16593 12 tủ tài liệu bách khoa

CHAPTER 12

CHEMICAL OXIDATION1
Philip C. Singer, Ph.D.
Professor, Department of Environmental
Sciences and Engineering
University of North Carolina
Chapel Hill, North Carolina

David A. Reckhow, Ph.D.
Professor of Civil and Environmental Engineering
University of Massachusetts
Amherst, Massachusetts

Chemical oxidation processes play several important roles in the treatment of drinking water. Chemical oxidants are used for the oxidation of reduced inorganic species,
such as ferrous iron, Fe(II); manganous manganese, Mn(II); and sulfide, S(-II); and
hazardous synthetic organic compounds such as trichloroethylene (TCE) and
atrazine. Oxidants can also be used to destroy taste- and odor-causing compounds
and eliminate color. In addition, in some cases, they may improve the performance
of, or reduce the required amount of, coagulants.
Because many oxidants also have biocidal properties, they can be used to control
nuisance aquatic growths, such as algae, in pretreatment basins, and may be used as
primary disinfectants to meet CT (disinfectant concentration times contact time)
requirements (see Chapter 14). These oxidants are often added at the head of the
treatment plant, prior to or at the rapid mix basin, but they can also be employed
after clarification, prior to filtration, after a substantial portion of the oxidant
demand has been removed.
The most common chemical oxidants used in water treatment are chlorine,
ozone, chlorine dioxide, and permanganate. Ozone is sometimes used in conjunction
with hydrogen peroxide or ultraviolet irradiation to produce radicals that have powerful oxidative properties. Mixed oxidant technologies are also available.
Free chlorine has traditionally been the oxidant (and disinfectant) of choice in

the United States, but concerns about the formation of potentially harmful halogenated disinfection by-products (DBPs) produced by reactions between free chlorine and natural organic material (NOM), exacerbated in some cases by the
presence of bromide, have caused many water systems to adopt alternative chemical
oxidants (and disinfectants) to lower halogenated DBP formation. These other oxidants may also react with NOM and bromide to various degrees, depending upon
1
Acknowledgment: We would like to thank Dr. William H. Glaze of the University of North Carolina at
Chapel Hill who wrote the earlier version of this chapter, which provided a starting point for the current
material.

12.1


12.2

CHAPTER TWELVE

the properties of the oxidant, to form oxidation by-products, some of which also
have adverse public health effects or result in downstream operational problems in
the treatment plant or distribution system.
This chapter reviews thermodynamic and kinetic principles associated with the
use of chemical oxidants in general, the types and properties of the chemical oxidants used in water treatment, specific applications of oxidation processes for the
treatment of drinking water, and the formation and control of oxidation and disinfection by-products. Comparisons among the different oxidant choices are presented where information is available.

PRINCIPLES OF OXIDATION
Thermodynamic Considerations
Thermodynamics establishes the bounds or constraints for oxidation reactions.
Chemical kinetics fills in much of the detail. In many cases there are simply no other
available data than the thermodynamic enthalpies and entropies of reaction.
Despite its limitations, the domain of thermodynamics is where one must begin the
task of characterizing and understanding oxidation reactions. In this section, the
most basic thermodynamic concepts relating to oxidation reactions will be presented. For a more comprehensive treatment of the subject, there are many excellent textbooks that can be consulted (e.g., Stumm and Morgan, 1996; Pankow, 1991).

Electrochemical Potentials. Oxidation reactions are often viewed as reactions
involving the exchange of electrons. Since acids are frequently defined as proton
donors and bases as proton acceptors, one can think of oxidants as electron acceptors and reductants as electron donors. In fact, it’s not quite this simple. Many oxidants actually donate an electron-poor element or chemical group, rather than
simply accept a lone electron. Nevertheless, it’s useful to treat all oxidation reactions
as simple electron transfers for the purpose of balancing equations and performing
thermodynamic calculations.
Thermodynamic principles can be used to determine if specific oxidation reactions are possible. This generally involves the calculation of some form of reaction
potential. Although in most cases oxidation equilibria lie very far to one side or the
other, it is sometimes instructive to calculate equilibrium concentrations of the reactants and products.
The first step is to identify the species being reduced and those being oxidized.
Appropriate half-cell reactions and their standard half-cell potentials (Eored and Eoox ,
respectively) are available in tables of thermodynamic constants (a few are listed in
Tables 12.1 and 12.2). These may be combined to get the overall standard cell potential Eonet (Eq. 12.1).
Eonet = Eoox + Eored

(12.1)

Much as a pKa describes the tendency of an acid to give up a hydrogen ion, an electrochemical potential E describes the tendency of an oxidant to take up an electron,
or a reductant to give one up. The standard-state Gibbs Free Energy of reaction ∆Go
is related to the standard electrochemical cell potential by Faraday’s constant F and
the number of electrons transferred n:
∆Go = −nFEonet

(12.2)

For a one-electron transfer reaction, this becomes:
∆Go (K cal) = −23Eonet (volts)

(12.3)

12.3

CHEMICAL OXIDATION

TABLE 12.1 Standard Half-Cell Potentials for Chemical Oxidants Used in Water Treatment
Oxidant

Eored, volts

Reduction half-reaction
1
⁄2O3(aq) + H + e → 1⁄2O2(aq) + 1⁄2H2O
OH + H+ + e− → H2O
1
⁄2H2O2 + H+ + e− → H2O
1
⁄3MnO4− + 4⁄3H+ + e− → 1⁄3MnO2(s) + 2⁄3H2O
ClO2 + e− → ClO2−
1
⁄2HOCl + 1⁄2H+ + e− → 1⁄2Cl− + 1⁄2H2O
1
⁄2OCl− + H+ + e− → 1⁄2Cl− 1⁄2H2O
1
⁄2HOBr + 1⁄2H+ + e− → 1⁄2Br− + 1⁄2H2O
1
⁄2NH2Cl + H+ + e− → 1⁄2Cl− + 1⁄2NH4+
1
⁄4NHCl2 + 3⁄4 H+ + e− → 1⁄2Cl− + 1⁄4NH4+
1

⁄4O2(aq) + H+ + e− → 1⁄2H2O

+

Ozone
Hydroxyl radical
Hydrogen peroxide
Permanganate
Chlorine dioxide
Hypochlorous acid
Hypochlorite ion
Hypobromous acid
Monochloramine
Dichloramine
Oxygen

−

2.08
2.85
1.78
1.68
0.95
1.48
1.64
1.33
1.40
1.34
1.23

Sources: Lide (1995); American Water Works Assoc. (1990); Stumm and Morgan (1996).

Classical thermodynamics indicates that reactions with a negative Gibbs Free
Energy (or a positive Eo) will spontaneously proceed in the direction as written (i.e.,
from left to right), and those with a positive value (or negative Eo) will proceed in
the reverse direction.
Consider a generic oxidation reaction:
aAox + bBred → aAred + bBox

(12.4)

where substance A picks up one electron from substance B. In order to determine
which substance is being reduced and which is being oxidized, one must calculate
and compare oxidation states of the reactant atoms and product atoms.
The equilibrium constant K for this reaction defines the concentration quotient
for the reactants and products at equilibrium:
[Ared]a[Box]b
K = ᎏᎏ
[Aox]a[Bred]b

(12.5)

The overall standard cell potential is then directly related to this equilibrium constant by:
RT
Eonet = ᎏ ln K
nF

(12.6)

TABLE 12.2 Standard Half-Cell Potentials for Some

Oxidation Reactions That Can Occur During Drinking
Water Treatment
E oox, volts

Oxidation half-reaction
⁄2Br + ⁄2H2O → ⁄2HOBr + ⁄2H + e
⁄2Mn+2 + H2O → 1⁄2MnO2(s) + 2H+ + e−
Fe+2 + 3H2O → Fe(OH)3(s) + 3H+ + e−
1
⁄8NH4+ + 3⁄8H2O → 1⁄8NO3− + 11⁄4H+ + e−
1
⁄2NO2− + 1⁄2H2O → 1⁄2NO3− + H+ + e−
1
⁄8H2S + 1⁄2H2O → 1⁄8SO4−2 + 11⁄4H+ + e−
1
⁄2H2S → 1⁄2S(s) + H+ + e−
1
⁄2HCOO− → 1⁄2CO2(g) + 1⁄2H+ + e−
1
1

−

1

1

1

+

−

−1.33
−1.21
−1.01
−0.88
−0.84
−0.30
−0.14
+0.29

Sources: Lide (1995); American Water Works Assoc. (1990); Stumm
and Morgan (1996).


