Aqueous chlorination kinetics and mechanism of substituted dihydroxybenzenes

Article abstract of DOI:10.1021/es950607t

The initial chlorination kinetics of several substituted dihydroxybenzenes, including chlorinated resorcinol compounds, was studied over the pH range of 2-12 at 22 °C. For each of the resorcinol substrates, the apparent chlorination rates are a minimum in the pH range of 3-6 and a maximum at pH values between 8-11. A mechanism that involves the reaction of HOCl with ArX(OH)₂, ArX(OH)O^-, and ArX(O^-)₂, and an acid-catalyzed pathway at pH < 4 was proposed to explain this pH dependence. Over natural water pH conditions, the reactions of HOCl with the anion and dianion forms of resorcinol groups are the most important. Although the intrinsic reactivity of HOCl with resorcinol substrates decreases with the extent of chlorine substitution on aromatic ring, the apparent reactivity of HOCl increases for more chlorinated resorcinols. In the presence of excess HOCl, monochloro- and dichloro resorcinol intermediates, therefore, should not accumulate when resorcinol groups undergo chlorine substitution. Linear free energy relationships for the reactivity of HOCl with resorcinols and phenols were developed. The sequential chlorination kinetics of resorcinol up to trichlororesorcinol can now be modeled. The initial chlorination kinetics of several substituted dihydroxybenzenes, including chlorinated resorcinol compounds, was studied over the pH range of 2-12 at 22°C. For each of the resorcinol substrates, the apparent chlorination rates are a minimum in the pH range of 3-6 and a maximum at pH values between 8-11. A mechanism that involves the reaction of HOCl with ArX(OH)₂, ArX(OH)O^-, and ArX(O^-)₂, and an acid-catalyzed pathway at pH < 4 was proposed to explain this pH dependence. Over natural water pH conditions, the reactions of HOCl with the anion and dianion forms of resorcinol groups are the most important. Although the intrinsic reactivity of HOCl with resorcinol substrates decreases with the extent of chlorine substitution on aromatic ring, the apparent reactivity of HOCl increases for more chlorinated resorcinols. In the presence of excess HOCl, monochloro- and dichloro resorcinol intermediates, therefore, should not accumulate when resorcinol groups undergo chlorine substitution. Linear free energy relationships for the reactivity of HOCl with resorcinols and phenols were developed. The sequential chlorination kinetics of resorcinol up to trichlororesorcinol can now be modeled.

Full text of DOI:10.1021/es950607t

Environ. Sci. Technol. 1996, 30, 2235-2242
TABLE 1
Aqueous Chlorination Kinetics
Rate Constants for Chlorination of Substituted
Phenolates
and Mechanism of Substituted
Dihydroxybenzenes
substrate
k2 (M-1 s-1 a
)
ref
phenolate
3.05 × 104
3.3 × 105
2150
16
15
16
15
16
15
L A U R E N C E M . R E B E N N E ,
A L I C I A C . G O N Z A L E Z , A N D
T E R E S E M . O L S O N *
2-chlorophenolate
4-chlorophenolate
5.9 × 104
2680
4.9 × 104
Department of Civil and Environmental Engineering,
University of California, Irvine, California 92717
a
Corresponds to the rate constant for the reaction of HOCl +
substrate. Values from ref 16 were obtained by fitting the reported pH
profiles of apparent rate constants (17).
The initial chlorination kinetics of several substituted
dihydroxybenzenes, including chlorinated resorcinol
compounds, was studied over the pH range of 2-12
at 22 °C. For each of the resorcinol substrates, the
apparent chlorination rates are a minimum in the
pH range of 3-6 and a maximum at pH values between
8-11. A mechanism that involves the reaction of
HOCl with ArX(OH)2, ArX(OH)O-, and ArX(O-)2, and an
acid-catalyzed pathway at pH < 4 was proposed
to explain this pH dependence. Over natural water pH
conditions, the reactions of HOCl with the anion and
dianion forms of resorcinol groups are the most
important. Although the intrinsic reactivity of HOCl
withresorcinolsubstrates decreases withthe extent
of chlorine substitution on aromatic ring, the apparent
reactivity of HOCl increases for more chlorinated
resorcinols. In the presence of excess HOCl, monochlo-
ro- and dichloro resorcinol intermediates, therefore,
should not accumulate when resorcinol groups
undergo chlorine substitution. Linear free energy
relationships for the reactivity of HOCl with resorcinols
and phenols were developed. The sequential
chlorination kinetics of resorcinol up to trichlororesor-
cinol can now be modeled.
later by Lee (16) have shown that the overall reaction is
second-order and proportional to the concentration of
aqueous chlorine and phenol at pH > 6. Both investigators
also observed that the rate was highly pH-dependent and
observed a maximum in the neutral or slightly alkaline pH
range. The mechanism over this pH range was found to
involve the reaction of the hypochlorous acid species with
the phenolate ion:
k2
HOCl + C6H5O- 98 C6H4Cl(O-) + H2O
(1)
Similar findings were reported by Brittain and de la Mare
(18).
Substantial differences exist, however, between the rate
constants for the reaction of HOCl with substituted
phenolates, as reported by Soper and Smith (15), and those
that can be obtained from Lee’s kinetic data (16). In the
latter study, only the apparent rate constants, kapp, were
reported as a function of pH where
d[phenol]
) kapp[phenol]TOT[chlorine]TOT
(2)
dt
Using Lee’s data over the range pH > 6 to determine k2,
the values shown in Table 1 are obtained. Soper and Smith’s
estimates of k2 for the same substituted phenols are
considerably larger (see Table 1).
