16593 06 tủ tài liệu bách khoa

CHAPTER 6

COAGULATION AND
FLOCCULATION
Raymond D. Letterman, Ph.D., P.E.
Professor, Department of Civil and Environmental
Engineering, Syracuse University, Syracuse, New York

Appiah Amirtharajah, Ph.D., P.E.
Professor, School of Civil and Environmental Engineering,
Georgia Institute of Technology,
Atlanta, Georgia

Charles R. O’Melia, Ph.D., P.E.
Abel Wolman Professor, Department of Geography and
Environmental Engineering, The Johns Hopkins University,
Baltimore, Maryland

Coagulation is a process for increasing the tendency of small particles in an aqueous
suspension to attach to one another and to attach to surfaces such as the grains in a
filter bed. It is also used to effect the removal of certain soluble materials by adsorption or precipitation. The coagulation process typically includes promoting the interaction of particles to form larger aggregates. It is an essential component of
conventional water treatment systems in which the processes of coagulation, sedimentation, filtration, and disinfection are combined to clarify the water and remove
and inactivate microbiological contaminants such as viruses, bacteria, and the cysts
and oocysts of pathogenic protozoa. Although the removal of microbiological contaminants continues to be an important reason for using coagulation, a newer objective, the removal of natural organic material (NOM) to reduce the formation of
disinfection by-products, is growing in importance.
Aluminum and ferric iron salts have long been used to remove color caused by
NOM. These organic substances are present in all surface waters and in many
groundwaters. They can be leached from soil, diffused from wetland sediments, and
released by plankton and bacteria. Natural organic material adsorbs on natural particles and acts as a particle-stabilizing agent in surface water. It may be associated
with toxic metals and synthetic organic chemicals (SOCs). Natural organic material
includes precursor compounds that form health-related by-products when chlorine

and other chemical disinfectants are used for disinfection and oxidation. For these
reasons, considerable attention is being directed at the removal of NOM by coagu6.1


6.2

CHAPTER SIX

lation in water treatment, even when color removal is not the principle objective. A
treatment technique requirement in the U.S. Environmental Protection Agency’s
(USEPA’s) Stage 1 Disinfection By-Products Rule requires NOM removal in conventional treatment systems by the practice of enhanced coagulation.
Coagulation has been an important component of high-rate filtration plants in
the United States since the 1880s. Alum and iron (III) salts have been employed as
coagulant chemicals since the beginning, with alum having the most widespread use.
In the 1930s, Baylis perfected activated silica as a “coagulant aid.” This material,
formed on site, is an anionic polymer or a small, negatively charged colloid. Synthetic organic polymers were introduced in the 1960s, with cationic polymers having
the greatest use. Natural starches were employed before the synthetic compounds.
Polymers have helped change pretreatment and filtration practice, including the use
of multimedia filters and filters with deep, uniform grain-size media, high-rate filtration, direct filtration (rapid mixing, flocculation, and filtration, but no sedimentation), and in-line filtration (rapid mixing and filtration only).
Coagulants are also being used to enhance the performance of membrane microfiltration systems (Wiesner et al., 1989) and in pretreatment that prolongs the bed
life of granular activated carbon (GAC) contactors (Nowack and Cannon, 1996).
The development of new chemicals, advances in floc removal process and filter
design, and particle removal performance standards and goals have stimulated substantial diversity in the design and operation of the coagulation process, and change
can be expected to continue into the future.
In evaluating high-rate filtration plants that were producing high-quality filtered
water, Cleasby et al. (1989) concluded, “Chemical pretreatment prior to filtration is
more critical to success than the physical facilities at the plant.” Their report recommends that plant staff use a well-defined coagulant chemical control strategy that
considers variable raw-water quality.There is no question that high-rate (rapid sand)
filtration plants are coagulant-based systems that work only as well as the coagulants that are used.

DEFINITIONS
Coagulation is a complex process, involving many reactions and mass transfer steps.
As practiced in water treatment the process is essentially three separate and sequential steps: coagulant formation, particle destabilization, and interparticle collisions.
Coagulant formation, particle destabilization, and coagulant-NOM interaction typically occur during and immediately after chemical dispersal in rapid mixing; interparticle collisions that cause aggregate (floc) formation begin during rapid mixing
but usually occur predominantly in the flocculation process. For example, using the
aluminum sulfate salt known as alum [Al2(SO4)3⋅14H2O] in coagulation involves
formation of an assortment of chemical species, called aluminum hydrolysis products, that cause coagulation. These species are formed during and after the time the
alum is mixed with the water to be treated. Coagulants are sometimes formed (or
partially formed) prior to their addition to the rapid-mixing units. Examples include
activated silica and synthetic organic polymers, and the more recently introduced
prehydrolyzed metal salts, such as polyaluminum chloride (PACl) and polyiron chloride (PICl).
The terminology of coagulation has not been standardized. However, in most of
the water treatment literature, coagulation refers to all the reactions and mechanisms that result in particle aggregation in the water being treated, including in situ


COAGULATION AND FLOCCULATION

6.3

coagulant formation (where applicable), particle destabilization, and physical interparticle contacts. The physical process of producing interparticle contacts is termed
flocculation.
These definitions of coagulation and flocculation are based on the terminology
used by early practitioners, such as Camp (1955). However, in the colloid science literature, LaMer (1964) considered only chemical mechanisms in particle destabilization and used the terms coagulation and flocculation to distinguish between two of
them. LaMer defined destabilization by simple salts such as NaCl (a so-called indifferent electrolyte) as “coagulation.” Destabilization of particles by adsorption of
large organic polymers and the subsequent formation of particle-polymer-particle
bridges was termed “flocculation.”
The water treatment literature sometimes makes a distinction between the terms
“coagulant” and “flocculant.” When this distinction is made, a coagulant is a chemical used to initially destabilize the suspension and is typically added in the rapid-mix
process. In most cases, a flocculant is used after the addition of a coagulant; its purpose is to enhance floc formation and to increase the strength of the floc structure.
It is sometimes called a “coagulant aid.” Flocculants are often used to increase filter

performance (they may be called “filter aids” in this context) and to increase the
efficiency of a sludge dewatering process. In any case, depending on how and where
it is used and at what dosage, a coagulant is sometimes a flocculant and vice versa. In
this chapter, no distinction is made between coagulants and flocculants. The term
“coagulant” is used exclusively.

Coagulants and Treatment Waste
The type and amount of coagulant or coagulants used in a water treatment facility
can have a significant effect on the type and amount of residue produced by the
plant. The amount of residue (weight and volume) impacts the cost of treatment and
the overall environmental significance of the plant. Because, in most water treatment facilities, coagulation is the process that generates the bulk of the residual
materials, their handling and disposal processes and costs must be considered in the
selection and use of coagulants. The use of enhanced coagulation is an important
example of this, because the higher coagulant dosages may produce residuals that
are much more difficult to dewater. Water treatment plant waste handling, treatment, and disposal are covered in Chapter 16.

