Environ. Sci. Technol. 2010, 44, 6443–6449
The Role of Polymer in Improving Floc Strength for Filtration LARA FABRIZI,† BRUCE JEFFERSON,† SIMON A. PARSONS,† ANDREW WETHERILL,‡ AND P E T E R J A R V I S * ,† Centre for Water Science, Cranfield University, Cranfield, Bedfordshire, MK43 0AL, U.K., and Yorkshire Water, Halifax Rd, Bradford, BD6 2LZ, U.K.
Received May 12, 2010. Revised manuscript received July 8, 2010. Accepted July 13, 2010.
Dosing polymer to improve floc characteristics is a widely practiced method in water treatment to improve floc strength, and there is strong operational evidence showing the benefit of polymer dosing. However, there is a paucity of information on how polymer operates in terms of quantifying the resulting floc size and strength over different size scales. A dual particle sizing approach was used to monitor large floc that contain most of the sludge volume and small floc that can cause downstream treatability problems for systems with and without polymer dosing. The polymer investigated was a slightly anionic polyacrylamide dosed in water collected post dissolved air flotation at concentrations of 0-0.03 mg L-1. With increasing polymer dose, median floc size increased from 228 to 325 µm. Floc responses to increased shear rate showed that polymer dosing increased resistance to floc break-up. While all of the flocs showed high potential to regrow, regrowth was greatest in polymer-dosed systems, where flocs exceeded the size that they had reached previously. Increasing the dose of polymer showed increased removal of small particles ( 0.99). From the curves, growth rates were established, showing clearly VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6445
FIGURE 4. Floc strength plots for aggregates formed with increasing polymer doses (mg/L). (a) Floc size distribution after 2 min into the jar test. (b) Floc size distribution after 17 min into the jar test. that dosing polymer increased floc growth. During the initial growth stage, after 1 min, the growth rates were 16.0, 21.8, 23.1, and 20.1 µm/min for 0, 0.01, 0.02, and 0.03 mg/L polymer concentration, respectively. After 10 min, growth rates had slowed to 10.4 to 17.7 µm/min from 0 to 0.03 mg/L polymer. The high propensity for flocs to regrow after breakage was supported by the observed growth rates. At 1 min, the growth rate increased from 14.4 µm/min when no polymer was added to 30.0 µm/min at 0.03 mg/L polymer. After 10 min, the growth rates had slowed to 6.8 µm/min (no polymer) and 14.7 µm/min (0.03 mg/L). When the D50 floc size was plotted against the velocity gradient (Figure 4), characteristic straight-line relationships were seen for all polymer systems investigated on a log-log scale of the form: log D50 ) log C - γ log G where D is the median floc diameter, C is the floc strength coefficient, γ is the stable floc exponent, and G is the average velocity gradient in the system. There was a clear difference in the size of the flocs at each breakage shear rate between polymer and nonpolymer-dosed systems with flocs formed with polymer being consistently larger and a general trend for higher polymer doses to have larger flocs for a given shear rate. For example, after breakage at 127.5 s-1, the size of flocs was 84 µm for no polymer to 129 µm at 0.01 mg/L and 137 µm at 0.03 mg/L. The stable floc exponent γ varied between 0.33-0.41 for all of the systems investigated, showing there was no significant difference in the relative resistance to increasing shear rate when more polymer was dosed. The PSD of the floc systems were compared to the Kolmogoroff’s microscale of turbulence (λ) from:
()
ν3 λ) ε
1/4
where ε ) G2ν
where ε is the average energy dissipation in the vessel, ν is the kinematic viscosity, and G is the average velocity gradient in the vessel. During floc growth at 7.5 s-1, the floc were predominantly smaller than λ, indicating that floc breakage would mainly be from erosion of small particles from the parent floc. With increasing polymer dose, a greater proportion of the PSD was >λ. For example, at the end of floc growth 12% of the floc PSD was >λ at 0.03 mg/L polymer concentration, while this was only 2% at 0 polymer dose. The implication is that dosing 6446
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010
polymer reduces the release of small ‘erosion’ particles that could cause operational difficulties. This was confirmed from the results obtained using the Spectrex instrument (Figure 5). Shortly after the flocculation period had begun, no difference was apparent in the particle size distribution for systems with and without polymer addition (Figure 5a). It was evident that there was a high concentration of particles λ. There was little change in the PSD with time during the floc breakage period indicating that when large-scale floc breakage occurs it does so instantaneously when turbulence conditions change during the jar test. All of the PSD for all polymer doses and breakage conditions are shown in Supporting Information D. There was a big difference in the size ranges of the particles observed using the two particle sizing instruments. This difference relates to the fact that the Malvern instrument measures particles based on a volume distribution, while the Spectrex instrument calculates the distribution by number frequency. For instance, a 200 µm
FIGURE 5. The distribution of small particles after floc growth with increasing polymer dose.
