Reaction Kinetics and The Kinetics-Based Interpretation of Equilibrium
Reaction Kinetics and The Kinetics-Based Interpretation of Equilibrium
Reaction Kinetics and The Kinetics-Based Interpretation of Equilibrium
77
Contents
78
reactions are proceeding at equal rates. In this chapter, we explore reaction kinetics
in greater detail.1
Chemical reactions happen when reactant molecules collide with sufficient en-
ergy to break existing bonds and rearrange into a different configuration. Therefore,
a complete description of the process requires knowledge about the frequency of
collisions, the energy involved in the collisions, the energy required for molecular
rearrangement, and the relative stability of the various configurations possible. Of-
ten, the net reaction that is observable at a macroscopic scale is actually the result
of several reactions. That is, when two reactants are mixed, they might form a new
species, which then reacts with others to form different species, which in turn might
react with yet others. Eventually, a stable suite of products is formed. In some cases,
the individual steps are slow enough to be readily identified and studied. In others,
though, the intermediate steps occur very quickly, and we might not even be aware
of their existence without the assistance of sophisticated analytical instrumentation.
The step-by-step sequence that leads to the overall reaction is referred to as the reac-
tion mechanism or reaction pathway.
Each step in the reaction pathway occurs as the result of individual, sufficiently
energetic collisions among the reactants for that step and is called an elementary
reaction. Theory suggests that collisions among more than two molecules are rare,
so that any overall reaction involving three or more reactants is virtually certain to be
nonelementary.
A hypothetical sequence of elementary reactions that leads to an overall, nonele-
mentary reaction is shown below. The designation 2 (X + C )* Y + Z) means that
the elementary reaction is between one X and one C molecule, but that two such
elementary reactions occur each time the overall sequence proceeds.
In this sequence, species X, Y, and Z are intermediates that are created and then
destroyed by the elementary reactions, so the concentrations of these species do not
change when the overall reaction occurs.
An elementary reaction is initiated when two reactant molecules approach one
another, causing the bonds in each to become strained. This strain increases the po-
1 Thecontents of this chapter draw heavily on the discussion in Chapter 3 of Benjamin, M.M. and
Lawler, D.F. Water Quality Engineering: Physical/Chemical Treatment Processes, J. Wiley & Sons,
Hoboken, NJ (2013).
79
tential energy stored in the bonds, much like compressing a spring increases the po-
tential energy stored in it. The intensity of the strain and the amount of energy stored
in the bonds increase dramatically with decreasing separation of the molecules. Since
energy is conserved during this interaction, the increased bond energy must be bal-
anced by a decrease in energy elsewhere in the system. In this case, most of the en-
ergy used to strain the bonds comes from molecular kinetic energy — the molecules
slow down as the distance between them decreases. In the absence of other factors,
the molecules would eventually stop their mutual approach and begin moving away
from one another, thereby relieving the strain and converting the bond energy back
into kinetic energy.
Although the molecules might indeed separate without undergoing any long-term
changes, it is also possible that, as the molecular structures adjust to the strain, new
bonds will begin to form. At some critical point, the original bonds may be suffi-
ciently distorted, and the new bonds may form to a sufficient extent, that the strain
can be relieved just as easily by rearrangements that form product molecules as by a
reversal of the process and re-formation of the original reactants. At this point, the
molecules are not identifiable as either reactants or products, but are an intermedi-
ate species of negligible stability, analogous to a ball perched at the crest of a hill
with equal likelihood of returning along the path of its climb or proceeding down the
other side. The amount of energy necessary to bring molecules from far apart (no
interaction) to the critical point is called the activation energy E ⇤ , and the process
of reaching the critical condition is referred to as overcoming the activation energy
barrier.
Catalysts operate by providing an alternative path by which a reaction can occur.
Specifically, they allow reactants to be converted to products via a route that has a
lower activation energy than the route that is taken in their absence. Extending the
analogy to a ball on a hill, a catalyst might be viewed as providing an alternative path
for the ball to follow that does not require quite so much of a climb before arriving
at a point on the downhill slope. In some cases, catalysts can increase the rate of an
overall reaction by many orders of magnitude. However, catalysts cannot alter the
overall energy change associated with the reaction or the ultimate mix of products
that forms.
Although the preceding discussion focuses on the changes in bond energy within
the reacting molecules as they approach one another, the amount of energy in the
molecules changes in a myriad of other ways as well during this process (e.g., the
bond strength between the reactants and their nonreacting neighbors also changes).
The sum of the energy terms that can change as the reaction proceeds is sometimes
referred to colloquially as the heat content of the molecules, or more formally as a
thermodynamic parameter known as the enthalpy, H. The enthalpy per mole of a
substance i (the molar enthalpy of i) is represented as H i , and the enthalpy change
accompanying the complete reaction, when normalized to the amount of material
80
reacting, is represented as DH r . The enthalpy and its relationship to other molecular
properties are discussed in detail in Chapter 4; here, we simply introduce the term
and use it without further explanation, understanding it as a composite measure of
various components of molecular energy that can change as a reaction proceeds.
In addition to the energy of the colliding molecules, their orientation when they
collide and the distribution of energy within the molecules can affect the likelihood
that a reaction will occur. For example, if the molecules are not spherically sym-
metric, only a fraction of all collisions can cause the bonds to distort in a way that
leads to the formation of product, even if the collisions involve an amount of en-
ergy greater than the activation energy. Thus, the overall rate of reaction depends
on the frequency with which reactant molecules collide, the likelihood that colliding
molecules are properly oriented and have sufficient kinetic energy to overcome the
activation barrier, and the rate at which the activated species is converted to products.
The most widely accepted mathematical model for chemical reaction kinetics is
called the activated complex or transition state model. In this model, the passage
from reactants to products is viewed as a continuum, with the reactant molecules be-
ing converted to products by proceeding along a path on the “energy landscape” de-
scribed above. Molecules that have acquired an amount of energy equal to or greater
than the activation energy are called activated complexes. For a generic elementary
reaction between A and B, the activated complexes are commonly represented as
AB⇤ ; that is, the complete elementary reaction is A + B )* AB⇤ ! P. The arrows
in this representation signify that, as they ‘climb’ the activation barrier, molecules
of A and B might reverse their approach and revert to separate molecules, but once
they overcome that barrier, they proceed to form P. A schematic of the key energy
relationships during the transition from reactants through the activated complex to
products according to this model is shown in Figure 3.1 for a system in which the
reactants and products are all in their standard states.2
Figure 3.1 includes three curves — one for the enthalpy (H) of the molecules,
one for the negative product of the system temperature and the entropy (S) of the
molecules, and one for their Gibbs energy (G). Like the enthalpy, the entropy and
Gibbs energy are thermodynamic parameters that are explained in detail in Chapter 4,
and we use them here without further definition. These three parameters are related
by
DG = DH T DS (3.1)
For this example reaction, the products have more enthalpy than the reactants,
2 The preceding discussion might give the impression that the reacting molecules engage in only one
collision as they climb the activation barrier. In truth, the activated complex model suggests that, as the
reactant molecules approach one another, each of them participates in many collisions that collectively
increase its energy content. Simultaneously, the mutual approach of the reactants strains the bonds
and converts some of the kinetic energy acquired in the collisions into chemical energy. The frequent
nonreactive collisions are essential for the molecules at all chemical energy levels to be at the same
temperature, which is a central assumption of the activated complex model.
81
so energy must be added to keep the system at constant temperature as the reaction
proceeds. Such a reaction is referred to as endothermic. A reaction in which the
products have less enthalpy than the reactants is referred to as exothermic. In the
activated complex model, the activation energy is associated with the change in Gibbs
energy between the reactants and the activated complexes. Note that the magnitude
of the activation energy required to induce the reaction is independent of the change
in Gibbs energy accompanying the overall reaction.3
Figure 3.1 Schematic representation of the transition from reactants to products according to
the activated complex model. H, G, and S are the enthalpy, Gibbs energy, and
entropy of the system at any point in the process (i.e., the sum of those values
for all the species that participate in the reaction), and DHr , DGr , and DSr are the
corresponding changes for the overall reaction.
The key result of the activated complex model is that the rate at which A and B
cross the activation barrier to form P is
3 Diagrams like Figure 3.1 are intended to convey a combination of qualitative and quantitative in-
formation, and it is important to understand which parts of the diagram fall into each category. Because
the properties of activated complexes are not measurable (indeed, the very existence of such species
is a theoretical construct), one cannot know the value of their enthalpy, entropy, or Gibbs energy. As
a result, any representation of the magnitude of the activation barrier is conceptual and, at best, semi-
quantitative. On the other hand, the reactants and products of the overall reaction are real species whose
thermodynamic properties are unambiguous. Thus, even though the heights of the activation barriers for
the forward and reverse reactions cannot be known with certainty, the difference in those two heights is
measureable.
82
kB T ⇤
rP = K cA cB (3.2a)
h
= kcA cB (3.2b)
EAr 1
ln k = ln kAr (3.4)
R T
Equation (3.4) indicates that a plot of ln k versus 1/T is expected to be a straight
line with slope EAr /R and intercept ln kAr , as shown in Figure 3.2.
Arrhenius’s result was empirical, and it is widely applied to characterize the tem-
perature dependence of rate constants for both elementary and nonelementary re-
actions. In an effort to blend this empirical relationship with the more theoretical
activated complex model, EAr is commonly referred to as an activation energy and
interpreted (not entirely correctly) as the enthalpy required to form the activated com-
plexes from the reactants.
83
Figure 3.2 Characteristic plot of the rate constant versus inverse absolute temperature, from
which the values of kAr and EAr can be computed; ln kAr is found by extrapolating
the straight line to the hypothetical condition 1/T = 0.