In addition to the large uncertainty that exists between
the reported chlorination rates of phenols in neutral or
alkaline solutions, the mechanism of phenol chlorination
in acidic solution (pH < 6) has not been determined. The
two existing kinetic studies were conducted in the presence
of high chloride ion concentrations (15, 16). When the
solution was acidified with HCl, the reaction was found to
be independent of phenol and proportional to the square
of the HCl concentration (15). It was proposed that the
reaction was limited by the formation rate of Cl2. The
reversible formation of molecular chlorine from chloride
and hypochlorous acid is shifted to the right in the presence
of chloride:
Introduction
The formation mechanisms ofchlorination byproducts have
been extensively investigated since humic substances were
first identified as precursors for trihalomethanes in natural
water disinfection processes (1). The majority of these
studies have focused on the identification of stable inter-
mediates and products with humic and fulvic acids (2-7)
or model compounds (8-13) as substrates. Dihydroxy-
benzene moieties in humic matter were hypothesized by
Rook (9) to be responsible for the formation of chloroform.
Later studies by de Leer et al. (5) corroborated this theory.
Norwood and Christman (14) have also demonstrated the
importance of phenol groups as reactive centers for
chlorination byproducts.
H
+ + Cl- + HOCl a Cl2 + H2O
(3)
Early kinetic studies of the chlorination of phenol and
chlorine-substituted phenols by Soper and Smith (15) and
Detailed kinetic studies of the bromination of phenols
have also been conducted in the presence of high con-
centrations of bromide (>0.1 M) (19). Under these condi-
tions, the rate of phenol bromination was governed by the
* Author to whom correspondence should be addressed; fax: (714)
824-3672; voice: (714) 824-7188; e-mail address: tolson@uci.edu.
9
S0013-936X(95)00607-9 CCC: $12.00
1996 Am erican Chem ical Society
VOL. 30, NO. 7, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 2 3 5

reactions of molecular bromine with phenol and the
phenoxide ion over the pH range between 0 and 7. In the
absence of bromide, however, the rate of bromination of
aromatic compounds by molecular hypobromous acid is
proportional to the acidity at low concentrations of acid
(20). It was hypothesized that the reaction pathwayinvolves
either (1) the pre-equilibrium protonation of hypobromous
acid followed by the rate-limiting reaction of the positive
bromine species with the substrate or (2) HOBr and the
substrate are first brought together, protonated, and the
rate-determining step is the decomposition of the proto-
nated transition state. Given that the chlorination kinetics
of phenols in acidic solution have only been characterized
in the presence of high chloride concentrations and that
phenol bromination is apparently catalyzed by acid in the
absence of bromide, additional kinetic studies of phenol
chlorination at pH < 6 are needed to determine whether
acid catalysis is observed in the absence or presence of
only small chloride concentrations.
Methods. All kinetic experiments were conducted in a
batch reactor at room temperature (22 °C). The pH of the
solution was kept constant by additions of H2SO4 and KOH
with a Radiometer pH-stat apparatus. Aluminum foil was
used to shield the reactor from extraneous photochemical
reactions. Reagents were added to the reactor with
micropipets.
Chlorination experiments were conducted under pseudo-
first-order conditions where HOClwas in excess. The molar
chlorine to substrate ratio varied from 20 to 50. In a typical
experiment, a 250-mL volume of HOCl solution was
prepared in a batch reactor, and the pH was adjusted to the
desired value. The dihydroxybenzene was added to the
rapidly stirred chlorine solution, and at fixed time intervals
the reaction was quenched with an excess of sodium
thiosulfate. Half-lives for the substituted resorcinols ranged
from 7 min to 5 s, and the uncertainty in measuring the
reaction times was estimated to be 1 s. The stirring rate
of the reactor was set by increasing the mixing speed until
the reaction rate was independent of the stirring rate.
Although no rate constant is available for the reaction of
chlorine with S2O3-, thiosulfate is typically used as a titrant
in chlorine determinations on the basis of its rapid reaction
rate. The reaction rate of thiosulfate with molecular iodine
is also known to be diffusion-limited (24). Given the excess
thiosulfate that was added and the high reactivity of
Few detailed quantitative studies of the chlorination
kinetics of dihydroxybenzene compounds are available.
Heasley et al. (21, 22) determined through the identification
of intermediates that successive electrophilic chlorination
of resorcinol from monochloro-, to dichloro-, to 2,4,6-
trichlororesorcinol occurs before ring-opening. They hy-
pothesized that a pentachloro intermediate was involved
in the ring-opening step. In these studies, only relative
chlorination rates of resorcinol and chlorine-substituted
resorcinols were determined at pH 7 however. It is expected
that the completely deprotonated resorcinol species (di-
anion) would be even more reactive than the singly
deprotonated species. The contribution ofthe dianion form
to the chlorination reaction at natural water pH conditions,
however, cannot yet be predicted.
-
thiosulfate with chlorine, quenching of HOCl by S2O3 is
expected to be significantly faster than the reaction of
resorcinol and HOCl.
