14 Coagulation and flocculation Update by Ken Ives Special contribution from Samia Al Azharia Jahn

14 Coagulation and flocculation

14.1 Introduction
Coagulation and flocculation provide the water treat-ment process by which finely divided suspended and colloidal matter in the water is made to agglomerate and form flocs. This enables their removal by sedi-mentation, dissolved air flotation or filtration. Colloidal particles (colloids) are midway in size1 between dissolved solids and suspended matter. Colloids are kept in suspension (stabilised) by electrostatic repulsion and hydration. Electrostatic repulsion occurs because colloids usually have a surface charge due to the presence of a double layer of ions around each particle. Thus, the colloid has an electric charge, mostly a negative one. Hydration is the reaction of particles at their surface with the surrounding water. The resulting particle-water agglomerates have a specific gravity, which differs little from that of water itself. The substances that frequently are to be removed by coagulation and flocculation are those that cause turbidity and colour. Surface waters in tropical countries often are turbid and contain colouring material. Turbidity may result from soil erosion, algal growth or animal/vegetable debris carried by surface run-off. Substances leached from decomposed organic matter, leaves, or soil such as peat may impart colour. Both turbidity and colour are mostly present as colloidal particles. The electrostatic repulsion between colloidal particles effectively cancels out the electronic attraction forces (Van der Waals forces) that would attach the particles together. Certain chemicals (called coagulating agents, coagulants) have the capacity to compress the double layer of ions around the colloidal particles. They reduce the range of the electrostatic repulsion, and thus enable the particles to flocculate, i.e. to form flocs. These flocs can grow to a sufficient size and specific weight to allow their removal by settling, flotation or filtration. Generally water treatment processes involving the use of chemicals are not so suitable for small community water supplies. They should be avoided whenever possible. Chemical coagulation and flocculation should only be used when the needed treatment result cannot be achieved with another treatment process using no chemicals. If the turbidity and colour of the raw water are not much higher than is permissible for drinking water, it should be possible to avoid chemical coagulation in the treatment of the water. A process such as slow sand filtration or multi-stage filtration would serve both to reduce the turbidity and colour to acceptable levels, and to improve the other water quality characteristics, in a single unit. A roughing filter can serve to reduce the turbidity load on the slow sand filter, if necessary.

Size range: 5 x 10-3 2.10-1 micrometer (0.2 microns)

296

Chapter 14

14.2 Coagulants Alum (Al2(SO4)3.nH20) where n=14, 16, or 18, depends on the form of alum supplied. This may be in liquid solution, broken crystalline granules 2-5 cm size (kibbled) or crystalline blocks. It is by far the most widely used coagulant. Iron salts (e.g. ferric chloride (FeCl3), or ferric sulphate (Fe2(SO4)3.9H2O) can be used as well and in some instances have advantages over alum. A significant advantage of iron salts over aluminium is the broader pH2 range for good coagulation. Thus, in the treatment of soft coloured waters where colour removal is best obtained at low pHs, iron salts may be preferred as coagulants. Iron salts should also be considered for coagulation at high pHs, since ferric hydroxide is highly insoluble in contrast to aluminium salts, which form soluble aluminate ions at high pHs. Sodium aluminate is mostly used for coagulation at medium pHs. Synthetic organic polyelectrolytes have become available as coagulants but are generally not economical for small water supply systems, nor are they readily available. The use of extracts from the crushed seeds of certain tropical trees is described in detail in the next section, 14.3. Coagulants such as soluble aluminium and iron salts react with the alkalinity of the water, and hydrolyse in it. For example, alum reacts to form aluminium-hydroxide floc, Al(OH)3, a gelatinous precipitate. The required alkalinity may be naturally present in the water or it has to be added through dosage of lime, Ca(0H)2 or sodium carbonate, Na2CO3 (also called soda ash). For good coagulation the optimal dose of coagulant should be fed into the water and quickly and properly mixed with it. The optimal dose will vary depending upon the nature of the raw water and its overall composition. It is not possible to compute the optimal coagulant dose for a particular raw water. A laboratory experiment called the jar test is generally used for the periodic determination of the optimal dose. The jar test may be briefly described as follows: A series of samples of water are placed on a special multiple stirrer and the samples, typically 800 ml in 1-litre beakers (or jars), are dosed with a range of coagulant, e.g. 10, 20, 30, 40 and 50 mg/l, and are stirred vigorously for about one minute. Then follows a gentle stirring (10 minutes), after which the samples are allowed to stand and settle for 30-60 minutes. The samples are then examined for colour and turbidity and the lowest dose of coagulant that gives satisfactory clarification of the water is noted.

