CHAPTER 14

DISINFECTION
Charles N. Haas, Ph.D.
LD Betz Professor of Environmental Engineering
Drexel University
Philadelphia, Pennsylvania

Disinfection is a process designed for the deliberate reduction of the number of
pathogenic microorganisms. While other water treatment processes, such as filtration or coagulation-flocculation-sedimentation, may achieve pathogen reduction,
this is not generally their primary goal. A variety of chemical or physical agents may
be used to carry out disinfection. The concept of disinfection preceded the recognition of bacteria as the causative agent of disease. Averill (1832), for example, proposed chlorine disinfection of human wastes as a prophylaxis against epidemics.
Chemical addition during water treatment for disinfection became accepted only
after litigation on its efficacy (Race, 1918). The prophylactic benefits of water disinfection soon became apparent, particularly with respect to the reduction of typhoid
and cholera.
While significant progress is being made in controlling the classic waterborne diseases, newly recognized agents have added to the challenge. These include viruses
(Melnick et al., 1978; Mosley, 1966), certain bacteria (Campylobacter, Palmer et al.,
1983; Yersinia, Brennhovd et al., 1992; Reasoner, 1991; or Mycobacteria, Geldreich,
1971; Iivanainen et al., 1993; Reasoner, 1991; for example), and protozoans (Giardia,
Brown et al., 1992; Le Chevallier et al., 1991; Miller et al., 1978; Reasoner, 1991; Renton et al., 1996; Rose et al., 1991; Cryptosporidium, Bridgman et al., 1995; Centers for
Disease Control and Prevention, 1995; Gallaher et al., 1989; Goldstein et al., 1996;
Hayes et al., 1989; Le Chevallier et al., 1991; Leland et al., 1993; Mac Kenzie et al.,
1994; Miller, 1992; Reasoner, 1991; Richardson et al., 1991; Rose et al., 1991; Rush et
al., 1990; Smith, 1992). Occasional outbreaks of drinking-water-associated hepatitis
have also occurred (Nasser, 1994; Rosenberg et al., 1980). In addition, new viral
agents are continually being found to be capable of waterborne transmission.
The state of disinfection practice in the United States in the late 1980s was summarized in a survey of the AWWA Disinfection Committee (Haas et al., 1992). Most
water utilities continue to rely on chlorine or hypochlorite as their primary disinfection chemicals (Table 14.1), although increasing numbers are using ammonia (for
pre- or postammoniation) or chlorine dioxide or ozone. With the increasing concern
for removing and inactivating some of the more resistant pathogens, such as Giardia
and Cryptosporidium, while minimizing disinfection by-products, options other than

14.1


14.2

CHAPTER FOURTEEN

TABLE 14.1 Water Utility Disinfection Practices
According to 1989 AWWA Survey (N = 267)

Chlorine alone
Gas
Hypochlorite
Chlorine + ClO2
Ozone
Other

No ammonia

Ammonia

67.42%
5.99%
3.37%
0.37%
0.75%

19.85%
0.75%
1.50%


Source: Haas et al., 1992.

traditional chlorination are gaining popularity. This chapter will cover the use of
chlorine, as well as the major alternative agents, for the purpose of disinfection.

HISTORY OF DISINFECTION
Chlorine
Chlorine gas was first prepared by Scheele in 1774, but chlorine was not regarded as a
chemical element until 1808 (Belohlav and McBee, 1966). Early uses of chlorine
included the use of Javelle water (chlorine gas dissolved in an alkaline potassium solution) in France for waste treatment in 1825 (Baker, 1926) and its use as a prophylactic
agent during the European cholera epidemic of 1831 (Belohlav and McBee, 1966).
Disinfection of water by chlorine first occurred in 1908 at Bubbly Creek (Chicago)
and the Jersey City Water Company. Within two years, chlorine was introduced as a
disinfectant at New York City (Croton), Montreal, Milwaukee, Cleveland, Nashville,
Baltimore, and Cincinnati, as well as other smaller treatment plants. Frequently, dramatic reductions in typhoid accompanied the introduction of this process (Hooker,
1913). By 1918, over 1000 cities, treating more than 3 billion gal/day (1.1 × 107 m3/day)
of water, were employing chlorine as a disinfectant (Race, 1918).
Chloramination, the addition of both chlorine and ammonia either sequentially
or simultaneously, was first employed in Ottawa, Canada, and Denver, Colorado, in
1917. Both of these early applications employed prereaction of the two chemicals
prior to their addition to the full flow of water. Somewhat later, preammoniation
(the addition of ammonia prior to chlorine) was developed. In both cases, the process was advocated for its ability to prolong the stability of residual disinfectant during distribution and for its diminished propensity to produce chlorophenolic taste
and odor substances. Shortages of ammonia during World War II, and recognition of
the superiority of free chlorine as a disinfectant, reduced the popularity of the chloramination process. Recent concerns about organic by-products of chlorination,
however, have increased the popularity of chloramination (Wolfe et al., 1984).

Chlorine Dioxide
Chlorine dioxide was first produced from the reaction of potassium chlorate and
hydrochloric acid by Davy in 1811 (Miller et al., 1978). However, not until the

industrial-scale preparation of sodium chlorite, from which chlorine dioxide may
more readily be generated, did its widespread use occur (Rapson, 1966).


DISINFECTION

14.3

Chlorine dioxide has been used widely as a bleaching agent in pulp and paper
manufacture (Rapson, 1966). Despite early investigations on the use of chlorine
dioxide as an oxidant and disinfectant (Aston and Synan, 1948), however, its ascendancy in both water and wastewater treatment has been slow. As recently as 1971
(Morris, 1971), it was stated that “. . . ClO2 has never been used extensively for water
disinfection.”
By 1977, 84 potable water treatment plants in the United States were identified
as using chlorine dioxide treatment, although only one of these relied upon it as a
primary disinfectant (Miller et al., 1978). In Europe, chlorine dioxide was being used
as either an oxidant or disinfectant in almost 500 potable water treatment plants
(Miller et al., 1978).

Ozone
Ozone was discovered in 1783 by Van Marum, and named by Schonbein in 1840. In
1857, the first electric discharge ozone generation device was constructed by
Siemens, with the first commercial application of this device occurring in 1893
(Water Pollution Control Federation, 1984).
Ozone was first applied as a potable water disinfectant in 1893 at Oudshoorn,
Netherlands. In 1906, Nice, France, installed ozone as a treatment process, and this
plant represents the oldest ozonation installation in continuous operation (Rice et
al., 1981). In the United States, ozone was first employed for taste and odor control
at New York City’s Jerome Park Reservoir in 1906. In 1987, five water treatment
facilities in the United States were using ozone oxidation primarily for taste and

odor control or trihalomethane precursor removal (Glaze, 1987). Since the 1993 Milwaukee Cryptosporidium outbreak, there has been an upsurge in interest in ozone
as a disinfectant.

UV Radiation
The biocidal effects of ultraviolet radiation (UV) have been known since it was
established that short-wavelength UV was responsible for microbial decay often
associated with sunlight (Downes and Blount, 1877). By the early 1940s, design
guidelines for UV disinfection were proposed (Huff et al., 1965). UV has been
accepted for treating potable water on passenger ships (Huff et al., 1965). Historically, however, it has met with little enthusiasm in public water supply applications
because of the lack of a residual following application. In wastewater treatment, in
contrast, over 600 plants in the United States are either using, currently designing, or
constructing UV disinfection facilities (Scheible et al., 1992).

Other Agents
A variety of other agents may be used to effect inactivation of microorganisms. These
include heat, extremes in pH, metals (silver, copper), surfactants, permanganate, and
electron beam irradiation. Heat is useful only in emergencies as in “boil water” orders,
and is uneconomical. An alkaline pH (during high lime softening) may provide some
microbial inactivation, but is not usually sufficient as a sole disinfectant. Potassium
permanganate has been reported to achieve some disinfecting effects; however, the
magnitudes have not been well characterized. High-energy electrons for disinfection
of wastewaters and sludges have also been studied (Farooq et al., 1993); however, their


14.4

CHAPTER FOURTEEN

feasibility in drinking water is uncertain. In this chapter, therefore, primary consideration will be given to chlorine compounds, ozone, chlorine dioxide, and UV.


