Technically speaking

Before resorting to new technologies for meeting more stringent standards, close attention should be paid to maximising existing processes. While it might be more fashionable to introduce membranes, resin-based adsorption such as magnetic ion exchange (Miex) or ozone and granular activated carbon (GAC), coagulation-based processes have undergone significant changes to achieve lower treated water turbidity and total organic carbon (TOC) removal to meet new regulatory requirements. While there is considerable interest in enhanced coagulation for TOC removal, it is also important to recognise effective turbidity removal is essential and cannot be compromised. This broader view of the coagulation process is presented to illustrate the importance of identifying coagulant adjustments, including changes in pH and coagulant dose, use of alternative coagulant chemicals and changes in the order of chemical addition.

Coagulation is essential for successful treatment with any clarification and rapid sand filtration. In fact, a 1989 survey of filtration plants around the US indicated the coagulation process was more critical than physical attributes for meeting filtration goals for particle removal. Since then the regulatory limit for filtered turbidity has been reduced from 1 NTU to 0.3 NTU and many utilities have a goal for maintaining filtered turbidity at levels less than 0.1 NTU. These objectives increase the importance of understanding and characterising coagulation as goals for both particulate and organic removal have changed. Coagulation has historically played a role in the removal of colour that is associated with natural organic material (NOM) at a number of locations. In the US, EPA rules on stage one disinfectants/disinfection byproducts enhanced removal requirements for total organic carbon (TOC) removal, making coagulation a multi-objective process in which stringent goals for both turbidity and TOC removal must now be met.

Process parameters

Several aspects of the coagulation process need to be examined to meet new multiple-objective treatment goals for turbidity and organic removals. Key process inputs for effective coagulation are coagulant dose and pH. Normally, pH is controlled by the addition of the coagulant and its interaction with alkalinity in the raw water, but more independent control of pH through separate addition of acids or bases can be considered, in accordance to water quality conditions and chemical characteristics, together with their seasonal variations such as temperature changes.

General characterisations of coagulant response: Bench-scale test results can be graphically interpreted in multiple-parameter 3D plotting and contour plotting formats, as shown in Figure 1. A 3D format provides useful general insight into the relationship between pH and coagulant dose as the key treatment input parameters, and settled turbidity as the selected treatment output parameter. In addition, this format allows the data fit to be scrutinised to determine how well it represents the data trends.

The contour plot provides a mapped format of the relationships between pH and coagulant dose as input parameters, and shows settled turbidity as contour levels. This type of visualisation allows regions of best performance to be more clearly identified.

Evaluation of coagulation as a multiple-goal process: The trends for the SJWD plant in South Carolina are representative of observations for particulate removal by alum at a number of other locations, which have been assessed by these methods. Effective conditions for turbidity removal by settling tend to overlap a region where the most effective DOC removal is also being achieved.

This condition of effective floc formation is a critical aspect of coagulation for meeting future regulations and goals because conditions for effective filtration tend to occur within the general region of efficient removal by settling. The precise optima may be slightly different but will not occur in regions of poor floc formation. Adjustments that attempt to optimise DOC removal, while failing to achieve effective particle removal, will compromise performance for meeting new turbidity removal goals.

Low alkalinity conditions in raw water that require the addition of a pH adjusting chemical in conjunction with the addition of alum, ferric sulfate or ferric chloride often result in the addition of at least two chemicals during coagulation. The decisions regarding the sequence of chemical addition can have a significant effect on coagulation, as shown in bench-scale results that were obtained from the SJWD plant during cold-water testing with alum at a temperature of 8°C. When alum was added prior to the addition of caustic soda, a 3min contact period was provided before the caustic soda was added. This simulated time that was available in piping and rapid mix before caustic soda was added to the rapid mix effluent in the full-scale plant.

When alum is added prior to caustic soda, the optimum condition for turbidity removal was at a lower pH and with a lower alum dose. As is a common practice, the plant had been adding caustic soda prior to alum and had been experiencing difficulty with plant performance in cold water. The applicability of
bench-scale results was confirmed when a change in the order of addition was made at full-scale operation. This change improved turbidity removal and also increased the capability for removal of TOC by allowing effective floc formation to occur at a lower pH. This approach offered an effective alternative for achieving an overlap of optimum conditions for organic and particulate removal for the cold water conditions. It should be noted there are conditions under which the type of pH adjusting chemicals can cause shifts in coagulation properties as well.

Alternative clarification processes

Results from cold water jar testing with alum at Pawtucket, Rhode Island, showed a degree of floc formation could occur in the region of a pH of 6 and an alum dose of 12 mg/l. Initial water quality conditions at the time of testing can be characterised by an alkalinity of 16mg/l as CaCO3, a turbidity of 1.5 NTU, a pH of 6.8 and a temperature of 3°C. However, a full contour characterisation indicated formation of a strong floc with much lower settled turbidities required a higher pH and alum dose. These data were developed using a 5min settling period that is generally representative of settling conditions in a conventional sedimentation basin as described previously.

The effect of these coagulation conditions on the performance of dissolved air flotation was examined in a parallel pilot plant study. The adverse effect of poor floc integrity, which was demonstrated in the jar test, is also observed for the dissolved air flotation system when results for the lower pH and coagulant dose condition are compared with performance at a higher pH and coagulant dose. Testing also indicated it was not possible to achieve low filtered turbidity levels (less than 0.1 NTU) unless the higher dose and pH condition was applied. These results illustrate the necessity for providing an appropriate combination of coagulant dose and pH for dissolved air flotation. Similar observations regarding appropriate coagulation conditions were made in a bench-scale simulation of the ballasted flocculation process. Clarification rapid sand filtration sequences depend on the formation of a well-formed floc that can be efficiently removed. The term ‘well-formed floc’ does not necessarily refer to floc size. Dissolved air flotation, for example, works best with small to moderate-sized flocs, as determined by the physical nature of the flocculation step rather than the chemical conditions that result from selection of coagulant dose and pH.

Polymers can also be applied to strengthen floc or reduce coagulant demand but carry-over of polymers into filtration has to be closely monitored. Floc integrity, irrespective of the target size, is a key to efficient removal. This is an essential condition for proper functioning of this type of treatment sequence and any adjustments to coagulation must overlap with conditions that meet this basic objective. Existing coagulant conditions may not provide effective overlap of regions for turbidity and DOC removal, thereby further increasing the need to consider application of alternative coagulation conditions. The extent of this type of overlap may vary seasonally as temperature and other water quality conditions change, as illustrated by the shift to a higher optimum pH for turbidity removal in cold water at some locations. This shift tends to separate the regions of effective turbidity and DOC removal, thereby constraining the capability to operate under conditions where effective DOC removal can be achieved.

Some coagulants are applied for achieving other objectives, which may include the removal of arsenic. Ferric salts often remove a higher level of DOC but tend to reduce pH. A key consideration is effective turbidity removal needs to occur under pH and dose conditions where effective DOC removal can also be achieved.
Bench-scale evaluations offer a convenient starting point and methods are available that allow a reasonable simulation of full-scale performance and determine conditions that can be expected for both organic and turbidity removal. For the final adjustment and confirmation at
full-scale other factors to be considered include the effect of alternate coagulation approaches on the management and disposal of solids residuals, effects on the pH profile within a plant that may affect other processes and total effect on cost – including costs associated with pH-adjusting chemicals to meet target levels in coagulation and for readjustment of pH from the level applied in coagulation to the level required to provide stability in finished water