84
metal ions (e.g., Zn2+ ) and dissolved anions (e.g., SO2–4 ) or neutral molecules (e.g.,
NH3 ). A simple way to think about such a reaction is that a weak bond exists be-
tween each of the reacting species and the surrounding water molecules. When the
two species approach one another, one of the water molecules adjacent to each of
them is released, and a bond between the two nonwater species forms in its place.
These kinds of reactions are described in more detail in Chapter 10. Like acid/base
reactions, they tend to proceed very quickly in aqueous solutions at normal tempera-
tures, and they therefore typically reach equilibrium in all systems of interest. Again,
a few exceptions exist. For example, the activation energy for stripping away water
molecules that surround Cr3+ ions is unusually large, so reactions like (3.5) are slow
if A or B is Cr3+ .
The only major category of aqueous-phase reactions that are frequently slow com-
pared to the amount of time available are those in which electrons are transferred from
one molecule to another. Such reactions are called oxidation/ reduction or ‘redox’ re-
actions; these reactions are the central topic of Chapter 12.
Not all redox reactions in solution are slow; some, such as the reaction of hydro-
gen sulfide (H2 S) with hypochlorous acid (HOCl) proceed very quickly. The rates of
many others depend strongly on the pH. For example, as shown in Figure 3.3a, the
reaction of dissolved ferrous iron (a term referring collectively to various chemical
species containing iron (Fe) atoms with a +2 charge) with dissolved oxygen is very
slow in acidic solutions but proceeds rapidly if the solution is made alkaline. The
reaction rates of many organic compounds with HOCl and OCl– also have a complex
dependence on pH, as exemplified by tetracycline in Figure 3.3b. On the other hand,
the redox reactions of many organic molecules with oxygen are extremely slow at
any pH, so that, in the absence of catalysts (e.g., enzymes), the reactants can coexist
and barely react at all over periods of centuries. The rates of many important envi-
ronmental redox reactions have been studied extensively and are discussed in detail
in many books and journal articles.4
The other broad category of relatively slow reactions that involve aqueous solu-
tions are those in which a species crosses a phase boundary (heterogeneous reac-
tions). This category includes transfer across gas/liquid and solid/liquid boundaries
when, for instance, carbon dioxide [CO2 (g)] dissolves into a solution or diffuses out
of it (described in Chapter 9), or solids such as calcium carbonate [CaCO3 (s)] or iron
hydroxide [Fe(OH)3 (s)] precipitate or dissolve (Chapter 11). In these cases, the actual
transfer from one phase to the other usually occurs rapidly, so that the two phases are
very nearly equilibrated right at the interface. However, concomitant processes that
occur adjacent to the interface are often slow, and they can place significant limita-
tions on the rate of the overall reaction. Such processes include diffusion of dissolved
species from the bulk solution through the liquid boundary layer and, in the case of
4 See,for example, Schwarzenbach, R., Gschwend, P., and Imboden, D. Environmental Organic
Chemistry, 2nd ed., J. Wiley & Sons, Hoboken, NJ (2003), or Stumm, W., and Morgan, J.J. Aquatic
Chemistry, 3rd ed., J. Wiley & Sons, New York (1996).
85
precipitation or dissolution reactions, the formation or breaking of bonds between the
reacting ions and the solid structure. More detailed discussions of these processes is
provided in Section 3.10 and in other texts.5
Figure 3.3 pH–dependence of some important environmental reaction rates. (a) Apparent
(observed) first-order reaction rate constant for oxidation of ferrous iron by dis-
solved oxygen, in a solution equilibrated with pure O2 (g) at 1 atm total pressure.
Open circles are from Singer and Stumm (1970), and filled circles from Millero
et al. (1987). From Benjamin, M.M. and Lawler, D.F. Water Quality Engineer-
ing: Physical/Chemical Processes, John Wiley & Sons, Hoboken, NJ (2013). (b)
Apparent second-order rate constant for oxidation of tetracycline by OCl species
(HOCl and OCl– ). After Wang, et al. (2011).6
Wiley & Sons, Hoboken, NJ (2009), or Benjamin, M.M., and Lawler, D.F. Water Quality Engineering:
Physical/ Chemical Processes, J. Wiley & Sons, Hoboken, NJ (2013).
6 Singer, P.C., and Stumm, W. (1970) “The solubility of ferrous iron in carbonate-bearing waters.”
Science 167, 1121-1123; Millero, F.J., Sotolongo, S., and Izaguirre, M. (1987) “The oxidation kinetics
of Fe(II) in seawater.” Geochim. Cosmochim. Acta 51, 793-801; Wang, P., He, Y.L, and Huang, C-H.
(2011) “Reactions of tetracycline antibiotics with chlorine dioxide and free chlorine.” Water Research
45, 1838-1846.
86
the key parameters. That is, for an elementary reaction between A and B (Equation
3.6), the reaction rate might be written as shown in either Equation (3.7a) or (3.7b):
A + B )* products (3.6)
rf = kf0 {A} {B} (3.7a)
rf = kf [A] [B] (3.7b)
where rf is the rate of the forward reaction, expressed as the mass or number of moles
of A or B reacting per unit volume per unit time (e.g., mol/L-s), and kf0 and kf are both
referred to as the forward reaction rate constant (kf0 when the rate is computed based
on reactant activities, and kf when it is computed based on concentrations). The rate
constant accounts for all factors that affect the reaction rate, other than the concentra-
tions or activities of the reactants. The primary such factors are the activation energy
and temperature, although ionic strength also plays a role for reactions in solution.
Because chemical activities are dimensionless, kf0 has the same units as rf (e.g.,
mol/L-s). On the other hand, for a two-molecule (bimolecular) reaction, kf has units
corresponding to rf divided by concentration squared (e.g., L/mol·s). The two forms
of the rate constant are related by:
g Ag B
kf = kf0 (3.8)
[A] [B]
2A )* products (3.9)
87
sufficient energy via a collision with a second molecule in the system. In such a case,
the reaction and rate expression might be written as follows:
A + X )* products + X (3.11)
rf = kf0 {A} {X} = kf [A] [X] (3.12)
where X represents a composite of all the molecules in the system (because any of
those molecules might collide with A and provide the activation energy needed for
the reaction). In aqueous solutions, water molecules account for the overwhelming
majority of X, and both [X] and {X} are essentially identical in all such solutions.
As a result, the rate can be expressed as the product of the activity or concentration
of A and an effective rate constant that incorporates the activity or concentration of
X:
0
rf = kf,eff {A} = kf,eff [A] (3.13)
0
where kf,eff = kf0 {X} and kf,eff = kf [X].
If the activation energy for a reaction is very small, the reaction occurs virtually
every time that the reactant molecules collide. The overall rate of such reactions
therefore approaches the collision rate, which in turn is limited by the rate at which
reactants can diffuse toward one another. Reactions that proceed at rates close to
these maximum predicted values are said to be diffusion-controlled. For bimolecu-
lar elementary reactions between uncharged solutes in aqueous solutions, diffusion-
controlled rate constants at room temperature are on the order of 1010 (mol/L) 1 s 1 ;
this rate constant can be up to about an order of magnitude larger if the reaction is be-
tween two ions that have opposite charges, and up to about two orders of magnitude
smaller if it is between ions that have like charges.
Although the reaction rates of some acid-base reactions approach those estimated
based on diffusion control, the vast majority of reactions in aquatic systems have
activation energy barriers that are much higher than the barrier imposed by diffusion;
the corresponding reaction rates are anywhere from a few orders of magnitude to tens
of orders of magnitude slower than the diffusion-controlled rates. In such cases, the
energy required for bond rearrangement provides the dominant impediment to the
progress of the reaction, and the reaction rate is said to be chemically controlled.
88
unlikely (e.g., if an elementary, bimolecular reaction generates three molecules as
products, the likelihood of the reverse reaction is very small). Regardless of its
likelihood, though, the reverse reaction is possible in principle. Such reversibility
is intimately tied to the idea of chemical equilibrium because, as noted previously,
equilibrium can be described as a condition in which a reaction is proceeding in the
forward and reverse directions at equal rates. We next derive the quantitative link
between these two important concepts using the following generic example reaction,
which we assume to be elementary in both directions:
f k
A+B ) * C+D (3.14)
kr
{C} {D}
Q {A} {B} kr0 {C} {D} rr
= = = (3.17)
K kf0 /kr0 kf0 {A} {B} rf
This result indicates that, if Q > K, the reverse reaction proceeds more rapidly
than the forward reaction, causing a net conversion of products to reactants. This
conversion decreases the numerator and increases the denominator of Q, so that its
89
value decreases. Similarly, if Q < K, reactants will be converted to products, causing
Q to increase. In either case, the reaction causes Q to approach K until, at equilibrium,
they are equal. The ratio Q/K is thus an indicator of how far a reaction is from
equilibrium. Because Q/K can vary over many orders of magnitude and equilibrium
is reached when Q/K = 1, the extent of disequilibrium is conveniently quantified as
log(Q/K), with a negative value implying that the forward reaction is favored, and a
positive value implying that the reverse reaction is favored.
If Q ⌧ K, the forward reaction proceeds much more rapidly than the reverse re-
action, and if Q K, the opposite is true. The net rate of the reaction, rnet , is rf rr ,
so if log Q/K is smaller than approximately 1.5 (i.e., Q/K < 0.03), rnet ⇡ rf , and if
log Q/K is larger than approximately +1.5 (i.e., Q/K > 30), rnet ⇡ rr . Under either
of these conditions, the reaction is said to be irreversible, meaning that the reaction
is proceeding so much faster in one direction than the other that only the faster re-
action need be considered when evaluating the net reaction rate. Many reactions of
interest in environmental systems meet this criterion. These ideas relating Q/K to the
direction and rate of the reaction are summarized in Table 3.1, which is an expanded
version of Table 2.3.