Measure of the acidity/alkalinity of water. Acid water has a pH below 7, the pH of alkaline water is higher than 7.

297

A second test involves the preparation of samples with the pH adjusted so that the samples cover a range (e.g. pH = 5, 6, 7 and 8). The coagulant dose determined previously is added to each beaker. Then follows stirring, flocculation and settlement as before. After this, the samples are examined and the optimum pH is determined. If necessary, a re-check of the minimum coagulant dose can be done.

Fig. 14.1. pH zone-coagulation relationship (Adapted from R.F. Packham)

As mentioned earlier, aluminium and iron salts have considerable differences in their pH zones of good coagulation. For alum the pH zone for optimum coagulation is quite narrow, ranging from about 6.5-7.5. The comparable range for ferric sulphate is considerably broader, a pH range of about 5.5-9.0 (Fig. 14.1). When the results of a jar test are plotted, this type of curve is typical. The most common method of dosing the alum or ferric sulphate is in the form of a solution. Such a solution (usually of 3-7% strength, that is, 30-70 kg alum Al2(SO4)3. 14H2O per m3 water) is prepared in special tanks with a holding capacity of ten or more hours coagulant feeding requirements. Two tanks are required, one in operation, while the solution is being prepared in the other. When using alum, one should keep in mind that in solutions of less than 1% strength, the chemical is hydrolysed (i.e. forms agglomerates with the chemical feed water) before it is dosed into the raw water. To prevent this, the solution should always have a strength of more than 1.5%.
298

Chapter 14

Fig. 14.2. Chemical feed arrangement for alum

Various chemical feed arrangements can be used. Figure 14.2 shows an example. The simplest method of using lime is in the form of a suspension led into a special tank (called a lime saturator), to produce a saturated solution of calcium hydroxide. The size of the tank depends on the required lime dosage. 14.3 Coagulation and flocculation with extracts of particular seeds Special contribution from Dr. Samia Al Azharia Jahn, Water Treatment Adviser Botanical identification Early attempts at water coagulation with plant materials arose through self-help by poor people who had no access to groundwater. In India the seeds of Strychnos potatorum (table 14.1) have been used to clarify turbid water since the first centuries AD. Where this wild tree was not available, kernels of Prunus species and seeds of legumes (Papilionoidea) became substitutes. In the Northern Sudan, seeds of Moringa oleifera were tried out instead of beans and groundnuts. A relationship between the botanical type and a content of coagulants (chemotaxonomy) could be detected for several plant genera and families.

299

Table 14.1

Seed extracts that contain outstanding natural coagulants1

Plant Chemical nature of coagulant Strychnos potatorum (clearing nut) mannogalactan, probably anionic polyelectrolyte basic polypeptides Raw water, turbidity range, application Jamuna River (India): ca. 200-2200 NTU, recommended as coagulant aid Moringa oleifera (horseradish tree)
White Nile (Sudan):

Reference Bulusu and Sharma 1965

Jahn 1986, 1988

cationic polyelectrolyte 50-250 NTU; Blue Nile:

> 30 NTU-several thousand NTU, fast acting primary coagulant

Moringa stenopetala five other species of Moringa Vicia faba (broad bean) Prunus armeniaca (apricot) , and P.amygdalis (almond) Prunus persica (peach)
2

similar basic polypeptides not yet known not known

performance like Moringa oleifera performance like Moringa oleifera Blue Nile: > 800 NTU slow acting primary coagulant

Jahn 1986, 1988 Jahn 1986, 1988 Jahn 1986, 1988

not known not known not known

River Nile, Egypt Changjiang, China Oranje River, Lesotho rivers of Potos Bolivia

