameter silica particles. The shear yield stresses of the consolidated sediment beds from settled and centrifuged flocs were determined via the vane technique. The polymer charge density plays an important role in influencing the shear yield stresses of sediment beds. The shear yield stresses of sediment beds from flocs induced by the 10% charged polymer were observed to increase with an increase in polymer dose, initial solid concentration and background electrolyte concentration at all volume fractions. In comparison, polymer dose has a marginal effect on the shear yield stresses of sediment beds from flocs induced by the 40% and 100% charged polymers. The shear yield stresses of sediments from flocs induced by the 40% charged polymer are independent of salt concentration whereas the addition of salt decreases the shear yield stresses of sediments from flocs induced by the 100% charged polymer. When flocculated at the optimum dose for each polymer (12 mg/g silica for the 10% charged polymer at 0.03 M NaCl, 12 mg/g for 40% and 2 mg/g for 100%), shear yield stress increases as polymer charge increases. The effects observed are related to the flocculation mechanism (bridging, patch attraction or charge neutralisation) and the magnitude of the adhesive force. Comparison of shear and compressive yield stresses show that the network is only slightly weaker in shear than in compression. This is different than many other systems (mainly salt and pH coagulation) which have shear yield stress much less than compressive yield stress. The existing models relating the power law exponent of the volume fraction dependence of the shear yield stress to the network fractal structure are not satisfactory to predict all the experimental behaviour.