Table 3.1 Relationships among Q, K, and the direction and reversibility of chemical reactions
Q<K log(Q/K) < 0 Not at equilibrium Forward rate > Reverse rate
Reactants ! Products
Q>K log(Q/K) > 0 Not at equilibrium Forward rate < Reverse rate
Reactants Products
Example 3.1
In Example 2.9, we determined that if the activities of H3 O+ , HOCl, and OCl–
in a solution were 10 7 , 10 3 , and 10 4 , respectively, HOCl would dissociate into
H+ and OCl– as equilibrium was approached. Assuming that the association and
90
dissociation reactions are elementary, would it be reasonable to treat the reaction as
irreversible under these conditions?
Solution
In the solution to Example 2.9, we determined that the activity quotient for HOCl
dissociation under the given conditions was Q = 10 8.0 . The equilibrium constant
for the reaction was given as K = 10 7.53 , so Q/K = 10 0.47 = 0.34. The forward
reaction is therefore proceeding ⇠3.0 (i.e., 1/0.34) times as fast as the reverse reac-
tion. This difference in rates is insufficient to justify ignoring the slower reaction,
so it would not be reasonable to consider the reaction irreversible under the given
conditions.
91
Recall that the empirical Arrhenius model for reaction kinetics could be linked,
at least qualitatively, to the changes in well-defined thermodynamic parameters in
the activated complex model. The thermodynamically based analysis of the effect of
temperature on equilibrium constants is presented in Chapter 4 and yields an equa-
tion almost identical to Equation (3.20), except that EAr is replaced by the normalized
(molar) enthalpy of the reaction when all the reactants and products are in their stan-
dard states (DH r ). Thus, the thermodynamic equation analogous to Equation (3.20),
known as the van’t Hoff equation, is
✓ ◆
K|T2 DH r 1 1
ln = (3.21)
K|T1 R T1 T2
Example 3.2
The reaction for the dissociation of water is shown below as it is conventionally
written. The equilibrium constant for this reaction, Kw , is 1.00 ⇥ 10 14 at 25 C, and
the standard molar enthalpy of the reaction is 55.81 kJ/mol. Compute the equilibrium
constant for the reaction at 4 C and determine the activity of OH– in equilibrium
solutions at 25 C and at 4 C, if {H+ } = 10 7.0 in both solutions.
H2 O )* H+ + OH
Solution
Substituting the given values into Equation (3.21), we find
✓ ◆✓ ◆
Kw |4 C 55.81 kJ/mol 1 1
ln = = 1.71
Kw |25 C 8.314 ⇥ 10 3 kJ/mol-K 298 K 277 K
14 15
Kw |4 C = exp ( 1.71) Kw |25 C = (0.18) 1.0 ⇥ 10 = 1.8 ⇥ 10
Thus, lowering the temperature substantially decreases the tendency for water
molecules to dissociate. The equilibrium constant for the water dissociation reaction
H+ OH
is Kw = . Since the activity of H2 O is 1.0 (because its mole fraction
{H2 O}
is very nearly 1.0), and the activity of H+ is given as 10 7.0 , the activity of OH– is
Kw
. Therefore, at 25 C, {OH– } is 1.0 ⇥ 10 7 , and at 4 C it is 1.8 ⇥ 10 8 .
10 7.0
92
the information necessary to interpret reactions at that level of detail. Furthermore, as
explained previously, most overall reactions are considerably more complicated than
the preceding description suggests, because they represent the net result of several
elementary reactions. Fortunately, even without knowing the molecular-level details
of reactions, it is possible to make some general observations about their tendency
to occur and the stable endpoint (i.e., the equilibrium condition) toward which they
progress. This section describes ways of thinking about and analyzing the kinetics
and equilibrium of nonelementary reactions.
{D} {F}2 0
kf,R1 0
kf,R2 0
kf,R3
Koverall = = KR1 KR2 KR3 = (3.22)
{A}3 {B} eq
kr,R1
0 kr,R2
0 kr,R3
0
The final equality in Equation (3.22) indicates that Koverall for the nonelemen-
tary Reaction (2.18) can be computed as the ratio of the product of the forward rate
constants to the product of the reverse rate constants for the constituent reactions.
However, the numerator and denominator of this expression cannot be interpreted as
forward and reverse rate constants for the overall reaction; i.e., one cannot express
0
the rate of the forward overall reaction as kf,R1 0
kf,R2 0
kf,R3 {A}3 {B}, nor the rate of the
0 k0 k0 2
reverse reaction as kr,R1 r,R2 r,R3 {D} {F} .
This point can be demonstrated by considering the rate at which the products of
the overall reaction are generated. Based on the reaction scheme, the rate of creation
0
of F at any instant is kf,R3 {C} {E}. Thus, increasing the concentrations of A and B
would not necessarily have any immediate effect on the rate of formation of F (e.g.,
0
if kf,R1 0
and kf,R2 were both very small, the added A and B would not be converted to
C and E for a long time). Since increasing the activities of A and B might or might
not have any immediate effect on the rate of formation of D and F, it is clear that the
93
0
rate of the overall forward reaction cannot be computed as kf,R1 0
kf,R2 0
kf,R3 {A}3 {B}; a
similar argument applies to the reverse reaction.
Generalizing this result, we conclude that, unlike the case for elementary reac-
tions, the forward and reverse rates of nonelementary reactions cannot be expressed
by any simple or universal relationship involving a rate constant and the activities of
the reacting species. Rather, detailed understanding of the reaction steps and their
individual rates is required to derive the rate expression for nonelementary reactions.
Nevertheless, for both elementary and nonelementary reactions, equilibrium is de-
fined by equal rates of the overall forward and reverse reactions (i.e., by rnet = 0), and
for both groups, the extent of disequilibrium can be quantified as log(Q/K). Thus,
regardless of whether a reaction is elementary or not, values of Q/K that are much
larger or much smaller than 1 indicate that the reaction is approximately irreversible.
Example 3.3
The conversion of Fe2+ to Fe3+ by reaction with dissolved oxygen can be repre-
sented by the following reaction, which has an equilibrium constant of 2.95 ⇥ 108 .
Solution
The molar concentrations of Fe2+ , Fe3+ , and O2 (aq) are
⇥ 2+ ⇤ 2.0 mg/L 5 mol
Fe = = 3.57 ⇥ 10
56, 000 mg/mol L
⇥ 3+ ⇤ 0.01 mg/L 7 mol
Fe = = 1.79 ⇥ 10
56, 000 mg/mol L
0.1 mg/L 6 mol
[O2 (aq)] = = 3.13 ⇥ 10
32, 000 mg/mol L
The assumption of ideal solute behavior means that the activity of each solute
can be equated with its molar concentration. In addition, the activity of H+ is, by
definition, 10 pH , and we will assume that the activity of water is ⇠1.0. The reaction
quotient for the given conditions is therefore
1.79 ⇥ 10 7
(1.0)0.5
= = 7.50 ⇥ 105
5 6 0.25 6.8
3.57 ⇥ 10 3.13 ⇥ 10 10
94
The reaction quotient is less than the equilibrium constant, indicating that the
reaction will proceed by conversion of the reactants to the products. Furthermore,
the ratio Q/K is only 0.0025, meaning that the reaction is far from equilibrium, so it
can be considered irreversible under the specified conditions.
where a, b, d, and f (i.e., the order of the reaction with respect to A, B, D, and F,
respectively) are treated as empirical constants that must be determined from exper-
iments and that have no fundamental relationship to the reaction stoichiometry.
Note that, although the exponents in Equation (3.23) are decoupled from the
reaction stoichiometry, the rates of destruction or formation of the reacting chemicals
must still conform to the stoichiometry. Thus, for example, the reactants and products
other than A must be generated at rates that are related to rA,net by
3
rA,net = 3rB,net = 3rD,net = rF,net (3.24)
2
If we write an expression analogous to Equation (3.24), but for the rate of reaction
of B, and then divide Equation (3.24) by that expression, we obtain
95
However, Equation (3.24) indicates that rA,net /rB,net equals 3. Combining that
observation with Equation (3.26), we conclude that kf,A = 3kf,B and kr,A = 3kr,B . This
result illustrates that, if a reaction includes species with stoichiometric coefficients
other than 1, the reaction rate and the rate constant must be referenced to a particular
species to avoid ambiguity. For example, a statement that “the forward rate constant
for Reaction (2.18) is xyz” is ambiguous, because such a statement does not make
clear whether the specified value is kf,A or kf,B [i.e., whether it applies to Equation
(3.23) or (3.25)].
Equations (3.23), (3.25), and analogous expressions for rD,net and rF,net all refer to
the same reaction. It is therefore both possible and convenient to identify a single rate
expression that applies to all of the reactants and products identically. To do this, we
define one mole of stoichiometric reaction as conversion of reactants to products in
amounts that correspond to one mole times the corresponding stoichiometric coeffi-
cients. For example, for Reaction (2.18), one mole of stoichiometric reaction occurs
when 3 moles of A and one mole of B are converted to one mole of D and two moles
of F. The net rate of the reaction could then be written as follows, without reference
to any particular species:
ri = ni r(2.18) (3.28)
where ni is the stoichiometric coefficient for i in the reaction.
Example 3.4
We found in Example 3.3 that, in a solution that has pH= 6.8 and contains
2.0 mg/L Fe2+ , 0.01 mg/L Fe3+ , and 0.1 mg/L O2 (aq), the reaction converting ferrous
ions and dissolved oxygen to ferric ions is out of equilibrium, with Q < K.