Alpino 1591, dArcet 1838, Lu Yu 1170 Offringa 1993, Soliz Rodriguez 1984

1 2

For details and bibliography cf. Jahn 1986, 1988, 1994 Apricot kernels were in Egypt, North Sudan and Sichuan (China) still in use in the 20th century

Chemical composition and toxicology Certain seeds contain substances for the growth of the seedling, which also have flocculating properties. The polyelectrolytes of Moringa oleifera, M. stenopetala and Strychnos potatorum are examples of such flocculating substances (table 14.1). The destabilisation of water colloids is due to positive or negative charges. The cationic electrolytes of Moringa are presumed to act like cations of metal coagulants. Alum and a seed extract different impacts on several chemical and physical characteristics of coagulation (table 14.2). Moringa seeds contain seed oil and a germicide. The smell and taste effect on the water is very small, even for the maximum Moringa seed dose (200-250 mg/l). In the low range of the optimal doses there is no germicide effect and there are no risks from toxic substances. The possible health effects from toxic substances in Strychnos potatorum seeds and Prunus kernels have not been investigated yet (Jahn, 1988, 1994).

300

Chapter 14

Table 14.2

Differences between water coagulation with alum and Moringa oleifera seed extract
Parameters Destabilisation Support of aggregation Efficiency at low turbidity Mineralogical nature of suspended matter Pollution by organic matter Floc Aluminium sulphate Al
3+

Moringa seed extract cationic polyelectrolytes none failing below 30 NTU montmorillonite clays are present (Blue Nile)

adsorption at Al(OH)3 flocs still satisfactory at 3 NTU

minor impact on optimal doses doses about 50% lower if

minor impact on optimal doses great impact on doses and additional organic load coarse due to aluminium hydroxides very fine; result is slower settling; further decrease of settling at lower temperature small
2-

Sludge volume pH of treated water Hardness

great increase of Ca2+ due to neutralisation

increase of acidity (effect of SO4 ) no change

no change

Preparation and application of Moringa seed coagulants Seed quality Only ripe Moringa oleifera seeds (with brown seed coat) give a high coagulant efficiency (Jahn, 1986). Improper treatment of the seeds leads to loss of coagulation capacity. It can be prevented if collection and treatment is done by agro-foresters or trained women and by using correct equipment. Turbidity range Experiences with seed coagulation relate to highly turbid waters only. If optimal doses and proper stirring are applied, seed extracts of Moringa oleifera can also reduce low turbidity levels to 5-10 NTU after one hour of settling. For raw water with turbidity below 30 NTU, alum is superior in performance (table 14.2). Thus, surface waters with varying turbidity (below 30 NTU for certain periods of the year) cannot always be treated adequately with seed extracts alone but need coagulant aids during such low turbidity periods. Use of coagulant aids is not feasible at household level. Chemical water quality Seed coagulants are more sensitive than alum to the mineralogical composition of suspended matter. Waters polluted by organic matter are only suitable for long-term treatment as long as the required seed coagulant dose is less than 250 mg/l. This is for both economic and microbiological considerations.
301

Climatic factors Seed coagulants are mainly applicable as primary coagulants for tropical and subtropical countries. At low temperatures the settling of the fine flocs (table 14.2) is too slow and incomplete, unless seed treatment is combined with other coagulants. Removal of pathogens Like alum, Moringa seed extracts achieve a removal of faecal coliforms of more than 98% under optimal conditions. However, if the residual turbidity is more than 30-40 NTU, the treated water still has the same contamination as raw water. Therefore, optimal dosing is required for small plants and for household application. Determination of optimal doses Simplified (village) jar tests are a convenient alternative where the usual measurements cannot be afforded or electricity is not available. Special devices have been designed for stirring the seed coagulant at plant level. A seed coagulant suspension has different reaction properties than a solution of a purified chemical such as alum. The coagulant efficiency is strongly reduced by shortening the time needed for thorough mixing of the seed extract with turbid water. Involvement of women, teachers and commerce Asian and African women from the lower classes are mainly responsible for the discovery of seed extracts as coagulants for turbid water. Lack of knowledge on application has limited the possible improvement of water quality. Water coagulation at household level cannot be learned from a demonstration alone. Therefore, the involvement of teachers or commercial outlets is required to help the women in determining the optimal seed coagulant dose, and to distribute or sell standard solutions of the seed coagulant. Socio-economic challenges Since Moringa oleifera is a multipurpose tree, certain economic interests may endanger long-term uses of seed extracts for water coagulation. Sometimes people prefer to have the fruit trees in their own gardens rather than on community lands (women then have no right to harvest the seeds). There may also be a commercial preference to sell green Moringa pods as a delicacy to urban customers or factories. On the other hand, removal of the seed oil for other uses and application of press cake extracts instead of seed extracts has been tried out in small plants of Malawi and Tanzania (Folkard et al., 1993). There are commercial developments too in the production of a completely biodegradable Moringa coagulant, known under the commercial name PhytoflocTM.