12.4

CHAPTER TWELVE

and for a one-electron-transfer reaction at 25°C, this simplifies to:
1
log K = ᎏ Eonet
0.059

(12.7)

Oxidation-Reduction Reactions
Oxidation State. Oxidation state is characterized by an oxidation number, which
is the charge one would expect for an atom if it were to dissociate from the surrounding molecule or ion (assigning any shared electrons to the more electronegative

atom). Oxidation number may be either a positive or a negative number—usually an
integer between −VII and +VII, although in their elemental forms, for example, S(s),
O2(aq), atoms have an oxidation number of zero. This concept is useful in balancing
chemical equations and performing certain calculations. The rules for calculating oxidation number are described in most textbooks on general chemistry.
Balancing Equations. The first step in working with oxidation reactions is to
identify the role of the reacting species. At least one reactant must be the oxidizing
agent (i.e., containing an atom or atoms that become reduced), and at least one must
be a reducing agent (i.e., containing an atom or atoms that become oxidized). The
second step is to balance the gain of electrons from the oxidizing agent with the loss
of electrons from the reducing agent. Next, oxygen atoms are balanced by adding
water molecules to one side or another, and hydrogens are balanced with H+ ions.
For a more detailed treatment on calculations using oxidation reactions, the reader
is referred to a general textbook on aquatic chemistry (e.g., Stumm and Morgan,
1996; Pankow, 1991).
As an example consider the oxidation of manganese by ozone (Eq. 12.8). The
substance being oxidized is manganese (i.e., the reducing agent) and the one doing
the oxidizing (i.e., being itself reduced) is ozone.
Mn + O3 → products

(12.8)

Next, the products formed need to be evaluated. It might be known from experience that reduced soluble manganese (i.e., Mn+2) can be oxidized in water to the relatively insoluble manganese dioxide. It might also be known that ozone ultimately
forms hydroxide and oxygen after it becomes reduced.
Mn+2 + O3 → MnO2 + O2 + OH−

(12.9)

The next step is to determine the oxidation state of all atoms involved (Eq. 12.10).
−II +I

}

0

}

+IV −II

}

}

}

0

}
}

+II

Mn + O3 → MnO2 + O2 + OH−
+2

(12.10)

From this analysis, it is clear that manganese is oxidized from +II to +IV, which
involves a loss of two electrons per atom. On the other side of the ledger, the ozone
undergoes a gain of two electrons per molecule, as one of the three oxygen atoms
goes from an oxidation state of 0 to −II. The two half-reactions can be written as single electron transfers. These half-reactions are balanced by adding water molecules

and H+ ions to balance oxygen and hydrogen, respectively.
Mn+2 + H2O → ᎏ12ᎏMnO2 + 2H+ + e−

1
ᎏᎏ
2

(12.11)

By convention, when hydroxide appears in a half-reaction, additional H+ ions are
added until all of the hydroxide is converted to water. This is done to the reduction
half-reaction.
O3 + H+ + e− → + ᎏ12ᎏO2 + ᎏ12ᎏH2O

1
ᎏᎏ
2

(12.12)


12.5

CHEMICAL OXIDATION

From this point, it is a simple matter of combining the equations and canceling
out terms or portions of terms that appear on both sides. At the same time, the standard electrode potentials can be combined to get the overall potential.
Mn+2 + H2O → ᎏ12ᎏMnO2(S) + 2H+ + e−

−1.21 V (Eoox)

O3(aq) + H+ + e− → ᎏ12ᎏO2(aq) + ᎏ12ᎏH2O

+2.04 V (Eored)

O3(aq) + ᎏ12ᎏMn+2 + ᎏ12ᎏH2O → ᎏ12ᎏO2(aq) + ᎏ12ᎏMnO2(S) + H+

+0.83 V (Eonet)

1
ᎏᎏ
2

1
ᎏᎏ
2
1
ᎏᎏ
2

(12.13)

Immediately, it is seen that this reaction will proceed toward the right (the Eonet is
positive). But how far to the right will it go? To answer this, Eq. 12.7 is rearranged to
get
o

K = e16.95E net

(12.14)

K = e16.95*0.83 = 1.29 × 106

(12.15)

So for this reaction

and using the concentration quotient from the reaction stoichiometry,
[O2(aq)]0.5[MnO2(s)]0.5[H+]
1.29 × 106 = ᎏᎏᎏ
[O3(aq)]0.5[Mn+2]0.5[H2O]0.5

(12.16)

Because the activity of solvents (i.e., water) and solid phases are, by convention,
equal to 1,
[O2(aq)]0.5[H+]
1.29 × 106 = ᎏᎏᎏ
[O3(aq)]0.5[Mn+2]0.5

(12.17)

Furthermore, if the pH is 7.0 and a dissolved oxygen concentration of 10 mg/L
and an ozone concentration of 0.5 mg/L is maintained in the contactor, an equilibrium Mn+2 concentration of 1.8 × 10−25 M or about 10−27 mg/L can be calculated.Thermodynamic principles therefore indicate that this reaction essentially goes to
completion.
Now, knowing that the Mn+2 should react essentially completely to form manganese dioxide, it might be desirable to determine if ozone can possibly oxidize the
manganese dioxide to a higher oxidation state, that is, to permanganate. To examine
this, the preceding ozone equation must first be combined with the reverse of the
permanganate reduction equation (from Table 12.1).
O3(aq) + H+ + e− → ᎏ12ᎏO2(aq) + ᎏ12ᎏH2O

1
ᎏᎏ
2

MnO2 + ᎏ23ᎏH2O → ᎏ13ᎏMnO4− + ᎏ43ᎏH+ + e−

1
ᎏᎏ
3

O3(aq) + ᎏ13ᎏMnO2 + ᎏ13ᎏH2O → ᎏ13ᎏMnO4− + ᎏ13ᎏH+ + ᎏ12ᎏO2(aq)

1
ᎏᎏ
2

(12.18)

This allows the net potential to be calculated:
Eonet = Eoox + Eored = (−1.68V) + (+2.04V) = +0.36V

(12.19)

Again, this is a favorable reaction. The equilibrium constant is:
1
1
log K = ᎏ Eonet = ᎏ (+0.36V) = 6.1
0.059
0.059

K = 1.26 × 106

(12.20)


12.6

CHAPTER TWELVE

The equilibrium quotient can now be formulated directly from the balanced
equation. Note that neither manganese dioxide (MnO2) nor water (H2O) appears in
this quotient. This is because both are presumed present at unit activity. Manganese
dioxide is a solid and as long as it remains in the system, it is considered to be in a
pure, undiluted state. The same may be said for water. As long as the solutes remain
dilute, the concentration of water is at its maximum and remains constant.
[MnO4−]0.33[H+]0.33[O2]0.5
K = ᎏᎏᎏ
= 106.1
[O3]0.5

(12.21)

So under typical conditions where the pH is near neutrality (i.e., [H+] = 10−7), dissolved oxygen is near saturation (i.e., [O2(aq)] = 3 × 10−4 M), and the ozone residual
is 0.25 mg/L (i.e., [O3(aq)] = 5 × 10−6 M), the expected equilibrium permanganate
concentration should be:
[MnO4−]0.33[10−7]0.33[3 × 10−4]0.5
K = ᎏᎏᎏᎏ
= 106.1
[5 × 10−6]0.5

(12.22)

and solving for permanganate
[MnO4−]0.33 = 3.5 × 107
[MnO4−] = 327

(12.23)

Obviously, one cannot have 327 mol/L of permanganate. Nevertheless, the system
will be forced in this direction so that all of the manganese dioxide would be converted to permanganate. Once the manganese dioxide is gone, the reaction must stop.
As already mentioned, the preceding thermodynamic analysis is quantitatively
accurate when all reactions are at equilibrium. However, this is rarely the case. Many
oxidation reactions are quite slow or, in some cases, kinetically unfavored, and the
actual concentrations of reactants and products observed during water treatment
are far from those predicted by classical thermodynamics. For this reason, oxidation
chemistry must rely heavily on kinetics.

Kinetics and Mechanism
Reaction Kinetics. Thermodynamics indicates whether a reaction will proceed as
written. However, it will not indicate whether this reaction will produce significant
change within milliseconds or thousands of years. For this, chemical kinetics must be
considered. As an example, consider the reaction between hypochlorous acid and
bromide ion.
HOCl + Br− = HOBr + Cl−

(12.24)

In order for a molecule of hypochlorous acid and a molecule of bromide to combine
to form products, the two molecules must come into contact with each other (contact
meaning approach within a certain distance so that bonding forces can play a role).