Regulatory Issues for Disinfection Processes
SWTR and GDR Requirements. Amendments to the Safe Drinking Water Act
require that all surface water suppliers in the United States filter and/or disinfect to
protect the health of their customers. The filtration and disinfection treatment
requirements for public water systems using surface water sources or groundwater
under the direct influence of surface water are included in what is called the Surface
Water Treatment Rule (SWTR, June 1989).
The SWTR requires that all surface water treatment facilities provide filtration and
disinfection that achieves at least (1) a 99.9 percent (3-log) removal-inactivation of
Giardia lamblia cysts and (2) a 99.99 percent (4-log) removal-inactivation of enteric
viruses. The SWTR assumes that for effective filtration, a conventional treatment
plant achieves 2.5-log removal of Giardia and a 2-log removal of viruses. Disinfection
is required for the remainder of the removal-inactivation. The amount of disinfection
credit to be awarded is determined with the CT concept, CT being defined as the
residual disinfectant concentration (C, mg/L) multiplied by the contact time (T, min)
between the point of disinfectant application and the point of residual measurement.
The SWTR Guidance Manual provides tables of CT values for several disinfectants,
which indicate the specific disinfection or CT credit awarded for a calculated value of
CT. A large safety factor is incorporated into the CT values included in the Guidance
Manual tables. In addition to relying on the CT tables to calculate disinfection credit,
the SWTR allows utilities to demonstrate the effectiveness of their disinfection systems through pilot-scale studies, which may be prohibitively expensive for smaller
operations. The SWTR is being revised to take into account knowledge developed
since the mid-1980s, and the anticipated formal promulgation of the Enhanced Surface Water Treatment Rule (ESWTR) will further affect the level of required disinfection. A more complete discussion of the SWTR is included in Chapter 1.
Furthermore, under the Safe Drinking Water Act, EPA is required to promulgate
rules for the disinfection of groundwaters. While the regulatory development of the
anticipated Groundwater Disinfection Rule is currently pending, this is expected to
require a level of disinfection either by chemical agents or by virtue of aquifer passage of all groundwaters being used in community water supply systems.
Disinfection By-product Requirements. Along with disinfection requirements,
since 1974 there have been explicit regulations on disinfection by-products—first
with respect to trihalomethanes, and more recently with respect to haloacetic acids,

bromate, and other possible by-products. The combination of the requirement to
achieve disinfection along with the requirement to minimize disinfection byproducts has led to an increasing spectrum of options being considered.

DISINFECTANTS AND THEORY OF DISINFECTION
Basic Chemistry
Chlorine and Chlorine Compounds. Chlorine may be used as a disinfectant in the
form of compressed gas under pressure that is dissolved in water at the point of


DISINFECTION

14.5

application, solutions of sodium hypochlorite, or solid calcium hypochlorite. The
three forms are chemically equivalent because of the rapid equilibrium that exists
between dissolved molecular gas and the dissociation products of hypochlorite compounds.
Elemental chlorine (Cl2) is a dense gas that, when subject to pressures in excess
of its vapor pressure, condenses into a liquid with the release of heat and with a
reduction in specific volume of approximately 450-fold. Hence, commercial shipments of chlorine are made in pressurized tanks to reduce shipment volume. When
chlorine is to be dispensed as a gas, supplying thermal energy to vaporize the compressed liquid chlorine is necessary.
The relative amount of chlorine present in chlorine gas, or hypochlorite salts, is
expressed in terms of available chlorine. The concentration of hypochlorite (or any
other oxidizing disinfectant) may be expressed as available chlorine by determining
the electrochemical equivalent amount of Cl2 to that compound. Equation 14.1
shows that 1 mole of elemental chlorine is capable of reacting with two electrons to
form inert chloride:
Cl2 + 2 e− = 2 Cl−

(14.1)


Equation 14.2 shows that 1 mole of hypochlorite (OCl−) may react with two electrons to form chloride:
OCl− + 2 e− + 2H+ = Cl− + H2O

(14.2)

Hence, 1 mole of hypochlorite is electrochemically equivalent to 1 mole of elemental chlorine, and may be said to contain 70.91 g of available chlorine (identical
to the molecular weight of Cl2).
Calcium hypochlorite (Ca(OCl)2) and sodium hypochlorite (NaOCl) contain 2
and 1 moles of hypochlorite per mole of chemical, respectively, and, as a result, 141.8
and 70.91 g available chlorine per mole, respectively. The molecular weights of
Ca(OCl)2 and NaOCl are, 143 and 74.5, respectively, so that pure preparations of the
two compounds contain 99.2 and 95.8 weight percent available chlorine; hence, they
are effective means of supplying chlorine for disinfection purposes.
Calcium hypochlorite is available commercially as a dry solid. In this form, it is
subject to a loss in strength of approximately 0.013 percent per day (Laubusch,
1963). Calcium hypochlorite is also available in a tablet form for use in automatic
feed equipment at low-flow treatment plants.
Sodium hypochlorite is available in 1 to 16 weight percent solutions. Higherconcentration solutions are not practical because chemical stability rapidly diminishes with increasing strength. At ambient temperatures, the half-life of sodium
hypochlorite solutions varies between 60 and 1700 days, respectively, for solutions of
18 and 3 percent available chlorine (Baker, 1969; Laubusch, 1963).
It should be noted that the loss of strength in sodium hypochlorite solutions may
also result in the formation of by-products that may be undesirable. Thermodynamically, the autodecomposition of hypochlorite to chlorate is highly favored by the following overall process (Bolyard et al., 1992):
3 ClO− → 2 Cl− + ClO3−

(14.3)

Measurements of sodium hypochlorite disinfectant solutions at water utilities
have revealed that the mass concentration of chlorate is from 1.7 to 220 percent of
the mass concentration of free available chlorine (Bolyard et al., 1992, 1993). The
concentration of chlorate present in these stock solutions is kinetically controlled



14.6

CHAPTER FOURTEEN

and may be related to the solution strength, age, temperature, pH, and presence of
metal catalysts (Gordon et al., 1993, 1995).
When a chlorine-containing compound is added to a water containing insignificant quantities of kjeldahl nitrogen, organic material, and other chlorine-demanding
substances, a rapid equilibrium is established among the various chemical species in
solution. The term free available chlorine is used to refer to the sum of the concentrations of molecular chlorine (Cl2), hypochlorous acid (HOCl), and hypochlorite
ion (OCl−), each expressed as available chlorine.
The dissolution of gaseous chlorine to form dissolved molecular chlorine is
expressible as a phase equilibrium, and may be described by Henry’s law:
Cl2(g) = Cl2(aq)

H(mol/L-atm) = [Cl2(aq)]/PCl2

(14.4)

where quantities within square brackets represent molar concentrations, PCl2 is the
gas phase partial pressure of chlorine in atmospheres, and H is the Henry’s law constant, estimated from the following equation (Downs and Adams, 1973):
H = 4.805 × 10−6 exp (2818.48/T) (mol/L-atm)

(14.5)

Dissolved aqueous chlorine reacts with water to form hypochlorous acid, chloride ions, and protons as indicated by Equation 14.6.
Cl2(aq) + H2O = H+ + HOCl + Cl−
[H+][HOCl][Cl−]
KH = ᎏᎏ

[Cl2(aq)]

΂

΃

2581.93
= 2.581 exp − ᎏ (mol2/L2)
T

(14.6)

This reaction typically reaches completion in 100 ms (Aieta and Roberts, 1985;
Morris, 1946) and involves elementary reactions between dissolved molecular chlorine and hydroxyl ions. The extent of chlorine hydrolysis, or disproportionation
(because the valence of chlorine changes from 0 on the left to +1 and −1 on the
right), as described by Equation 14.6, decreases with decreasing pH and increasing
salinity; hence, the solubility of gaseous chlorine may be increased by the addition of
alkali or by the use of fresh, rather than brackish, water.
Hypochlorous acid is a weak acid and may dissociate according to Equation 14.7:
HOCl = OCl− + H+
Ka = [OCl−][H+]/[HOCl]

(14.7)

The pKa of hypochlorous acid at room temperature is approximately 7.6 (Brigano et al., 1978). Morris (1966) has provided a correlating equation for Ka as a function of temperature:
ln(Ka) = 23.184 − 0.0583 T − 6908/T