302

Chapter 14

14.4 Rapid mixing Rapid mixing aims at the immediate dispersal of the entire dose of chemicals throughout the mass of the raw water. To achieve this, it is necessary to agi-tate the water violently and to inject the chemical in the most turbulent zone, in order to ensure its uniform and rapid dispersal. The mixing has to be rapid, because the hydrolysis of the coagulant is almost instantaneous (within a few seconds). The destabilisation of colloids also takes a similar short time. The location of the rapid mixer should be near to the chemical house where solutions of chemicals are prepared. The feeding pipes then will be of short length. It is also desirable to place the rapid mixing device close to the flocculators. To combine both these requirements in the layout of a treatment plant is often quite difficult. Many devices are used to provide rapid mixing for the dispersal of chemicals in water. Basically, there are two groups: Hydraulic rapid mixing Mechanical rapid mixing

Hydraulic rapid mixing For hydraulic rapid mixing, arrangements are used such as channels or chambers with baffles producing turbulent flow conditions, overflow weirs, and hydraulic jumps (Figs. 14.3; 14.4; 14.5). Rapid mixing may also be achieved by feeding the chemicals at the suction side of pumps. With a good design, a hydraulic mixer can be as effective as a mechanical mixing device.

Fig. 14.3. Baffled channel for rapid mixing (not very effective)

303

Fig. 14.4. Overflow weir

Fig. 14.5. Hydraulic jump

304

Chapter 14

Mechanical rapid mixing With mechanical mixing the power required for agitation of water is imparted by impellers, propellers or turbines (rapid mixers, flash mixers, and turbo mixers). See figure 14.6.
Generally mechanical rapid mixers are less suitable for small treatment plants than hydraulic ones since they require a reliable and continuous supply of power and maintenance.

Fig. 14.6. Mechanical mixer

14.5 Flocculation Flocculation is the process of gentle and continuous stirring of coagulated water for the purpose of forming flocs through the aggregation of the minute particles present in the water. It is thus the conditioning of water to form flocs that can be readily removed by settling, dissolved air flotation or filtration. The efficiency of the flocculation process is largely determined by the number of collisions between the minute coagulated particles per unit of time. There are mechanical and hydraulic flocculators. In mechanical flocculators the stirring of the water is achieved with devices such as paddles, paddle reels or rakes.

305

These devices can be fitted to a vertical or horizon-tal shaft. Vertical shaft flocculators are usually placed in a square tank with several chambers (four or more). They have the advantage of having only bearing in the water, and no gland is necessary as the motor and gearing are above the water. With horizontal shaft flocculators having a traverse flow, one should provide at least four rows of shafts, with partitions of baffles (stop logs), so as to avoid short-circuiting. In hydraulic flocculators the flow of the water is so influenced by small hydraulic structures that a stirring action results. Typical examples are channels with baffles, flocculator chambers placed in series (e.g. Alabama-type flocculator) and gravel bed flocculators. The main shortcomings of hydraulic flocculators are: No adjustment is possible to changes of raw water composition. No adjustment is possible to the water production rate of the treatment plant. The head loss is often appreciable. They may be difficult to clean.

