Particle aspects of precipitative softening: experimental measurement and mathematical modeling of simultaneous precipitation and flocculation

Precipitative coagulation processes (i.e., alum or iron sweep-floc coagulation and lime softening) are used nearly ubiquitously in the treatment of drinking water from surface water sources. Although the benefits of such processes are well known, the ability to predict the particle size distributions from such processes is virtually non-existent. The objective of this research was to improve the quantitative understanding of how particle size distributions change due to simultaneous precipitation and flocculation in water treatment through experimental investigation and mathematical modeling. This study focused on one such process, precipitative softening, as an example system. The experimental research used bench-scale calcium carbonate precipitation experiments under a variety of conditions to elucidate how the particle size distribution changes due to simultaneous precipitation and flocculation and to identify the controlling variables. Independent variables included the saturation ratio, pH, ratio of precipitating ions, initial seed type and concentration, and the mixing intensity. Experiments were performed under conditions of constant and declining solution composition, and particle size distributions were measured using a Coulter Counter. Trends in the experimental results were clear; particle size distributions changed dramatically by nucleation, particle growth and flocculation. The saturation ratio, initial seed type and concentration, and the mixing intensity were identified as the most important variables. Where possible, relationships linking the changes in the particle size distribution with the independent variables were delineated. An existing mathematical model for flocculation was modified to include mathematical expressions describing nucleation and particle growth. Flocculation in the revised model was described using three plausible rate expressions. An underlying hypothesis of the research was that the incorporation of the mechanisms of precipitation into the existing flocculation model would allow the prediction of particle size distributions from softening processes. Model predictions were tested against experimental data to determine which representation of flocculation during precipitation was most appropriate. It was found that at high mixing intensities flocculation during precipitation was most accurately modeled using a size-independent flocculation expression. At lower mixing intensities, the shortrange force model was a better predictor of the particle size distribution. The implications of these findings are discussed.