The probability that a single HOCl:Br− molecular encounter will occur within any
fixed time period is directly proportional to the number of molecules of each type in
the system. It will also depend on the rate of movement of each of the reactant
molecules. As a consequence, the rate of formation of products—for example,


CHEMICAL OXIDATION

12.7

HOBr—will be dependent on a number of factors, including the concentration of
hypochlorous acid and the concentration of bromide in the reacting solution. This is
the kinetic law of mass action, which is expressed mathematically in Eq. 12.25.
d[HOBr]
ᎏᎏ = kf [HOCl][Br−]
dt

(12.25)

The reactants and products are expressed in molar units of concentration and kf
is called the forward reaction rate constant. The units for kf are liters/mole per unit
time. The reaction rate constant is going to be a function of such things as the rate of
movement of the molecules and the probability of HOBr formation, given that a collision between hypochlorous acid and bromide has already occurred. Because the
concentrations of HOCl and Br− that appear in Eq. 12.25 are raised to the first
power, it is said that this rate law is first order in both reactants. The overall order of
the reaction is the sum of the individual orders (i.e., second order in this case).
In a more general sense, Eq. 12.26 is the rate law for any elementary reaction of
the type described by Eq. 12.27.
d[A]
− ᎏ = kfa[A]a[B]b

dt

(12.26)

aA + bB → cC + dD

(12.27)

where the capital letters represent chemical species participating in the reaction and
the small letters are the stoichiometric coefficients (i.e., the numbers of each
molecule or ion required for the reaction). The overall order describes the extent of
dependence of the reaction rate on reactant concentrations. For the reaction in Eq.
12.27, it is equal to (a + b). The order with respect to species A is a, and the order
with respect to species B is b. Thus, the reaction in Eq. 12.27 is first order in both
reactants and second order overall.
Chemical reactions may be either elementary or nonelementary. Elementary
reactions are those reactions that occur exactly as they are written, without any
intermediate steps. These reactions almost always involve just one or two reactants.
The number of molecules or ions involved in elementary reactions is called the
molecularity of the reaction. Thus, for all elementary reactions, the overall order
equals the molecularity. Nonelementary reactions involve a series of two or more
elementary reactions. Many complex environmental reactions are nonelementary.
In general, reactions with an overall reaction order greater than 2 or reactions with
some noninteger reaction order are nonelementary.
Reaction rate constants for the various oxidants with similar solutes are often
positively correlated. In other words, a compound favored for oxidation by one oxidant is generally favored by others as well. Those that are relatively resistant to oxidation by one will likewise be unreactive with others. A good case study is the
extensive research done on the oxidation of phenolic compounds, as presented by
Tratnyek and Hoigne (1994).These data highlight the similarities between the chemical structure of a reactant and its reactivity with various oxidants. Chemists have
used such relationships to develop quantitative structure-activity relationships
(QSARs). The Hammett equations are one of the most widely used QSARs (see

Brezonik [1994] for more detail on this subject).
Temperature Dependence. As mentioned previously, the reaction rate constant k is
a function of temperature. The Arrhenius equation (Eq. 12.28) is the classic model
that describes this relationship:
k = koe−Ea /RT

(12.28)


12.8

CHAPTER TWELVE

where ko is called the frequency factor or the preexponential factor, Ea is the activation energy, R is the universal gas constant (199 cal/°K-mole), and T is the temperature in °K. The values for ko and Ea may be either found in the literature or
determined from experimental measurements.
Types of Reactions. To this point, considerations have addressed whether or not a
certain oxidation reaction can occur, and perhaps how fast it can occur. However, it
is sometimes quite useful to know how the reaction occurs on a molecular scale. In
other words, by what mechanism or pathway does it go from reactants to products?
For example, the problem of disinfection by-products is one of chemical pathways.
There is no inherent problem with oxidizing natural organic matter using chlorine.
However, when that reaction occurs through addition and substitution reactions
(see below) rather than simple electron transfer reactions, chlorinated organic byproducts such as the trihalomethanes (THMs) are obtained.
Oxidation reactions can generally be categorized as those involving electron
transfer and those involving transfer of atoms and groups of atoms.They may also be
characterized as reactions involving species with paired electrons (ionic) and those
involving unpaired electrons (radical). Aqueous chlorine presents a wide array of
ionic reactions (e.g., see Morris [1975]) that will serve as illustrative examples for
this discussion.
Table 12.3 presents a summary of the major types of ionic reactions occurring in

drinking water. Hypochlorous and hypobromous acid can be added to olefinic bonds
(i.e., carbon-carbon double bonds), forming halohydrins. This is an electrophilic
reaction where the initial attack is by the halogen atom (on the positive side of the
HOX dipole). The most stable configuration places the halogen on the carbon with
the most hydrogen atoms (producing the most stable carbonium ion: Markovnikov’s
rule). The other carbon becomes a carbonium ion, which subsequently reacts with
the HO portion of the HOX species or with water.
Activated ionic substitution can occur with both aromatic and aliphatic compounds. As with the addition reactions, this type of reaction will also lead to the formation of organohalide compounds. Aromatic substitution reactions occur readily
when an electron-donating substituent is bound to the ring. Functional groups on
the aromatic ring, such as OH and NH2, can be thought of as creating a partial negative charge on the ortho and para positions (second closest and farthest carbon
atoms from the functional group, respectively). The halogen end of the HOX
molecule attacks one of these carbons. Next, there is a loss of the OH end of the
molecule, and displacement of the H atom from the carbon under attack. Substitution on aliphatic species is also a multistep reaction, as exemplified by the haloform
reaction (see next section on pathways).
When substitution (or transfer) of a halogen occurs onto a nitrogen atom, a relatively reactive N-halo organic compound results. These compounds retain some of
the oxidizing capabilities of hypohalous acid and, consequently, the reactions are not
considered to cause an oxidant demand. The rates of substitution reactions with
nitrogenous organic compounds generally increase as the basicity of the nitrogen
atom increases.
Oxidation reactions with halogens are characterized by the formation of the inorganic halide ion, and an oxidized (nonhalogenated) form of the reacting compound.
With organic compounds, it is quite common to observe addition of an oxygen atom.
For example, oxidation reactions transform unsaturated hydrocarbons into alcohols,
then to aldehydes and ketones, and finally to carboxylic acids. Some oxidations do
not result in a net transfer of atoms. For these electron transfer reactions, it is common to form free radical intermediates. When this happens, chain reactions can
occur, sometimes leading to the types of reactions listed in Table 12.4.


12.9

CHEMICAL OXIDATION

TABLE 12.3 Major Types of Ionic Reactions
Reaction type

Example

1. Addition to an olefinic bond

2. Activated aromatic substitution

3. Substitution onto nitrogen

4. Oxidation with oxygen transfer

5. Oxidation with electron transfer

In addition to these ionic reactions, there are several reactions involving free radical species that can occur following addition of drinking water oxidants. These types
of reactions are commonly encountered with ozone, chlorine dioxide, and especially
the advanced oxidation processes (see below). For example, addition of ozone will
always lead to some decomposition and subsequent formation of hydroxyl radicals
(•OH). These reactive species engage in reactions that generally lead to the formation of new free radical species. The most common types are addition reactions,
hydrogen abstractions, and single electron transfers.
Catalysis. Many types of oxidation reactions are strongly affected by the presence
of catalysts. These are compounds that alter reaction rates without being formed or
consumed in the reaction. They typically participate in some key, rate-limiting step
and are regenerated during some later step. Catalysts generally provide an alternative pathway to a reaction with a lower activation energy.
Probably the most important catalytic processes involve the participation of acids
and bases. Specific acid and specific base catalysis involve H+ and OH−, respectively.
General acid and base catalysis involves any electron acceptor (e.g., a proton donor)
and electron donor (e.g., a proton acceptor), respectively.A good example of general

base catalysis is the classic haloform reaction (Figure 12.1). Here the rate-limiting


12.10

CHAPTER TWELVE

TABLE 12.4 Major Types of Radical Reactions
Reaction type

Example

6. Radical addition reaction

7. Hydrogen abstraction reaction

8. Radical oxidation reaction
with single electron transfer

step is the loss of a proton giving the enol. While it is shown in Figure 12.1 as occurring by reaction with hydroxide, any strong base will participate. Also the base (or
hydroxide) consumed in the first step is regenerated in the second.
Another type of catalysis that is important in oxidation processes comes from the
initiation of a free radical chain reaction. Examples include the decomposition of
ozone by hydroxide and the decomposition of chlorine by iron (e.g., see Brezonik
[1994]). In either case, the original oxidant will not react appreciably with recalcitrant compounds such as oxalate. However, in the presence of sufficient catalyst,
decomposition is initiated, leading to a series of chain propagation reactions
whereby oxalate can be easily converted to carbon dioxide.
Reaction Pathways. Oxidation reactions in drinking waters can be very complex.
They may begin with one of the mechanisms discussed here, but then may be followed by a wide range of nonoxidation processes, such as elimination reactions,
hydrolysis reactions, radical chain reactions, and rearrangement reactions.