(14.8)

where T is specified in degrees Kelvin (K = °C + 273). Figure 14.1 illustrates the

effect of pH on the distribution of free chlorine between OCl− and HOCl.
One practical consequence of the reactions described by Equations 14.4 through
14.8 is that the chlorine vapor pressure over a solution depends on solution pH,
decreasing as pH increases (because of the increased formation of nonvolatile
hypochlorite acid). Therefore, the addition of an alkaline material such as lime or


DISINFECTION

14.7

FIGURE 14.1 Effect of pH on relative amount of hypochlorous acid and hypochlorite ion at 20°C.

sodium bicarbonate will reduce the volatility of chlorine from accidental spills or
leaks and thus minimize danger to exposed personnel.
The acid-base properties of gaseous chlorine, or the hypochlorite salts, will also
result in a loss or gain, respectively, of alkalinity, and a reduction or increase, respectively, in pH. For each mole of free chlorine (i.e., 1 mole of Cl2, or of NaOCl or 0.5
mole of Ca(OCl)2), there will be a change of one equivalent of alkalinity (increase
for sodium and calcium hypochlorite, and decrease for chlorine gas).
The solution produced by a gas chlorinator contains 3500 mg/L
available chlorine at a pH of 3. What is the equilibrium vapor pressure of this solution at 20°C (given that the value of the hydrolysis constant KH is 4.5 × 104 at this
temperature)?
EXAMPLE 14.1

1. The pH is sufficiently low that the dissociation of hypochlorous acid to form
hypochlorite can be ignored. Therefore, a balance over chlorine species yields:
[Cl2] + [HOCl] = (3500 × 10−3)/71
2. The factor of 71 reflects the fact that 1 mole of either dissolved chlorine or
hypochlorous acid contains 71 g of available chlorine.
3. The hydrolysis equilibrium constant can be used to develop an additional equation:

4.5 × 104 = [H+][Cl−][HOCl]/[Cl2]
or, because the pH is given,
4.5 × 107 = [Cl−][HOCl]/[Cl2]


14.8

CHAPTER FOURTEEN

4. Because chlorine gas was used to generate the dissolved free chlorine, the disproportionation reaction requires that for each mole of HOCl produced, 1 mole
of Cl− must have been produced. If the initial concentration of chloride (in the
feedwater to the chlorinator) was minimal, then a third equation results:
[Cl−] = [HOCl]
5. These three equations can be manipulated to produce a quadratic equation in the
unknown [Cl2]1/2:
[Cl2] + 6708[Cl2]1/2 − 0.05 = 0
6. The single positive root is the only physically meaningful one, hence:
[Cl2]1/2 = 7.45 × 10−6
or
[Cl2] = 5.55 × 10−11
7. The Henry’s law constant can be computed from Equation 14.5 as 0.072 moles/
L-atm, and therefore the partial pressure of chlorine gas is found:
PCl2 = 5.55 × 10−11/0.072 = 7.7 × 10−10 atm
= (0.77 ppb)
8. The OSHA permissible exposure limit (PEL) is reported as 1 ppm (ACGIH,
1994). Therefore, this level is of no apparent health concern to the workers.
Chlorine Dioxide. Chlorine dioxide (ClO2) is a neutral compound of chlorine in
the +IV oxidation state. It has a boiling point of 11°C at atmospheric pressure. The
liquid is denser than water and the gas is denser than air (Noack and Doeff, 1979).
Chemically, chlorine dioxide is a stable free radical that, at high concentrations,

reacts violently with reducing agents. It is explosive, with the lower explosive limit in
air variously reported as 10 percent (Downs and Adams, 1973; Masschelein, 1979b)
or 39 percent (Noack and Doeff, 1979). As a result, virtually all applications of chlorine dioxide require the synthesis of the gaseous compound in a dilute stream (either
gaseous or liquid) on location as needed.
The solubility of gaseous chlorine dioxide in water may be described by Henry’s
law, and a fit of the available solubility data (Battino, 1984) results in the following
relationship for the Henry’s law constant (in units of atm−1):
ln(H) = mole fraction dissolved ClO2(aq)/PClO2
= 58.84621 + (47.9133/T) − 11.0593 ln(T)

(14.9)

Under alkaline conditions, the following disproportionation into chlorite (ClO2−)
and chlorate (ClO3−) occurs (Gordon et al., 1972):
2 ClO2 + 2 OH− = H2O + ClO3− + ClO2−

(14.10)

In the absence of catalysis by carbonate, the reaction (Equation 14.10) is governed by parallel first- and second-order kinetics (Gordon et al., 1972; Granstrom
and Lee, 1957). The half-life of aqueous chlorine dioxide solutions decreases substantially with increasing concentration and with pH values above 9. Even at neutral
pH values, however, in the absence of carbonate at room temperature, the half-life


DISINFECTION

14.9

of chlorine dioxide solutions of 0.01, 0.001, and 0.0001 mol/L is 0.5, 4, and 14 h,
respectively. Hence, the storage of stock solutions of chlorine dioxide for even a few
hours is impractical.

The simple disproportionation reaction to chlorate and chlorite is insufficient to
explain the decay of chlorine dioxide in water free of extraneous reductants. Equation 14.10 predicts that the molar ratio of chlorate to chlorite formed should be 1:1.
Medir and Giralt (1982), however, found that the molar ratio of chlorate to chlorite
to chloride to oxygen produced was 5:3:1:0.75, and that the addition of chloride
enhanced the rate of decomposition and resulted in the predicted 1:1 molar ratio of
chlorite to chlorate. Thus, the oxidation of chloride by chlorate, and the possible formation of intermediate free chlorine, may be of significance in the decay of chlorine
dioxide in demand-free systems (Gordon et al., 1972).
The concentration of chlorine dioxide in solution is generally expressed in terms
of g/L as chlorine by multiplying the molarity of chlorine dioxide by the number of
electrons transferred per mole of chlorine dioxide reacted and then multiplying this
by 35.5 g Cl2 per electron mole. Conventionally, the five-electron reduction (Equation 14.11) is used to carry out this conversion.
ClO2 + 5e− + 4H+ = Cl− + 2 H2O

(14.11)

Note, however, that the typical reaction of chlorine dioxide in water, being
reduced to chlorite, is a one-electron reduction as follows:
ClO2 + e− = ClO2−

(14.11a)

Hence, according to Equation 14.11, 1 mole of chlorine dioxide contains 67.5 g of
mass, and is equivalent to 177.5 (=5 × 35.5) g Cl2. Therefore, 1 g of chlorine dioxide
contains 2.63 g as chlorine. In examining any study on chlorine dioxide, due care with
regard to units of expression of disinfectant concentration is warranted.
Ozone. Ozone is a colorless gas produced from the action of electric fields on oxygen. It is highly unstable in the gas phase; in clean vessels at room temperature the
half-life in air is 20 to 100 h (Manley and Niegowski, 1967).
The solubility of ozone in water can be described by a temperature- and pHdependent Henry’s law constant. The following provisional relationship (H in atm−1)
has been suggested (Roy, 1979):
H = 3.84 × 107 [OH−] exp (−2428/T)


(14.12)

Practical ozone generation systems have maximum gaseous ozone concentrations of about 50 g/m3; thus, the maximum practical solubility of ozone in water is
about 40 mg/L (Stover et al., 1986). Upon dissolution in water, ozone can react with
water itself, with hydroxyl ions, or with dissolved chemical constituents, as well as
serving as a disinfecting agent. Details of these reactions will be discussed later in
this chapter and in Chapter 12.