(b) If 90% of the available dissolved oxygen is consumed in the reaction, how
many moles of stoichiometric reaction have occurred?
96
Solution
(a) According to the reaction stoichiometry shown in Example 3.3, 0.25 mole of
dissolved oxygen reacts per mole of Fe2+ reacting, so
1
rO2 (aq) = rFe2+
4
(b) For the given reaction, one mole of stoichiometric reaction corresponds to con-
version of 1 mole of Fe2+ , 0.25 mole of O2 (aq), and 1 mole of H+ to 1 mole of
Fe3+ and 0.5 mole of H2 O. The solution described in Example 3.3 contained
3.13 ⇥ 10 6 mol/L O2 (aq), so consumption of 90% of this oxygen corresponds
to
✓ ◆✓ ◆
6 mol O2 mol stoichiometric reaction
(0.90) 3.13 ⇥ 10 1
L 0.25 mol O2
5 mol stoichiometric reaction
= 1.13 ⇥ 10
L
rA = kn [A]n (3.29)
where A is the one reactant whose concentration changes significantly during the
experiment. When this approach is used, the concentration of A is typically analyzed
in a batch reactor as a function of time, and the data are analyzed using one of two
methods.
97
In the integral method, a mass balance on the reactant is written and integrated.
Assuming that the only process that alters the concentration of A is the reaction, the
rate of change of the concentration of A (i.e., d [A]/dt ) can be equated with the
reaction rate, so7
d [A]
= rA (3.30)
dt
Z[A]t Zt
d [A]
= dt (3.31)
rA
[A]0 0
Next, a guess can be made for the order of the reaction with respect to A, and the
corresponding rate expression can be substituted into Equation (3.31). For instance,
if we guess that the reaction is first order with respect to A, we can substitute k1 [A]
for rA and integrate the equation to obtain
Z[A]t Zt
d [A]
= dt (3.32)
k1 [A]
[A]0 0
[A]t
ln = k1 t (3.33a)
[A]0
The two forms of Equation (3.33) show that, if the guessed rate expression is correct,
the concentration of A will decline exponentially over time, and a plot of ln [A]t /[A]0
versus t will be linear with a slope of k1 , as shown in Figure 3.4.
There is, of course, no guarantee that the hypothesized rate expression will ac-
curately reproduce the experimental observations. If it does not, we would conclude
that the rate expression is not first order in A, make an alternative guess, and re-
peat the process. Expressions for the concentration of A over time for any nth -order
irreversible reaction proceeding in a batch reactor are summarized in Table 3.2.
While first- and second-order kinetics are commonly reported, reactions can be
7 Formally, d[A]/dt is the rate of change of the concentration of A in the reactor. This rate depends
not only on the reaction rate, but also on all other processes that bring A into, or take it out of, the reactor
(e.g., advection, transfer from the gas phase, etc.). The term rA , on the other hand, refers solely to the
rate of change of the concentration of A in the reactor due to chemical reactions. Therefore, rA equals
d[A]/dt if and only if the only process affecting the concentration of A is the reaction. For a thorough
discussion of mass balances in aquatic systems and ways to evaluate rate expressions using data from
batch and other types of reactors, see Benjamin, M.M. and Lawler, D.F. Water Quality Engineering:
Physical/Chemical Processes, John Wiley & Sons, Hoboken, NJ (2013).
98
Figure 3.4 Characteristic profiles for decay of a reactant undergoing an irreversible, first-
order reaction in a batch reactor.
of other (and sometimes even fractional) orders. In such cases, guessing the correct
value of n is difficult, if not impossible. However, it is sometimes possible to de-
termine both n and kn by an analysis of the time required for a given fraction of the
reactant to disappear. For example, by substituting [A]t = 0.5[A]0 into either of the
integrated rate expressions in the final row of Table 3.2, we can solve for the amount
of time needed for the concentration of A to decline by 50%; this value of t is re-
ferred to as the half-time of the reaction or the half-life of A, and is designated t1/2 .
The result is
8 n 1
> 2 1
>
>
> [A]1 n if n 6= 1 (3.34a)
< kn (n 1) 0
t1/2 =
>
>
>
> ln 2
: if n = 1 (3.34b)
k1
2n 1 1
lnt1/2 = ln + (1 n) ln [A]0 if n 6= 1 (3.35)
kn (n 1)
These results indicate that, if a reaction proceeds according to the rate law rA =
kn [A]n , a plot of ln t1/2 versus ln [A]0 will yield a straight line with a slope of 1 n.
Thus, the reaction order with respect to A can be determined from the slope of such
a plot without guessing the value of n in advance. If n 6= 1, the y-intercept of such a
2n 1 1
plot equals ln , so once n is determined from the slope, kn can be evaluated
kn (n 1)
from the intercept. If n = 1, Equation (3.34b) indicates that t1/2 is independent of the
initial concentration of A, and k can be evaluated as (ln 2)/t1/2 .
99
Table 3.2 Expressions for the concentration of reactant A as it undergoes a reaction with rate expression rA = kn [A]n in a batch reactor
Rate Expression
Reaction Differential form Integral form Integrated expression for kit Integrated expression for [A]t
order
d [A]
Zero rA = = k0 d [A] = k0 dt k0t = [A]0 [A]t [A]t = [A]0 k0t
100
dt
Z Z
d [A] d [A] 1 1 1
Two rA = = k2 [A]2 = k 2 dt k2t = [A]t = + k2 t
dt [A]2 [A]t [A]0 [A]0
Z Z ✓ ◆ 1
n 6= 1
1 ⇣ 1 n ⌘ ⇣ ⌘ 11n
An alternative approach that can be used to characterize the rate expression with-
out postulating a value for n is known as the differential method. This method relies
on the following logarithmic transformation of Equation (3.29):
ln ( rA ) = ln kn + n ln [A] (3.36)
Equation (3.36) indicates that, if the reaction is nth order with respect to A, a plot
of ln ( rA ) versus ln [A] should be linear, with slope n and intercept ln kn . Therefore,
if the reaction rate is determined experimentally at several values of [A] and the data
are plotted as indicated by Equation (3.36), a linear plot implies that the reaction is
indeed nth order in A and allows n and kn to be evaluated; if the plot is not linear, the
conclusion is that the dependence of the rate on A cannot be represented by a simple
power-law expression.
The rate expressions developed above are for irreversible reactions in which A is
the only species whose concentration changes significantly during the reaction. The
conditions that must be met for the reaction to be approximately irreversible have
been described previously (viz., it must be far from equilibrium). The approaches for
maintaining approximately fixed concentrations of other species include buffering
the solution composition and starting with a concentration of the target constituent
that is much less than that of the other reactants. The former approach is often used
to maintain a constant concentration of H+ in solutions, while the latter is used in
most other circumstances.8
A study that investigated the reaction of bisphenol A (BPA) with hypochlorous
acid illustrates both approaches.9 BPA is used in the production of various plastics
and can leach from those plastics into water. It has been found in trace concentrations
in natural waters, drinking water, wastewater, and landfill leachate, and is of concern
because it can interfere with the endocrine systems of animals by mimicking the
biological activity of estrogen. The structure of BPA is shown in Figure 3.5.
When BPA is contacted with HOCl, chlorine attacks the aromatic rings to form a
variety of reaction products (some of which are of equal or more health concern than
unreacted BPA). If the reaction proceeded to equilibrium under conditions typically
encountered in disinfection processes, essentially all of the BPA would be destroyed.
Thus, as long as any BPA remains, the reaction is far from equilibrium and can be
considered irreversible.
In the study, batch experiments were first conducted with initial concentrations
of 1 µmol/L total BPA (T OT BPA) and 38 µmol/L total OCl (T OT OCl). It was an-
ticipated that up to four HOCl molecules would react per BPA molecule. As a result,
8 Buffering is achieved by adding chemicals to the system that rapidly replenish certain species if
they are depleted, or consume those species if they are generated. Approaches for buffering the H+
activity (and therefore the pH) are described in Chapter 8.
9 Gallard, H., Leclercq, A., and Croué, J P. (2004) “Chlorination of bisphenol A: Kinetics and by-
101
Figure 3.5 Molecular structure of BPA
even if all the BPA reacted, T OT OCl would decline by only about 10%; T OT OCl
could therefore be treated as approximately constant under the experimental condi-
tions. The speciation of both T OT BPA and T OT OCl is sensitive to pH (as shown
for OCl in Example 2.6), so the solutions were buffered to maintain a constant pH
during each experiment. Several experiments were conducted at each pH of interest.
To model the reaction kinetics, the authors postulated the power-law rate expres-
sion
rBPA = kf [T OT BPA]a [T OT OCl]b (3.37)
However, based on the assumption that [T OT OCl] was approximately constant,
this expression could be simplified to
3 1
kobs, pH 6.75 = slopepH 6.75 = 1.43 ⇥ 10 s (3.39)
Experiments at other TOTOCl values (still at the same pH, and still with TOTOCl
much greater than TOTBPA) were then conducted to assess the dependence of the
reaction rate on TOTOCl. In each case, a plot of ln ([T OT BPA]t /[T OT BPA]0 ) vs.
t was linear, and kobs was determined from the slope. These kobs values were then
plotted against TOTOCl, as shown in the inset in Figure 3.6. The linearity of this
plot indicates that kobs is directly proportional to TOTOCl, implying that the reaction
is first order with respect to TOTOCl and allowing kf to be evaluated from the slope.