The formation of trihalomethanes may occur through many different reaction
mechanisms. One of the most widely discussed is the haloform reaction (Figure 12.1),
which involves the stepwise chlorine substitution of the enolate form of a methyl
ketone.This classic reaction begins with a base-catalyzed halogenation ultimately leading to a carboxylic acid and chloroform. It is base-catalyzed because the species that
reacts with hypochlorous acid is the enol form of the methyl ketone. This undergoes
electrophilic substitution, forming a monohalogenated intermediate. The presence of
halogens on this carbon speeds subsequent enolization, which leads to complete halogenation of the α-carbon. The resulting intermediate (a trihalogenated acetyl compound) is subject to base-catalyzed hydrolysis,giving a trihalomethane and a carboxylic
acid. If hypochlorous acid is the only halogenating species, chloroform is the result.
Many early studies with acetone (propanone) indicated that the rate-limiting
step was the initial enolization. Once the enolate was formed, the molecule quickly


12.11

CHEMICAL OXIDATION

FIGURE 12.1

The haloform reaction.

proceeded through the entire reaction pathway. Thus, the reaction rate expression
often cited in the chemical literature was
d[CH3COCH3]
ᎏᎏ = −k[CH3COCH3]
dt

(12.29)

However, under drinking water conditions (i.e., neutral pH, low chlorine residual),
other steps may be rate limiting. This changes the rate law and complicates attempts

to characterize it. In addition, a competing pathway exists that leads to the formation
of dichloroacetic acid. Trichloropropanone may undergo further base-catalyzed
chlorine substitution to form pentachloropropanone (Reckhow and Singer, 1985).
This compound will rapidly hydrolyze to form chloroform and dichloroacetic acid.
The rate law proposed for the loss of trichloropropanone is

Ά

·

k1k2[HOCl][PT]
d[CH3COCCl3]
−ᎏᎏ = ᎏᎏᎏ + k3[OCl−] + k4[OH−] [CH3COCCl3]
k−1[PT] + k2[HOCl]
dt
(12.30)
This complicated rate law reflects catalysis by hypochlorite (OCl−), hydroxide (OH−),
and phosphate (PT).

OXIDANTS USED IN WATER TREATMENT
Chlorine
Chlorine is the most widely used oxidant (and disinfectant; see Chapter 14) in water
treatment practice. Chlorine is available in gaseous form, as Cl2; as a concentrated


12.12

CHAPTER TWELVE

aqueous solution, sodium hypochlorite, NaOCl (e.g., bleach); or as a solid, calcium

hypochlorite, Ca(OCl)2.
When chlorine gas is added to water, the chlorine rapidly disproportionates to
form hypochlorous acid (HOCl) and the chloride ion (Cl−):
Cl2 + H2O → HOCl + H+ + Cl−

(12.31)

The equilibrium constant for Eq. 12.31 is 5 × 10−4 at 25°C, indicating that the reaction
goes relatively far to the right, as shown. The residual concentration of molecular Cl2
in solution will represent only a small fraction of the total chlorine concentration,
except at very low pH conditions or at high chloride concentrations. For example, in
water at pH 2 with a 10−3 M chloride concentration, only 2 percent of the total chlorine will exist as molecular Cl2. The chloride produced by Eq. 12.31 is essentially
inert with respect to its oxidizing and disinfecting properties.
Hypochlorous acid is a weak acid (pKa = 7.5 at 25°C), which partially dissociates
to hypochlorite ion (OCl−):
HOCl = H+ + OCl−

(12.32)

The ratio of hypochlorous acid to hypochlorite may be calculated from the expression:

΂

΃

[HOCl]
log ᎏ
= 7.5 − pH
[OCl−]

(12.33)

The sum of the three species, Cl2, HOCl, and OCl−, is commonly referred to as free
available chlorine (FAC), and the concentrations of the individual species and their
sum are usually expressed in units of mg/L as Cl2.
Equations 12.31 and 12.32 show that the species of chlorine present in water will
depend on the total concentration of chlorine, pH, and temperature. Figure 12.2 is a
diagram of the relative amounts of the three species as a function of pH. At 25°C,
hypochlorous acid is the predominant species between pH 1 and pH 7.5, and
hypochlorite ion is predominant at pH values greater than 7.5.The concentrations of
the two species are equal at pH 7.5 (the pKa value). The distribution shifts somewhat
with changing temperature because the equilibrium constants for Eqs. 12.31 and
12.32 are temperature dependent.

FIGURE 12.2 Distribution diagram for molecular chlorine, hypochlorous acid, and hypochlorite
ion in water as a function of pH ([Cl−] = 10−3M). (Source: Snoeyink and Jenkins, 1980.)


CHEMICAL OXIDATION

12.13

If chlorine is added to water as liquid sodium hypochlorite, the following reactions occur:
NaOCl → Na+ + OCl−

(12.34)

OCl− + H2O = HOCl + OH−

(12.35)

If calcium hypochlorite granules are dissolved in water, they also form the hypochlorite ion, which, like the hypochlorite formed from the addition of NaOCl, reacts with
water to form hypochlorous acid:
Ca(OCl)2 → Ca2+ + 2OCl−

(12.36)

OCl− + H2O = HOCl + OH−

(12.37)

Just as in the case of the species formed by the addition of chlorine gas, the relative distribution of HOCl and OCl− that result from the addition of sodium
hypochlorite or calcium hypochlorite will be determined by pH, temperature, and
the total chlorine concentration. Again, this distribution will be in accordance with
Eqs. 12.31 and 12.32 and Figure 12.2. No matter in what form the chlorine is added
to water, hypochlorite, hypochlorous acid, and molecular chlorine will be formed as
described by the chemical equilibria represented by the previous reactions. The only
difference is that chlorine gas produces an acidic reaction which lowers the pH of
the solution, while sodium hypochlorite and calcium hypochlorite are both bases
which will raise the pH of the water. The extent of the pH change will depend on the
alkalinity of the water.
As shown in Table 12.1, hypochlorous acid and hypochlorite ion are both strong
oxidizing agents, but HOCl is the stronger of the two. Hence, oxidation reactions of
chlorine tend to be more effective at low pH values. This is also true of the disinfecting properties of chlorine (see Chapter 14).
Molecular chlorine is typically provided for water treatment applications in pressurized tanks so that the chlorine exists as a liquid under pressure. The chlorine is
then added to water by reducing the pressure in the tank and releasing the chlorine
as a gas. Various types of gas-feeding equipment are available to introduce the
gaseous chlorine to a sidestream of water, after which the resulting concentrated
sidestream solution of chlorine is blended with the main flow through a diffuser at
the desired point of application.

Alternatively, if liquid sodium hypochlorite solution is used, it is usually added
from a concentrated NaOCl bulk solution directly to the main plant flow with the
help of a liquid metering pump. Solid granules of calcium hypochlorite can be added
to the water directly or a concentrated solution of calcium hypochlorite can be prepared, from which the hypochlorite is metered into the main flow, as in the case of
sodium hypochlorite.
Historically, gaseous chlorine in pressurized tanks has been the most common
method of applying chlorine in municipal water treatment practice, but, more
recently, because of concerns about transport and handling of hazardous chemicals,
the use of liquid sodium hypochlorite is becoming more widely practiced despite its
higher cost. A potential concern involving the use of sodium hypochlorite is its stability. The hypochlorite tends to degrade over time, particularly when it is stored at
high temperatures and/or exposed to sunlight. One of the degradation products is
chlorate (ClO3−), a potential health concern (see subsequent discussion of chlorine
dioxide). Recent reports also indicate the possible presence of bromate, another inorganic disinfection by-product usually associated with ozonation, in sodium hypo-


12.14

CHAPTER TWELVE

chlorite feedstocks at levels which may exceed regulatory limits when the hypochlorite is fed at the typical doses used in practice.
Free chlorine can also be generated electrolytically on-site from brine (sodium
chloride, NaCl) solutions. The molecular chlorine generated from the electrolysis
reaction can be dissolved in sodium hydroxide (NaOH), which is also a by-product
of the electrolysis reaction, to produce a concentrated sodium hypochlorite solution:
Cl2 + NaOH → NaOCl + H+ + Cl−

(12.38)