Their advantages are that they have no motor power, electric cables switchgear, etc. to maintain and general maintenance is easier. Design of flocculators In the design of a flocculator installation not only the velocity gradient (G) should be taken into account, but also the detention time (t). The product G.t gives a measure for the number of particle colli-sions, and thus for the floc formation action5. The equation for computing the velocity gradient is: G = (P/(V.m))1/2 in which G = velocity gradient (s-1) P = power transmitted to the water (watt = rghQ) r g h = density of water = 1000 kg/m3 = gravitational acceleration = 9.81 m/s2 = head loss of the water passing through the flocculator in m

Q = volumetric flow rate in m3/s V = volume of water to which the power is applied; where applicable, the volume of the mixing tank or basin (m3) m = dynamic viscosity of water (kg/m.s), see table 14.3

Scientifically the complete dimensionless group is G.t.C. where C = concentration volume/volume of floc participles in the water. But in the types of mechanical and hydraulic flocculators described here, which are flow-through (that is, no flocs are retained) the value of C is very small and usually ignored. C becomes more important in flocculators where the floc is retained by recirculation or in suspension as a floc blanket. Operationally these are not suitable for small community water supplies.

306

Chapter 14

Table 14.3

Dynamic viscosity of water m (kg/m.s)

Temperature t(C) Dynamic viscosity m (kg/m.s) 15 1.14 x 10
-3

20 1.01 x 10
-3

25 0.96 x 10-3

Table 14.4

Flocculator design criteria

Design factor Range Typical value G (s-1) 10-100 45-90 t (s) 1 200-1 800 1 800 G.t 30 000-150 000 50 000-100 000

For each individual flocculator the optimal G.t value should be carefully selected, and taken as high as is consistent with the optimal formation of flocs without causing disruption or disintegration of the flocs after they have formed. Chemicals such as activated silica or polyelectrolytes (coagulant aids) can improve the internal cohesion of the flocs. 14.6 Hydraulic flocculators Baffled channel flocculators For horizontal-flow baffled flocculation channels (Fig. 14.7), the design water velocity usually is in the 0.10-0.30 m/s range. Detention time normal-ly is 15-20 minutes.

Fig. 14.7. Horizontal-flow baffled channel flocculator (plan)

307

This type of flocculator is well suited for very small treatment plants. The efficiency, however, is highly dependent on the depth of water in the baffled channel. Flocculators with vertical flow through baffled chambers (Fig. 14.8) are mostly used for medium and larger-size water treatment plants. The water flow velocity range is 0.1-0.2 m/s. Detention time is 10-20 minutes. Cleaning arrangements are called for because of deposits in the flocculator.

Fig. 14.8. Vertical-flow baffled chamber flocculator (cross-section)

Alabama-type flocculators Alabama-type flocculators are hydraulic floccula-tors having separate chambers placed in series through which the water flows in two directions (Fig. 14.9). The water flows from one chamber to the next, entering each adjacent partition at the bottom end through outlets turned upwards. This type of flocculator was initially developed and used in the state of Alabama (U.S.A.) and later introduced in Latin America. For effective flocculation in each chamber, the outlets should be placed at a depth of about 2.50 m below the water level. Common design criteria are: Rated capacity per unit chamber: Velocity at turns Length of unit chamber (L) Width (B) Depth (h) Detention time (t)
308

25-50 l/s per square metre 0.40-0.60 m/s 0.75-1.50 m 0.50-1.25 m 2.50-3.50 m 15-25 minutes

Chapter 14

Fig. 14.9 .Alabama-type flocculator

The loss of head in this type of flocculator normally is about 0.35-0.50 m for the entire unit. The velocity gradient is usually in the 40-50 s-1 range. Table 14.5 provides practical guidance for the design of an Alabama-type flocculator. Example: Flow Q = 1.2 m3/minute. Detention time = 15 mm. Size of curved pipe: 250 mm (10). Unit chamber measures 0.60 x 0.75 m2. Volume of unit chamber: 1.3 m3. Total volume required. 15 x 1.2=18 m3 Number of chambers 18/1.3 = 14. Hydraulic jet mixer and flocculator In a jet flocculator the coagulant (alum) is injected in the raw water using a special orifice device. The water is then jetted into a tapered cylinder placed above the nozzle. The resulting jet pump action gives a gentle stirring of the water for floc formation, and part of the formed flocs are recycled (Fig. 14.10). Through the combination of these two actions excellent flocculation results can be achieved.

