Break-up and Re-formation of Floes Formed by Hydrolyzing Coagulants and Polymeric Flocculants

摘要

The results presented here highlight very significant aspects of floc formation, breakage and re-growth. Under the conditions used, the metal-based coagulants showed a more rapid onset of flocculation, but significantly smaller floes than with the polyelectrolytes. The lag time in the latter case is due to the relatively slow adsorption of these materials on the clay particles. With the dilute suspensions used and the low polymer concentrations it is expected that adsorption of sufficient polymer to give particle destabilization can be quite long compared to the particle collision rate (Gregory, 1988). The low molecular weight polyDADMAC shows a shorter lag time, which may be a result of its more rapid diffusion than the much larger molecules of the other two polymers. For the hydrolyzing coagulants, hydroxide precipitation occurs quite rapidly and so flocculation begins soon after coagulant dosing. It is noticeable that the pre-hydrolyzed products show an earlier onset of flocculation than with alum and ferric sulphate, most likely because of a more rapid precipitation. The experiments revealed that iron-based coagulants give larger floes than those based on aluminium. The behaviour of iron (III) is similar to that of aluminium in that sweep flocculation by the hydroxide is fast and dominates at around neutral pH. Because of the much lower solubility of ferric hydroxide, rather more precipitate would be expected than with alum for equimolar dosages (as in the present case). For both aluminium and iron, the pre-hydrolyzed product gave significantly increased floc size, but only marginal differences in residual turbidity before floc breakage. However, the re-formed floes showed better removal by sedimentation. The similar form of the FI curves in Figures 1 and 2 strongly suggests that floes are of broadly similar nature for all the metal-based coagulants, but are significantly stronger with the pre-hydrolyzed products. The underlying mechanisms are not well understood and it is unlikely that simple charge effects can explain the results. The polymeric flocculants gave significant differences in floc size, with polyDADMAC giving larger FI values. This finding is difficult to explain in terms of the different molecular weights, since it is expected that higher molecular weight polymers should give stronger floes, either as a result of polymer bridging or 'electrostatic patch' effects (Gregory, 1996). For the dilute suspensions used here and the relatively low particle collision rates, it is expected that adsorbed polymers would adopt a rather flat configuration, which makes bridging unlikely. It has been shown (Kam and Gregory, 2001) that high charge density poly electrolytes are more effective in removing humic substances from water. Since the suspensions used are stabilized by humic acid and polyDADMAC has a much higher charge density than Zetag 64, charge effects may partly explain the observed behaviour. The fact that an anionic polyelectrolyte (Magnafloc 156) gives similar results is surprising. The high level of calcium in the tap water used may play a significant part. Although there have been previous reports of the irreversible nature of floc breakage with hydrolysing coagulants, the reasons remain unclear. Irreversible breakage suggests that chemical bonds are broken. When particle interaction is of a physical nature (such as van der Waals or electrostatic attraction) there is no obvious reason why aggregates should not re-form after breakage. With polymeric flocculants, irreversible floc breakage is well known, but it is usually associated with bridging interactions (Ditter et al, 1982). High shear rates (especially with turbulence) may cause scission of polymer chains (Horn and Merrill, 1984) and adsorbed polymer could adopt a more flat configuration during the breakage phase. However, these considerations do not apply when charge neutralization or 'electrostatic patch' effects are responsible for flocculation, so tha