Based on all the experimental results, the overall reaction rate at pH 6.75 was reported
as
L
rBPA,pH 6.75 = 36 [T OT BPA] [T OT OCl] (3.40)
mol-s
102
Figure 3.6 Kinetics of the loss of TOTBPA from aqueous solutions initially containing
1.0 µmol/L TOTBPA and 38 µmol/L TOTOCl, with the pH buffered at various
values. Inset is for experiments at various initial TOTOCl concentrations at pH
6.75. From Gallard et al. (2004) (see footnote 9).
Example 3.5
Plot the expected TOTBPA as a function of time in a batch reactor for initial con-
centrations of 0.1 µmol/L TOTBPA and 15 µmol/L TOTOCl at pH 7.5. Gallard et
al. (2004) estimated kf to be 130 L/(mol·s) at this pH. How long would be required
to destroy 99% of the BPA in this solution?
Solution
As in the study described previously, the initial value of TOTOCl is much larger
than that of TOTBPA, so TOTOCl can be considered constant, and the BPA destruc-
tion reaction can be considered pseudo-first order:
✓ ◆
L 6 mol
rBPA,pH7.5 ⇡ 130 [T OT BPA] 15 ⇥ 10 = 1.95⇥10 3 s 1 [T OT BPA]
mol·s L
The BPA concentration therefore declines exponentially, consistent with Equa-
tion (3.33b):
⇥ ⇤
[T OT BPA] = [T OT BPA]0 exp 1.95 ⇥ 10 3 s 1 t (3.41a)
⇥ ⇤
= [T OT BPA]0 exp 0.117 min 1 t (3.41b)
103
Figure 3.7 Decay of TOTBPA by reaction with chlorine at pH 7.5, under the example condi-
tions.
The time required for 99% destruction of TOTBPA can be found by rearranging
Equation (3.33a) and setting [T OT BPA] equal to 0.01[T OT BPA]0 :
[T OT BPA] 1
ln = ln 0.01 = 0.117 min t
[T OT BPA]0
ln 0.01
t= 1
= 39.4 min
0.117 min
104
in that tank and, to a lesser extent, in the downstream tanks (Figure 3.8c). If the dif-
ference in resistance of the valves was even greater than in the illustration, at some
point essentially all the water remaining in the system would be in the tank with the
high-resistance outlet valve, and the rate at which water passed through the whole
network would be controlled by the rate at which it flowed out of that tank. That is,
increasing or reducing the resistance of that step would affect the rate at which wa-
ter appeared at the outlet, whereas altering the resistance of other steps would have
almost no effect.
Figure 3.8 Water flowing through a series of tanks. (a) Initial condition, with the outlet from
the first tank closed; (b) A short time later, with all valves open, but with the
resistance imposed by the valve downstream of the third tank much larger than
that of any of the other valves; (c) Later, after the first two tanks have drained.
105
stayed at its initial value continuously. Based on the equations in Table 3.2, this
characteristic time can be computed as
1
tchar = (3.42)
kn [A]n0 1
Like the half-time, tchar for first-order reactions is independent of the concentration
of A, but tchar for reactions of any other order depends on the initial concentration of
the reactant.
The idea of the rate-limiting step can be demonstrated for a hypothetical sequence
of a second-order reaction followed by a first order reaction, with both reactions
irreversible:
A !B !P
rA = kA [A]2 ; rB = kA [A]2 kB [B] ; rP = kB [B] (3.43)
Note that the first reaction in the sequence is second-order in A, even though the
stoichiometric coefficient of A is 1; because the rate expression does not conform to
the stoichiometry, we infer that the reaction is not elementary.
Assume that kA is 500 (mol/L) 1 min 1 , kB is 0.01 min 1 , and that we wish to
predict the reaction progress in a batch reactor that initially contains 10 3 mol/L
A and no B or P. The rate of conversion of A to B is independent of whether or
not B participates in subsequent reactions, so the decline in the concentration of
A conforms to the equation in Table 3.2 for a second-order irreversible reaction.
Changes in the concentration of B, on the other hand, reflect a balance between its
formation from A and its conversion to P. Because the reaction is taking place in a
batch reactor, d [B] /dt can be equated with rB , so
d [B]
= kA [A]2 kB [B] (3.44)
dt
Similarly, the rate of formation of P at any instant can be found from the mass
balance on P. Because P can be formed only by destruction of B and does not partic-
ipate in any other reactions, this mass balance is simply
d [P]
= kB [B] (3.45)
dt
The results from the simultaneous solution of the three mass balance equations in
conjunction with the specified initial condition are shown in Figure 3.9.
The analysis indicates that A is depleted quickly and that B builds up to approx-
imately 0.8 mmol/L within 15 min, before decaying to P over the next few hours.
According to Equation (3.42), the characteristic times of the two reactions can be
computed as 1/(kA [A]0 ) and 1/kB , with values of 2 min and 100 min, respectively.
These characteristic times suggest that the second reaction imposes more resistance
106
Figure 3.9 Concentration versus time profiles for the example, sequential reactions.
and is therefore rate-limiting. Applying this idea, we can make the approximation
that A is converted to B instantaneously, simplifying the overall reaction to the single
reaction B ! P, with [B]0 = 0.001 mol/L. The predicted concentration of P when
that approximation is applied is shown in Figure 3.9 by the curve labeled P⇤ . The
curve closely follows that obtained from the more complete analysis, suggesting that,
by identifying the rate-limiting step, we can significantly simplify the evaluation of
the reaction kinetics without much loss of accuracy.
Note that, in an analysis like the preceding one, the characteristic reaction time
for each step is calculated assuming that that reaction step is rate-limiting. This as-
sumption implies that essentially all of the initial reactant has been converted to the
reactant for the step being investigated. As a result, the appropriate values to use for
the initial concentrations in the calculation are not the actual initial concentrations of
the reactants, but rather the hypothetical concentrations that would exist if all the up-
stream reactions had proceeded to completion, and all the reactants were “detained”
just upstream of the step being evaluated. The benefits of carrying out such calcu-
lations and identifying the rate-limiting step increase dramatically with increasing
complexity of the overall reaction.
Example 3.6
The widely prescribed anti-bacterial agent ciprofloxacin (‘cipro’) is incompletely
metabolized by humans and is therefore routinely found at low concentrations in do-
mestic wastewater. Dodd et al. (2005)10 reported that the reaction of this compound
10 Dodd, M.C., Shah A.D., von Gunten, U., and Huang, C.H. (2005) “Interactions of fluoroquinolone
antibacterial agents with aqueous chlorine: Reaction kinetics, mechanisms, and transformation path-
ways.” Environ. Sci. Technol. 39, 7065-7076.
107
with HOCl can be modeled as occurring in two steps, comprising incorporation of a
chlorine atom into an intermediate species and subsequent cleavage of that species to
release small organic molecules and a Cl– ion.
RI
z }| {
Cipro + HOCl )* Intermed-Cl )* Organic Products + Cl
| {z }
RII
Both reactions are first-order with respect to each of the reacting species and, at
pH 5.5, the rates of the reactions are consistent with the following expressions:
Solution
The reaction of Cipro is first order with respect to Cipro and HOCl, and second
order overall. We can develop a spreadsheet to compute the rate of this reaction at
any instant, assume that that reaction rate applies over a short time step Dt, and then
compute the decline in the concentrations of the two reactants as the product rRI Dt.
A new rate can then be determined for the next time step based on the concentrations
at the end of the first time step, and the process can be repeated for the duration of
interest. Alternatively, if we wish to simplify the calculations, we can assume that
the concentration of HOCl is approximately constant throughout the reaction, since
its initial concentration is 20 times that of the Cipro. A similar process can be carried
out to compute the changes in the concentrations of the intermediate species and the
products, except that in these cases the changes during any time step are (rRI rRII ) Dt
and rRII Dt, respectively.
The results of the calculations are shown in Figure 3.10. Essentially all the Cipro
disappears within the first 5 seconds, but the intermediate reacts much more slowly.
As a result, after the reaction has proceeded for 5 s, the concentration of the interme-
diate is very close to the initial concentration of Cipro, i.e., 10 6 M. It then decays
gradually over the next few hours, and the product accumulates. Clearly, in this case,
reaction RII is rate limiting, and the time required for reaction RI to proceed is virtu-
ally irrelevant if the primary interest is the generation of the product.
108
Figure 3.10 Concentrations of the reactants, intermediate, and products of the reaction of
Cipro with HOCl under conditions described in the problem statement. (a)
Changes during the first minute of reaction; (b) Longer-term changes.
Another reaction sequence that has been widely applied to model nonelementary
reactions was first developed for enzyme-based reactions. This sequence begins with
a reversible, rapidly equilibrating reaction between an enzyme (E) and a substrate
(S) to generate an intermediate species (ES⇤ ). The intermediate is then irreversibly
converted to the product (P) by a reaction that is first order in ES⇤ , re-generating the
enzyme in the process. This sequence can be represented as
K k1
S + E )* ES⇤ ! P+E (3.46)
When this reaction model is applied, the reacting species are invariably assumed
to behave ideally, so the equilibrium constant for the first reaction (K) is numerically
109
equal to the concentration ratio of the products to the reactants; we represent this
concentration ratio as [K]. In addition, because the enzyme is not consumed, its total
concentration (T OT E, the sum of [E] and [ES⇤ ]) remains constant as S is converted
to P, so we can write
[ES⇤ ] [ES⇤ ]
[K] = = (3.47)
[E] [S] ([T OT E] [ES⇤ ]) [S]
[T OT E] [K] [S]
[ES⇤ ] = (3.48)
1 + [K] [S]
Inserting this expression for ES⇤ into the rate expression for formation of P yields
[T OT E] [K] [S]
rP = k1 [ES⇤ ] = k1 (3.49)
1 + [K] [S]
[T OT E] [S]
rS = k1 (3.51)
Km + [S]
The dependence of rP on [S] for such a reaction and the concentration changes
over time in a batch system are shown in Figure 3.11.