Analyses for the resorcinol substrates were performed
by HPLC with a Dionex Series 4000i chromatograph
equipped with a Supelcosil LC-18 column and a Dionex
variable UV/ Vis wavelength detector. Samples were eluted
with a methanol/ water/ acetic acid mobile phase at a flow
rate of 1.5 mL/ min. Chromatographic peaks for the
substituted resorcinols were well-resolved, and hence the
chlorinated products did not interfere with substrate
concentration measurements.
In this study, the chlorination kinetics of several
substituted resorcinols were studied over a wide range of
pH conditions. A series of chlorine-substituted resorcinol
substrates were selected to obtain a stepwise evaluation of
the successive chlorination of resorcinol through dichlo-
roresorcinol. Free energy correlations and structure-
reactivity relationships were also examined. The results
are useful for predicting the extent of accumulation of
chlorinated phenols and resorcinols under varying HOCl:
phenol ratios and for comparing the relative reaction rates
of chlorine with phenolic compounds and other substrates.
Final concentrations of HOCl were measured in parallel
experiments, in which potassium iodide was used to quench
the reaction. The I2 was quantitatively titrated with a
standard sodium thiosulfate solution in the presence of a
starch indicator, after acidifying the solution with acetic
acid to the pH range of 3-4 (25).
Acid dissociation constants Ka1 and Ka2 for orcinol,
4-chlororesorcinol, and 4,6-dichlororesorcinol were de-
termined by potentiometric titration according to the
method of Albert and Serjeant (26).
Experimental Procedures
Materials. Reagents including resorcinol (99% Baker),
orcinol or 5-methylresorcinol (ACS reagent grade, Fisher
Scientific), 4-chlororesorcinol (98% Aldrich), 4,6-dichlo-
roresorcinol (97% Aldrich), NaOCl (5% solution, Baker),
Na2S2O3 (Fisher), H2SO4 (Fisher), KOH (Fisher), and NaClO4
(99% Aldrich) were used without further purification. Stock
solutions (0.2 M) ofthe dihydroxybenzene compounds were
prepared in HPLC grade methanol. All other solutions were
prepared with deionized water from a Millipore purification
system.
Results and Discussion
The measured acid dissociation constants for each of the
resorcinol substrates and several reported pKa values for
resorcinol are tabulated in Table 2. A relatively wide range
of published pKa1 values are available for resorcinol and
hence the average of three reported values, 9.43, was used
in modeling the proposed mechanism herein. The average
literature value is also relatively consistent with the value
ofpKa1 )9.33 obtained from available Hammett correlations
for the acidity of phenols, i.e., pKa ) 9.92-2.23Σσ (29). With
the exception of the second pKa for 4,6-dichlororesorcinol,
all of the measured pKa values in Table 2 agree closely with
the published Hammett correlations of phenol acid dis-
sociation constants (see Figure 1).
The commercial NaOCl reagent contained equimolar
amounts of OCl- and Cl- ions. Based on reported values
of the equilibrium constant for the hydrolysis of Cl2 (23),
the maximum concentration of Cl2 (at pH 2) in our study
was approximately 2.5 × 10-8 M, since the reactions were
conducted with [HOCl]TOT ) [Cl-] e 3.5 × 10-5 M.
9
2 2 3 6 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 7, 1996

TABLE 2
Acid Dissociation Constants for Substituted
Resorcinolsa
substrate
pKa1 ((σ)
pKa2 ((σ)
resorcinol
9.43 (0.34)b
8.09 (0.08)
7.53 (0.09)
9.35 (0.03)
11.21 (0.16)c
10.75 (0.15)
10.35 (0.11)
11.50 (0.02)
4-chlororesorcinol
4,6-dichlororesorcinol
orcinol
a
Values determ ined experim entally in this study unless specified
b
otherwise. Value is the average of three reported values, 9.15 (27),
9.32 (26), and 9.81 (28). Value is the average of two reported values,
c
11.1 (26) and 11.32 (27).
FIGURE 2. Psuedo-first-order kinetic plot of the chlorination of
resorcinol at pH 4. Initial concentrations of resorcinol and HOCl
were 2.0 × 10-7 and 2.0 × 10-5 M, respectively. The solid line is
a linear least-squares regression of the data. Dotted lines show
intervals of 1 SD.
the resorcinol concentration as follows:
[ArH(OH)2]T,0
ln
(
)
[ArH(OH)2]T,t
kobs
)
(5)
∆t
where ∆t was the time period of reaction before quench-
ing. At least three determinations of kobs were performed
at each pH. Only reaction extents, ê ) 1 - ([ArH(OH)2]/
[ArH(OH)2]0), less than 0.8 were considered in calculating
kobs values.
FIGURE 1. Hammett correlation for resorcinol acid dissociation
constants, pK and pK . The solid line is a linear least-squares fit
a1
a2
to the data, corresponding to the equation in the graph. Dashed line
is a reported Hammett correlation for phenols (29).
No temperature dependence of kapp for resorcinol
chlorination was observed at pH 5. Based on the standard
error of kapp measurements made between 4 and 25 °C, an
upper limit activation energy for the reaction is therefore
e2.5 kJ/ mol. Allother experiments were performed at room
temperature, 22 °C.