Again, while the use of electrolysis cells to produce chlorine gas on-site tends to
be more costly than purchasing chlorine gas directly in pressurized tanks, the hazards associated with the transport and handling of pressurized chlorine containers is

avoided.
Reactions of Chlorine with Organic Compounds. Chlorine reacts with organic
material by a combination of oxidation and substitution reactions. For example,
chlorine reacts with amino acids to produce nonhalogenated oxidation by-products,
such as aldehydes and organic acids. Conversely, chlorine reacts with phenol to produce chlorinated phenolic compounds, such as ortho- and parachlorophenol. Formation of the latter contributes to an enhancement of taste and odor in chlorinated
water containing phenol (see below). In the case of aromatic compounds, the presence of electron-donating substituents on the aromatic ring facilitates both substitution reactions and oxidative ring cleavage reactions.
Chlorine reacts with natural organic material (e.g., humic substances) to produce
a variety of chlorine-substituted halogenated disinfection by-products, such as chloroform, di- and trichloroacetic acid, and chloropicrin. In view of the importance of
these reactions to water quality and treatment, the subject of disinfection byproduct formation and control is presented in a separate section.
Reactions of Chlorine with Ammonia (Formation of Chloramines). Chloramines
are formed by the reaction of aqueous chlorine with ammonia. Chloramine formation may be done purposefully by adding ammonia to a water containing free available chlorine if it is desirable to form monochloramine for maintenance of a stable
disinfectant residual in the distribution system (see Chapter 14). Conversely, chloramine formation may occur during the course of water treatment when chlorine is
added to a source water containing ammonia. The mixture that results from reaction
between free chlorine and ammonia may contain monochloramine (NH2Cl), dichloramine (NHCl2), and trichloramine or nitrogen trichloride (NCl3):
NH3 + HOCl = NH2Cl + H2O

(12.39)

NH2Cl + HOCl = NHCl2 + H2O

(12.40)

NHCl2 + HOCl = NCl3 + H2O

(12.41)

The sum of the concentrations of the three chloramine species is typically referred to
as combined chlorine and is often expressed in units of mg/L as Cl2. Combined chlorine is analytically distinguishable from free chlorine (Standard Methods, 1995). The
sum of the free and combined chlorine concentrations is referred to as total chlorine.
The relative amounts of the three species formed will depend primarily on pH,

temperature, and the ratio of chlorine to ammonia. At a chlorine:ammonia molar
ratio less than 1:1 or a weight ratio (Cl2:N) less than 5:1 and a pH of about 8, conditions that are common during the chloramination of drinking water, the principal


CHEMICAL OXIDATION

12.15

chloramine species formed is monochloramine. At higher chlorine:ammonia ratios
and lower pH values, dichloramine formation becomes important and, at mildly
acidic pH values in the range of 6 to 6.5, nitrogen trichloride can be formed. Because
dichloramine is unstable at pH 8 and decomposes essentially to nitrogen gas and
chloride, the addition of excess chlorine to an ammonia-containing water results in
the following overall reaction:
2NH3 + 3HOCl → N2(gas) + 3H+ + 3Cl− + 3H2O

(12.42)

This equation illustrates the so-called breakpoint phenomenon whereby chlorine
applied in sufficient doses (at a Cl2:N molar ratio greater than 1.5 or a weight ratio
greater than 7.5) will oxidize ammonia, resulting in the formation of a free chlorine
residual. This equation can be used to estimate the chlorine demand of ammonia,
although in practice the molar ratio tends to be closer to 2:1 (10:1 by weight).
If it is desirable to produce monochloramine to serve as a stable disinfectant
residual in the distribution system, this is usually done by adding more than 0.2 mg/L
ammonia as N for each mg/L of free Cl2 in order to ensure that the molar ratio of
Cl2:N is less than 1 so that only monochloramine is formed (Eq. 12.39). This is usually done because monochloramine, being a weaker oxidant, is a more stable and
persistent species than free chlorine. This contributes to its desirability as a secondary disinfectant in water distribution systems.
From the standpoint of its oxidation potential, monochloramine is too weak to
oxidize reduced iron and manganese and most taste- and odor-causing organics, and

it is too weak to eliminate natural organic color. While monochloramine will oxidize
natural organic material to some degree and produce some halogenated organic
material, it does not generally produce trihalomethanes (see below).
Part of the oxidation capacity of the chloramines derives from the hydrolysis
reactions represented by the reverse reactions of Eqs. 12.39 to 12.41. The free chlorine liberated, although often in small amounts, may be responsible in part for the
formation of some halogenated by-products associated with chloramines. Likewise,
the free ammonia liberated may contribute to nitrification problems in water distribution systems.
In general, the principal value of chloramines in water treatment is not as an oxidant but as a secondary disinfectant (see Chapter 14).
Reactions of Chlorine with Bromide. In drinking waters containing bromide,
chlorine will oxidize the bromide to produce hypobromous acid (HOBr):
HOCl + Br− → HOBr + Cl−

(12.43)

This reaction is very fast, with a second-order rate constant of 2.95 × 103 M−1 sec−1.
Hence, bromide represents another source of chlorine demand, and the presence of
bromide contributes to the depletion of free chlorine.
The resulting hypobromous acid, like hypochlorous acid, is a weak acid (pKa = 8.7
at 25°C) and will dissociate to some degree to form the hypobromite ion (OBr−),
depending upon pH:
HOBr → OBr− + H+

(12.44)

As noted in Table 12.1, hypobromous acid is a somewhat weaker oxidant than
hypochlorous acid. The formation of hyprobromous acid is important, however,
because hypobromous acid is capable of reacting with natural organic material to produce undesirable brominated disinfection by-products, such as bromoform (CHBr3)


12.16

CHAPTER TWELVE

and dibromoacetic acid. The presence of bromide and its subsequent oxidation to
hyprobromous acid is what contributes, along with hypochlorous acid, to the formation of mixed brominated and chlorinated DBPs such as bromodichloromethane
(CHBrCl2) and bromochloroacetic acid (see below). Because of their relative acidity
constants, more hypobromous acid remains in the undissociated HOBr form than
hypochlorous acid under most pH conditions encountered in practice. Accordingly,
because the undissociated forms of these species are stronger oxidants than their
deprotonated counterparts, hypobromous acid behaves as a stronger substituting
agent than hypochlorous acid, which results in the formation of greater amounts of
halogenated disinfection by-products in bromide-containing waters.
Chlorine Dioxide
Chlorine dioxide (ClO2) is a greenish-yellow gas at room temperature and has an
odor similar to that of chlorine. It is unstable at high concentrations and can explode
upon exposure to heat, light, electrical sparks, or shocks. Accordingly, it is never
shipped in bulk, but instead is generated on-site. Aqueous solutions are usually prepared from the gaseous chlorine dioxide generated, as chlorine dioxide is very soluble in water. It does not hydrolyze in water like chlorine does and remains in its
molecular form as ClO2. It is much more volatile than chlorine, and can easily be
stripped from aqueous solution if not properly handled.
Chlorine dioxide is usually generated by reacting sodium chlorite (NaClO2) with
either gaseous chlorine (Cl2) or hypochlorous acid (HOCl) under acidic conditions,
in accordance with the following reactions:
2NaClO2 + Cl2(g) → 2ClO2(g) + 2Na+ + 2Cl−

(12.45)

2NaClO2 + HOCl → 2ClO2(g) + 2Na+ + Cl− + OH−

(12.46)

In order to drive the reaction to completion and avoid the presence of unreacted
chlorite (ClO2−) in the product stream, some generators use excess chlorine,
although several variations in generator design have been developed which allow
for the sodium chlorite to be completely converted to chlorine dioxide without using
excess chlorine. Additionally, to overcome the basicity of sodium chlorite and the
hydroxide produced by Eq. 12.46, acid is sometimes added along with the hypochlorous acid to maintain the optimal pH for chlorine dioxide generation. Typically, pH
values in the range of 3.5 to 5.5 are preferred; more acidic pH values lead to the formation of chlorate (ClO3−):
NaClO2 + Cl2 + H2O → ClO3− + 2Cl− + 2H+ + Na+

(12.47)

Chlorine dioxide may also be prepared by acidification of a sodium chlorite solution:
5NaClO2 + 4H+ → 4ClO2(g) + 5Na+ + Cl− + 2H2O

(12.48)

Once generated, chlorine dioxide can be dissolved in water and is stable in the
absence of light and elevated temperatures. In the presence of the latter, or at high
pH values, it disproportionates to form both chlorite and chlorate:
2ClO2 + 2OH− → ClO2− + ClO3− + H2O
both of which are undesirable in drinking water (see below).