In systems where the substrate is being consumed by a microbial culture (as op-
posed to an abiotic system with purified enzymes), the enzyme concentration is typ-
ically assumed to be proportional to the biomass concentration, [X]. In such cases,
Equation (3.51) is commonly rewritten as follows:
rS kmax [S]
= (3.52)
[X] Km + [S]
where kmax = k1 [T OT E] / [X]. The left side of Equation (3.52) represents the rate
of substrate consumption per unit amount of biomass present and is known as the
specific rate of substrate utilization. Inspection of the right side of the equation
shows that, when [S] Km , the specific rate of substrate utilization is maximized and
equals kmax . Also, when [S] = Km , the specific rate of substrate utilization is 0.5kmax ;
for this reason, Km is called the half-saturation constant.
110
Figure 3.11 Classic kinetics of enzyme-based reactions, based on the sequence shown in
Reaction (3.46). (a) Product formation rate as a function of substrate con-
centration; (b) Changes in concentration in a batch system. The parameters
used in these simulations are Km = 1/K = 8.3 ⇥ 10 4 mol/L; k1 = 3 min 1 ;
T OT E = 10 6 mol/L; in part (b), [S]0 = 10 3 mol/L.
Both Equations (3.51) and (3.52) are known as the Michaelis-Menten equation.
When Michaelis-Menten kinetics apply, the rate of the overall reaction is first order
in T OT E or X under all conditions, but the order with respect to S varies from ap-
proximately zero (when [S] Km ) to approximately one (when [S] ⌧ Km ). Because
the equilibration of S and E with ES⇤ is assumed to be rapid and therefore to impose
negligible resistance on the progress of the reaction, the conversion of ES⇤ to P is
always rate-limiting.
111
Example 3.7
The rate of an enzyme-mediated reaction is consistent with the Michaelis-Menten
equation, with Km = 3 ⇥ 10 3 M, [T OT E]= 10 6 M, and k1 = 85 min 1 .
(a) How long would it take to convert 99% of the initial substrate S to product P
in a batch reactor if the initial concentration of S was 10 2 M?
(b) Develop a curve showing the expected value of [S] during the first four hours
of reaction, for the system in part (a).
Solution
(a) We wish to determine the time required to reduce the substrate concentration
from 10 2 M to 10 4 M. Because the reaction is taking place in a batch reactor,
we can substitute d[S]/dt for rS in Equation (3.51). Making that substitution
and integrating the resulting expression, we obtain
d[S] [T OT E] [S]
= k1
dt Km + [S]
[S(t)]
Z Zt
Km + [S]
d[S] = k1 [T OT E] dt
[S]
[S(0)] 0
[S(t)]
Z [S(t)]
Z
d[S]
Km + d[S] = k1 [T OT E]t
[S]
[S(0)] [S(0)]
[S(t)]
Km ln + {[S(t)] [S(0)]} = k1 [T OT E]t
[S(0)]
⇢
1 [S(t)]
t= Km ln + {[S(t)] [S(0)]}
k1 [T OT E] [S(0)]
112
The preceding discussion of irreversible reactions in series and Michaelis-Menten
kinetics describes only two of a virtually infinite variety of possible combinations
of elementary reactions that lead to the overall reactions observed at a macroscopic
scale. The particular examples presented have been chosen not because they are more
plausible than other combinations, but because they are relatively simple to develop,
they lead to interesting patterns of change in constituent concentrations over time,
and they have been found to characterize many overall reactions in important practi-
cal systems. Regardless of the reaction scheme or rate expression that is ultimately
adopted to describe an overall reaction, the procedure is as described here – collect
experimental data, postulate a rate expression, and iterate on the functional form of
that expression until it reproduces the data satisfactorily.
113
The model based on phase transfer as the rate-limiting step is applied primarily
to precipitation and dissolution of solids, in which case phase transfer requires the
formation or breaking of a bond with the solid. In this model, transport between the
bulk solution and the interface is assumed to proceed very rapidly (e.g., due to intense
mixing), so the composition of the solution is assumed to be uniform from the bulk
liquid right up to the interface. We next develop the key equations for this model
using an example system in which particles of SiO2 (s) are added to an intensely
mixed, ideal solution that initially contains no dissolved silica.
In such a system, a driving force would exist for SiO2 molecules to be released
from the surface and enter the solution. Assume, for now, that the rate-limiting el-
ementary reaction in the dissolution process is collision of a water molecule with a
molecule of SiO2 at the surface of the solid, causing the surface molecule [which we
will designate as SiO2 (surf )] to enter solution as H2 SiO3 (aq). When the SiO2 (surf )
molecule dissolves, the SiO2 (s) molecule “below” it in the solid phase is converted
to an SiO2 (surf ) molecule, so the overall reaction can be represented as follows:11
Although the reaction does not affect the concentration of SiO2 (surf ) in the sys-
tem (because one such molecule is covered up and a new one is generated), we show
that species on both the reactant and product sides of the reaction to emphasize that it
does participate in the reaction (and because its concentration is a key parameter in an
equation that is derived shortly). However, because Reaction (3.54) is an immediate
and direct consequence of Reaction (3.53), the rate of the overall reaction equals the
rate of just Reaction (3.53), so we can write:
where [SiO2 (surf )] is the molar concentration of SiO2 (surf ) molecules in the system
(i.e., with units of moles of SiO2 (surf ) per liter of solution). This concentration can be
expressed as the product of the concentration of SiO2 in the surface layer (sSiO2 (surf) ,
with units like moles of SiO2 (surf ) per m2 ) and the concentration of surface area in
the system, ([ASiO2 (s) ], m2 per liter of solution):
h i
[SiO2 (surf)] = sSiO2 (surf) ASiO2 (s) (3.57)
11 Throughout this section, the parameters r
f , rr , rnet , kf , kr , and K, when shown without any additional
designation, all refer to Equation (3.55).
114
Substituting this expression into Equation (3.56b), we obtain
h i
rf = kf sSiO2 (surf) ASiO2 (s) (3.58)
Equation (3.58) suggests that the rate of solid dissolution will be proportional to
the total surface area of SiO2 (s) particles dispersed in the solution. For example, it
suggests that SiO2 (s) will enter solution much more rapidly if the solid is present in
the form of many 0.1-mm particles than if it is present as a single 1.0-mm particle,
even if both systems contain the same mass of SiO2 (s) per liter of solution.
Essentially the same logic applies to the precipitation reaction [i.e., the reverse of
Reaction (3.55)]. That process can be envisioned as being initiated by a collision of
a dissolved H2 SiO3 (aq) with a surface molecule. The collision causes the H2 SiO3 to
split apart, with the SiO2 portion attaching to the solid (and thereby covering up a sur-
face molecule), and the remaining portion (an H2 O molecule) returning to solution.
The reaction rate is therefore
h i
rr = kr (H2 SiO3 (aq))sSiO2 (sur f ) ASiO2 (s) (3.59)
where the assumption that the solution behaves ideally allows us to use either the
molar concentration or activity of H2 SiO3 (aq) in the rate expression.
The net solid dissolution rate is
rnet = rf rr
h i h i
= kf sSiO2 (sur f ) ASiO2 (s) kr (H2 SiO3 (aq))sSiO2 (sur f ) ASiO2 (s)
h i
= kf kr (H2 SiO3 (aq)) sSiO2 (sur f ) ASiO2 (s) (3.60)
At equilibrium, the forward and reverse reaction rates are equal, and the net reac-
tion rate is zero. Therefore, the equilibrium activity of dissolved H2 SiO3 is
kf
(H2 SiO3 (aq))eq = (3.61)
kr
Substituting this result into Equation (3.60) shows that, like first-order reactions
that occur entirely in one phase, phase-transfer reactions proceed at rates that are
proportional to the extent of disequilibrium. However, in phase-transfer reactions,
the rate is also proportional to the concentration of interfacial area in the system:
⇢ h i
kf
r = kr (H2 SiO3 (aq)) sSiO2 (sur f ) ASiO2 (s)
kr
h i
= kr (H2 SiO3 (aq))eq (H2 SiO3 (aq)) sSiO2 (sur f ) ASiO2 (s) (3.62)
115
Equation (3.61) demonstrates that the concentration of H2 SiO3 (aq) in equilib-
rium with the solid is independent of the concentration of either bulk solid ([SiO2 (s)])
or SiO2 surface species ([SiO2 (surf )]). Thus, although the rate of the phase transfer
reaction depends on how much surface area is available to react, the equilibrium con-
dition does not.
The preceding discussion relied on an analysis of the reaction between SiO2
molecules at the solid surface and those in solution, and did not directly address
those in the bulk solid. However, we can incorporate the activity of the bulk solid
into the analysis by writing the equilibrium constant for Reaction (3.55), as follows:
Equation (3.63b) indicates that the activity of SiO2 (s) is the same in any system
where it is present. In particular, its activity is independent of the amount of solid
present, consistent with the discussion in Chapter 2 about the activity of any pure
solid. The numerical value of the fixed activity of SiO2 (s) is not established by Equa-
tion (3.63b) and depends on the choice of the standard state conditions. However, as
explained in Chapter 2, the universal convention is to make choose the standard state
in a way that causes the constant value of the solid activity to be 1.0.