Reaction Rate Orders. Chlorination experiments, which
were conducted at pH 4 with a large excess of HOCl relative
to the dihydroxybenzene, exhibited a pseudo-first-order
dependence on the substituted resorcinol (ArX(OH)2), as
demonstrated by the linear plot of ln([ArX(OH)2]/ [ArX-
2
(OH2)]0) vs time in Figure 2. These plots were linear (r
>
pH Dependence and Mechanism . The pH dependence
of kapp is shown in Figures 3-6 for each of the resorcinol
substrates. In each pH profile, the second-order apparent
rate constant exhibits a minimum between pH 3 and pH
6 and a maximum between pH 8 and pH 11. The minimum
rate constant for the reaction is also displaced to higher pH
values as the substrate is more extensively chlorinated. The
shape of the rate profile could be partially explained in
terms of the reactivity of the ionized and un-ionized forms
of the resorcinol substrates. Upon deprotonation, the O-
substituent is more activating than OH toward electrophilic
substitution. The shift of the minimum to lower pH values
with successive chlorination would be consistent with this
0.99) for reaction extents of at least 87%. The reaction order
with respect to chlorine was examined at pH 5 by varying
the chlorine concentration and monitoring the disappear-
ance of resorcinol. These experiments were limited to the
range of [HOCl]T < 3.5 × 10-5 M, since the reaction was too
rapid to monitor at higher chlorine concentrations. Plots
2
of the pseudo-first-order rate constant were linear (r
>
0.99) with respect to [HOCl]T. The rate of resorcinol
disappearance, therefore, is first-order in the concentrations
of chlorine and resorcinol:
d[ArH(OH) ]
2 T ) -kapp[ArH(OH)2]T[HOCl]T (4)
dt
explanation, since pKaresorcinol > pKamonochlororesorcinol
>
pKadichlororesorcinol
.
In addition, the maxima in the rate
where kapp ) kobs/ [HOCl]T is an apparent second-order
constant, and kobs is the observed pseudo-first-order rate
constant. Similar rate laws were found to describe the
disappearance of4-chlororesorcinol, 4,6-dichlororesorcinol,
and orcinol.
constant profiles might be explained if HOCl is the only
active electrophile and the reactivity of OCl- is negligible.
The increase in reaction rate in acidic solutions (pH < 4)
suggests that the reaction is subject to acid catalysis.
At all other pH conditions, the pseudo-first-order rate
constants were calculated from a single determination of
One mechanism that incorporates the above features is
proposed as follows:
9
VOL. 30, NO. 7, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 2 3 7

FIGURE 3. Apparent rate constant-pH dependence for the attack
of HOCl on resorcinol at 22 °C. The solid line is a non-linear least-
squares regression fit of k2, k3, and k4 to the data.
FIGURE 4. Apparent rate constant-pH profile for the attack of HOCl
on 4-chlororesorcinol at 22 °C. The solid line is a non-linear least-
squares regression fit of k2, k3, and k4 to the data.
Cl
Ka2
HOCl y z OCl- + H+
(fast)
(fast)
(fast)
(6)
(7)
(8)
R
Ka1
ArX(OH)2 y z ArX(OH)O- + H+
R
Ka2
ArX(OH)O- y z ArX(O-)2 + H+
k1
ArX(OH)2 + HOCl
9
8 products
(slow) (9)
(slow) (10)
(slow) (11)
k2
ArX(OH)O- + HOCl
98 products
k3
ArX(O-)2 + HOCl
98 products
k4
ArX(OH)2 + HOCl + H3O+ 9
8
products
(slow) (12)
and the resulting general rate expression for the above
reaction mechanism is
FIGURE 5. Apparent rate constant-pH dependence for the attack
of HOCl on 4,6-dichlororesorcinol at 22 °C. The solid line is a non-
linear least-squares regression fit of k1, k2, k3, and k4 to the data.
d[ArX(OH)2]T
ν ) -
ν ) (k1RoR + k2R1R + k3R2R + k4R0 [H+])R1
×
R
Cl
dt
) k1[ArX(OH)2][HOCl] + k2[ArX(OH)O-][HOCl] +
k3[ArX(O-)2][HOCl] + k4[ArX(OH)2][HOCl][H+] (13)
[ArX(OH)2]T[HOCl]T (16)
and based on eq 16, the kapp term in eq 4 can be rewritten
By introducing the following notation
as
[ArX(OH)2]T ) [ArX(OH)2] + [ArX(OH)O-] +
kapp ) (k1{H+}2 + k2Ka1R{H+} +
[ArX(O-)2] (14)
[
{H+}
+
[HOCl]T ) [HOCl] + [OCl-]
(15)
k3Ka1 Ka2R + k4{H+}3)
R
Cl
(
)
]
{H } + Ka2
/
where RiCl is the ionization fraction of hypochlorous acid
species, with i ) 0 or 1, for HOCl and OCl-, respectively,
and RiR is the ionization fraction of resorcinol substrate
species, with i ) 0, 1, or 2, for ArX(OH)2, ArX(OH)O-, and
ArX(O-)2, respectively, the rate expression can be rewritten
as follows:
[{H+}2 + Ka1R{H+} + Ka1 K12R] (17)
R
The intrinsic constants k1, k2, k3, and k4 were fit to the
data in Figures 3-6 with a non-linear least-squares regres-
sion and Marquardt-Levenberg algorithm solution method.