FIGURE 6. Proportion of particle size distribution greater than the Kolmogoroff’s microscale with increasing polymer dose after 15 min exposure to increasing breakage velocity gradients (s-1). particle occupies a volume that is 1 000 000 times larger than a 2 µm particle. The volumetric measurement is therefore biased toward large particles in a heterogeneous system while the frequency-based measurement will bias toward whichever particles are most common. However, the results from the Spectrex were consistent with the results from the Mastersizer, showing that as polymer was added to the system smaller particles shifted toward larger size ranges.
Discussion The impact of polymer dosing on floc structure and strength has been demonstrated in this work, with incremental increases in acrylamide polymer resulting in larger and stronger flocs. Translating this directly to practical impacts on the operation of water treatment systems is complex, but there is a clear link between the strength of a floc measured from laboratory testing and its removal in solid-liquid VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6447
separation processes, such that strong flocs can be removed more effectively and will break less, resulting in lower particle loads onto filtration processes (2). The explanation for increased strength of polymer-dosed flocs can be linked to increased bridging between primary particles and microfloc within the overall floc structure. Slightly charged polymers, such as the polyacrylamide used in this study, are advantageous for the purposes of bridging particles. Polymers with higher charge density polymers adsorb more strongly to particles, preventing the evolution of bridging mechanisms (25). Polymer dosing has been widely studied previously, and the dose conditions used in this study are in agreement with typical ranges used for this purpose when dosing into postclarified water of between 0.005 and 0.5 mg/L (16). Floc strength has been considered here in two main ways: (1) the absolute size of the particles (before, during, and after breakage); (2) the relative change in floc size when flocs are exposed to higher shear rates (measured from γ). For higher polymer doses, flocs were larger than in nonpolymer-dosed systems. The fact that γ remained constant was indicative of a similar strength of the individual connection points within the floc, but the greater frequency of connection points provided by increasing polymer concentration meant that these flocs maintained a larger overall size. The results of this work have also shown that by comparing the size of flocs to λ, an indication of the fine particle generation can be established. The practical importance of such a difference is that those flocs that have a substantial portion of their PSD smaller than λ will generate more fine particles as a result of surface erosion. The mechanism by which aggregation takes place following addition of nonionic or weakly charged polymers is well understood (26). The polymer molecule adsorbs to a particle while part of the macromolecule extends into the bulk solution where it is available to bond to other particles in suspension such that interparticular molecular bridges form (15). The observation that floc size increased with higher polymer dose was therefore not surprising because with more polymer dosed there were more opportunities for capture and bridging of particles. The larger floc size for a given breakage shear force was indicative of a greater number of connection points within the floc provided by the polymer which enabled the floc size to be less affected by the high shear rates. This is a view consistent with other work that has shown that polymer dosing increases linkages within flocs (27). The fact that both systems with and without polymer dosing resulted in significant floc regrowth indicated that the forces bonding the flocs together were able to reform after breakage. Reversible floc breakage is a phenomenon seen elsewhere for flocs formed from charge neutralization mechanisms (24) and biological wastewater flocs (28) and when using certain polymers (29). For the nonpolymer-dosed system, adsorbed NOM may bridge components of the floc together in a way similar to that of high MW man-made polymers (30). However, while NOM may act in this way, NOM lies relatively flat on the particle to which it is