(12.49)


12.17

CHEMICAL OXIDATION

The primary application of chlorine dioxide in the past has been for taste and odor

control, although it is also an effective oxidant for reduced iron and manganese, and
is a good primary disinfectant (see Chapter 14). One of its principal advantages is that
it does not react with ammonia. Hence, much lower doses of chlorine dioxide are
required for most oxidative applications compared to chlorine dosage requirements.
Another advantage is that chlorine dioxide does not enter into substitution reactions
with NOM to the same degree that free chlorine does and, accordingly, does not form
trihalomethanes, haloacetic acids, or most other commonly observed halogenated
disinfection by-products that result from chlorination. Richardson et al. (1994) and
others have identified a number of disinfection/oxidation by-products of chlorine
dioxide treatment—for example, aldehydes, carboxylic acids, and some halogenated
compounds—but most of the latter were present at extremely low concentrations.
Because it does not tend to form halogenated DBPs to any significant degree, chlorine dioxide is enjoying renewed interest as a water treatment oxidant. An additional
benefit is that chlorine dioxide reacts only very slowly with bromide (Hoigne and
Bader, 1994). Hence, brominated by-products, either organic or inorganic (e.g., bromate), are not a concern following treatment with chlorine dioxide.
Hoigne and Bader (1994) and Tratnyek and Hoigne (1994) provide a listing and
discussion of the rate constants for reactions between chlorine dioxide and a variety
of organic and inorganic compounds, a number of which are of relevance to the
treatment of drinking water.
Chlorine dioxide typically reacts with most reducing agents (e.g., ferrous iron,
natural organic material) through a one-electron transfer (see Table 12.1):
ClO2 + e− → ClO2−

(12.50)

As a result, chlorite is considered to be the principal oxidation by-product of chlorine dioxide usage. Most researchers (e.g. Werdehoff and Singer, 1987) have
reported that approximately 50 to 70 percent (by mass) of the chlorine dioxide
applied during the course of drinking water treatment ends up as chlorite.
Chlorite has been demonstrated to exhibit a number of potential adverse health
effects based on studies with laboratory animals. Because of this, the Disinfectants/Disinfection By-Products Rule (USEPA, 1998) establishes a maximum contaminant level (MCL) of 1.0 mg/L for chlorite at representative locations in the
distribution system and a maximum residual disinfectant level (MRDL) of 0.8 mg/L

for chlorine dioxide at the point of entry to the distribution system. No regulations
exist for chlorate at this time. Given the observations that chlorine dioxide is rapidly
consumed during the course of water treatment and that up to 70 percent of the
applied chlorine dioxide is reduced to chlorite, the practical upper limit for chlorine
dioxide doses would be approximately 1.4 to 1.5 mg/L unless chlorite is removed
(see below).
A second concern associated with residual chlorite in the distribution system is that
it reacts with free chlorine, producing low levels of chlorine dioxide and/or chlorate:
HOCl + 2ClO2− → 2ClO2 + Cl− + OH−

(12.51)

HOCl + ClO2 → ClO3 + Cl + H

(12.52)

−

−

−

+

If present in tap water, the chlorine dioxide, being volatile, can be released into the
home or office environment when customers open their taps. This can lead to offensive chlorinous odors or, if new carpeting has recently been installed, the escaping
chlorine dioxide can react with organic compounds released from the carpeting to
produce other offensive odors.

12.18

CHAPTER TWELVE

Chlorite, whether it is present as a result of incomplete oxidation of sodium
chlorite in the chlorine dioxide generator (Eqs. 12.45 and 12.46) or by chemical
reduction of the chlorine dioxide during the course of treatment (Eq. 12.50), can be
removed from water by the application of ferrous iron salts (Iatrou and Knocke,
1992) or reduced sulfur compounds. No practical method for the removal of chlorate is available; hence, its formation during chlorine dioxide generation should be
minimized.

Ozone and Advanced Oxidation Processes
Ozone is an unstable gas that must be generated on-site. A simplified representation
of the chemistry involved in the formation of ozone (O3) is as follows:
O2 + energy → O + O

(12.53)

O + O2 → O3

(12.54)

The energy required to produce nascent or elemental oxygen (O) from molecular
oxygen (O2) is usually supplied by an electric discharge with a peak voltage from 8
to 20 kV, depending on the apparatus used. Dry, refrigerated, particle-free air, oxygen, or oxygen-enriched air is passed through a narrow gap between two electrodes
and a high-energy discharge is generated across the gap between the electrodes. This
corona or cold plasma discharge is induced by an alternating current that creates a
voltage cycle between the two electrodes.The yield of ozone will depend on the voltage, the frequency, the design of the ozone generator, and the type and quality of the
feed gas used. Ozone streams containing up to 14 percent ozone by volume can be
produced. Current ozone generators are available as low-frequency (50 to 60 Hz),

medium-frequency (400 to 1000 Hz), and high-frequency (2000 to 3000 Hz) systems.
Once generated, the ozone-enriched air or oxygen gas is passed through a gas
absorption device to transfer ozone into solution. This can be achieved through
either a countercurrent multistage bubble contactor, an in-line gas injection system,
or other such gas transfer devices.
Ozone is very unstable in aqueous solution. It is very reactive with a number of
common constituents in drinking water (e.g., NOM), and it also undergoes a spontaneous decomposition process, sometimes referred to as auto-decomposition. The
auto-decomposition of ozone is a complex chain reaction process involving several
free radical species. Decomposition may be initiated by a number of different water
constituents, such as hydroxide ion (e.g., high pH values), natural organic material,
and ferrous iron, or it may be initiated by the addition of hydrogen peroxide or by
irradiation with ultraviolet light. The reactions shown in Eqs. 12.55 through 12.60
illustrate the auto-decomposition scheme when hydroxide ion is the initiator.
OH− + O3 → HO2 + O2−

(12.55)

HO2 = H + O2

(12.56)

O2− + O3 → O2 + O3−

(12.57)

+

O3 + H → HO3

(12.58)

HO3 → O2 + OH

(12.59)

OH + O3 → HO2 + O2

(12.60)

+

−

−


CHEMICAL OXIDATION

12.19

These reactions constitute a chain mechanism because the hydroperoxyl radical
(HO2) and the superoxide ion (O2−) produced by the initiation reaction (Eq. 12.55)
generate new chain reactions that further contribute to ozone decomposition. In pure
water, the chain may be very long; that is, hundreds of ozone molecules may be
decomposed by a single initiation step. In natural waters, the lifetime of ozone
depends on several variables, including pH, temperature, total organic carbon (TOC)
concentration, and bicarbonate and carbonate concentrations. Bicarbonate and carbonate increase the lifetime of ozone by reacting with the hydroxyl radical (OH)
OH + HCO3− → OH− + HCO3

(12.61)

OH + CO3−2 → OH− + CO3−

(12.62)

and thereby decelarating the chain mechanism shown in Eqs. 12.56 to 12.60.
The bicarbonate and carbonate radicals (HCO3 and CO3−, respectively) are relatively unreactive intermediates that cannot propagate the chain. Thus, waters high in
bicarbonate and carbonate alkalinity will retain an ozone residual for longer periods
of time than low-alkalinity waters. This is especially important when ozone is used as
a disinfectant (see Chapter 14). Additionally, the radical scavenging activity of the
carbonate species increases with increasing pH because carbonate is a more effective scavenger than bicarbonate. This partially offsets the more rapid rate of the
hydroxide-induced initiation reaction (Eq. 12.55) at higher pH values (higher
hydroxide ion concentrations).
The hydroxyl radical (OH), one of the intermediates produced by the decomposition of ozone, is one of the strongest chemical oxidants known and is capable of
rapidly reacting with a myriad of organic and inorganic compounds (see below).
Accordingly, the oxidative properties of ozone depend significantly on the oxidative
characteristics of this free radical species.
Aieta et al. (1988) have presented an illustrative schematic of the behavior of
ozone in aqueous solution, based on the fundamental work of Hoigne and his
coworkers (e.g., Hoigne and Bader, 1976; Staehelin and Hoigne, 1982). Figure 12.3
shows that ozone reacts by two distinct types of pathways: a direct pathway involving molecular ozone (O3) and an indirect pathway originating with the decomposition of ozone to produce the hydroxyl free radical (OH). Direct reactions involving
molecular ozone are very selective; ozone reacts very rapidly with some species—for

FIGURE 12.3

Reaction pathways for ozone. (Source: Aieta et al., 1988.)