DISINFECTANT DEMAND REACTIONS
Chlorine
Reactions with Ammonia. In the presence of certain dissolved constituents in
water, each of the disinfectants may react and transform to less biocidal chemical


14.10

CHAPTER FOURTEEN

forms. In the case of chlorine, these principally involve reactions with ammonia and
amino nitrogen compounds. In the presence of ammonium ion, free chlorine reacts
in a stepwise manner to form chloramines. This process is depicted in Equations
14.13 through 14.15:
NH4+ + HOCl = NH2Cl + H2O + H+

(14.13)

NH2Cl + HOCl = NHCl2 + H2O

(14.14)


NHCl2 + HOCl = NCl3 + H2O

(14.15)

These compounds, monochloramine (NH2Cl), dichloramine (NHCl2 ), and
trichloramine (NCl3), each contribute to the total (or combined) chlorine residual in
a water.The terms total available chlorine and total oxidants refer, respectively, to the
sum of free chlorine compounds and reactive chloramines, or total oxidating agents.
Under normal conditions of water treatment, if any excess ammonia is present, at
equilibrium the amount of free chlorine will be much less than 1 percent of total
residual chlorine. Each chlorine atom associated with a chloramine molecule is
capable of undergoing a two-electron reduction to chloride; hence, each mole of
monochloramine contains 71 g available chlorine; each mole of dichloramine contains 2 × 71 or 142 g; and each mole of trichloramine contains 3 × 71 or 223 g of available chlorine. Inasmuch as the molecular weights of mono-, di-, and trichloramine
are 51.6, 86, and 110.5, respectively, the chloramines contain, respectively, 1.38, 1.65,
and 2.02 g available chlorine per gram. The efficiency of the various combined chlorine forms as disinfectants differs, however, and thus the concentration of available
chlorine does not completely characterize process performance. On an approximate
basis, for example, for coliforms, the biocidal potency of HOCl:OCl−:NH2Cl:NHCl2
is approximately 1:0.0125:0.005:0.0166; and for viruses and cysts, the combined chlorine forms are considerably less effective (Chang, 1971). As Equation 14.12 indicates, the formation of monochloramine is accompanied by the loss of a proton,
because chlorination reduces the affinity of the nitrogen moiety for protons (Weil
and Morris, 1949a).
The significance of chlorine speciation on disinfection efficiency was graphically
demonstrated by Weber et al. (1940) as shown in Figure 14.2. As the dose of chlorine
is increased, the total chlorine residual (i.e., remaining in the system after 30 min)
increases until a dose of approximately 50 mg/L, whereupon residual chlorine
decreases to a very low value, and subsequently increases linearly with dose indefinitely. The “hump and dip” behavior is paralleled by the sensitivity of microorganisms to the available chlorine residual indicated by the time required for 99 percent
inactivation of Bacillus metiens spores. At the three points indicated, the total available chlorine is approximately identical at 22 to 24 mg/L, yet a 32-fold difference in
microbial sensitivity occurred.
The explanation for this behavior is the “breakpoint” reaction between free chlorine and ammonia (Figure 14.3). At doses below the hump in the chlorine residual
curve (zone 1), only combined chlorine is detectable. At doses between the hump

and the dip in the curve, an oxidative destruction of combined residual chlorine
accompanied by the loss of nitrogen occurs (zone 2) (Taras, 1950). One possible
reaction during breakpoint is:
2 NH3 + 3 HOCl = N2 + 3 H+ + 3 Cl− + 3 H2O

(14.16)

This reaction also may be used as a means to remove ammonia nitrogen from
water or wastewaters (Pressley et al., 1972). Finally, after the ammonia nitrogen has


DISINFECTION

14.11

FIGURE 14.2 Effect of increased chloride dosage on residual chlorine and germicidal
efficiency; pH 7.0, 20°C, NH3 10 mg/L. (Source: Adapted from Weber et al., 1940.)

been completely oxidized, the residual remaining consists almost exclusively of free
chlorine (zone 3). The minimum in the chlorine residual–versus–dose curve (in this
case Cl2:NH4+ − N weight ratio of 7.6/1) is called the breakpoint and denotes the
amount of chlorine that must be added to a water before a stable free residual can
be obtained.

FIGURE 14.3 Schematic idealization of breakpoint curve. (Source:
Adapted from G. C.White, Disinfection of Wastewater and Water for Reuse,
Van Nostrand Reinhold, New York. Copyright 1978.)


14.12


CHAPTER FOURTEEN

In their investigations of the chlorination of drinking water, Griffin and Chamberlin (1941a,b) observed that:
1. The classical hump and dip curve is only seen at water pHs between 6.5 and 8.5.
2. The molar ratio between chlorine and ammonia nitrogen dose at the breakpoint
under ideal conditions is 2:1, corresponding to a mass dose ratio (Cl2:NH4+ − N)
of 10:1.
3. In practice, mass dose ratios of 15:1 may be needed to reach breakpoint.
The breakpoint reaction may also affect the pH of a water. If sodium hypochlorite is used as the source of active chlorine, as breakpoint occurs, the pH decreases
due to an apparent release of protons during the breakpoint process (Equation
14.16). If gaseous chlorine is used, this effect is reinforced by the release of protons
by hydrolysis of gaseous chlorine according to Equations 14.6 and 14.7 (McKee,
1960).
The oxidation of ammonia nitrogen by chlorine to gaseous nitrogen at the breakpoint would theoretically require 1.5 mol of chlorine (Cl2) per mole of nitrogen oxidized according to Equation 14.16.The observed stoichiometric molar ratio between
chlorine added and ammonia nitrogen consumed at breakpoint is typically about
2:1, suggesting that more oxidized nitrogen compounds are produced at breakpoint
rather than N2 gas. Experimental evidence (Saunier and Selleck, 1979) indicates that
the principal additional oxidized product may be nitrate formed via Equation 14.17:
NH4+ + 4 HOCl = NO3− + 4 Cl− + 6 H+ + H2O

(14.17)

Depending upon the relative amount of nitrate formed in comparison to nitrogen
at breakpoint, between 1.5 and 4.0 mol of available chlorine may be required, which
is consistent with the available data.
Below the breakpoint, inorganic chloramines decompose by direct reactions with
several compounds. For example, monochloramine may react with bromide ions to
form monobromamine (Trofe, 1980). If trichloramine is formed, as would be the case
for applied chlorine doses in excess of that required for breakpoint, it may decompose either directly to form nitrogen gas and hypochlorous acid or by reaction with

ammonia to form monochloramine and dichloramine (Saguinsin and Morris, 1975).
In distilled water, the half-life of monochloramine is approximately 100 h (Kinman
and Layton, 1976). Even in this simple circumstance, however, the decomposition
products have not been completely characterized. Valentine (1986) found that the
decomposition of pure solutions of monochloramine produces an unidentified product that absorbs UV light at 243 nm and is capable of being oxidized or reduced.
Where the pH is below 9.0 (so that the dissociation of ammonium ion is negligible), the amount of combined chlorine in dichloramine relative to monochloramine
after the reactions in Equations 14.13 and 14.14 have attained equilibrium is given
by the following relationship (McKee, 1960):
BZ
A = ᎏᎏᎏ − 1
1 − ͙ෆ1ෆ
−ෆ
Bෆ
Zෆ
(2ෆ
−ෆ
Zෆ)

(14.18)

In Equation 14.18, A is the ratio of available chlorine in the form of dichloramine
to available chlorine in the form of monochloramine, Z is the ratio of moles of chlorine (as Cl2) added per mole of ammonia nitrogen present, and B is defined by Equation 14.19:
B = 1 − 4 Keq[H+]

(14.19)


DISINFECTION

14.13


The equilibrium constant in Equation 14.19 refers to the direct interconversion
between dichloramine and monochloramine as follows:
H+ + 2 NH2Cl = NH4+ + NHCl2
Keq = [NH4+][NHCl2]/[H+][NH2Cl]2

(14.20)

At 25°C, Keq has a value of 6.7 × 105 L/mol (Gray et al., 1978). From these relationships, determination of the equilibrium ratio of dichloramine to monochloramine as a function of pH and applied chlorine dose ratio is possible (assuming no
dissipative reactions other than those involving the inorganic chloramines). As pH
decreases and the Cl:N dose ratio increases, the relative amount of dichloramine
also increases (Figure 14.4). As the Cl:N molar dose ratio increases, the relative
amount of dichloramine also increases. As the Cl:N molar dose ratio increases
beyond unity, the amount of dichloramine relative to monochloramine rapidly
increases as well. For the conversion from dichloramine to trichloramine, the equilibrium constant given at 0.5 M ionic strength and 25°C indicates that the amount of
trichloramine to be found in equilibrium with di- and monochloramine at molar
dose ratios of up to 2.0 is negligible (Gray et al., 1978). This agrees with experimental measurement of the individual combined chlorine species as a function of
approach to breakpoint (White, 1972).
These findings, coupled with the routine observation of the breakpoint at molar
doses at or below 2:1 (Cl2-to-N weight ratios below 10:1), indicate that trichloramine
is not an important species in the breakpoint reaction. Rather, the breakpoint reaction leading to oxidation of ammonia nitrogen and reduction of combined chlorine
is initiated with the formation of dichloramine.