A second model for the rate of phase-transfer reactions is based on the assump-
tion that equilibrium is reached very rapidly at the interface, and that the rate of solute
transfer between the two phases depends primarily on the rate at which the transfer-
able species migrate through the boundary layer between the interface and the bulk
fluid phase.12 This migration is assumed to proceed by molecular diffusion and there-
fore to be characterized by Ficks law:
dci
Ji,x = Di (3.64)
dx
where Ji,x is the flux (mass transported per unit area of interface per unit time) of
species i in the x direction, Di is the diffusion coefficient of i, and dci /dx is the local
concentration gradient of i.13
12 When this model is applied to transfer between solution and a gas phase, boundary layers exist on
both sides of the interface, but if the resistance to transport is much greater in one boundary layer than
in the other, the overall kinetics can be modeled considering only the high-resistance layer. We make
that assumption and consider only a single boundary layer in this analysis. Some discussion of systems
where both boundary layers contribute to the overall resistance is provided at the end of this section.
13 Fick’s law actually predicts that the flux is proportional to the gradient in the activity, not the
116
The earliest model of this process assumed that the fluid in the boundary layer
was stagnant, that ci at the interface (ci,int ) was controlled by the rapid equilibration
with the other phase at that location, and that ci at the outer edge of the boundary layer
equaled that in the bulk fluid (ci,b ). In that case, if the concentrations at the edges of
the boundary layer change slowly compared to the time frame needed for diffusion
through the layer, a steady-state concentration profile develops in the boundary layer,
in which the concentration changes linearly across the layer (i.e., the concentration
gradient in the boundary layer is constant). In that case, Equation (3.62) indicates that
the flux of i is constant through the layer. Defining x to be positive in the direction
from the interface to the bulk fluid, the flux into the fluid is
ci,b ci,int ci,int ci,b
Ji,x = Di = Di (3.65)
dBL dBL
where dBL is the thickness of the boundary layer, which is expected to depend on
factors like the geometry of the interface and the intensity of fluid mixing in the
system.
Equation (3.65) is commonly written as follows:
117
Figure 3.12 The changing concentration profile of a dissolved species in a packet in the liquid
boundary layer adjacent to a gas/liquid or solid/liquid interface. In this hypothet-
ical system, species are migrating from the nonaqueous phase into solution, so
the concentration declines from the interface to the bulk solution phase. If the
packet stays in the boundary layer long enough, a linear profile develops, as
predicted by the stagnant-layer model. (After Benjamin and Lawler, 2013 [see
footnote 1]).
r " ⇢ #
•
Di j2 dBL
2
kmt,i = 1 + 2 Â exp (3.67)
pt j=1 Dit
where t is the time that the packet spends at the interface.
Now, consider how we would apply model for transport-limited phase transfer
reactions to the SiO2 (s) dissolution reaction that we evaluated previously. In this
scenario, the interfacial concentration of H2 SiO3 (aq) would be the fixed value in
equilibrium with the solid. Inserting that value into Equation (3.66), we obtain
n o
JH2 SiO3 (aq) = kmt,H2 SiO3 (aq) H2 SiO3 (aq) eq H2 SiO3 (aq) (3.68)
The rate at which H2 SiO3 (aq) enters the solution is the product of the flux into
solution and the concentration of interfacial area, so
h i
r = JH2 SiO3 (aq) ASiO2 (aq)
n oh i
= kmt,H2 SiO3 (aq) H2 SiO3 (aq) eq H2 SiO3 (aq) ASiO2 (aq) (3.69)
Comparing Equations (3.61) and (3.69), we see that both models predict that the
rate of dissolution is directly proportional to the extent of disequilibrium between
the two phases and to the volume-based concentration of interfacial area. Where the
118
models differ is in the interpretation of the coefficients that appear in the governing
equations. Equation (3.61) includes two such coefficients - the rate constant of the
precipitation reaction and the concentration of SiO2 (surf ) molecules in the surface
layer - whereas Equation (3.69) includes just one - the mass transfer coefficient for
i in the boundary layer [defined either as Di /dBL or by the expression in Equation
(3.67)].
Modeling of gas-transfer processes almost always assumes that transport through
the liquid and/or gas boundary layer (as opposed to transfer across the gas/liquid
interface) is rate-limiting. Transport through the two individual layers is linked by
the assumption of equilibrium at the interface and by the fact that the flux out of
either phase must equal the flux into the other. When this assumption and constraint
are combined with two applications of an equation like Equation (3.69) (one written
for each boundary layer), the following expressions are obtained for the flux and
the overall transfer rate of a species from the bulk gas to the bulk aqueous solution,
assuming ideal solute behavior:14
⇣ ⌘
Ji = KL,i c⇤i(aq) ci(aq) (3.70)
⇣ ⌘
rgt,i = KL,i aL c⇤i(aq) ci(aq) (3.71)
In these equations, KL,i is the overall gas transfer coefficient, with dimensions
of length per time (e.g., cm/s); c⇤i(aq) is the hypothetical concentration of dissolved
i that would be in equilibrium with the bulk gas phase; ci(aq) is the concentration
of dissolved i that is actually present in the solution; and aL is the volume-based
concentration of interfacial area in the system.15 KL,i is related to the mass transfer
coefficients through the individual boundary layers and the gas/liquid equilibrium
constant by
and the mathematics of models that account for fluid exchange between a bulk phase and a boundary
layer are provided in Chapter 5 of Water Quality Engineering: Physical/Chemical Treatment Processes
by Benjamin, M.M., andhLawler, D.F.,i J. Wiley & Sons, Hoboken, NJ (2013)
15 a is analogous to ASiO2 (s)(aq) in Equation (3.57); by convention, this parameter is always
L
(aq)
represented as a or aL in the gas transfer literature.
119
gas transfer systems (especially for species that are only slightly soluble in solution,
like oxygen). In cases where kL,i kG,i Keq , the overall gas transfer coefficient is
approximately equal to kG,i Keq ; this situation can arise for highly soluble gases (like
ammonia), and it implies that most of the resistance to gas transfer is in the gas-phase
boundary layer.
Example 3.8
Wastewater is being treated in a pond that is aerated by a fountain aerator that
sprays droplets with an average diameter of 0.4 cm into the air. The pond water
contains 1.5 mg/L dissolved oxygen, the liquid-phase mass transfer coefficient is
0.02 cm/s, and almost all the resistance to gas transfer is in the liquid boundary layer.
What will the dissolved oxygen concentration be just before a droplet enters the
pond after spending 5 s in the air, if the equilibrium dissolved oxygen concentration
is 10.5 mg/L?
Solution
While the droplet is in the air, it can be treated as a batch reactor, so rgt,O2 (aq) =
dcO2 (aq) /dt. Also, because the gas transfer resistance is almost all in the liquid
boundary layer, KL ⇡ kL . The ratio of interfacial area to liquid volume in the drop is
pd 2 6 6 1
aL = = = = 15 cm
pd /6 d 0.4 cm
3
We can insert these values and the other, given information into Equation (3.72)
to find
dcgt,O2 (aq) ⇣ ⌘
= KL,O2 (aq) aL c⇤O cO2 (aq)
dt 2 (aq)
Z dcgt,O2 (aq) Z
= KL,O2 (aq) aL dt
c⇤O cO2 (aq)
2 (aq)
⇣ ⌘ 5s 5s
ln c⇤O cO2 (aq) = KL,O2 (aq) aLt
2 (aq) 0 0
mg
10.5 cO2 (aq) |5 s ⇣ cm ⌘
ln L = 0.02 15 cm 1
(5 s)
mg mg s
10.5 1.5
L L
mg
cO2 (aq) |5 s = 8.5
L
120
3.11 SUMMARY
All observable chemical reactions are either elementary reactions – that is, they oc-
cur in a single step via a collision among the reacting molecules – or combinations
of elementary reactions. Elementary reactions require an initial input of energy to
destabilize existing chemical bonds, after which those bonds might re-form or break
and allow new bonds to form. The minimum energy input needed to destabilize the
bonds and allow reactants to be converted to products is called the activation energy
of the reaction. The lower the activation energy, the faster a reaction proceeds. In
addition, all such reactions proceed at a rate that is proportional to the activities of
the reacting species. However, it is common to write the rate expression in terms
of the concentrations of the reactants, thereby incorporating the activity coefficients
and standard state concentrations into the rate constants. Nonelementary overall re-
actions are the net result of a group of elementary reactions that proceed in parallel
and/or series.
The rates of all reactions increase with increasing temperature. This relationship
can be quantified by the Arrhenius equation. When the Arrhenius equation is ap-
plied to both the forward and reverse directions of a reaction, the result is the van’t
Hoff equation, in which the key parameter is the standard molar enthalpy of reaction
(DH r ).
In principle, all chemical reactions are reversible. The net effect of the simulta-
neous forward and reverse reactions is that all reactions proceed toward equilibrium,
a condition that is characterized by the absence of any net production of either re-
actants or products by the reaction. Such a scenario can be interpreted as the result
of a balance in which the forward and reverse reactions proceed at equal rates, or a
condition of minimum chemical potential energy.
The ratio of the activities of products to reactants for a reaction is called the
reaction quotient, Q. Based on the recognition that the forward and reverse reaction
rates must be equal for a reaction at equilibrium, the value of Q for a given reaction
must always be the same once the reaction has reached equilibrium, although that
value can be achieved via many different combinations of activities for the individual
reacting species. The value of Q at equilibrium is called the equilibrium constant, K.
For elementary reactions, K equals the ratio of the forward to the reverse activity-
based rate constants; for nonelementary reactions, this latter relationship does not
apply, but K can still be computed as the ratio, at equilibrium, of the activities of the
products to those of the reactants. For reactions that are not at equilibrium, Q does
not equal K, and the reaction proceeds in the direction that causes Q to approach K.