CONTAMINANTS
Natural Organic Material
Humic substances are typically the major component of NOM in water supplies.
They are derived from soil and are also produced within natural water and sediments by chemical and biological processes such as the decomposition of vegetation.
Humic substances are anionic polyelectrolytes of low to moderate molecular weight;
their charge is primarily caused by carboxyl and phenolic groups; they have both
aromatic and aliphatic components and can be surface active; they are refractive and
can persist for centuries or longer. Humic substances are defined operationally by
the methods used to extract them from water or soil. Typically, they are divided into


6.4

CHAPTER SIX

the more soluble fulvic acids (FAs) and the less soluble humic acids (HAs), with FAs
predominating in most waters (Christman, 1983).
The concentration of NOM in water is typically expressed using the amount of
organic carbon. Organic carbon that passes through a 0.45 µm pore-size membrane
filter is defined as dissolved organic carbon (DOC), and the amount that does not is
known as particulate organic carbon (POC). Total organic carbon (TOC) is the sum
of DOC and POC. Most groundwaters have a DOC of less than 2 mg C/L, whereas
the DOC of lakes ranges from 2 mg C/L or less (oligotrophic lakes) to 10 mg C/L
(eutrophic lakes) (Thurman, 1985). The DOC of small, upland streams will typically
fall in the range 1 to 3 mg C/L; the DOC of major rivers ranges from 2 to 10 mg C/L.
The highest DOC concentrations (10 to 60 mg C/L) are found in wetlands (bogs,
marshes, and swamps). The DOC concentration in upland lakes has been shown to
be directly related to the percentage of the total watershed area that is near-shore
wetlands (Driscoll et al., 1994). The median raw water TOC concentration for U.S.
plants treating surface water is approximately 4 mg C/L (Krasner, 1996).
Disinfection By-Products. The amount of by-products formed by disinfectant
chemicals such as chlorine is proportional to the amount of organic carbon in the
water. A number of relationships between organic carbon and disinfection byproduct concentration have been presented in the literature. For example, Chapra,
Canale, and Amy (1997) used data from groundwater, agricultural drains, and surface waters (rivers, lakes, and reservoirs) to show a highly significant correlation
(r 2 = 0.936, n = 133) between the TOC and the trihalomethane formation potential
(THMFP). The relationship is given by
THMFP = 43.78 TOC1.248

(6.1)

where THMFP is in µg/L and TOC is in mg C/L. The data gathered by Chapra et al.
(1997) are consistent with the frequent observation that high-TOC waters (with
their higher fraction of humic acids) yield a greater amount of THMs per amount of
TOC than do low-TOC waters.The yield was 20 to 50 µg THMFP/mg C for low-TOC

waters and 50 to 100 µg THMFP/mg C for high-TOC waters. Disinfection byproduct formation is covered in detail in Chapter 12.
Specific Ultraviolet Light Absorbance (SUVA). Organic compounds that are aromatic in structure or that have conjugated double bonds absorb light in the ultraviolet wavelength range.The higher molecular weight fraction of NOM (the fraction that
tends to be removed by coagulation and that has the greater yield of disinfection byproducts) absorbs UV light, and consequently, UV light absorbance (typically at a
wavelength of 254 nm) can be used as a simple surrogate measure for DOC. Also, the
ratio of the UV absorbance to the DOC concentration (called the specific UV
absorbance, or SUVA) can be used as an indicator of the molecular weight distribution of the NOM in the water. Based on the absorbance (at 254 nm), expressed as the
reciprocal of the light path length in meters, divided by the DOC concentration in mg
C/L, the units of SUVA are L/mg C⋅m−1. Waters with a low humic acid fraction (generally low-DOC waters) tend to have SUVAs that are less than 2 L/mg C⋅m−1,
whereas waters with a high humic acid fraction have SUVAs between 3 and 5 L/mg
C⋅m−1. A higher SUVA value means that the DOC of the water will tend to control
the coagulant dosage and relatively high removals of DOC can be expected (50 to 80
percent). When the SUVA is less than 3 L/mg C⋅m−1, the effect of the DOC on the
coagulant dosage may be negligible, and relatively low removal percentages (20 to 50
percent) are likely (Edzwald and Van Benschoten, 1990).


6.5

COAGULATION AND FLOCCULATION

USEPA’s Enhanced Coagulation Requirement. The USEPA’s 1998 Stage 1 Disinfection By-Products Rule (DBPR) requires the use of an NOM removal strategy
called “enhanced coagulation” to limit the formation of all DBPs. The requirement
applies to conventional water treatment facilities that treat surface water or groundwater that is under the influence of surface water. The amount of TOC a plant must
remove is based on the raw water TOC and alkalinity.
Enhanced coagulation ties the TOC removal requirement to the raw water alkalinity to avoid forcing a utility to add high dosages of hydrolyzing metal salt (HMS)
coagulants to reduce the pH to between 5 and 6, a range where HMS coagulants frequently appear to be most efficient. It also recognizes that higher TOC removal is
usually possible when the raw water TOC concentration is relatively high and the
fraction of the NOM that is more readily removed by HMS coagulants is typically
greater. The matrix in Table 6.1 gives Stage 1 DBPR’s required TOC removal percentages.
The application of Table 6.1 is illustrated by the following example. A plant’s

source water has a TOC of 3.5 mg C/L and an alkalinity of 85 mg/L as CaCO3.
According to the table, the required TOC removal is 25.0 percent. The TOC of the
water before the application of chlorine would have to be less than 2.6 mg C/L, calculated using the relationship 2.6 = 3.5 × (1 − 0.25).
The regulatory negotiators who formulated the enhanced coagulation requirement were concerned that coagulant chemical costs might be excessive for utilities
that treat water with a high fraction of NOM that is not amenable to removal by
coagulants. For these plants, the removal requirements of Table 6.1 may be infeasible. The Rule allows them to use a jar test procedure to determine an appropriate,
alternative TOC removal requirement (White et al., 1997).
The alternative TOC removal requirement is determined by performing jar tests
on at least a quarterly basis for one year. In these tests, alum or ferric coagulants are
added in 10 mg/L increments until the pH is lowered to a target pH value that varies
with the source water alkalinity. For the alkalinity ranges of 0 to 60, more than 60 to
120, more than 120 to 240, and more than 240 mg/L as CaCO3, the target pH values
are 5.5, 6.3, 7.0, and 7.5, respectively. When the jar test is complete, the residual TOC
concentration is plotted versus the coagulant dosage (in mg coagulant/L) and the
alternative TOC percentage is found at the point called the “point of diminishing
returns,” or PODR. The PODR is the coagulant dosage where the slope of the TOCcoagulant dosage plot changes from greater than 0.3 mg C/10 mg coagulant to less
than 0.3 mg C/10 mg coagulant. If the plot does not yield a PODR, then the water is
considered to be not amenable to enhanced coagulation and the primary agency
may grant the system a waiver from the enhanced coagulation requirement. Details
of the jar test procedure are given in the USEPA’s Guidance Manual for Enhanced
Coagulation and Enhanced Precipitative Softening (USEPA, 1999).
TABLE 6.1 Required Percent Removals of Total Organic Carbon by Enhanced Coagulation
in the 1998 Stage 1 Disinfection By-Products Rule
Source water alkalinity
(mg/L as CaCO3)
Source water total organic carbon (mg C/L)