FIGURE 14.4 Effect of pH and Cl2:NH4+ molar ratio on dichloramine-tomonochloramine ratio (25°C).


14.14

CHAPTER FOURTEEN


The kinetics of formation of chloramine species have been investigated by various researchers since initial attempts by Weil and Morris (1949b). The formation of
monochloramine is a first-order process with respect to both hypochlorous acid and
un-ionized ammonia. However, determining whether this, or a process involving
hypochlorite ions reacting with ammonium cations, is the actual mechanism of reaction is not possible solely through kinetic arguments. If the neutral species are
selected as the reactants, then the rate of formation of monochloramine (r) may be
described by (Morris and Isaac, 1983):
r (mol/L-s) = 6.6 × 108 exp(−1510/T) [HOCl][NH3]

(14.21)

Because hypochlorous acid dissociates into hypochlorite with a pKa of approximately 7.4 and ammonia is able to associate with a proton to form the ammonium
cation, with the pKa for the latter of approximately 9.3, for a constant chlorine:nitrogen dose ratio, the maximum rate of monochloramine formation occurs at a pH
where the product HOCl × NH3 is maximized, which is at the midpoint of the two pK
values or 8.4. At this optimum pH and the usual temperatures encountered in practice, the formation of monochloramine attains equilibrium in seconds to 1 min; however, at either a higher or lower pH, the speed of the reaction slows.
A number of the other reactions in the chlorine-ammonia system may be kinetically limited. These have recently been reviewed; Table 14.2 is a compilation of the
known reaction kinetics involving chlorine, ammonia, and intermediate species.
The reaction of NH2Cl with HOCl to form NHCl2 is catalyzed by a number of
acidic species that may be present in water (Valentine and Jafvert, 1988). Possibly, a
number of the other reactions in Table 14.2 can also be catalyzed in a similar manner; however, insufficient data are available to evaluate this possibility.
When free chlorine is contacted with a water containing ammonia, the initial
velocity of monochloramine formation is substantially greater than the velocity of
the subsequent formation of dichloramine. Hence, relative to equilibrium levels, an
initial accumulation of monochloramine will occur if large dose ratios are employed,
until the dichloramine formation process can be driven (Palin, 1983).
TABLE 14.2 Summary of Chlorine Reaction Kinetics
Reaction

Forward rate expression

Reverse rate expression


NH3 + HOCl ⇔ NH2Cl + H2O

1510
6.6 × 108 exp − ᎏ
T

8800
1.38 × 108 exp − ᎏ
T

NH2Cl + HOCl ⇔ NHCl2 + H2O

2010
3 × 105 exp − ᎏ
T

΃

7.6 × 10−7 L/mol-s*

NHCl2 + HOCl ⇔ NCl3 + H2O

3420
2 × 105 exp − ᎏ
T

΃

5530

5.1 × 103 exp − ᎏ
T

2NH2Cl ⇔ NHCl2 + NH3

2160
80 exp − ᎏ
T

΂

΂
΂

΂

Rates are in units of L/mol-s.
Concentrations are in mol/L.
Reactions are elementary and water is at unit activity.
* Rate constant at 25°C.
Source: Morris and Isaac, 1983.

΃

΃

΂

΂


24.0 L/mol-s*

΃

΃


DISINFECTION

14.15

The kinetic evolution of the chlorine-ammonia speciation process in batch systems is described by a series of coupled ordinary differential equations. While these
are highly nonlinear, various authors have applied numerical integration techniques
for their solutions and, below the breakpoint, have found reasonable concordance
between model predictions and experimental measurements (Haag and Lietzke,
1980; Isaac et al., 1985; Saunier and Selleck, 1979; Valentine and Jafvert, 1988).
The breakpoint process involves a complex series of elementary reactions, of
which Equations 14.16 and 14.17 are the net results. Saunier and Selleck (1979) proposed that hydroxylamine (NH2OH) and NOH may be intermediates in this reaction. However, sufficient evaluation of their proposed kinetic scheme for the
breakpoint process has not yet been achieved to justify its use for design applications.
EXAMPLE 14.2 A water supply is to be postammoniated. If the water has a pH of
7.0, a free chlorine residual of 1.0 mg/L, and a temperature of 25°C, how much
ammonia should be added such that the ratio of dichloramine to monochloramine is
0.1 (assume that, upon the addition of ammonia, none of the residual dissipates)?

1. From Equations 14.17 and 14.18, the following is determined:
B = 1 − 4 Keq (10−7) = 1 − 4 (6.7 × 105)(10−7)
= 0.732
2. From Equation 14.16, noting that the problem condition specifies A = 0.1, the following equation is to be solved:
0.732 Z
0.1 = −1 + ᎏᎏᎏ

1 − ͙ෆ1ෆ
−ෆ0ෆ
.7ෆ3ෆ2ෆ
(2ෆ
−ෆ
Zෆ
)Z
This can be rearranged into a quadratic equation
−0.289 Z2 + 0.134 Z = 0
3. The single nonzero root gives Z = 0.463, which is molar ratio of chlorine (as Cl2)
to ammonia nitrogen. Because chlorine has a molecular weight of 70, 1 mg/L of
free chlorine has a molarity of 1.43 × 10−5. Therefore, 3.09 × 10−5 molarity of
ammonia is required, or (multiplying by the atomic weight of nitrogen, 14) a concentration of 0.43 mg/L as N of ammonia must be added.
Reactions with Organic Matter. Morris (1967) has determined that organic
amines react with free chlorine to form organic monochloramines. The rate laws for
these reactions follow patterns similar to the inorganic monochloramine formation
process, except that the rate constants are generally less. In addition, the rate constants for this process correlate with the relative basicity of the amine reactant.
Organic chloramines may also be formed by the direct reaction between monochloramine and the organic amine, and this is apparently the most significant mechanism
of organic N-chloramine formation at higher concentrations such as might exist at
the point of application of chlorine to a water (Isaac and Morris, 1980). Pure solutions of amino acids and some proteins yield breakpoint curves identical in shape to
those of ammonia solutions (Baker, 1947; Wright, 1936).
Free chlorine reacts with organic constituents to produce chlorinated organic byproducts. Murphy (1975) noted that phenols, amines, aldehydes, ketones, and pyrrole
groups are readily susceptible to chlorination. Granstrom and Lee (1957) found that
phenol could be chlorinated by free chlorine to form chlorophenols of various
degrees of substitution. The kinetics of this process depend upon both phenolate


14.16

CHAPTER FOURTEEN


ions and hypochlorous acid. If excess ammonia was present, however, the formation
of chlorophenols was substantially inhibited.
More recently, DeLaat (1982) determined that polyhydric phenols are substantially more reactive than simple ketones in the production of chloroform, and that
the rates of these processes are first order with respect to the phenol concentration
and the free chlorine concentration. More significantly, the reactivity of these compounds was observed to be greater than the reactivity of ammonia with hypochlorous acid. Therefore, even if subbreakpoint chlorination is practiced, some
chloroform may be formed rapidly prior to the conversion of free to combined chlorine. Chapter 12 presents additional discussion of the formation of trihalomethanes
and other disinfection by-products that can arise from reactions with naturally
occurring dissolved humic substances.
The reactivity of the chlorine species with compounds responsible for taste and
odor depends on the predominant form of chlorine present. In field tests, Krasner
(1986) determined that free chlorine, but not combined chlorine, could remove
tastes and odors associated with organic sulfur compounds.
Reactions with Other Inorganic Compounds. The rates of reaction between free
chlorine residuals and other inorganic compounds likely to be present in water are
summarized in Table 14.3 (Wojtowicz, 1979). These reactions are generally first
order in both the oxidizing agent (hypochlorous acid or hypochlorite anion) and the
reducing agent.
Nitrites present in partially nitrified waters react with free chlorine via a complex, pH-dependent mechanism (Cachaza, 1976). While combined chlorine residuals
were generally thought to be unreactive with nitrite, Valentine (1985) has found that
the rate of decay of monochloramine in the presence of nitrite was far greater than
would be predicted based on reaction of the equilibrium free chlorine, implicating a
direct reaction between NH2Cl and NO2ᎏN.
Overall Chlorine Demand Kinetics. Chlorine demand is defined as the difference
between the applied chlorine dose and the chlorine residual measured at a particular time. The rate of exertion of chlorine demand in complex aqueous solutions has
been the subject of numerous studies. The most systematic work has been that of
TABLE 14.3 Summary of Kinetics of HOCl and OCl− Reduction by Miscellaneous Reducing Agents after Wojtowicz (1979)
Oxidizing agent
−


OCl
OCl−
OCl−
OCl−
HOCl
HOCl
HOCl
HOCl
HOCl
HOCl
Source: Feng, 1966.