12.20

CHAPTER TWELVE

example, phenol and mercaptans—but very slowly with other species—for example,
benzene and tetrachloroethylene (PCE). Conversely, the OH radical is nonselective
in its behavior, reacting rapidly with a large number of species. Additionally, the OH
radical reacts rapidly with molecular ozone (see Eq. 12.60 and Figure 12.3), thereby
contributing to the autocatalytic rate of ozone decomposition. Hydroxyl radical
scavengers, such as the bicarbonate and carbonate ion, react with the hydroxyl radical (see also Eqs. 12.61 and 12.62), removing it from the cycle and, in so doing, decelerating the kinetics of ozone decomposition, which promotes the stability of
molecular ozone in solution, as already noted.
Hydroxyl free radicals can be produced through a number of other pathways in
addition to the hydroxide-induced ozone decomposition chain previously described.
For example, the addition of both hydrogen peroxide and ozone to water accelerates
the decomposition of ozone and enhances production of the hydroxyl radical. The
hydrogen peroxide (H2O2) dissociates into the hydroperoxide ion (HO2−):
H2O2 = HO2− + H+

(12.63)

The hydroperoxide ion then reacts with molecular ozone to produce the superoxide
ion (O2−) and the hydroxyl radical, plus molecular oxygen:
HO2− + O3 → OH + O2− + O2

(12.64)

The hydroxyl radical and the superoxide ion then participate in the ozone
decomposition cycle depicted by Eqs. 12.57 to 12.60, leading to an accelerated production of more hydroxyl free radicals. Because reactions 12.63 and 12.64 tend to be
appreciably faster than reaction 12.55 under most conditions, the conjunctive use of
O3 and H2O2 tends to be a more effective method of generating the highly reactive,
nonselective hydroxyl radicals for chemical oxidation reactions.
Another method for generating hydroxyl radicals is by ultraviolet (UV) irradiation of hydrogen peroxide. The UV irradiation provides the energy to split the

hydrogen peroxide into two hydroxyl radicals:
H2O2 + UV → 2 OH

(12.65)

This process tends to be slower than reactions 12.63 and 12.64 and is therefore a less
effective method of generating hydroxyl radicals.
Ultraviolet irradiation of waters containing dissolved molecular ozone leads to
the formation of hydrogen peroxide. The hydrogen peroxide then reacts with molecular ozone in the same manner as it does through Eqs. 12.63 and 12.64, which produces the hydroxyl radical and other species that enter the ozone decomposition
cycle depicted by Eqs. 12.57 to 12.60.
These reactions, all of which involve the accelerated production of the hydroxyl
free radical, are termed advanced oxidation processes (AOPs). Often, when applying
ozone to water, it is difficult to distinguish between reactions that are attributable to
molecular ozone and those that are attributable to the hydroxyl radical. It is
believed that hydroxyl radical reactions are at the heart of all reactions involving
molecular ozone, except perhaps for disinfection reactions, reactions with some
solutes that have very high reaction rate constants, and reactions that take place in
the presence of high concentrations of hydroxyl radical scavengers.
Kinetics of Ozone Reactions. The kinetics of reactions of molecular ozone and
hydroxyl radicals have been studied extensively. The most comprehensive listing of
reaction rate constants is that provided by Hoigne and coworkers (e.g., Hoigne and


12.21

CHEMICAL OXIDATION

Bader, 1983a,b; Hoigne et al., 1985). As previously noted, some solutes react very
quickly with molecular ozone. For example, the oxidation of sulfide by molecular ozone is extremely rapid, with a second-order rate constant on the order of
109 M−1 sec−1. The second-order rate constants for phenol and naphthalene oxidation

by molecular ozone are 1300 and 3000 M−1 sec−1, respectively, also relatively high.
Conversely, the second-order rate constants for benzene and tetrachloroethylene
oxidation are 2 and <0.1 M−1 sec−1, respectively. The oxidation of trichloroethylene
and atrazine, two common contaminants of drinking water, are also relatively low,
being in the range of 10 to 20 M−1 sec−1.Also, ammonia reacts very slowly with molecular ozone in the pH range of interest in drinking water treatment. Amines and
amino acids react quickly when the amino group is not protonated.
In general, compounds that dissociate in water react with molecular ozone at
rates that depend on the individual species present. For example, the phenolate
anion reacts much faster than its protonated counterpart phenol. The same is true
for formate relative to formic acid. Hence, the rates of ozonation may increase or
decrease as pH changes, depending on the specific compound being oxidized. Also,
while benzene itself is relatively unreactive toward molecular ozone, derivatives of
benzene tend to be much more reactive, especially with the addition of electronwithdrawing substituents on the aromatic ring.
In the case of the hydroxyl radical, the magnitude of the second-order reaction
rate constants with various solutes is much more uniform, tending to range from 107
to 1010 M−1 sec−1. It is for this reason that the hydroxyl radical is often called nonselective with respect to its reactivity, in contrast to molecular ozone, which is highly
selective in that its rate constants with various solutes range over more than 12
orders of magnitude. Atrazine, trichloroethylene, and tetrachloroethylene, all relatively inert toward molecular ozone, are oxidized at appreciable rates by the
hydroxyl radical, and therefore by AOPs. Because of the large magnitude of the rate
constants for oxidation reactions involving the hydroxyl radical, the ratedetermining step in these reactions is the rate at which the radicals are generated
either by ozone auto-decomposition or by the various AOPs previously discussed.
Hence, in many cases (e.g., atrazine or trichloroethylene), the oxidation of these
solutes is enhanced at elevated pH values because of the more rapid generation of
hydroxyl radicals by hydroxide-induced ozone decomposition.
Formation of Biodegradable Organic Material. When ozone reacts with organic
contaminants in water, including natural organic material, it partially oxidizes them
to lower molecular weight, more polar species, including a variety of aldehydes and
organic acids (see below). These oxidation by-products, while not believed to be
harmful in themselves, tend to be biodegradable and may contribute to biofouling
problems in the water distribution system if not properly controlled. Often, ozonation is followed by a biologically active filtration process to remove these biodegradable organic materials. Ozonation by-products are discussed in greater detail later.

Reactions of Ozone with Bromide. While ozone will not react with natural
organic material to directly produce halogenated DBPs, it may result in the formation of brominated DBPs in waters containing bromide. Ozone is capable of oxidizing bromide to hypobromous acid in the same manner that chlorine does:
O3 + Br− + H+ → HOBr + O2

(12.66)
−1

−1

This is a very fast reaction, with a reaction rate constant of 160 M sec . The hypobromous acid can then react with natural organic material to produce brominated


12.22

CHAPTER TWELVE

DBPs, such as bromoform, dibromoacetic acid, or bromopicrin. The extent of formation of these DBPs depends upon the bromide ion and TOC concentration of the
water and its pH. In general, the concentration of these halogenated DBPs produced
indirectly by ozonation is at least an order of magnitude lower than those formed by
chlorination. This is discussed in greater detail later.
A particular concern associated with the ozonation of bromide-containing waters
is the formation of bromate (BrO3−). Bromate is a possible human carcinogen and is
currently regulated at 10 µg/L, the practical quantitation level (PQL) for bromate,
although it is expected that this level will decrease in the near future as more sensitive analytical methodologies are developed. Bromate is produced by a number of
possible pathways involving molecular ozone and the hydroxyl radical. One such
pathway (unbalanced stoichiometrically) proceeds through the formation of hypobromite (OBr−) and bromite (BrO2−):
O3 + Br− → OBr−

(12.67)

OBr + O3 → BrO2

(12.68)

−

−

BrO2 + O3 → BrO3
−

−

(12.69)
−

In the pH range of most natural waters, a large portion of the OBr is protonated
as HOBr (the pKa for hypobromous acid is 8.7 at 25°C), so that the rate of formation
of bromite by reaction 12.68 is slowed considerably with decreasing pH. If the rate
of formation of hydroxyl radicals in the water is high, bromate may also be formed
through the hypobromite radical (BrO) as follows:
HOBr/OBr− + OH → BrO

(12.70)

BrO + O3 → BrO3

−

(12.71)

Techniques for controlling bromate formation most often involve ozonation at
slightly acidic pH values, multistage ozonation in which the ozone is added at several
different application points, and the use of ammonia to tie up the hypobromous acid
produced:
NH3 + HOBr → NH2Br + H2O

(12.72)

Additionally, once formed, bromate can be removed by chemical reduction using
reduced sulfur compounds, such as bisulfite (HSO3−) or ferrous iron. Granular activated carbon is also capable of adsorbing bromate, albeit to a limited degree, and
medium-pressure ultraviolet irradiation decomposes bromate to bromide. Bromate
can also be reduced under anaerobic conditions.