Any overall reaction can be derived as the sum of a group of other (elementary
or nonelementary) reactions; the equilibrium constant for the overall reaction equals
the product of the K values of the reactions in that summation. The derivation of this
relationship emphasizes that chemical activities and equilibrium constants are simply
121
numbers that can be manipulated according to the rules of algebra, independent of
the fact that they have chemical significance.
The ratio Q/K [or log(Q/K)] is an indicator of the extent of disequilibrium. If
log(Q/K) is greater than ⇠1.5, the reaction is proceeding much faster in one direc-
tion than the other, and the overall reaction rate can be computed considering only
the faster reaction. In such cases, the reaction is effectively irreversible.
The rates of nonelementary reactions can depend in complex ways on the reac-
tant concentrations and the rate constants for the constituent elementary reactions.
As a result, the rate expressions for such reactions cannot be predicted without de-
tailed knowledge of the reaction pathway and the rates of the individual steps, in-
formation that is frequently unavailable for reactions of environmental importance.
In these cases, rate expressions for the overall reactions are derived empirically, of-
ten by studying the reaction kinetics under simplified conditions where the reaction
is effectively irreversible and the rate depends on the concentration of a single con-
stituent.
When reactions proceed in series, one reaction often provides the majority of
the resistance, and that reaction is referred to as the rate-limiting step in the overall
reaction. If the resistance associated with that step is much larger than the resis-
tance of the rest of the sequence, the kinetics of the overall reaction can be closely
approximated by considering the kinetics of that step alone.
3.12 PROBLEMS
1. An elementary reaction A )* B has an equilibrium constant of 10, and a
forward rate constant of 1.0 d 1 .
122
Hydroxide ion has been found to catalyze (i.e., to increase the rate of) the
forward reaction. How would the following parameters change when the pH
is increased? Briefly explain your reasoning.
A + B )* 2 C )* D + E
The rate constants for the reactions are as follows, where ‘1’ and ‘2’ refer to
the first and second reaction, respectively:
1 1 1 1
k1,f = 0.04 M ·s k1,r = 0.01 M ·s
1 1 1 1
k2,f = 0.10 M ·s k2,r = 0.10 M ·s
(a) Write the overall reaction corresponding to the sum of these elementary
reactions and determine the equilibrium constant that applies to that re-
action.
123
(b) A system that contains all the species that participate in these reactions
has reached equilibrium. Do you think any of the elementary reactions
can be treated as irreversible under these conditions? Why or why not?
6. The data below summarize the loss of a solute from water due to a chemical
reaction in a batch reactor. Determine whether the data are best fit by a zero-
order, first-order, or second-order rate expression, and report the rate constant
with proper units.
8. The radionuclide 32 P has a half-life of 14.3 days. How long would an aque-
ous waste containing 1.0 mg/L of this nuclide have to be stored to reduce the
concentration to 0.03 mg/L through radioactive decay?
124
Explain why the reaction is pseudo-first-order, even though it requires a colli-
sion between two molecules. Present an appropriate equation to support your
explanation.
10. The decomposition of ozone in water fits the general rate equation
⇥ ⇤a ⇥ ⇤b
rO3 (aq) = kd O3 (aq) OH
Staehelin and Hoigne (1982)16 reported the data in the following table for the
decrease in the O3 (aq) concentration over time at pH=10 in batch reactors.
125
12. At pH below about 3.5, ferrous iron (Fe2+ ) oxidizes in streams according to
the following overall reaction:
Fe2+ + 0.25 O2 (aq) + H+ )* Fe3+ + 0.5 H2 O
The rate law for the abiotic oxidation of ferrous iron under these conditions
and at 20 C is ⇥ ⇤
rFe2+ = k Fe2+ PO2
where PO2 is the partial pressure of gas-phase oxygen that is in equilibrium
with the solution, and k = 10 3.2 atm 1 ·d 1 .
Calculate the time required for the concentration of Fe2+ to be reduced to 1%
of its initial concentration in a batch reactor where the solution is in continuous
equilibrium with the atmosphere.
13. When chlorine is added to water containing ammonia, chloramine compounds
(NHx Cl3 x , with x=0, 1, or 2) form. Monochloramine (NH2 Cl) can split (‘dis-
proportionate’) into dichloramine and ammonia according to the following re-
action and rate expression:
2 NH2 Cl ! NHCl2 + NH3 rNH2 Cl = k(NH2 Cl)2
Two experiments are carried out with solutions that initially contain 0.01 M
Fe2+ , no Fe3+ , and 10 3 M HOCl. In one, conducted at 10 C, the concentra-
tion of Fe3+ generated in 1 minute is 1.0 ⇥ 10 5 M. When the temperature is
increased to 20 C, 2.4 ⇥ 10 5 M Fe3+ is generated in one minute. Assuming
the rate expression has the following form, calculate the rate constant at each
temperature and estimate the activation energy for the reaction.
⇥ ⇤2
rFe3+ = k Fe2+ [HOCl]
126
(a) If, at a given pH, the half-life of pyrophosphate is 140 h at 75 C and 13 h
at 100 C, what is EAr for the reaction?
(b) Estimate the time required for 50% hydrolysis of pyrophosphate in a
solution at the same pH as in part (a), but at 20 C.
16. In the presence of dissolved oxygen, microorganisms can oxidize nitrite ions
(NO–2 ) to nitrate ions (NO–3 ) via the following reaction:
NO2 + 0.5 O2 (aq) ! NO3
(a) This reaction is occurring in a batch bioreactor at a rate that is first or-
der with respect to both NO–2 and O2 (aq), with k2 = 3 ⇥ 104 L/mol·d.
How much time is required for the nitrite concentration to be reduced
from 0.3 mM to 0.1ṁM if O2 is resupplied continuously to maintain the
O2 (aq) concentration at 4 mg/L?
(b) Repeat part (a), but assume that O2 is not resupplied, so that the O2 (aq)
concentration is initially 4 mg/L but declines as the reaction proceeds.
17. The stoichiometry of the reaction of hydrogen sulfide with dissolved oxygen
is given by
H2 S(aq) + 2 O2 (aq) ! SO24 + 2 H+
Millero et al. (1987)17 reported that, at pH < 7 and 25 C, this reaction is first
order with respect to both H2 S and O2 , with a rate constant of k = 11 L/mol·h.
Consider a solution containing 10 5 M H2 S(aq) and a constant dissolved oxy-
gen concentration of 0.25 mM (8.0 mg/L). If the given reaction is the only way
that sulfide is lost from the system, how much dissolved sulfide will remain
after one day at pH 5? (As is explained in Chapter 5, at pH 5.0, H2 S(aq)
accounts for virtually all the dissolved sulfide in the system.)
18. In the biochemical oxygen demand (BOD) test, a sample of water is inoculated
with microorganisms and then incubated with essential nutrients to determine
how much dissolved oxygen is consumed by the constituents in the sample un-
der conditions favorable for microbial growth. The rate of oxygen consump-
tion is typically modeled as first-order with respect to the concentration of
dissolved oxygen that could be consumed by reaction with other constituents
present in the solution; this parameter is conventionally represented as L. Thus:
rO2 = k1 L
127
the treated wastewater have the potential to consume a maximum of 15 mg/L
of oxygen. (This value is known as the ultimate BOD of the sample, L0 .) How
long can the solution be incubated before the dissolved oxygen concentration
decreases to 1.0 mg/L if the constituents in the dilution water consume almost
no oxygen, the mixture of sample and dilution water initially contains 10.5
mg/L dissolved oxygen, and the rate constant for oxygen consumption is 0.45
d 1?
Figure 3.13 Reaction pathway for formation of NDMA from DMA and monochloramine.
128
of 10 5 M DMA, 10 4 M NH2 Cl, 2 ⇥ 10 5 M NH3 , and zero for the other
species shown. Explain the concentration trends qualitatively.
(a) For a water at pH 7.5 that contains 10 4.5 M T OT OCl, calculate the con-
centrations of HOCl and OCl– in mol/L and in mg/L as Cl2 . (Hint: use
the approach shown in Example 2.6 to determine the relative concentra-
tions of HOCl and OCl– .)
(b) Plot log(N/N0 ) vs. time (in seconds) for E.coli and for Poliovirus 1, for
the solution described in part (a). Assume that n = 1, the disinfectant
concentrations are approximately constant and that the effects of HOCl
and OCl– are additive. How long is required to reduce the population of
each organism by a factor of one million?
21. Fine particles suspended in water can collide as a result of differential settling
rates, fluid mixing, and/or Brownian motion. Under certain conditions, these
collisions cause the particles to aggregate into larger particles which are more
likely than the original particles to settle out of the water column. By itself,
the aggregation process reduces the number concentration of suspended par-
ticles (i.e., number of particles per unit volume of solution), but not the mass
concentration.
19 Scarpino, P.V., Lucas, M., Dahling, D.R., Berg, G., and Chang, S.L. “Effectiveness of hypochlorous
acid and hypochlorite ion in destruction of viruses and bacteria.” Chapter 15 in Chemistry of Water
Supply, Treatment, and Distribution. A.J. Rubin, ed. Ann Arbor Science (Ann Arbor, MI) (1974).
129
Aggregation is sometimes modeled as an irreversible rate process with second
order dependence on the number concentration of suspended particles. Con-
sider a suspension that initially contains 108 particles per liter and in which the
rate constant for aggregation is 5 ⇥ 10 10 L/particle·s.
22. In the early sanitary engineering literature, there was a great deal of interest
in the effect of temperature on the rates of biodegradation and oxygen transfer
in aquatic systems, where the range of temperatures is limited to around 20 ±
20 C (293±20 K). Based on that work, an empirical model was developed
suggesting that
kT2
= q T2 T1
kT1
130