9
2 2 3 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 7, 1996

in the ratio k4/ k1. These ratios are smaller for dichlorore-
sorcinol and orcinol than the other substrates when the
upper-limit estimates of k1 for resorcinol and 4-chlorore-
sorcinol are assumed. The variability in these ratios also
appears to be driven by differences in the sensitivity of the
reaction to acid catalysis. It is hypothesized that the
simultaneous reaction of HOCl and H3O+ with orcinol and
4,6-dichlororesorcinol is sterically hindered, since the
reaction centers of these substrates have ortho substituents
on either side.
The proposed mechanism of eqs 6-12 invokes a
termolecular reaction to describe the increased chlorination
rates of resorcinols at pH < 4. Similar acid-catalyzed
reactions of HOCl with other nucleophiles have been
proposed (23, 30, 31). An alternative mechanism, however,
can also explain the acid-catalyzed reaction rate at pH <
4 if the k step is replaced with the following reaction:
4
k4′
ArX(OH)2 + ClOH2+ 98 products
(slow) (18)
FIGURE 6. Apparent rate constant-pH profile for the attack of HOCl
on orcinol at 22 °C. The solid line is a non-linear least-squares
regression fit of k1, k2, k3, and k4 to the data.
which in turn requires an additional equilibrium:
Cl
1/ Ka1
HOCl + H+ y
z ClOH2
(fast)
(19)
+
Acid dissociation constants were held fixed at the values
Cl
given in Table 2 for the resorcinol substrates, and pKa2
)
7.54 (26). Fitted values of all of the intrinsic rate constants
are presented in Table 3. Reasonably close fits were
obtained for each of the substrates, as shown by the solid
lines in Figures 3-6, and hence the postulated mechanism
is consistent with the data.
Given the large difference between the measured pKa2
for 4,6-dichlororesorcinol and the value predicted by
published LFERs (see Figure 1), a non-linear regression
analysis of the data in Figure 5 was also performed wherein
Ka2 was fixed at its “predicted” value. Very poor fits for
The magnitude of Ka1Cl is not known, although Arotsky and
Symons (32) have estimated that -4 < pKa1Cl < -3. Over
the range of pH studied here Ka1Cl . {H+}, so consequently
+
the ionization fraction corresponding to ClOH2 can be
Cl
approximated as R0 ≈ {H+}/ Ka1Cl when pH e pKa2Cl. An
“alternate” kineticallyequivalent rate expression is obtained
Cl
by replacing the k4R0RR0Cl term in eq 16 by k4′R0R{H+}/ Ka1
when pH e pKa2Cl. The constant k4 obtained previously in
Table 3 is related to k4′ by k4 ) k4′/ Ka1Cl. Given that the
Cl
largest value of k4′/ Ka1 in Table 3 is 9.8 × 106 M-2 s-1
,
k
app, however, were obtained in this case. The reaction
pKa1Cl values smaller than approximately -3 would require
that the k4′ step be faster than diffusion-controlled. This
alternate mechanism is possible, therefore, only if the value
of pKa1Cl is close to the high end of the range estimated by
Arotsky and Symons. Thus, there is insufficient evidence
to rule out the alternate k4′ reaction pathway.
rate pH dependence, therefore, is further evidence that the
pKa2 value of 4,6-dichlororesorcinol is much closer to our
measured value of 10.35 than the 9.1 value predicted by the
regression in Figure 1.
Since the chlorination rate-pH dependence ofresorcinol
and 4-chlororesorcinol could be fit satisfactorily with or
without the k1 step of the mechanism, the solid lines in
Figures 3 and 4 correspond to the fit obtained with only the
k2, k3, and k4 steps. While the magnitude of k1 cannot be
determined for resorcinol and 4-chlororesorcinol, upper-
limit estimates of k1 e 330 M-1 s-1 for resorcinol and k1 e
65 M-1 s-1 for 4-chlororesorcinol were obtained as k1 + σ,
when the k1 step was included in the regression analysis.
Modeling the rate constant-pH profiles for the chlorination
of 4,6-dichlororesorcinol and orcinol required the inclusion
of the k1 step to fit the data, as might be expected from the
broader shape of the minima in Figures 5 and 6.
Based on the earlier chlorination studies of phenols by
other investigators (15, 16), it is unlikely that the above
mechanism would be valid in acidic solution if high
concentrations of chloride ions were present. Since Cl2 is
more electrophilic than HOCl and Cl2 formation is favored
under acidic conditions, the chlorination ofphenolbecomes
rate-limited by the reaction shown in eq 3 and independent
of the phenol concentration at low pH. The commercial
hypochlorite reagent used in our study contained equimolar
concentrations of HOCl and Cl-, and hence the maximum
chloride concentration was 3.5 × 10-5 M. This Cl-
concentration, however, did not apparently affect the rate
of resorcinol decay since the reaction rates were first-order
One explanation for the differences in the importance
of eq 9 among these substrates is the apparent variability
TABLE 3
Rate Constants for Proposed Reaction Mechanism (eqs 9-12) at 22 °C
substrate
k1 ((σ) (M-1 s-1
)
k2 ((σ) (M-1 s-1
)
k3 ((σ) (M-1 s-1
)
k4 ((σ) (M-2 s-1
)
resorcinol
<330
<65
47 (17)
1250 (160)
1.36 (0.26) × 106
1.43 (0.16) × 105
3.21 (0.76) × 104
5.18 (0.34) × 106
1.15 (0.10) × 108
6.73 (0.53) × 107
5.91 (0.81) × 107
4.20 (0.04) × 108
8.5 (1.8) × 106
1.19 (0.15) × 106
2.6 (1.2) × 104
9.8 (1.1) × 106
4-chlororesorcinol
4,6-dichlororesorcinol
orcinol
9
VOL. 30, NO. 7, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 2 3 9

with respect to each of the substituted resorcinols, even at
pH 2.