Potassium Permanganate
Potassium permanganate (KMnO4) has been used as a water treatment oxidant for
decades. It is commercially available in crystalline form and either it is fed into solution directly using a dry chemical feeder or a concentrated solution is prepared onsite from which the desired dose is metered into the water.
Permanganate contains manganese in the +VII oxidation state. Under most
treatment applications, oxidation by permanganate involves a three-electron transfer, with the permanganate (MnO4−) being reduced to insoluble manganese dioxide,
MnO2(s):
MnO4− + 4H+ + 3e− → MnO2(s) + 2H2O

(12.73)


12.23

CHEMICAL OXIDATION

The manganese dioxide produced is a black precipitate that, if not properly removed

by a suitable solid-liquid separation process, will create black particulate deposits in
the distribution system and on household plumbing fixtures. Most often, removal of
MnO2(s) is achieved by conventional clarification or filtration processes. Because
most operators are fearful of seeing pink water (reflecting unreacted permanganate)
coming through their filters, permanganate is commonly added at the head of the
treatment plant, as close to the intake as possible. This allows sufficient time for the
permanganate to perform its oxidative function and to be reduced completely to
solid manganese dioxide prior to filtration.
The kinetics of oxidation reactions involving permanganate tend to be more
rapid with increasing pH values. Hence, in some cases, addition of a base prior to filtration may be desirable to hasten the reduction of permanganate.
The manganese dioxide that results from permanganate reduction (Eq. 12.73)
may have some beneficial attributes. MnO2(s) is an effective adsorbent for ferrous
iron (Fe2+), manganous manganese (Mn2+), radium (Ra2+), and other trace inorganic
cationic species. Accordingly, some additional removal of these contaminants occurs
as a result of permanganate treatment beyond that achieved simply by oxidation. In
fact, the adsorptive behavior of MnO2(s) is the principle underlying the historic
manganese greensand process, in which the filter media is coated with manganese
dioxide, which subsequently serves as an adsorbent for Fe2+, Mn2+, and Ra2+ in the filter influent. The filter backwash water is treated with permanganate, or low doses of
permanganate are applied to the filter influent to oxidize the adsorbed metals,
thereby creating additional adsorption sites.
Solid manganese dioxide is also capable of adsorbing natural organic material
that serves as DBP precursors.This benefit is particularly pronounced in hard waters
(Singer and Colthurst, 1982), presumably because of the bridging action of calcium
and magnesium.

Mixed Oxidants
The electrolysis of brine has been used since the nineteenth century to produce
chlorine on an industrial scale. This technology has recently been modified and
adopted to electrochemically produce a mixture of free chlorine and other powerful
oxidants and disinfectants for drinking water applications for small, rural water supplies. Both liquid- and gas-phase generators are available.

The underlying principle is based on fundamental electrochemical theory. At the
anode of the electrochemical cell, chloride is oxidized to chlorine, which subsequently hydrolyzes to hypochlorous acid:
2Cl− → Cl2 + 2e−

(12.74)

Cl2 + H2O → HOCl + Cl + H
−

+

(12.75)

The pH of the anodic stream tends to be on the order of 3 to 5. At the cathode, the
water is reduced to hydrogen gas, producing a strongly alkaline solution with a pH
of about 10 to 11:
2H2O + 2e− → H2 + 2OH−

(12.76)

The anodic and cathodic streams are separated by a semipermeable barrier.
While free chlorine is the primary oxidant produced by these mixed oxidant generators, other reactions occur at the anode, purportedly resulting in the formation of


12.24

CHAPTER TWELVE

ozone, chlorine dioxide, hydrogen peroxide, other short-lived reactive oxidant
species, and a number of inorganic by-products such as chlorite, chlorate, and bromate. However, the composition of the oxidant streams from these mixed oxidant

generators has not been fully characterized. The output from one of these generators has been shown to produce on the order of 200 to 400 mg/L of free available
chlorine (Dowd, 1994; Gordon, 1998), but no ozone, chlorine dioxide, or hydrogen
peroxide was detected in the anodic stream (Gordon, 1998). These findings appear
to contradict those of Venczel et al. (1997), who found that the oxidant product had
disinfecting properties distinctly different than those of free chlorine alone. More
research is needed to characterize these product streams before widespread use of
these generators is recommended.

APPLICATIONS OF OXIDATION PROCESSES
Application of Oxidants to Water Treatment Practice
This section describes typical applications of oxidants and the role they play in overall water treatment practice. These oxidants are most often used for the oxidation of
reduced iron and manganese, destruction of taste- and odor-causing organic contaminants, elimination of color, and the destruction of synthetic organic chemicals of
public health concern.Additionally, many of these oxidants act as coagulant aids and
are also employed as part of an overall program for the control of potentially harmful disinfection by-products. The formation and control of oxidation and disinfection
by-products are discussed in a separate section. Many of these oxidants are also
powerful disinfectants and therefore serve the dual purposes of oxidation and disinfection. Disinfection is discussed in detail in Chapter 14.
Control of Iron and Manganese. Iron and manganese are relatively soluble under
reducing conditions, for example, in groundwaters, stagnant surface waters, and in
the hypolimnetic waters of eutrophic lakes, reservoirs, and impoundments. Correspondingly, they are quite insoluble under oxidizing conditions, for example, in flowing streams and in the epilimnetic waters of lakes or impoundments, or in
hypolimnetic waters that have been subject to hypolimnetic aeration to maintain
oxidizing conditions. The reduced forms of iron and manganese—ferrous iron,
Fe(II), and manganous manganese, Mn(II)—may occur as the free metal ions, Fe2+
and Mn2+, which is often the case in most groundwaters, or they may be found complexed to various degrees with natural organic material, as is often the case in surface waters and highly colored groundwaters. During and immediately following
lake overturn—that is, when the iron- and manganese-rich hypolimnetic water is
mixed with the remainder of the lake water—dissolved iron and manganese levels in
the upper portions of the lake can increase appreciably.
The primary concern with elevated levels of dissolved iron and manganese in
water is that when they become oxidized to insoluble ferric hydroxide, Fe(OH)3(s),
and manganese dioxide, MnO2(s), they precipitate and cause reddish-orange or
black deposits, respectively, to appear on plumbing fixtures and to create stains during laundering operations:

Fe2+ + ᎏ14ᎏO2 + ᎏ52ᎏH2O → Fe(OH)3(s) + 2H+

(12.77)

Mn2+ + ᎏ12ᎏO2 + H2O → MnO2(s) + 2H+

(12.78)


12.25

CHEMICAL OXIDATION

Dissolved Fe(II) and Mn(II) are usually removed from water by oxidizing them
under engineered conditions to their insoluble forms through the addition of an oxidant and then removing the precipitated ferric hydroxide and manganese dioxide by
sedimentation and filtration. The oxidants used most often for this purpose are oxygen, chlorine, permanganate, chlorine dioxide, and ozone.
The kinetics of oxidation of Fe(II) by oxygen are relatively rapid at pH values
above 7, provided that the ferrous iron is not complexed by organic material. This is
often the case in the removal of iron from groundwater. The rate expression for the
oxygenation of ferrous iron is:
d[Fe(II)]
− ᎏᎏ = k[Fe(II)][OH−]2PO2
dt

(12.79)

The kinetics are first order with respect to the concentration of ferrous iron and the
partial pressure of oxygen (PO2 ) and second order with respect to the hydroxide ion
concentration (Stumm and Morgan, 1996); the latter illustrates the strong pH dependency of Fe(II) oxidation by oxygen. If the ferrous iron is complexed by organic
material, the rate of oxidation by oxygen can be very slow and a stronger oxidant,

such as chlorine or permanganate, may be needed.
In the case of Mn(II), the kinetics of oxidation by oxygen have been shown to
conform to the following rate expression:
d[Mn(II)]
− ᎏᎏ = ko[Mn(II)] + k[Mn(II)][MnO2(s)])
dt

(12.80)

where k = k1[OH−]2 PO2. This is an autocatalytic reaction whereby the by-product of
the oxidation reaction, manganese dioxide or MnO2(s), catalyzes the oxidation of
reduced manganese (Stumm and Morgan, 1996). Because the initial oxidation step
represented by the first term in Eq. 12.80 is relatively slow at pH values below 9, a
strong chemical oxidant, such as chlorine or permanganate, is often used in place of
oxygen. Alternatively, because the reaction is catalyzed by manganese dioxide, and
manganese dioxide strongly adsorbs Mn2+, a common practice for manganese oxidation and removal is to allow the filter media to be coated by manganese dioxide.
Reduced manganese then adsorbs to the manganese dioxide coating and can be oxidized within the filter bed using chlorine or permanganate. This sequence of reactions can be represented as:
Mn2+ + MnO2(s) → MnO2(s) − Mn2+

(12.81)

MnO2(s) − Mn + HOCl + H2O → 2 MnO2(s) + Cl + 3H
2+

−

+

(12.82)

Commercially available glauconite (greensand) can also be used as a filter medium
in the same manner. It has an affinity for the adsorption of manganese.
Last, it should be noted that, because ozone is such a strong oxidant, it is capable
of oxidizing reduced manganese to permanganate (see illustrative calculations at the
beginning of this chapter):
Mn2+ + ᎏ52ᎏO3 + ᎏ32ᎏH2O → MnO4− + ᎏ52ᎏO2 + 3H+

(12.83)

As a result, it is not uncommon to observe pink water if excess ozone is added to a
water containing reduced manganese.
Destruction of Tastes and Odors. Tastes and odors occur in water from a variety
of sources, most notably algae, actinomycetes, organic and inorganic sulfides (e.g.,