I1

FIGURE 5.4 Images of aggregates from Buffalo River suspensions at different times of aggregation.

stages described in Figure 5.1. During the initial period of coagulation, aggregates quickly grow and become more convoluted and porous. Later, after about 50 or 60 min, aggregates appear to be more regular, with a smoother surface. This condition corresponds with the later stages described in Figure 5.1, where more loosely bound clusters have been broken off and the aggregates become denser as inner pore spaces become occupied. The period between about 40 and 60 min shows only minor changes, which should correspond with the state indicated around point A in Figure 5.1. For times greater than about 90 min, the aggregate size seems to decrease slightly, again consistent with the later stages described in Figure 5.1.

5.4.1.2 Particle Size and Shape

Geometric information from the image analysis was used to characterize particles. Temporal changes in aggregate size distribution for experiments with latex particles mixed with G = 20 sec-1 and G = 10 sec-1 are shown in Figure 5.5, where frequency of occurrence (number) in each size class is plotted against the aggregate characteristic length (major ellipse axis) using a logarithmic scale. Test results using alum (Experiment Set 2) and polymer (Experiment Set 3) are shown. In both cases the gradual movement of the peak of the distribution toward larger size is clearly seen, indicating that aggregate growth was the primary mechanism during this period (also shown in Figure 5.4). In this period (0 to 30 min) conditions are such that the model of Equation (5.10) is applicable, that is, breakup is negligible. Compared with the results using polymer, alum treatment seems to result in more peaked size distributions, with less spread about the mode. The peak size for the polymer treated experiment showed a more significant increase with time and the peak had smaller magnitudes than with the alum treated test. The higher peak associated with the alum

Log l

FIGURE 5.5 Temporal plot of particle size distribution for latex suspensions mixed at G = 20 sec-1 using alum, and mixed at G = 10 sec-1 using polymer (Poly).

Log l

FIGURE 5.5 Temporal plot of particle size distribution for latex suspensions mixed at G = 20 sec-1 using alum, and mixed at G = 10 sec-1 using polymer (Poly).

treated tests may be related to additional solids added by the alum, and to different particle concentrations used in the two sets of experiments. It is difficult to compare the two suspensions directly since different coagulants were used and the zeta potentials for the two suspensions at charge neutralization were different, as were the mixing speeds. Keeping these differences in mind, the polymer treated coagulation produced larger (and possibly stronger) particles more quickly than coagulation by alum at the charge neutralization stage.

Temporal changes in the peak of the particle size distribution for Buffalo River suspensions mixed with polymer at G = 10 sec-1 are illustrated in Figure 5.6. The peak size gradually increases (until about 50 to 60 min) and then it decreases. The decrease at later times is thought to be due to the breakup effect, which has been described in several previous studies. For example, Williams et al.29 reported a breakup of particle size after reaching a peak for silica particles mixed at various speeds. Their study suggested that smaller G would induce a larger peak and a relatively smaller decrease of peak size over time, compared to a higher mixing rate. A similar observation was reported by Selomulya et al.,30 who found that the average aggregate size (for latex particles) decreased with time after reaching a peak, using a range of shear rates (40 to 80 sec-1).

Further evidence of this type of behavior is seen in Figure 5.7 and Figure 5.8, where temporal changes in median aggregate size and D2 are plotted for both latex (LT) and Buffalo River (BR) suspensions treated with polymer under three different mixing rates. Although not shown here, similar results were found with D3 as with D2. In general, slower mixing produced larger aggregates (Figure 5.7) and lower D2 (Figure 5.8) for both these experiments, and changes in the Buffalo River suspensions were relatively more pronounced. This may be due to higher solids concentration for the Buffalo River suspension, or to the presence of organic material, which was not a factor in the latex tests. In addition, there was greater heterogeneity in aggregate size and shape for Buffalo River suspensions. In both cases there is a gradual increase in size and decrease in D2, followed by the attainment of approximately steady-state

0 0

Post a comment