0–60

>60–120

>120

>2.0–4.0
>4.0–8.0
>8.0

35
45
50

25
35
40

15
25
30


6.6

CHAPTER SIX

The Stage 1 DBPR provides alternative compliance criteria for the enhanced coagulation, treatment technique requirement. The six criteria are listed below:
1. The system’s source water TOC is less than 2.0 mg C/L.
2. The system’s treated-water TOC is less than 2.0 mg C/L.
3. The system’s source water TOC is less than 4.0 mg C/L, the source water alkalinity is more than 60 mg/L as CaCO3, and the system is achieving TTHM less than
40 µg/L and HAA5 (haloacetic acids) less than 30 µg/L.
4. The system’s TTHM is less than 40 µg/L, HAA5 is less than 30 µg/L, and only

chlorine is used for primary disinfection and maintaining a distribution system
residual.
5. The system’s source water SUVA prior to any treatment is less than or equal to
2.0 L/(mg⋅m−1).
6. The system’s treated-water SUVA is less than or equal to 2.0 L/(mg⋅m−1).
The measurements used to test compliance with criteria 1, 2, 5, and 6 are made
monthly and a running annual average is calculated quarterly. Compliance with criteria 3 and 4 is based on monthly measurements of TOC, alkalinity, quarterly measurements of TTHMs, and HAA5, and the running annual average is calculated
quarterly.
Particles
Particles in natural water vary widely in origin, concentration, size, and surface
chemistry. Some are derived from land-based or atmospheric sources (e.g., clays and
other products of weathering, silts, pathogenic microorganisms, asbestos fibers, and
other terrestrial detritus and waste discharge constituents), and some are produced
by chemical and biological processes within the water source (e.g., algae, precipitates
of CaCO3, FeOOH, MnO2, and the organic exudates of aquatic organisms). Certain
toxic metals and SOCs are associated with solid particles, so coagulation for particle
aggregation can be important in the removal of soluble, health-related pollutants.
Particle size may vary by several orders of magnitude, from a few tens of
nanometers (e.g., viruses and high-molecular-weight NOM) to a few hundred
micrometers (e.g., zooplankton). All can be effectively removed by properly
designed and operated coagulation, floc separation, and filtration facilities. The very
important cysts and oocysts of pathogenic protozoa (e.g., Giardia and Cryptosporidium) are ovoid particles with overall dimensions in the 4 to 12 µm (micrometer)
range. A comparison of the size spectra of waterborne particles and filter pores is
shown in Figure 6.1.
The smallest particles, those with one dimension less than 1 µm, are usually called
“colloidal,” and those that are larger than this limit are said to be “suspended.” The
operational definition of “dissolved” and “suspended” impurities is frequently
established by a 0.45 µm pore-size membrane filter, but colloidal particles can be
smaller than this dimension. The effect of gravity on the transport of colloidal particles tends to be negligible compared with the diffusional motion caused by interaction with the fluid (Brownian motion) and, compared with suspended particles,
colloidal particles have significantly more external surface area per unit mass.

Measuring Particle Concentration. The principal methods for measuring the performance of particle removal processes in water treatment systems are turbidity and


COAGULATION AND FLOCCULATION

6.7

FIGURE 6.1 Size spectrum of waterborne particles and filter
pores (from Stumm and Morgan, 1981).

particle counting. Both techniques have limitations, and, consequently, a single
method may not provide all the information needed to successfully monitor and
control process performance.
Turbidity is measured using an instrument called a turbidimeter, or nephelometer, that detects the intensity of light scattered at one or more angles to an incident
beam of light. Light scattering by particles is a complex process and the angular distribution of scattered light depends on a number of conditions including the wavelength of the incident light and the particle’s size, shape, and composition (Sethi et
al., 1996). Consequently, it is difficult to correlate the turbidity with the amount,
number, or mass concentration of particles in suspensions.
When the turbidity measurement is used for regulatory purposes, it should theoretically be possible to take a given suspension and measure its turbidity at any
water treatment facility and obtain an unbiased result that is reasonably close to the
average turbidity measured at all other facilities. To achieve reasonable agreement,
three factors must be considered: the design of the instruments, the material used to
calibrate the instrument, and the technique used to make the measurement. Given
this need, turbidimeter design, calibration, and operation criteria have been developed using the consensus process by a number of organizations, including Standard
Methods for the Examination of Water and Wastewater (Section 2130), the American
Society for Testing Materials (ASTM, Method D 1889), the International Standards
Organization (ISO 7027-1948E), and the United States Environmental Protection
Agency (Method 180.1). However, standardization is a difficult and imperfect process, and it has been shown (Hart et al., 1992) that instruments designed and calibrated using the criteria of these standards can give significantly different responses.
Turbidity measurements were first used to maintain the aesthetic quality of
treated water. In 1974, after the passage of the Safe Drinking Water Act, the USEPA
lowered the limit for filtered water to one nephelometric turbidity unit (1 NTU)

with the explanation that particles causing turbidity can interfere with the disinfec-


6.8

CHAPTER SIX

tion process by enmeshing and, therefore, protecting microbiological contaminants
from chemical disinfectants such as chlorine. Today, the turbidity measurement is
also used to assess filter performance. It is viewed as an important indicator of the
extent to which disinfectant-resistant pathogens have been removed by the filtration
process. Filtered-water turbidity criteria must be met before the protozoan
cyst/oocyst and virus removal credits allowed by the Surface Water Treatment Rule
of the 1986 Amendments of the Surface Water Treatment Rule can be applied (Pontius, 1990).
Particle-counting instruments are becoming widely used in the drinking water
industry, especially for monitoring and controlling filtration process performance.
Plants use them to detect early filter breakthrough and to maintain plant performance at a high level. On-line devices that continuously measure particle concentrations in preselected size ranges at various points in the treatment system are
especially important. Batch sampling devices are also used. Two types of particlecounting/sizing sensors are important in water treatment applications: light-blocking
(light-obscuration) devices and light-scattering devices (Hargesheimer et al., 1992;
Lewis et al., 1992). At the present time, instruments with light-blocking-type sensors
are more common.
The types of particle-counting instruments used in water treatment plants have
limitations (Hargesheimer et al., 1992). Most do not detect particles smaller than
about 1 µm, and therefore, they must be used in conjunction with turbidimeters that
do detect these smaller particles. Differences in the optical characteristics of the sensors make achieving direct count and size agreement between instruments difficult.
There are no industrywide standards for sensor resolution or for particle counting
and sizing accuracy. For a given particle suspension, it is not possible to make similar
sensors yield identical particle counts and sizes. Until this is feasible, the regulatory
use of particle count measurements will be limited.

STABILITY OF PARTICLE SUSPENSIONS
In water treatment, the coagulation process is used to increase the rate or kinetics of
particle aggregation and floc formation. The objective is to transform a stable suspension [i.e., one that is resistant to aggregation (or attachment to a filter grain)]
into an unstable one. Particles that may have been in lake water for months or years
as stable, discrete units can be aggregated in an hour or less following successful
destabilization. The design and operation of the coagulation process requires proper
control of both particle destabilization and the subsequent aggregation process.
As particles in a suspension approach one another, or as a particle in a flowing
fluid approaches a stationary surface such as a filter grain, forces arise that tend to
keep the surfaces apart. Also, there are forces that tend to pull the interacting surfaces together. The most well-known repulsive force is caused by the interaction of
the electrical double layers of the surfaces (“electrostatic” stabilization). The most
important attractive force is called the London–van der Waals force. It arises from
spontaneous electrical and magnetic polarizations that create a fluctuating electromagnetic field within the particles and in the space between them. These two types
of forces, repulsive and attractive, form the basis of the Derjaguin, Landau, Verwey,
and Overbeek (DLVO) theory of colloid stability. Other forces include those associated with the hydration of ions at the surfaces (a repulsive force) and the presence
of adsorbed polymers, which can cause either repulsion (“steric” interaction) or
attraction (“polymer bridging”). As particles approach one another on a collision


6.9

COAGULATION AND FLOCCULATION

course, the fluid between them must move out of the way. The repulsive force caused
by this displacement of fluid is called hydrodynamic retardation.