The rate-limiting steps in eqs 9-12 do not involve
reactions of more than one charged species, and hence the
intrinsic rate constants k1, k2, k3, and k4 should not vary
with ionic strength according to classical Debye-Hu¨ ckel
theory (33). The apparent rate constant, kapp, on the other
hand is expected to vary with ionic strength since it is a
complex function of equilibrium constants (see eq 17).
However, at pH 2 and pH 12 where the ionic strength was
a maximum (0.01 M), the correction to kapp is estimated to
be at most 10%, which is within the error of measurement.
The ionic strength was varied using sodium perchlorate up
to 0.1 M at pH 5 with resorcinol, although little variation
in kapp was observed. According to the proposed mech-
anism, the k2 and k4 steps contribute primarily to kapp at
pH 5, and using the Davies equation to calculate activity
corrections, the expected change in kapp between 0 and 0.1
M ionic strengths is only between 10 and 11%. The
chlorination rate of phenol at pH > 6 has also been reported
to be independent of ionic strength, although the salt
concentration range used was not given (16).
While the relative order of the intrinsic reactivities of
chlorine with the four substrates in Table 3 is as expected,
the order of apparent reactivity, as indicated in Figures
3-6, is reversed between pH 6 and pH 10, with the most
apparently reactive substrate being 4,6-dichlororesorcinol.
This reversal can be explained in terms of the relative acidity
ofthe substrates. Chlorine-substituted resorcinols are more
acidic, and the deprotonated forms are more activated for
electrophilic attack. An important implication of this result
is that chlorinated resorcinol intermediates up to dichlo-
roresorcinols should not accumulate at natural water pH
conditions when HOCl is in excess. Comparisons of the
chlorination kinetics ofdihydroxybenzene-type groups with
other competing reactant centers in natural organic matter,
therefore, are best made on the basis of the initial reactivity
with these moieties.
Reaction Products. The initial reaction products were
identified by performing chlorination experiments under
pseudo-first-order conditions in the substrate at pH 5.
Chlorination of resorcinol gave two products that were
discernible by HPLC. One was 4-chlororesorcinol, as
demonstrated by matching its retention time with a pure
standard. The second product eluted immediately after
4-chlororesorcinol and before 4,6-dichlororesorcinol. While
a positive identification ofthis compound against a standard
was not performed, it is likely to be 2-chlororesorcinol, in
accordance with a studyconducted under similar conditions
(21).
Two chlorination products of 4-chlororesorcinol were
also observed by HPLC, and one of the chromatogram
peaks matched the retention time of 4,6-dichlororesorcinol.
The second product was not identified; however, it is
probably 2,4-dichlororesorcinol, as previously reported by
Heasley (21).
Preferred sites for chlorine substitution on the benzene
ring depend usually on the combined resonance and
inductive effects of the ring substitutents. For phenolic
compounds, OH and O- substituents are activating and
ortho- and para-directing. A Cl substitutent on the other
hand deactivates the ring. The immediate expected
products upon electrophilic substitution of chlorine onto
resorcinol, 4-chlororesorcinol, and 4,6-dichlororesorcinol
are as follows:
The rate constants in Table 3 for resorcinol, 4-chlorore-
sorcinol, and orcinol therefore correspond to the sum of
rate constants for the A and B pathways shown above. The
orientation of resorcinol OH groups renders the 2, 4, and
6 positions equivalent in terms of reactivity and their
reaction-directing effect. Consequently, the proportion of
each of the products can be estimated on the basis of a
statistical distribution. In the absence of steric hindrance,
the product distribution for the chlorination of resorcinol
should be approximately 33% 2-chlororesorcinol and 66%
4-chlororesorcinol. Heasley et al. (21) found a product ratio
of 2- and 4-chlororesorcinol of approximately 1:3, respec-
tively, for resorcinol chlorination over the pH range of 2-12.
The reduced reactivity measured by Heasley at the ortho
position may be due to steric hindrance or to deactivation
caused by inductive effects of the neighboring OH group.
In Heasley’s study, however, significant fractions of the
resorcinol were already converted to dichloro- and trichlo-
roresorcinol when the products were determined. There-
fore, the actual product formation rate ratio may be closer
to the statistically predicted ratio of 1:2. Nevertheless,
Heasley’s study is evidence that the three ring positions are
similarly reactive toward HOCl.
Chlorine substitution rates at either the 2 or 6 ring
position of 4-chlororesorcinol should also be similar, and
hence equal fractions of the products 2,4- and 4,6-
dichlororesorcinol are expected. Deviations from this
statistical distribution are possible ifsteric hindrance effects
are important. Such effects would likely favor 4,6-dichlo-
roresorcinol.
The introduction of a methyl group to resorcinol at the
C 5 position serves to weakly activate meta positions, which
are already occupied by OH substitutents. Similar product
isomer distributions should therefore be expected for
resorcinol and orcinol.
9
2 2 4 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 7, 1996

Linear Free Energy Relationships. Substituent effects
on the second-order rate constants in Table 3 appear to
follow expected trends. Chlorination rates increase by
almost 5 orders of magnitude, for example, when both OH
groups are replaced with O- (k1 vs k3). The rates also
decrease in the following order: orcinol > resorcinol >
4-chlororesorcinol > 4,6-dichlororesorcinol, as the weakly
deactivating Cl group is added to the benzene ring. Rate
constants reported by Soper and Smith (15) for the reaction
of HOCl with substituted phenolates similarly decrease as
chlorine substituents are added to the ring.
Hammett-type linear free energy relationships (LFER)
are often used to correlate effects of substituent groups for
aromatic compounds. Using the rate constants k1, k2, and
k3 determined in this study and the reported rate constants
for the chlorination of substituted phenolate compounds
(15), a Hammett-type correlation was tested. For substrates
in which electrophilic attack by HOCl occurs at more than
one site on the ring, an assumption of approximately equal
reactivity was made for each of the reaction centers in
estimating Hammett σ parameters. When the chlorination
reactions occur at n positions on the ring, the measured
rate constant is
n
k ) ki
∑
i)1
where ki values correspond to the rate constants of the
reaction pathways A, B, etc., described above. Furthermore,
if all the ki are nearly equal so they can be approximated
by ki,av and the rates are well-correlated with the Hammett
σ parameter, then
n
log k ) log
ki ≈ log(nki,av
)
(20)
∑
(
)
i)1
k
log ) log ki,av
(21)
(22)
n
FIGURE 7. Hammett-type correlation of log (k/n) vs (∑σ)av for the
reaction of substituted resorcinols and phenols with HOCl. Filled
circles correspond to the rate constants for substituted resorcinols
determined in this study. Open circles correspond to previously
determined rate constants for substituted phenols (15). The solid
line is a linear least-squares fit to all of the data.
k
log ) log k0 + F(Σσ)av
n
where k0 is the rate constant for the reaction of HOCl with
benzene, F is the Hammett slope, and (Σσ)av is the average
of the (Σσ)i constants for the n sites of attack. As shown
in Figure 7, a relatively linear plot of log(k/ n) vs (Σσ)av is
obtained for substituted resorcinols and phenols, using the
substituent Hammett σ parameters tabulated by Perrin et
al. (29). The magnitude and sign of the slope reflect the
sensitivity of the reaction to the electron density. Negative
values of F are typical of electrophilic substitution reactions.
While the rate constants obtained by Soper and Smith
(15) for substituted phenolates are consistent with the
Hammett correlation for the resorcinol compounds studied
here, rate constants obtained from an analysis of Lee’s
phenol chlorination study are not (16). In the latter kinetic
study, the pH dependence of phenol chlorination was
examined over the range of pH 5-12. While a mechanism
was not fit to the data, it can be shown that the observed
chlorination rates can be modeled by the following rate
determining steps (17):
k2
HOCl + C6H4X(O-)
98 products
(24)
The resulting values of k2 for phenolate, 2-chlorophenolate,
and 4-chlorophenolate, however, are more than an order
of magnitude smaller than those reported by Soper and
Smith (15). The reason for these differences is not clear,
but the rate constants obtained by Soper and Smith seem
more reasonable given the agreement oftheir rate constants
with our Hammett correlation for resorcinols.
Relatively close correlations of the chlorination rates of
phenolates (k2) with the pKa of the phenol group have also
been demonstrated using Soper and Smith’s data (34),
where log k2 ) -3.86 + 0.95 pKa. As shown in Figure 8, the
analogous LFERfor resorcinolrate constants k2 and k3 yields
a very strong correlation with a slope that is similar to the
LFER slope obtained for phenols but a different intercept.
This LFERsuggests that an even higher degree ofcorrelation
exists between chlorination rate constants and thermo-
dynamic acid dissociation constants than between chlo-
k1
HOCl + C6H4X(OH)
98 products
(23)
9
VOL. 30, NO. 7, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 2 4 1

(7) Hanna, J. V.; Johnson, W. D.; Quezada, R. A.; Wilson, M. A.;
Xiao-Qiao, L. Environ. Sci. Technol. 1991, 25, 1160-1164.
(8) Smith, J. G.; Lee, S-F.; Netzer, A. Water Res. 1976, 10, 985-990.
(9) Rook, J. Environ. Sci. Technol. 1977, 11, 478-482.
(10) Larson, R. A.; Rockwell, A. L. Environ. Sci. Technol. 1979, 13,
325-329.
(11) Norwood, D. L.; Johnson, J. D.; Christman, R. F.; Hass, J. R.;
Bobenrieth, M. J. Environ. Sci. Technol. 1980, 14, 187-189.
(12) Boyce, S. D.; Hornig, J. F. Environ. Sci. Technol. 1983, 17, 202-
211.
(13) Tretyakova, N. Y.; Lebedev, A. T.; Petrosyan, V. S. Environ. Sci.
Technol. 1994, 28, 606-613.
(14) Norwood, D. L.; Christman, R. F. Environ. Sci. Technol. 1987, 21,
791-798.
(15) Soper, F. G.; Smith, G. F. J. Chem. Soc. 1926, 1582-1591.
(16) Lee, F. C. In Principles and Applications of Water Chemistry;
Faust, S. D., Hunter, J. V., Eds.; Wiley: New York, 1967; pp 54-
74.
(17) Rebenne, L. M. Kinetic and Mechanistic Study of the Aqueous
Chlorination of Dihydroxybenzenes. M.S. Thesis, University of
California, Irvine, 1994.
FIGURE 8. LFER correlation of log k2 and log k3 with pK and pK ,
respectively, for the chlorination of substituted resorcinols. The
solid line is a linear least-squares fit to all the data.
(18) Brittain, J. M.; de la Mare, P. B. D. In The Chemistry of Halides,
Pseudohalides, and Azides, Part I, Supplment D; Patai, S., Ed.;
Wiley: New York, 1983.
(19) Tee, O. S.; Paventi, M.; Bennett, J. M. J. Am. Chem. Soc. 1989,
111, 2233-2240.
a1
a2
rination rates and the Hammett σ parameter. Where pKa
values for substituted resorcinols or phenols are available,
these LFERs are likely to give better predictions of the rate
constants. Since the correlation in Figure 8 spans more
than a 4-order of magnitude range in rate constants,
however, significant errors may still be associated with
estimates obtained using this LFER.
In the pH range of most natural waters, the contribution
of the ArX(OH)O- + HOCl and ArX(O-)2 + HOCl pathways
are typically the most important. Rate constants for these
steps of some isomer intermediates formed during the
chlorination of resorcinol have not been experimentally
determined, such as 2-chlororesorcinol and 2,4-dichlo-
roresorcinol. Using the LFERs developed in this study,
however, it is now possible to estimate their reactivity and
thus to completely model the chlorination sequence of
resorcinol up to the formation of 2,4,6-trichlororesorcinol
for the pH range of most aquatic systems.
(20) de la Mare, P. B. D. Electrophilic Halogenation; Cambridge
University Press: London, 1976; pp 125-126.
(21) Heasley, V. L.; Burns, M. D.; Kemalyan, N. A.; McKee, T. C.;
Schroeter, H.; Teegarden, B. R.; Whitney, S. E.; Wershaw, R. L.
Environ. Toxicol. Chem. 1989, 8, 1159-1163.
(22) Heasley, V. L.; Anderson, M. E.; Combes, D. S.; Elias, D. S.;
Gardner, J. T.; Hernandez, M. L.; Moreland, R. J.; Shellhamer, D.
F. Environ. Toxicol. Chem. 1993, 12, 1653-1659.
(23) Wang, T. X.; Margerum, D. W. Inorg. Chem. 1994, 33, 1050-
1055.
(24) Scheper, W. M.; Margerum, D. W. Inorg. Chem. 1992, 31, 5466-
5473.
(25) APHA, AWWA, and WEF. Standard Methods for the Examination
of Water and Wastewater, 18th ed.; American Publich Health
Association: Washington, DC, 1992; Chapter 4, p 38.
(26) Albert, A.; Serjeant, E. P. The Determination of Ionization
Constants: A Laboratory Manual, 3rd ed.; Chapman and Hall:
New York, 1984.
(27) Kortum, G.; Vogel, W.; Andrussow, K. Dissociation Constants of
Organic Acids in Aqueous Solutions; International Union of Pure
and Applied Chemistry: Butterworth: London, 1961.
(28) Lide, D., Ed.; Handbook of Chemistry and Physics, 72nd ed.; CRC
Press: Boca Raton, FL, 1991-1992.
(29) Perrin, D. D.; Dempsey, B.; Serjeant, E. P. pKa Prediction for
Organic Acids and Bases; Chapman and Hall: New York, 1981.
(30) Kumar, K.; Day, R. A.; Margerum, D. W. Inorg. Chem. 1986, 25,
4344-4350.
(31) Kumar, K.; Margerum, D. W. Inorg. Chem. 1987, 26, 2706-2711.
(32) Arotsky, J.; Symons, M. C. R. Q. Rev. Chem. Soc. 1962, 16, 282-
297.
Acknowledgments
The authors gratefully thank the National Science Founda-
tion (Contract BES-9257896) and the Irvine Ranch Water
District (Project IRWD-15021) for their financial support of
this work.
(33) Laidler, K. J. Chemical Kinetics, 3rd ed.;Harper Collins: NewYork,
1994; pp 197-201.
(34) Larson, R. A.; Weber, E. J. Reaction Mechanisms in Environmental
Organic Chemistry; Lewis Publishers: Boca Raton, FL, 1994; p
280.
Literature Cited
(1) Rook, J. J. Water Treat. Exam. 1974, 23, 234-243.
(2) Peters, C. J.; Young, R. J.; Perry, R. Environ. Sci. Technol. 1980,
14, 1391-1394.
(3) Christman, R. F.; Norwood, D. L.; Millington, D. S.; Johnson, J.
D. Environ. Sci. Technol., 1983, 17, 625-628.
(4) Miller, J. W.; Uden, P. C. Environ. Sci. Technol. 1983, 17, 150-
157.
Received for review August 15, 1995. Revised manuscript
received January 17, 1996. Accepted March 25, 1996.X
(5) de Leer, E. W. B.; Damste´, J. S. S.; Erkelens, C.; de Galan, L.
Environ. Sci. Technol. 1985, 19, 512-522.
ES950607T
(6) Reckhow, D. A.; Singer, P. C.; Malcolm, R. L. Environ. Sci. Technol.
1990, 24, 1655-1664.
X Abstract published in Advance ACS Abstracts, May 15, 1996.
9
2 2 4 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 7, 1996