Reducing agent
−

IO
OCl−
ClO2−
SO3
NO2−
HCOO−
Br−
OCN−
HC2O4−
I−

Oxidation product
−

IO4

ClO2−
ClO3−
SO42−
NO3−
H2CO
BrO−
HCO3−, N2
CO2
IO−

Log k, L/m-s, 25°C
−5.04
−7.63
−5.48
3.93
0.82
−1.38
3.47
−0.55
1.20
8.52


14.17

DISINFECTION

Taras (1950), who chlorinated pure solutions of various organic compounds and
found that chlorine demand kinetics could be described by Equation 14.22:
D = kt n


(14.22)

where t is the time in hours, D is the chlorine demand, and k and n are empirical constants. In subsequent work, Feben and Taras (1950, 1951) found that chlorine
demand exertion of waters blended with wastewater could be correlated to Equation 14.22, with the value of n correlated to the 1-h chlorine demand.
Haas and Karra (1984) developed Equation 14.23 to describe chlorine demand
exertion kinetics.
D = Co{1 − [x exp (−k1t) + (1 − x) exp (−k2t)]}

(14.23)

where x is an empirical parameter, typically 0.4 to 0.6, k1 and k2 are rate constants,
typically 1.0 min−1 and 0.003 min−1, respectively, and Co is the chlorine dose in mg/L.
Dugan et al. (1995) developed a Monod (Langmuir Hinshelwood) model for
describing free chlorine decay in drinking water in the absence of ammonia. It
describes chlorine decay as a reaction with total organic carbon (TOC) in water
according to the following differential equation:
dC
k(TOC)C
ᎏ = − ᎏᎏ
K(TOC) + C
dt

(14.24)

where TOC (assumed constant) is in mg C/L and C is the free chlorine concentration
in mg Cl2/L. Equation 14.24 can be integrated to the following implicit equation for
chlorine concentration at time t (Ct):

΂ ΃


C0
Ct = K (TOC) ln ᎏ − k (TOC) t + C0
Ct

(14.25)

Furthermore, k and K were correlated with the initial chlorine dose (C0) and
TOC concentration. In tests conducted on a variety of waters, the constants were
found to be given by the following equations (it should be noted that the pH and
temperature were fixed at 8 and 20°C, and the waters were of relatively low ionic
strength, so that the applicability of these relationships under other conditions is
unclear):

΂

C0
K = −0.85 ᎏ
TOC

΃
΂

C0
k = 0.030 − 0.0060 ᎏ
TOC

΃

(14.26)


Dechlorination. When the chlorine residual in a treated water must be lowered
prior to distribution, the chlorinated water can be dosed with a substance that reacts
with or accelerates the rate of decomposition of the residual chlorine. Compounds
that may perform this function include thiosulfate, hydrogen peroxide, ammonia,
sulfite/bisulfite/sulfur dioxide, and activated carbon; however, only the latter two
materials have been widely used for this purpose in water treatment (Snoeyink and
Suidan, 1975).


14.18

CHAPTER FOURTEEN

Chlorine Dioxide
The reaction of chlorine dioxide with material present in waters containing chlorine
dioxide demand appears to be less significant than in the case of chlorine. Rather,
the dominant causes of loss of chlorine dioxide during disinfection may be the direct
reactions with water and interconversions to chlorite and chloride, as outlined in
Equations 14.10 and 14.11. At mg/L concentrations, ammonia nitrogen, peptone,
urea, and glucose have insignificant chlorine dioxide demand in 1 h (Ingolls and
Ridenour, 1948; Sikorowska, 1961). However, a variety of inorganic and biological
materials will react (Werderhoff and Singer, 1987).
Masschelein (1979a) concluded that only the following organic-ClO2 reactions
are of significance to water applications:
1.
2.
3.
4.


Oxidation of tertiary amines to secondary amines and aldehydes
Oxidation of ketones, aldehydes, and (to a lesser extent) alcohols to acids
Oxidation of phenols
Oxidation of sulfhydryl-containing amino acids

Wajon et al. (1982) found a reaction stoichiometry of 2 mol of chlorine dioxide
consumed per mole of phenol (or hydroquinone) consumed. Products formed
included chlorophenols, aliphatic organic acids, benzoquinone, and (in the case of
phenol) hydroquinone. The mechanism appeared to include the possible formation
of hypochlorous acid as an active intermediate, and the rate of this process was
found to be base catalyzed and first order in each of the reactants.
In general, chlorine dioxide itself has been found to produce fewer organic byproducts with naturally occurring dissolved organic material, although some nonpurgeable organic halogenated compounds are formed (Rav-Acha, 1984). In
practice, however, chlorine dioxide may be generated in a manner in which chlorine is present as an impurity. Therefore, the reactions of such a stream may also
include those discussed earlier regarding chlorine reactions. The inorganic byproducts consist of chloride, chlorate, and chlorite; specific ratios may depend
on the precise application conditions (Noack and Doeff, 1981; Werderhoff and
Singer, 1987).
Ozone
Upon addition to water, ozone reacts with hydroxide ions to form hydroxyl radicals
and organic radicals. These radicals cause increased decomposition of ozone, and
also are responsible for nonselective (compared to the direct ozone reaction) oxidation of a variety of organic materials. Carbonate, and possibly other ions, may act as
radical scavengers and slow this process (Hoigne and Bader, 1975, 1976).
Gurol and Singer (1982) determined that ozone decomposition kinetics in various aqueous solutions are second order in ozone concentration and base promoted.
Some systematic difference between various buffer systems employed does occur,
with borate giving higher decomposition rates than phosphate, and phosphate at
higher ionic strength giving lower decomposition rates than phosphate at lower
ionic strength (1 versus 0.1 M). This effect was suggested as being caused by phosphate being a radical scavenger (and by radical decomposition being important at
higher pH values).
As a result of these decomposition processes, the half-life of ozone in water, even
in the absence of other reactive constituents, is quite short, on the order of seconds



DISINFECTION

14.19

FIGURE 14.5 Schematic of bromate formation pathways. Solid lines: direct ozone
reactions. Dashed lines: radical reactions (Source: Reprinted with permission from
von Gunten, U., and J. Hoigne, 1994. Bromate formation during ozonation of bromidecontaining waters: interaction of ozone and hydroxyl radical reactions. Environmental
Science and Technology 28(7): 1234–1242. Copyright 1994 American Chemical Society.)

to minutes. Water chemistry may exert a strong influence on the rate and extent of
ozone demand in a given application. Reactions of ozone in aqueous solution are
discussed further in Chapter 12.
Bromide reacts with ozone under aqueous conditions typical of drinking water
disinfection. Products of the reaction may be hypobromous acid, hypobromite,
and/or bromate. Higher concentrations of bromide can reduce the rate of ozone
decomposition. Under alkaline conditions, this may be influenced by trace metal
catalysts and organic sinks for radicals and oxidized bromide species (Cooper et
al., 1985).
Ozone will react with cyanides at a very fast rate. The mechanism involves reaction of the cyanide ion (to form unknown products), and the process is inhibited by
iron complexes but catalyzed by copper complexes of cyanide (Gurol et al., 1985).
The reaction of ozone with bromide may proceed to the further product of bromate (BrO3−) by a complex process that involves direct reaction as well as hydroxyl
radical mediation (von Gunten and Hoigne, 1994). The overall process is summarized in Figure 14.5. The formation of bromate by ozonation is highly important in
view of the potential carcinogenicity of bromate in disinfected waters (Bull and
Kopfler, 1991).