Electrostatic Stabilization
Origins of Surface Charge. Most particles in water, mineral and organic, have electrically charged surfaces, and the sign of the charge is usually negative (Niehof and
Loeb, 1972; Hunter and Liss, 1979). Three important processes for producing this
charge are considered in the following discussion. First, surface groups on the solid

may react with water and accept or donate protons. For an oxide surface such as silica, the surface site might be indicated by the symbol ϵSiOH and the surface site
ionization reactions by
ϵSiOH2+ ⇔ ϵSiOH + H+

(6.2a)

ϵSiOH ⇔ ϵSiO− + H+

(6.2b)

An organic surface can contain carboxyl (COO−) and amino (NH3+) groups that
become charged through ionization reactions as follows:
COO−

COOH
⇔

R

R

(6.3a)

NH3+

NH3+

COO−

COO−

⇔

R
NH3+

R

(6.3b)
NH2

In these reactions, the surface charge on a solid particle depends upon the concentration of protons ([H+]) or the pH (= −log [H+]) in the solution. As the pH increases
(i.e., [H+] decreases), Eqs. 6.2 and 6.3 shift to the right and the surface charge
becomes increasingly negative. Silica is negatively charged in water with a pH above
2; proteins contain a mixture of carboxyl and amino groups and usually have a negative charge at a pH above about 4. The adsorption of NOM onto particles can be
responsible for site behavior like that shown previously.
Second, surface groups can react in water with solutes other than protons. Again,
using the ϵSiOH surface groups of silica,
ϵSiOH + Ca2+ ⇔ ϵSiOCa+ + H+

(6.4)

ϵSiOH + HPO4 ⇔ ϵSiOPO3 H + OH
2−

−

−

(6.5)

These surface complex formation reactions involve specific chemical reactions
between chemical groups on the solid surface (e.g., silanol groups) and adsorbable
solutes (e.g., calcium and phosphate ions). Surface charge is again related to solution
chemistry.


6.10

CHAPTER SIX

Third, a surface charge may arise because of imperfections within the structure of
the particle; this is called isomorphic replacement, or substitution. It is responsible
for a substantial part of the charge on many clay minerals. Clays are layered structures and in these structures sheets of silica tetrahedra are typically cross-linked
with sheets of alumina octahedra. The silica tetrahedra have an average composition
of SiO2 and may be depicted as shown in Figure 6.2(a). If an Al atom is substituted
for an Si atom during the formation of this lattice, a negatively charged framework
results [Figure 6.2(b)].
Similarly, a divalent cation such as Mg(II) or Fe(II) may substitute for an Al(III)
atom in the aluminum oxide octahedral network, also producing a negative charge.
The sign and magnitude of the charge produced by such isomorphic replacements
are independent of the characteristics of the aqueous phase after the crystal is
formed.
The Electrical Double Layer. In a colloidal suspension, there can be no net imbalance in the overall electrical charge; the primary charge on the particle must be
counterbalanced in the system. Figure 6.3 shows schematically a negatively charged
particle with the counterbalancing cloud of ions (the “diffuse layer”) around the
particle. Because the particle is negatively charged, excess ions of opposite charge
(positive) accumulate in this interfacial region. Ions of opposite charge accumulating in the interfacial region, together with the primary charge, form an electrical
double layer. The diffuse layer results from electrostatic attraction of ions of opposite charge to the particle (“counterions”), electrostatic repulsion of ions of the same
charge as the particle (“coions”), and thermal or molecular diffusion that acts
against the concentration gradients produced by the electrostatic effects. The formation of diffuse layers is shown in Figures 6.3 and 6.4(a).

Because of the primary charge, an electrostatic potential (voltage) exists between
the surface of the particle and the bulk of the solution. This electric potential can be
pictured as the electric pressure that must be applied to bring a unit charge having

(a)

(b)

FIGURE 6.2 SiO2 structure. (a) With no net charge.
(b) With −1 net charge.


COAGULATION AND FLOCCULATION

6.11

FIGURE 6.3 Negatively charged particle, the diffuse double layer, and the location of the zeta
potential.

the same sign as the primary charge from the bulk of the solution to a given distance
from the particle surface, shown schematically in Figure 6.4(b). The potential has a
maximum value at the particle surface and decreases with distance from the surface.
This decrease is affected by the characteristics of the diffuse layer, and, thus, by the
type and concentration of ions in the bulk solution.


6.12

CHAPTER SIX

(a)

(c)

(b)

(d)

FIGURE 6.4 Schematic representations of (a) the diffuse double layer; (b)
the diffuse layer potential; (c and d) two cases of particle-particle interaction
energies in electrostatically stabilized colloidal systems.

Dynamic Aspects of the Electrical Double Layer. The interaction of two particles
can be evaluated in terms of potential energy (i.e., the amount of energy needed to
bring two particles from infinite separation up to a given separation distance). If the
potential energy is positive, the overall interaction is unfavorable (repulsive)
because energy must be provided to the system. If the potential energy is negative,
the net effect is attractive.
When particles are forced to move in a fluid or when the fluid is forced to move
past a stationary particle, some of the charge that balances the surface charge moves
with the fluid and some of it does not. Thus, the electrical potential at the plane of
shear (which is the boundary between the charge that remains with the particle and
the charge that does not) is less than the electrostatic potential at the surface of the
particle.The exact location of the slipping plane is not known, and, therefore, it is difficult to relate the potential measured at the plane of shear (by electrokinetic techniques such as electrophoresis or streaming current) to the surface potential.
Lyklema (1978) locates the slipping plane at the outer border of the Stern layer as
shown in Figure 6.3.


COAGULATION AND FLOCCULATION

6.13

The mathematical treatment of the double layer, shown schematically in Figures
6.3 and 6.4, was developed independently by Gouy and Chapman. The result is
called the Gouy-Chapman model; details of it have been presented by Verwey and
Overbeek (1948). In this model, electrostatic or coulombic attraction and diffusion
are the interacting processes responsible for the formation of the diffuse layer. Ions
are treated as point (dimensionless) charges and have no other physical or chemical
characteristics.
When two similar colloidal particles approach each other, their diffuse layers
begin to overlap and to interact. This comingling of charge creates a repulsive potential energy, ΨR, that increases in magnitude as the distance separating the particles
decreases. Figure 6.4 (c and d) plots the repulsive energy of interaction as a function
of the distance between the interacting surfaces.
Attractive forces (van der Waals forces) exist between all types of particles, no
matter how dissimilar they may be. These forces arise from interactions among
dipoles, either permanent or induced, in the atoms composing the interacting surfaces and the water. The magnitude of the attractive energy of interaction, ΨA,
decreases with increasing separation distance. Schematic curves of the van der Waals
attractive potential energy of interaction are shown in Figure 6.3 (c and d). Unlike
electrostatic repulsive forces, the van der Waals attractive forces are essentially independent of the composition of the solution; they depend upon the kind and number
of atoms in the colloidal particles and the continuous phase (water).
Summation of ΨR and ΨA yields the net interaction energy between two colloidal
particles. This sum with the proper signs is (ΨR − ΨA) and is shown schematically as
a function of separation distance in Figure 6.4 (c and d). The force acting on the particles is the derivative d(ΨR − ΨA)/ds, where s is the separation distance. When electrostatic repulsion dominates during particle-particle interactions, the suspension is
said to be electrostatically stabilized and to undergo only “slow” coagulation. Electrostatic stabilization is fundamental to the current understanding of coagulation in
water treatment. For additional insight into electrostatic stabilization, the mathematical treatment of Verwey and Overbeek (1948) and a summary by Lyklema
(1978) should be consulted. Valuable summaries of the methods, mathematics, and
meanings of models for electrostatic stabilization are also contained in texts by
Morel (1983), Stumm and Morgan (1980), and Elimelech et al. (1995).
The Gouy-Chapman model provides a good qualitative picture of the origins of
electrostatic stabilization. It does not provide a quantitative, predictive tool for such