309

Table 14.5

Guidance for Alabama-type flocculator design

Flow Rate Q (l/s) 10 20 30 40 50 60 70 80 90 100 Width B (m) 0.60 0.60 0.70 0.80 0.90 1.00 1.05 1.15 1.20 1.25 Length L (m) 0.60 0.75 0.85 1.00 1.10 1.20 1.35 1.40 1.50 1.60 Diameter D (mm) 150 250 300 350 350 400 450 450 500 500 Unit chamber area (m ) 0.35 0.45 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
2

Unit chamber volume (m3) 1.1 1.3 1.8 2.4 3.0 3.6 4.2 4.8 5.4 6.0

Fig. 14.10. Hydraulic jet flocculator

310

Chapter 14

Bibliography American Water Works Association. (1969). Water treatment plant design. Denver, CO, USA, American Water Works Association. Camp, T.R. (1955). Flocculation and flocculation basins. In: Transactions of the American Society of Civil Engineers, vol. 120, p. 1-16. Fair, G N; Geyer, J C. and Okun, D.A. (1968). Water and wastewater engineering. Vol. 2. Water purification and wastewater treatment and disposal. New York, NY, USA, John Wiley. Folkard, G.K.; Sutherland, J.P. and Grant, W.D. (1993) Natural coagulants at pilot scale In: Pickford, J. (ed.). Water, environment and management: proceedings of the 18th WEDC Conference, Kathmandu, Nepal, 30 August - 3 September 1992. Loughborough, UK, Water, Engineering and Development Centre, Loughborough University of Technology. Gomella, C. and Guerre, H. (1973). La traitement des eaux de distribution. Paris, France, Editions Eyrolles. Hudson, H.E. and Wolfner, J P. (1967). Design of mixing and flocculating basins. In: Journal of the American Water Works Association, vol. 59, no. 10, p. 1257-1267. Jahn, S.A. (1986). Proper use of African natural coagulants for rural water supplies: research in the Sudan and a guide for new projects. (Schriftenreihe der GTZ; no. 191). Eschborn, Germany, Deutsche Gesellschaft fr Technische Zusammenarbeit (GTZ). Jahn, S.A. (1988). Using Moringa seeds as coagulants in developing countries. In: Journal of the American Water Works Association, vol. 80, no. 6, p. 43-50. Jahn, S.A. (1994). On the history of water coagulation: transfer of ancient Hindu practises to the valleys of the Yangtze River and the Nile. In: ASIEN, no. 50, p. 45-57. McConnachie, H.L.; Folkard, G.K.; Mtwali, M.A. and Sutherland, J.P. (1999). Field trials of appropriate flocculation processes In: Water research, vol. 33, no. 6, p. 1425-1434. Olsen, A. (1978) Low technology water purification by bentonite clay and Moringa Oleifera seed flocculation as performed in Sudanese villages: effects on Schistosoma Mansoni Cerceriae. In: Water research, vol. 21, May, p. 517-522. ONelma, C.R. (1969). A review of the coagulant process. In: Journal of public work, no. 3.

311

Singley, J.E. (1972). State-of-the-art of coagulation. In: PAITO symposium on modern water treatment methods, Asuncion, Paraguay. Stumm, W. and Morgan, J.J. (1971) Chemical aspects of coagulation. In: Journal of the American Water Works Association, vol. 63, no. 8, p. 931-994. Twort, A.C.; Ratnayaka, A.D. and Brandt, M.J. (2000). Water supply. 5th ed. London, UK, Edward Arnold Publishers and IWA Publishing. Chapter7, part I, p. 267-317. Vral, L. and Jordan, R N. (1971). Rapid-mixing in water treatment. In: Journal of the American Water Works Association, vol. 63, no. 8, p. 52-63. Web sites Water Treatment Plant Operation (training programme): http://www.owp.csus.edu/WTPO1.html

312