Demand for UV
For UV disinfection, the “dose” may be described in terms of the emitted lamp
power in the germicidal range per unit volume of fluid under irradiation, for example, W/m3. This can also be expressed as an integral over the disinfection reactor volume of the surface intensity (in W/m2, for example) (Severin et al., 1983a, 1984b).
With ultraviolet light disinfection systems, the equivalent of demand results from

dissolved and suspended materials, such as proteins, humic material, and iron compounds, that absorb radiation and thus shield microorganisms. Huff et al. (1965)
found that intensity monitoring within the reactor itself could be used to correct for
such effects.
One particular problem unique to physical systems such as UV is the need to
assure complete mixing in the transverse direction so that all microorganisms may
come equally close to the UV source. Cortelyou (1954) analyzed this effect for batch
UV reactors, and the analysis was extended to flow-through reactors by Haas and
Sakellaropoulous (1979). This phenomenon results in the desirability to achieve turbulent flow conditions in a UV reactor.


14.20

CHAPTER FOURTEEN

ASSESSMENT OF MICROBIAL QUALITY
(INDICATORS)
The microbial quality of a source water, or the efficacy of a treatment system for
removing microorganisms, can be assessed either by direct monitoring of pathogens
or by the use of an indicator system. Because pathogens are a highly diverse group,
generally requiring a highly specialized (and often insensitive and expensive) analytical technique for each pathogen, the use of indicator organisms is a more popular technique.
An indicator group of organisms can be used either to assess source water contamination or degree of treatment; however, the same indicator group is often used
to assess both properties. This places severe constraints on the group of indicator
organisms chosen. Bonde (1966) has proposed that an ideal indicator must:
1. Be present whenever the pathogens concerned are present
2. Be present only when the presence of pathogens is an imminent danger, that is,
be unable to proliferate to any greater extent in the aqueous environment
3. Occur in much greater numbers than pathogens
4. Be more resistant to disinfectants and to the aqueous environment than pathogens
5. Grow readily on relatively simple media
6. Yield characteristic and simple reactions enabling, as far as possible, an unambiguous identification of the group

7. Be randomly distributed in the sample to be examined, or be able to be uniformly
distributed by simple homogenization procedures
8. Grow widely independent of other organisms present when inoculated in artificial media, that is, not be seriously inhibited in growth by the presence of other
bacteria
The use of coliforms as indicator organisms stems from the pioneering work of
Phelps (1909).The basic rationale was that coliforms and enteric bacterial pathogens
originate from a common source—namely human fecal contamination. Subsequent
work by Butterfield et al. (1943, 1946), Kabler (1951), and Wattie and Butterfield
(1944) confirmed that these organisms were at least as resistant to free or combined
chlorine as enteric bacterial pathogens.
The coliform group is a heterogeneous conglomerate of microorganisms, including forms native to mammalian gastrointestinal tracts as well as a number of exclusively soil forms. The common fermentation tube (FT) and membrane filter (MF)
procedures are subtly different in the organisms they enumerate. Classically, coliforms have been defined as “Gram-negative, non-sporeforming bacteria which [sic]
ferment lactose at 35–37°C, with the production of acid and gas” (APHA, AWWA,
and WPCF, 1989). The FT procedure, however, ignores anaerogenic and lactosenegative coliforms, and the MF procedure ignores non-lactose-fermenting strains
(Clark and Pagel, 1977).
Furthermore, interferences can selectively reduce coliforms as measured by one
or the other method. Allen (1977), for example, found that high concentrations
(>500 to 1000/mL) of standard plate count (SPC) organisms appeared to reduce the
recovery of coliforms by the MF technique when compared to the FT technique.
The fecal coliform group of organisms is that subset of coliforms that are capable
of growing at elevated temperature (44.5°C).The original rationale for development
of this test was to provide a more selective indicator group, excluding mesophilic


DISINFECTION

14.21

coliforms primarily indigenous to soils. Total coliforms, however, continue to be the
basic U.S. microbiological standard for drinking water because the absence of coliforms ensures the absence of fecal coliforms, which is a conservative standard.

While coliforms, either fecal or total, may be reasonably good indicators of fecal
contamination of a water supply, reservations were expressed as early as 1922
(Anonymous, 1922) about the relative resistance of coliforms to chlorine vis-à-vis
pathogenic bacteria and the resulting adequacy of the coliform test as an indicator
of disinfection efficiency. In more recent work, coliforms have been found to be
more sensitive to disinfection by one or more forms of chlorine than various human
enteric viruses (Grabow et al., 1983; Kelly and Sanderson, 1958) and the protozoan
pathogens Naegleria (Rubin et al., 1983), Giardia (Jarroll, 1981; Korich et al., 1990;
Leahy, 1985; Rice et al., 1982), and Cryptosporidium (Kovich et al., 1990). In addition, viruses (Scarpino et al., 1977) and protozoan cysts (Leahy, 1985) have been
found to be more resistant to ClO2 inactivation than coliforms. Farooq (1976) has
determined that coliforms are more resistant to ozone than viruses. Rice and Hoff
(1981) found that Giardia lamblia cysts survived exposure to UV doses sufficient to
effect over 99.99 percent inactivation of E. coli. Human enteric viruses have been
isolated in full-scale water treatment plants practicing conventional treatment, and
meeting turbidity and coliform standards in the presence of free residual chlorine
(Payment et al., 1985; Rose et al., 1986).
As a result of the problems with the coliform group of organisms, a number of
workers have investigated alternative indicator systems with greater resistance to
disinfectants than coliforms. Among the most successful of these are the acid-fast
bacteria and yeasts studied by Engelbrecht et al. (1977, 1979) and Haas et al. (1983a,
b; 1985a, b). In addition, work using endotoxins (Haas and Morrison, 1981),
Clostridia (Cabelli, 1977; Payment and Franco, 1993), and bacteriophage (Abad et
al., 1994; Grabow, 1968; Grabow et al., 1983; Payment and Franco, 1993) has been
carried out. In addition, to some degree, heterotrophic plate count (HPC) organisms
may provide a conservative indicator of treatment efficiency. Despite these studies,
however, in U.S. practice, no alternative to the total coliform group of organisms has
yet found widespread application.

PATHOGENS OF CONCERN
A variety of pathogenic organisms capable of transmission by the fecal-oral route

may be found in raw wastewaters. Waterborne outbreaks of shigellosis, salmonellosis,
and various viral agents have been reported, in many cases associated with sewagecontaminated water supplies (Blostein, 1991; Drenchen and Bert, 1994; Haas, 1986;
Herwaldt et al., 1991, 1992; Levine et al., 1990; Reeve et al., 1989; Rosenberg et al.,
1976, 1980). Among the bacteria, Salmonella, Shigella, and Vibrio cholerae organisms
are the classical agents of concern (Mosley, 1966). In more recent times, concern has
expanded to other agents that have been found in wastewater—viruses and protozoa.
Among the viruses, enteroviruses (ECHO virus, Coxsackievirus), rotavirus,
reovirus, adenovirus, and parvovirus have been isolated from wastewater (Melnick
et al., 1978). New viruses that are suspected of waterborne transmission have been
identified at the rate of about one organism per year (Gerba, personal communication). Among the more important of these newly identified agents may be Norwalk
virus and calicivirus.
Over the past 15 years, significant concerns have increased over the risk from
pathogenic protozoa in drinking water, particularly Giardia and Cryptosporidium


14.22

CHAPTER FOURTEEN

(Gallaher et al., 1989; Goldstein et al., 1996; LeChevallier et al., 1991; Leland et al.,
1993; Richardson et al., 1991; Rose et al., 1991; Smith, 1992). The SWTR arose, to a
significant extent, from concerns over Giardia (Regli et al., 1988). Revisions to
drinking water regulations presently under discussion are concerned with assuring
an adequate degree of protection from Cryptosporidium.

DISINFECTION KINETICS
The information needed for the design of a disinfection system includes knowledge
of the rate of inactivation of the target, or indicator, organism(s) by the disinfectant.
In particular, the effect of disinfectant concentration on the rate of this process will
determine the most efficient combination of contact time (i.e., basin volume at a

given design flow rate) and the dose to employ.