important characteristics as coagulant requirements and coagulation rates. Its principal drawback lies in the characterization of all ions as point charges. This allows for
physical description of electrostatics, but omits description of chemical interactions.
For example, the sodium ion (Na+), the dodecylamine ion (C12H25NH3+), and the proton (H+) have identical charges but are not identical coagulants.
Quantitative models for the surface chemistry of oxides and other types of surfaces with ionizable surface sites are available. Some have been incorporated into
widely used computer programs such as MINEQL, allowing self-consistent calculations to be made simultaneously for surface and solution. The most popular are the
surface complexation models, which describe the formation of charge, potential, and
the adsorption of ions at the particle-water interface.The fundamental concept upon
which all surface complexation models are based is that adsorption takes place at
defined coordination sites (surface hydroxyl groups present in finite number) and
that adsorption reactions can be described quantitatively by mass action expressions
(Goldberg, Davis, and Hem, 1996). The various models use different assumptions for
the electrostatic interaction terms. For example, some use a simple linear relationship between charge and potential at the surface [the “constant capacitance model”


6.14

CHAPTER SIX

of Schindler and Stumm (Stumm, 1992)], and others use more elaborate relationships such as the triple-layer model of Davis and coworkers (1978).All of the surface
complexation models reduce to a similar set of simultaneous equations that are typically solved numerically: (1) mass action equations for all surface reactions, (2) mass
balance equations for surface hydroxyl groups, (3) equations for calculation of surface charge, (4) a set of equations that describes the charge and potential relationships of the electrical double-layer. Attempts have been made, with limited success,
to use these models to predict the effects of solution and suspension variables on
flocculation efficiency when aluminum salt coagulants are used (Letterman and
Iyer, 1985) and on the rate of coagulation of iron oxide particles (Liang and Morgan,
1988).
Secondary Minimum Aggregation. In certain systems, the kinetic energy and electrical double-layer characteristics of the interacting particles may be such that aggregation takes place in what is called a secondary minimum [see Figure 6.4 (d)]. Under
this condition, the particles are held in proximity (at least momentarily) by van der
Waals attraction but remain nanometers apart due to the repulsive force associated
with the interacting electrical double layers. If the kinetic energy of the interacting
particles is increased, the repulsive force may not be great enough to limit the close

approach of the particles, and aggregation will tend to occur in the primary minimum,
with the surfaces of the particles close together. In this case, the separation distance
in Figure 6.4 (c) is reduced until the quantity (ΨR − ΨA) becomes less than Ψmax.

Steric Stabilization
Steric stabilization can result from the adsorption of polymers at solid-water interfaces. Large polymers can form adsorbed segments on a solid surface with loops and
tails extending into the solution (Lyklema, 1978) as illustrated in Figure 6.5.
Adsorbed polymers can be either stabilizing or destabilizing, depending on the relative amounts of polymer and solid particles, the affinities of the polymer for the solid
and for water, electrolyte type and concentration, and other factors. A stabilizing
polymer may contain two types of groups, one of which has a high affinity for the

FIGURE 6.5 Illustration of adsorbed polymer configuration with
loops, trains, and tails (from Lyklema, 1978).


COAGULATION AND FLOCCULATION

6.15

FIGURE 6.6 Two possible repulsive interactions of adsorbed polymer
layers in sterically stabilized colloidal systems (Gregory, 1978). Shaded
areas are zones occupied by polymers. Zone thickness relative to particle
size is arbitrary.

solid surface and a second, more hydrophilic group, that extends into the solution
(Gregory, 1978). The configuration of such an interfacial region is difficult to characterize either theoretically or experimentally. This, in turn, prevents quantitative
formulation of the interaction forces between two such interfacial regions during a
particle-particle encounter in coagulation. Some useful qualitative descriptions can,
however, be made.
Gregory (1978) summarized two processes that can produce a repulsive force

when two polymer-coated surfaces interact at close distances.These are illustrated in
Figure 6.6. First, the adsorbed layers can each be compressed by the collision, reducing the volume available for the adsorbed molecules. This reduction in volume
restricts the movement of the polymers (a reduction in entropy) and causes a repulsion between the particles. Second, and more frequently, the adsorbed layers may
interpenetrate on collision, increasing the concentration of polymer segments in the
mixed region. If the extended polymer segments are strongly hydrophilic, they can
prefer the solvent to reaction with other polymer segments. An overlap or mixing
then leads to repulsion. These two processes are separate from and in addition to the
effects of polymer adsorption on the charge of the particles and the van der Waals
interaction between particles. Charged polymers (polyelectrolytes) can alter particle
charge; organic polymers can also reduce the van der Waals attractive interaction
energy. Steric stabilization is widely used in the manufacture of industrial colloids,
such as paints and waxes. Natural organic materials such as humic substances are
ubiquitous in water supplies. They are anionic polyelectrolytes, absorb at interfaces,
can be surface active, and may contribute to particle stability by steric effects
(O’Melia, 1995).

COAGULANTS
Introduction
Coagulants are widely used in water treatment for a number of purposes. Their principal use is to destabilize particulate suspensions and enhance the rate of floc formation. Hydrolyzing metal salt coagulants are also used to form flocculent
precipitates that adsorb NOMs and certain inorganic materials, such as phosphates,
arsenic compounds, and fluoride.
For many years most water treatment plants that required a coagulant used the
hydrolyzing metal salt, aluminum sulfate (alum). Organic polymers (polyelectrolytes) came into widespread use in the 1960s. In the last 10 years, the number of


6.16

CHAPTER SIX

coagulant products used in water treatment has grown substantially. Today, the list

includes ferric iron salts and prehydrolyzed metal salts plus an assortment of chemical mixtures and products supplemented with additives. In some countries, organic
compounds derived from natural materials are used as coagulants.

Destabilization Mechanisms
Destabilization is the process in which the particles in a stable suspension are modified to increase their tendency to attach to one another (or to a stationary surface
like a filter grain or deposit).The aggregation of particles in a suspension after destabilization requires that they be transported toward one another. In filtration they
must be transported to the stationary filter surface. Transport processes in suspensions are discussed in this chapter in the section on flocculation. Transport in filtration is discussed in Chapter 8.
Double-Layer Compression. The classical method of colloid destabilization is
double-layer compression. To affect double-layer compression, a simple electrolyte
such as NaCl is added to the suspension. The ions that are opposite in sign to the net
charge on the surface of the particles enter the diffuse layer surrounding the particle. If enough of these counterions is added, the diffuse layer is compressed, reducing the energy required to move two particles of like surface charge into close
contact. Destabilization by double-layer compression is not a practical method for
water treatment, because the salt concentrations required for destabilization may
approach that of seawater and, in any case, the rate of particle aggregation would
still be relatively slow in all but the most concentrated suspensions. Double-layer
compression, however, is an important destabilization mechanism in certain natural
systems [e.g., estuaries (O’Melia, 1995)].
Surface Charge Neutralization. Destabilization by surface charge neutralization
involves reducing the net surface charge of the particles in the suspension.As the net
surface charge is decreased the thickness of the diffuse layer surrounding the particles is reduced and the energy required to move the particles into contact is minimized.
Two processes are used to accomplish surface charge neutralization. In the first,
coagulant compounds that carry a charge opposite in sign to the net surface charge
of the particles are adsorbed on the particle surface. (In some cases, the coagulant is
a very small particle that deposits on the particle surface.) The coagulants used to
accomplish this usually have a strong tendency to adsorb on (attach to) surfaces.
Examples include the synthetic and natural organic polyelectrolytes and some of the
hydrolysis products formed from hydrolyzing metal salt coagulants. The tendency
for these compounds to adsorb is usually attributable to both poor coagulantsolvent interaction and a chemical affinity of the coagulant, or chemical groups on
the coagulant, for the particle surface. Most of the coagulants that are used for
charge neutralization can adsorb on the surface to the point that the net surface

charge is reversed and, in some cases, increased to the point that the suspension is
restabilized.
Adjustment of the chemistry of the solution can be used to destabilize some common types of suspensions by reducing the net surface charge of the particle surfaces.
For example, when most of the surface charge is caused by the ionization of surface
sites (see Eq. 6.1), pH adjustment with acid or base may lead to destabilization. For
some surfaces, such as positively charged oxides and hydroxides, the adsorption of


COAGULATION AND FLOCCULATION

6.17

simple multivalent anions (such as sulfate and phosphate) or complex polyvalent
organic compounds (such as humic materials), will reduce the positive charge and
destabilize the suspension.
Heterocoagulation is a destabilization mechanism that is similar to the process of
surface charge neutralization by the adsorption of oppositely charged soluble
species. However, in this case, the process involves one particle depositing on
another of opposite charge. For example, large particles with a high negative surface
charge may contact smaller particles with a relatively low positive charge. Because
the particles have opposite surface charge, electrostatic attraction enhances particleparticle interaction. As the stabilizing negative charge of the larger particles is
reduced by the deposited positive particles, the suspension of larger particles is
destabilized.
Adsorption and Interparticle Bridging. Destabilization by bridging occurs
when segments of a high-molecular-weight polymer adsorb on more than one particle, thereby linking the particles together. When a polymer molecule comes into
contact with a colloidal particle, some of the reactive groups on the polymer
adsorb on the particle surface and other portions extend into the solution. If a second particle with open surface is able to adsorb the extended molecule, then the
polymer will have formed an interparticle bridge. The polymer molecule must be
long enough to extend beyond the electrical double layer (to minimize doublelayer repulsion when the particles approach) and the attaching particle must have
available surface. The adsorption of excess polymer may lead to restabilization

of the suspension. Ions such as calcium are known to affect the bridging process, apparently by linking sites on interacting polymer chains (Black et al., 1965;
Lyklema, 1978; Dentel, 1991).

Hydrolyzing Metal Salt (HMS) Coagulants
The most widely used coagulants in water treatment are sulfate or chloride salts that
contain the metal ions Al3+ or Fe3+. In aqueous solutions, these small, highly positive
ions form such strong bonds with the oxygen atoms of six surrounding water
molecules that the oxygen-hydrogen atom association in the water molecules is
weakened, and hydrogen atoms tend to be released to the solution (see Figure 6.7).
This process is known as hydrolysis and the resulting aluminum and ferric hydroxide

FIGURE 6.7 Deprotonation of the aquo aluminum ion, initial
step in aluminum hydrolysis (from Letterman, 1991).


6.18

CHAPTER SIX

FIGURE 6.8 Aluminum hydrolysis products (from Letterman, 1991).

species are called hydrolysis products. The water molecules are frequently omitted
when writing the chemical formula for these species, but their role is important in
determining species behavior.
The chemistry of aluminum and iron hydrolysis reactions and products is complex and not completely understood.As hydrolysis proceeds, and if there is sufficient
total metal ion in the system, simple mononuclear products can form complex
polynuclear species, which in turn can form microcrystals and the metal hydroxide
precipitate (see Figure 6.8). Hydrolysis products can adsorb (and continue to
hydrolyze) on many types of particulate surfaces. Hydrolysis products tend to react
with the higher-molecular-weight fraction of NOM. Soluble hydrolysis products may

bind with NOM functional groups and small, positively charged microcrystalline
metal hydroxide particles can be stabilized by the adsorption of a negatively charged
coating of NOM (Kodama and Schnitzer, 1980; Bertsch and Parker, 1996).
The solubility of the metal hydroxide precipitate is one factor that must be considered in maximizing coagulant performance and in minimizing the amounts of
residual Al and Fe in treated water. At low pH, the dissolution of the metalhydroxide precipitate produces positively charged, soluble hydrolysis products and
the aquo-metal ion (Al3+ and Fe3+)1. At high pH, the negatively charged, soluble
hydrolysis products Al(OH)4− and Fe(OH)4− are formed. These species are tetrahedral rather than octahedral, so no further deprotonation can occur. The minimumsolubility pH of aluminum hydroxide precipitate at 25°C is approximately 6.3
(Figure 6.9); for ferric hydroxide it is about 8 (Figure 6.10). The pH of minimum solubility increases with decreasing temperature. At 4°C, the pH of minimum solubility
of freshly precipitated aluminum hydroxide is approximately 6.8 (Van Benschoten
and Edzwald, 1990).
1

From this point on, water molecules are omitted from the formulas for hydrolysis products.


COAGULATION AND FLOCCULATION

6.19

FIGURE 6.9 Solubility diagram for amorphous, freshly precipitated aluminum hydroxide. Water molecules are omitted in species notation.

The solubility diagrams of Figures 6.9 and 6.10 were plotted using the formation
constants (βx,y) and solubility constants (Ksp) listed in Table 6.2. The ionic strength
was assumed to be 0.001 M. The formation constants in Table 6.2 are for reactions
and mass action expressions of the form
x M3+ + y H2O ⇔ Mx(OH)y(3x − y)+ + y H+

FIGURE 6.10 Solubility diagram for amorphous, freshly precipitated ferric hydroxide. Water molecules are omitted in species notation.

(6.6)

6.20

CHAPTER SIX

and
− y)+
[Mx(OH)(3x
] [H+]y
y
βx,y = ᎏᎏᎏ
[M3+]x

(6.7)

The solubility constants are for reactions and solubility product relationships of the
form
M(OH)3(s) + 3 H+ ⇔ M3+ + 3 H2O

(6.8)

[M3+]
Ksp = ᎏ
[H+]3

(6.9)

and

A number of researchers have assumed that the important mononuclear Al hydrolysis products include Al(OH)2+, Al(OH)2+, Al(OH)30, and Al(OH)4− (Bertsch and
Parker, 1996). Others, including Letterman and Driscoll (1994), Hayden and Rubin
(1974), and VanBenschoten and Edzwald (1990), have reported that the effect of pH
on the solubility of freshly precipitated aluminum hydroxide can be accurately predicted using just the aquo-Al species (Al3+) and two mononuclear hydrolysis products [Al(OH)2+, Al(OH)4−]. This assumption was used to plot Figure 6.9.
The dotted lines plotted in Figure 6.9 give the concentrations of each Al species
when it is in equilibrium with the Al(OH)3 precipitate. The equation of each line was
derived by combining Equations 6.6 and 6.8 and using the appropriate equilibrium
constants corrected for ionic strength. For example, the equation for log [Al(OH)2+]
when the ionic strength I = 0.001 M is
log[Al(OH)2+] = logβ1,1 + logKsp + 2logH+ = 5.57 − 2pH

(6.10)

Values of log β1,1 and log Ksp for I = 0.001 M are listed in Table 6.2.
As an Al salt solution is titrated with strong base, hydrolysis of the Al increases,
forming a varying combination of hydrolysis products.The distribution of the Al(III)
species at equilibrium depends on the pH and the total Al concentration (AlT). The
effect of pH on the fractional distribution of Al hydrolysis products was calculated
TABLE 6.2 Formation Constants and Solubility Constants Used to Plot
Solubility Diagrams for Amorphous Aluminum Hydroxide and Amorphous
Ferric Hydroxide Precipitates, T = 25°C and I = 0*
Hydrolysis product

pβx,y and pKsp

Reference

Amorphous Aluminum Hydroxide
Al(OH)2+
Al(OH)4−

Al(OH)3(s)

4.99 (5.07)
23.00 (23.12)
−10.50 (10.64)

Ball, Nordstrom, and Jenne, 1980
Ball, Nordstrom, and Jenne, 1980
Ball, Nordstrom, and Jenne, 1980

Amorphous Ferric Hydroxide
2+

Fe(OH)
Fe(OH)2+
Fe(OH)4−
Fe(OH)3(s)

2.2
5.7
21.6
−3.6

Stumm and Morgan, 1981
Stumm and Morgan, 1981
Stumm and Morgan, 1981
Stumm and Morgan, 1981

* The values in parentheses have been corrected for ionic strength, I = 0.001 M.

COAGULATION AND FLOCCULATION

6.21

using Equations 6.7 and 6.9 and the Al hydrolysis products and equilibrium constants (I = 0.001 M) listed in Table 6.2. The results are plotted in Figures 6.11 (a and
b) for total Al concentrations (AlT) of 1 × 10−6 M [0.03 mg Al/L or 0.33 mg alum
(Al2(SO4)3 14H2O)/L] and 3 × 10−4 M [8.1 mg Al/L or 89 mg alum/L], respectively.
According to Figure 6.11(a) for AlT = 1 × 10−6 M, a small amount of Al(OH)3 precipitate forms between pH = 5.8 and 6.5.At pH = 6.1, where the amount of precipitate
is maximum, Al(OH)4− and Al(OH)2+ are the principal hydrolysis products, representing over 70 percent of the total Al in the system. For the higher AlT (3 × 10−4 M)
of Figure 6.11(b), Al(OH)3 precipitate begins to form at pH = 4.7 and by pH = 9.0 all
of it has dissolved to form Al(OH)4−. Over most of this pH interval (from pH 5.5 to
8.5), the precipitate is the predominant hydrolysis product.

(a)

(b)

FIGURE 6.11 Effect of pH and total aluminum concentration on
the speciation of Al(III): (a) AlT = 1 × 10−6 M, and (b) AlT = 3 × 10−4 M.


6.22

CHAPTER SIX

TABLE 6.3 Examples of Polynuclear Al(III) Species
Polynuclear species
Al8(OH)204+
Al13O4(OH)247+

Al14(OH)3210+

[OH]/[Al]

Reference

2.5
2.46
2.29

Hayden and Rubin (1974)
Bertsch (1986)
Turner (1976)

The literature contains convincing evidence that polynuclear hydrolysis products
are formed under certain conditions. Table 6.3 lists three examples of the many
polynuclear forms that have been suggested. Most have molar ratios of OH to Al in
the 2.3 to 2.5 range. Nuclear magnetic resonance (NMR) spectroscopy (Bertsch and
Parker, 1996) has provided strong evidence for the existence of Al13O4(OH)247+.
Bertsch and Parker conclude that polynuclear species are metastable intermediates
in solutions that reach or exceed a critical degree of supersaturation with respect to
a solid phase such as gibbsite (Al2O3) or amorphous Al(OH)3.
In polynuclear (and microcrystalline) hydrolysis products most of the metal
atoms are interconnected by double-hydroxide bonds. Double-hydroxide bonds are
not easily broken. Therefore, when complex metal hydroxide species are transferred
from the solution in which they were formed to one in which they could not be
formed and are not stable, the transition to the new set of species is typically slow.
For example, if a concentrated solution of polynuclear species is added to a suspension as a coagulant, it is likely that a significant amount of time (days to months) will
be required for a new stable distribution of hydrolysis products to form in the system; that is, the polynuclear species will tend to persist. Consequently, particle or
NOM removal diagrams prepared using preformed polynuclear species (Dempsey,

Ganho, and O’Melia, 1984) may show effective coagulation in areas where the corresponding solubility diagram based on equilibrium conditions [e.g., Figure 6.9 for
Al(OH)3] does not show the formation of destabilizing hydrolysis products.
Bertsch and Parker (1996) have argued that it is not appropriate to explain
polynuclear species formation and stability using complex formation reactions (such
as Eq. 6.6). These reactions use electrolyte theory and conventions to describe
single-ion activities. It is not correct to use electrolyte theory to describe the activity
of large, polyvalent, polynuclear species, and, consequently, it is not thermodynamically correct to include polynuclear species in the calculations used to plot solubility
diagrams such as Figure 6.9 and 6.10.
Polynuclear species are typically detected at high total Al concentrations
(>1 × 10−3 M) and at pH values between 7 (on the right) and the acid boundary of the
Al(OH)3 precipitate region of Figure 6.9 (on the left). Dispersed polynuclear species
cannot be detected in the presence of significant concentrations of destabilizing
multivalent anions, such as sulfate, or in solutions with high ionic strengths (de Hek,
Stol, and deBruyn, 1976).
Types of HMS Coagulants Used in Water Treatment
The essential HMS product groups are described below. The list is not comprehensive because new types of products are constantly entering the water treatment
market in this highly competitive business.
Simple Metal Salts. The simple HMS coagulants are aluminum sulfate (alum), ferric sulfate, and ferric chloride. These are sold as dry crystalline solids and as concen-


COAGULATION AND FLOCCULATION

6.23

trated (∼2 M) aqueous solutions. Alum is still the predominant HMS coagulant;
however, the iron salts are growing in importance.
Prehydrolyzed Metal Salts. As noted above, when HMSs are added to and diluted
in the water to be treated, the hydrolysis reaction produces hydrogen ions that react
with alkalinity species in the solution. If some of this acid is neutralized with base
when the coagulant is manufactured, the resulting product is a prehydrolyzed metal

salt coagulant solution. The degree to which the hydrogen ions produced by hydrolysis are preneutralized is called the basicity. The basicity is given by

΂ ΃

100
[OH]
Basicity(%) = ᎏ × ᎏ
3
[M]

(6.11)

where [OH]/[M] is the weighted average of the molar ratio of the bound hydroxide
to metal ion for all the metal hydrolysis products in the undiluted coagulant solution. For example, if a hypothetical product solution contained just one hydrolysis
product, the polynuclear species Al13O4(OH)247+, [OH]/[Al] would be effectively
equal to 2.46 = [24 + (2 × 4)]/13 and the basicity would be 82 percent = (100/3) × 2.46.
For commercial coagulant solutions, the basicity varies from ∼10 (low prehydrolysis)
to around 83 percent. As the basicity is increased beyond about 75 percent, it
becomes increasingly difficult to keep the metal hydroxide precipitate from forming
in the product solution during shipping and extended storage. AWWA Standard
B408-98, Liquid Polyaluminum Chloride, provides a laboratory method for determining the basicity of product solutions.
To avoid forming a precipitate in prehydrolyzed product solutions, the higher
basicity products are usually made with chloride as the dominant anion; multivalent
anions such as sulfate (SO4=) tend to destabilize positively charged polynuclear and
microcrystalline hydrolysis products. Prehydrolyzed metal salt coagulants made
with aluminum chloride are called polyaluminum chloride, or sometimes, polyaluminum hydroxychloride, or simply PACl. Use of the term polyaluminum (or polyiron for Fe salt products) is based on the assumption that the product solution
contains significant amounts of polynuclear metal hydrolysis products, which tends
to be true only for the higher basicity (70-plus percent) solutions. Some investigators
(Bottero et al., 1987) believe that the principal hydrolysis product in PACl solutions
with basicities greater than about 75 percent is Al13O4(OH)247+. Prehydrolyzed iron

solutions exist but are still a relatively uncommon commercial product. The significance of the basicity of iron salt coagulant solutions has been discussed by Tang and
Stumm (1987).
Metal Salts Plus Strong Acid. Several coagulant manufacturers prepare coagulant
solutions that contain the metal salt (e.g., alum) and an amount of strong acid, typically sulfuric acid. The typical acid-supplemented alum product (also called acidulated alum or acid alum) contains 5 percent to 20 percent (weight basis) 93 percent
sulfuric acid. Iron salt solutions are available that contain supplemental sulfuric acid.
For a given amount of metal ion, Al or Fe, added to the water, strong acid-fortified
products react with more alkalinity and depress the pH to a greater extent than the
nonfortified metal salt solutions.
Metal Salts Plus Additives. Metal salt coagulant solutions are available with special additives including phosphoric acid, sodium silicate, and calcium salts. Alum
with phosphoric acid has some of the characteristics of acid-supplemented alum, but
AlPO4 precipitate is formed when the solution is added to the water. Metal salt solu-


6.24

CHAPTER SIX

tions are also sold premixed with polyelectrolyte coagulant compounds such as
epichlorohydrin dimethylamine (epiDMA) and polydiallyl dimethylammonium
chloride (polyDADMAC).
Sodium Aluminate. Sodium aluminate (NaAlO2), a common chemical in papermaking, is sometimes used as a coagulant in water treatment, typically in combination with alum to treat waters with low alkalinity. The chemical properties of sodium
aluminate make it more difficult to handle than alum and the other metal salt products and this—and cost—have limited its use. It is a basic aluminum salt; when added
to water and aluminum hydroxide precipitate forms, the alkalinity and pH of the
solution tend to increase.
AlO2− + 2H2O ⇔ Al(OH)3(s) + OH−
A sodium aluminate dosage of 1 mg Al/L increases the alkalinity 0.037 milliequivalents per liter (meq/L) or 1.9 mg/L as CaCO3.

Impurities in HMS Coagulant Solutions
Most of the impurities in HMS coagulants are derived from the raw materials used
to make them. For example, alum is usually made by digesting an aluminum source

in sulfuric acid. Typical aluminum sources are bauxitic or high-aluminum clays, aluminum trihydrate, and high-purity bauxite. Impurities in the aluminum source tend
to appear in the alum product. The most significant contaminant in aluminum salt
coagulants is iron. Standard alum solution (4.2 percent Al) may contain 1000 ppm
Fe. Small amounts of heavy metals such as chromium and lead can be found in
standard-grade alum solution (Lind, 1995).
Iron salt coagulants are manufactured by dissolving various iron sources (iron ores
and scrap iron) in sulfuric or hydrochloric acid or by reprocessing materials such as
acidic iron salt solutions from iron mills and foundries. Ferric chloride is made from
reprocessed titanium dioxide liquors. Like alum, the iron salt coagulants typically contain metal contaminants, usually Mn, Cu,V, Zn, Pb, and Cd.The amount varies with the
source of the product so checking metal concentration is a prudent procedure.
In most cases, the low amounts of heavy metal contaminants in HMS coagulants
will not have a significant effect on metal concentrations in the treated water. The
metals may already be in an insoluble form or they are likely to precipitate or adsorb
on the floc when the coagulant is added to the water. The metals may, however,
increase the heavy metal content of treatment residue.
AWWA standards for HMS coagulants (e.g., AWWA B403-98, Aluminum Sulfate—Liquid, Ground, or Lump) include only a general statement about limits on
impurity levels in the products. The statement says that the contaminants should not
be present in quantities that will cause “deleterious or injurious effects” on human
health when the product is used properly. For alum products the only specific impurity limit is for iron.
Aluminum in drinking water has been implicated as a contributing factor in
Alzheimer’s disease. However, to date, researchers have been unable to verify or
refute these claims. An AWWA Executive Committee paper (Anonymous, 1997) on
aluminum salt coagulants states that research results are not sufficiently consistent
or accurate to support concerns about aluminum in general or aluminum in drinking
water as causal agents for Alzheimer’s disease. The amount of residual Al in treated
water can be minimized by optimizing filtration to maximize the removal of partic-


6.25

COAGULATION AND FLOCCULATION

ulate matter and by keeping the pH during coagulation and flocculation between 6
and 6.5 [i.e., near the pH of the minimum solubility of aluminum hydroxide (Letterman and Driscoll, 1993)].

Acidity of Hydrolyzing Metal Coagulants
When an acidic HMS coagulant solution, such as alum or PACl, is diluted in the
water to be treated, the hydrolysis reaction produces hydrogen ions that decrease
the alkalinity of the solution and tend to lower the pH, for example:
H+ + HCO3− → H2CO3 (CO2 + H2O)

(6.12)

The strong acid content or acidity of commercial coagulant solutions depends on the
basicity of prehydrolyzed products (Eq. 6.10) and the acid content of acidsupplemented products. Letterman, Chappell, and Mates (1996) have shown that the
effective acidity can be estimated using the following expression:
A
300 − 3B
Effective acidity (meq/mg metal) = ᎏᎏ + ᎏᎏ
100 (AW) 52.7 (MC)
where

(6.13)

B = basicity of the prehydrolyzed products (percent)
A = weight percent of 93 percent sulfuric acid solution in acidsupplemented products
AW = atomic weight of the metal; Al = 27 and Fe = 55.9
MC = metal concentration in the product solution (weight percent)

Equation 6.13 is based on the assumption that the predominant hydrolysis product

present after coagulant addition is the metal hydroxide precipitate, Al(OH)3 or
Fe(OH)3, with a molar ratio of hydroxide to metal of 3. Because it is very unlikely
that a manufacturer would add supplemental sulfuric acid to a prehydrolyzed product solution, it can be assumed that when B > 0, then A = 0, and, conversely, when
A > 0, then B = 0. Example values of coagulant solution acidities are listed in Table
6.4 for four commercial HMS products.
The effective acidity of a coagulant product can be used to determine the relationship between the coagulant dosage and the pH after flocculation and floc separation.
TABLE 6.4 Calculated Effective Acidities of Selected Commercial Coagulant Products*

Type of coagulant solution
Aluminum sulfate
(alum)
Polyaluminum chloride
(PACl)
Acid-supplemented
alum
Ferric sulfate

Sulfuric acid
content (%, 93%
sulfuric acid)

Metal concentration
in product solution
(weight %)

Calculated acidity
(meq/mg M)‡

0

0

4.3

0.111

75

0

12.3

0.028

Basicity
(%)

NA†
7

10
0

2.95
12

* The solution attributes used in the calculations are from the manufacturer’s product data sheets.
†
NA = not applicable.
‡

M = aluminum or iron.

0.168
0.051