adhered, which limits the bridging potential. The bridging bonds resulting from adsorbed NOM are therefore considered to be much less frequent when compared with artificial polymers. Charge neutralization mechanisms of reaggregation were therefore considered to be the main mode by which these flocs regrew. Conditions were favorable for short-range attractive forces (van der Waals bonds) to exist between flocs, given that the charge on the aggregates in the system had been minimized (zeta potential of -2.1 mV). These short-range forces can reform after breakage (28). When polymer was dosed, faster floc growth rates and larger floc sizes were observed during reaggregation. 6448
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010
Greater polymer bridging potential with increased polymer dose is an explanation for this observation, indicating that polymer-floc bonds could reform again after breakage (29). A further interesting observation was the continued growth of flocs under high shear rate conditions for systems dosed with polymer. Under any given shear rate, a pseudoequilibrium exists between floc breakage and growth such that a stable floc size is reached quickly for the hydrodynamic conditions. For polymer-dosed systems, there was evidence that after the initial floc breakage under high shear rate conditions, the additional bonding provided by the polymer enabled the bulk of the floc in the system to begin to slowly regrow, such that a stable floc size was not reached. The explanation for this observation is linked to increasing the number of the bridging bonds between attached floc when exposed to stresses, which results in stronger flocs and the fact that adsorbed polymers remain available to reattach after breakage (26). Polymer therefore offers significant treatment benefits as a result of improvements to particle characteristics. The two particle sizing methods adopted here have provided complementary information that is directly relevant to water treatment practice. Small particles are challenging for filters to remove and therefore have the highest probability of passing into final drinking water. The bias of the Mastersizer instrument enables quantification of the floc that are important for removal purposes because the bulk of the contaminants are in the large, high volume floc while the Spectrex instrument is able to show how small particles are generated or removed.
Acknowledgments The authors thank Yorkshire Water for the financial support provided for this project as well as access to treatment works for sampling.
Supporting Information Available (A) Differences between particle sizing instruments; (B) table of raw and clarified water characteristics; (C) particle size distributions of flocs relative to the Kolmogoroff microscale; (D) table of floc strength and regrowth parameters. This material is available free of charge via the Internet at http:// pubs.acs.org.
Literature Cited (1) Kim, J.; Tobiason, J. E. Particles in filter effluent: The roles of deposition and detachment. Environ. Sci. Technol. 2004, 38, 6132–6138. (2) Jarvis, P.; Mergen, M.; Banks, J.; Mcintosh, B.; Parsons, S. A.; Jefferson, B. Pilot scale comparison of enhanced coagulation with magnetic resin plus coagulation systems. Environ. Sci. Technol. 2008, 42, 1276–1282. (3) Bache, D. H.; Gregory, R. Flocs and separation processes in drinking water treatment: a review. J. Water Supply Res. Technol. AQUA 2010, 59 (1), 16–30. (4) Levich, V. G. Physico-Chemical Hydrodynamics; Prentice Hall: Englewood Cliffs, NJ, 1962. (5) Parker, D. S.; Kaufman, W. J.; Jenkins, D. Floc breakup in turbulent flocculation processes. J. Sanit. Eng. Div., Am. Soc. Civ. Eng. 1972, SA 1, 79–99. (6) Hendricks, D. Fundamentals of Water Treatment Unit Processes: Physical, Chemical, and Biological; CRC Press: Boca Raton, FL, 2009. (7) Clark, M. M.; Flora, J. R. Floc re-structuring in varied turbulent mixing. J. Colloid Interface Sci. 1991, 147, 407–421. (8) Boller, M.; Blaser, S. Particles under stress. Water Sci. Technol. 1998, 37 (10), 9–29. (9) Ducoste, J. J.; Clarke, M. M. The influence of tank size and impeller geometry on turbulent flocculation: I. Experimental. Environ. Eng. Sci. 1998, 15 (3), 215–224. (10) Bridgeman, J.; Jefferson, B.; Parsons, S. A. Assessing floc strength using CFD to improve organics removal. Chem. Eng. Res. Des. 2008, 86 (8), 941–950.
(11) O’Melia, C. R. Particles, pretreatment, and performance in water filtration. J. Environ. Eng. 1985, 111 (6), 874–890. (12) Jegatheesan, V.; Vigneswaran, S. Deep bed filtration: Mathematical models and observations. Crit. Rev. Env. Sci. Technol. 2005, 35, 515–569. (13) Tobiason, J. E.; Edzwald, J. K.; Levesque, B. R.; Kaminski, G. K.; Dunn, H. J.; Galant, P. B. Full-scale assessment of waste filter backwash recycle. J. Am. Water Works Assoc. 2003, 95 (7), 80– 93. (14) Bolto, B. Soluble polymers in water purification. Prog. Polym. Sci. 1995, 20, 987–1041. (15) Bolto, B.; Gregory, J. Organic polyelectrolytes in water treatment. Water Res. 2007, 41, 2301–2354. (16) Crittenden, J. C.; Trussel, R. R.; Hand, D. W.; Howe, K. J.; Tchobanoglous, G. Water Treatment: Principles and Design, 2nd ed.; John Wiley & Sons: New York, 2005. (17) Saltnes, T.; Eikebrokk, B.; Ødegaard, H. Contact filtration of humic waters: Performance of an expanded clay aggregate filter (Filtralite) compared to a dual anthracite/sand filter. Water Sci. Technol. Water Supply 2002, 2 (5-6), 17–23. (18) Rawlings, M. M.; Fitzpatrick, C. S. B.; Gregory, J.; Wetherill, A. The effect of polymeric flocculants on floc strength and filter performance. Water Sci. Technol. Water Supply 2006, 53 (7), 77–85. (19) Cleasby, J. L.; Dharmarajah, A. H.; Sindt, G. L.; Baumann, E. R. Design and operations guidelines for optimization of the highrate filtration process: Plant survey results; American Water Works Association Research Foundation: Denver, CO, 1989. (20) Rajagopalan, R.; Tien, C. Trajectory analysis of deep-bed filtration with the sphere in cell porous media model. AIChE J. 1976, 22 (3), 523–533.
(21) Jarvis, P.; Jefferson, B.; Parsons, S. A. How the natural organic matter to coagulant ratio impacts on floc structural properties. Environ. Sci. Technol. 2005, 39, 8919–8924. (22) Sharp, E.; Jarvis, P.; Parsons, S. A.; Jefferson, B. The impact of zeta potential on the physical properties of ferric-NOM flocs. Environ. Sci. Technol. 2006, 40, 3934–3940. (23) Henderson, R.; Sharp, E.; Jarvis, P.; Parsons, S.; Jefferson, B. Identifying the linkage between particle characteristics and understanding coagulation performance. Water Sci. Technol. Water Supply 2006, 6 (1), 31–38. (24) Jarvis, P.; Jefferson, B.; Parsons, S. A. Breakage, re-growth and fractal nature of natural organic matter flocs. Environ. Sci. Technol. 2005, 39, 2307–2314. (25) Nasser, M. S.; James, A. E. Effect of polyacrylamide polymers on floc size and rheological behaviour of kaolinite suspensions. Colloids Surf., A 2007, 301 (1-3), 311–322. (26) Muhle, K.; Domasch, K. Floc strength in bridging flocculation. In: Chemical Water and Wastewater Treatment; Hahn, H. H.; Klute, R., Eds.; Springer-Verlag: Berlin, Germany, 1990; p 105. (27) Chang, E. E.; Chiang, P. C.; Tang, W. T.; Chao, S. H.; Hsing, H. J. Effects of polyelectrolytes on reduction of model compounds via coagulation. Chemosphere 2005, 58, 1141–1150. (28) Chaignon, V.; Lartiges, B. S.; El Samrania, A.; Mustin, C. Evolution of size distribution and transfer of mineral particles between flocs in activated sludges: an insight into floc exchange dynamics. Water Res. 2002, 36 (3), 676–684. (29) Gregory, J. Fundamentals of Flocculation. CRC Crit. Rev. Environ. Control 1989, 19, 185–230. (30) Walker, H. W.; Bob, M. M. Stability of particle flocs upon addition of natural organic matter under quiescent conditions. Water Res. 2001, 35 (4), 875–882.
ES101543H
VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6449