Chick’s Law and Elaborations
The major precepts of disinfection kinetics were enunciated by Chick (1908), who
recognized the close similarity between microbial inactivation by chemical disinfectants and chemical reactions. A good overview of the principles of kinetic modeling
of disinfection has been presented by Gyurek and Finch (1998). Disinfection is analogous to a bimolecular chemical reaction, with the reactants being the microorganism and the disinfectant, and can be characterized by a rate law as are chemical
reactions:
r = −kN

(14.27)

where r is the inactivation rate (organisms killed/volume-time) and N is the concentration of viable organisms. In a batch system, this results in an exponential decay in
organisms, because the rate of inactivation equals dN/dt, assuming that the rate constant k is actually constant (e.g., the disinfectant concentration is constant).
Watson (1908) proposed Equation 14.28 to relate the rate constant of inactivation k to the disinfectant concentration C:
k = k′Cn

(14.28)

where n is termed the coefficient of dilution and k′ is presumed independent of disinfectant concentration, and, by virtue of Equation 14.27, microorganism concentration.
From the Chick-Watson law, when C, n, and k′ are constant (i.e., no demand, constant concentration), the preceding rate law may be integrated so that in a thoroughly mixed batch system,
ln(N/N0) = −k′Cnt

(14.29)

where N and N0 are, respectively, the concentrations of viable microorganisms at
time t and time 0. When disinfectant composition changes with time, or when a configuration other than a batch (or plug flow) system is used, the appropriate rate laws
characterizing disinfectant transformation (Haas and Karra, 1984b) along with the
applicable mass balances must be used to obtain the relationship between microbial
inactivation and concentration and time.



DISINFECTION

FIGURE 14.6

14.23

Chick’s law and its deviations.

Inactivation of microorganisms in batch experiments, even when disinfectant
concentration is kept constant, does not always follow the exponential decay pattern predicted by Equation 14.29. Indeed, two common types of deviations are
noted (Figure 14.6). In addition to the linear Chick’s law decay, the presence of
“shoulders” or time lags until the onset of disinfection is often observed. Also,
some microorganisms and disinfectants exhibit a “tailing” in which the rate of
inactivation progressively decreases. In some cases, a combination of both of these
behaviors is seen.
Even if deviations from Chick-Watson behavior are observed, plotting combinations of disinfectant concentration and time to produce a fixed percent inactivation
is generally possible. Such plots tend to follow the relationship Cnt = constant, where
the constant is a function of the type of organism, pH, temperature, form of disinfectant, and extent of inactivation. Such plots are linear on a log-log scale (Figure
14.7). If the value of n is greater than 1, a proportionate change in disinfectant concentration produces a greater effect than a proportionate change in time. In many
cases (Hoff, 1986), the Chick-Watson law n value is close to 1.0, and hence a fixed
value of the product of concentration and time (CT product) results in a fixed
degree of inactivation (at a given temperature, pH, etc.).
In the chemical disinfection of a water, the concentration of disinfectant may
change with time, and particularly during the initial moments of contact the chemical form(s) of halogens such as chlorine undergo rapid transformations from the
free to the combined forms. Because C would thus not be a constant, typically disinfection results obtained in batch systems exhibit tailing, the degree of which may
depend on the demand and the concentration of reactive constituents (such as
ammonia) in the system (Olivieri et al., 1971). Determination of the disinfectant
residual (and its chemical forms) is more critical than the disinfectant dose in these
systems.



14.24

CHAPTER FOURTEEN

FIGURE 14.7 Concentration-time relationships for 99 percent inactivation of various microorganisms by various disinfectants. (1) Giardia lamblia; free chlorine, 5°C
(Source: Hoff and Akin, 1986). (2) E. coli; free chlorine, 2 to 5°C, pH 8.5 (Source: Haas
and Karra, 1984a). (3) E. coli; free chlorine, 20 to 25°C, pH 8.5 (Source: Haas and
Karra, 1984a). (4) Poliovirus 1 (Mahoney); free chlorine, 2°C, pH 6 (Source: Haas and
Karra, 1984a). (5) E. coli; combined chlorine, 3 to 5°C, pH 7 (Source: Haas and Karra,
1984a). (6) Poliovirus 1 (Mahoney); ozone, 20°C, pH 7.2 (Source: Roy et al., 1981a).
(7) Giardia muris; ozone, 5°C, pH 7 (Source: Wickramanayake et al., 1985).

In the chlorine system, for example, knowing the rate laws for inactivation by
individual separate species and the dynamics of chlorine species interconversions as
described previously enables an overall model for chlorine inactivation to be formulated (Haas and Karra, 1984b). In doing this computation, the individual rates are
usually assumed to be additive (Fair et al., 1948), although this assumption has not
yet been experimentally verified.
The presence of shoulders in inactivation curves is often seen in organisms that
form clumps.This means that more than one cell must be inactivated to achieve inactivation of a colony or plaque-forming unit. For example, Rubin et al. (1983) found
that cysts of Naegleria gruberii in demand-free water showed shoulder-type inactivation to free chlorine. Similarly, when cells of E. coli were agglutinated, they displayed shoulder-type inactivation, which was absent in unagglutinated cultures
(Carlson et al., 1975). Severin et al. (1984a) found shoulder-type inactivation curves
in the case of E. coli with preformed chloramines (i.e., solutions of ammonia and
chlorine prereacted to form combined chlorine prior to addition of microorganisms), Candida parapsilosis (a yeast organism proposed as a possible disinfectionresistant indicator) with both preformed chloramines, and free chlorine and
poliovirus with iodine.
Shoulder inactivation curves may be explained by a multitarget model (Hiatt,
1964), by a series event model (Severin et al., 1984a), or by a diffusional model



14.25

DISINFECTION

(Haas, 1980). Tailing inactivation curves may be explained either by a vitalistic
hypothesis in which individuals in a population are nonidentical, and their inherent
resistance is distributed in a permanent (time-independent) manner, or by a mechanistic concept (Cerf, 1977). In the latter case, four particular mechanisms have been
advanced leading to tailing:
1.
2.
3.
4.

Conversion to resistant form during inactivation (hardening)
Existence of genetic variants of differing sensitivity
Protection of a subpopulation, or variations in received dose of disinfectant
Clumping of a subpopulation

The hardening process and resultant tailing have received wide attention, following discoveries of apparent hardening in the formaldehyde inactivation of poliovirus
prepared for the Salk vaccine (Nathanson and Langmuir, 1963). Gard (1960) has
proposed an empirical rate law for this behavior, which has been used by Selleck et
al. (1978) in the analysis of wastewater chlorination kinetics. Tailing behavior has
been found for viral and coliform inactivation by ozone (Katzenelson et al., 1974)
and for coliform inactivation by free chlorine (Haas and Morrison, 1981; Olivieri et
al., 1971).
Hom (1972) developed a flexible but highly empirical kinetic formulation for the
inactivation rate based on modifying Equations 14.27 and 14.28 to the following
form:
r = −k′mN tm − 1Cn


(14.30)

This equation is difficult to use as a rate model since it contains time as an explicit
variable. A formulation leading to the classical Hom integrated relationship can be
written as (Haas and Joffe, 1994):

΄ ΂ ΃΅

N
r = −mN(kCn)1/m −ln ᎏ
N0

(1 − 1/m)

(14.31)

Upon integration, if C is constant, this results in the following relationship:

΂ ΃

N
ln ᎏ = −k′Cnt m
N0

(14.32)

Depending upon the value of m, both shoulders and tailing may be depicted
by Equation 14.32. In early work, Fair et al. (1948) used a model of the form of
Equation (14.32) with m = 2 to analyze E. coli inactivation by free and combined
chlorine.

EXAMPLE 14.3 A certain water supply has operational problems due to high levels
of HPC organisms. To maintain adequate system water quality, a decision has been
made to keep the concentration of HPC organisms below 10/mL at the entry point
to the distribution system (i.e., following disinfection). Disinfection using free residual chlorine is practiced. As part of the laboratory investigation to develop design
criteria for this system, the inactivation of the HPC organisms is determined in batch
reactors (beakers). The pH and temperature are held constant at the expected final
water conditions. Using water with an initial HPC of 1000/mL, the following data are
taken: