Mechanistic implications of the intrahumic dechlorination of mirex

Article abstract of DOI:10.1021/es960581l

The hydrophobic chlorocarbon mirex readily binds to aqueous humic acids (HAs). In HA solutions irradiated at sunlight wavelengths (λ ≤ 290 nm), mirex molecules bound to HAs are transformed to photomirex by dechlorination. The mechanism of this intrahumic dechlorination was investigated both in HA solutions and in model systems simulating select photochemical capabilities of HAs. In HA solution, the reaction was unaffected by methanol and pentanol; was accelerated by hydroxide; and was inhibited by oxygen, 2-chloroethanol, nitrate, and hydrogen ion, all four of which can scavenge hydrated electron. Additional experiments probing for involvement of hydrated electron were consistent with it as the reactant. In irradiated N,N-dimethylaniline solution, a model system for generating hydrated electron, mirex was dechlorinated to form photomirex, the same product that is found in HA solution. Relative rate experiments in HA solution, while inconsistent with the reaction with hydrated electron in homogeneous solution, were consistent with reaction with a purely intrahumic hydrated electron. However, other humic-generated reductants cannot be eliminated as possible reactants. The potential confounding influence of hydrophobic partitioning to HAs on investigations using molecular probes is evaluated in terms of current data and previous reports.

Full text of DOI:10.1021/es960581l

Environ. Sci. Technol. 1997, 31, 1365-1371
Mechanistic Implications of the
Intrahumic Dechlorination of Mirex
S U S A N E . B U R N S , *
J O H N P . H A S S E T T , A N D
M A U R A V . R O S S I †
Department of Chemistry, State University of New York,
College of Environmental Science and Forestry,
Syracuse, New York 13210
The hydrophobic chlorocarbon mirex readily binds to
aqueous humic acids (HAs). In HA solutions irradiated at
sunlight wavelengths (λ g 290 nm), mirex molecules bound
to HAs are transformed to photomirex by dechlorination. The
mechanism of this intrahumic dechlorination was inves-
tigated both in HA solutions and in model systems simulating
select photochemical capabilities of HAs. In HA solution,
the reaction was unaffected by methanol and pentanol;
was accelerated by hydroxide; and was inhibited by oxygen,
2-chloroethanol, nitrate, and hydrogen ion, all four of which
can scavenge hydrated electron. Additional experiments
probing for involvement of hydrated electron were
consistent with it as the reactant. In irradiated N,N-dimethyl-
aniline solution, a model system for generating hydrated
electron, mirex was dechlorinated to form photomirex, the
same product that is found in HA solution. Relative rate
experiments in HA solution, while inconsistent with the
reaction with hydrated electron in homogeneous solution,
were consistent with reaction with a purely intrahumic
hydrated electron. However, other humic-generated reduc-
tants cannot be eliminated as possible reactants. The
potential confounding influence of hydrophobic partitioning
to HAs on investigations using molecular probes is evaluated
in terms of current data and previous reports.
FIGURE 1. Structures of mirex, photomirex (8-monohydromirex), and
10-monohydromirex.
hydroxyl radical (15), hydrogen peroxide (16), and hydrated
electron (11). While many organic contaminants undergo
phototransformations mediated by DOM (17-19), mirex is
unique because it is the first organic compound identified as
transformed entirely via intrahumic reaction; that is, only
mirex molecules bound to DOM undergo reaction (8). As the
greatest fraction of mirex in Lake Ontario water (36-64%) is
bound to DOM (4), this reaction has the potential to contribute
significantly to the environmental fate of mirex.
Several transient reactants produced by DOM may be
capable of dechlorinating mirex, including carbon-centered
radicals (20, 21) and hydrated electron (11, 22). DOM also
could dechlorinate mirex through electron transfer or other
reduction reactions (14, 23, 24). In this work, we report the
results of experiments designed to elucidate the mechanism
by which DOM transforms mirex to photomirex. Experiments
were performed both in humic acid solution, looking at the
effect of scavengers of various transient reactants on the
reaction of mirex, and in model systems using compounds
known to generate individual transient reactants.
Introduction
The chlorocarbon pesticide, mirex (Figure 1), is held to be a
persistent, unreactive pollutant because it resists biodegra-
dation (1) and partitions readily to sediment (2), natural
dissolved organic matter (3, 4) and biota (5). The lifetime of
mirex in Lake Ontario, where mirex is a contaminant, is
estimated at 10 years and is thought to be limited mostly by
sediment burial (6). Mirex nevertheless is readily transformed
in sunlit Lake Ontario water by dechlorination to form
8-monohydromirex (7), commonly called photomirex (Figure
1). The same reaction occurs in humic acid solutions
irradiated at sunlight wavelengths (7, 8) with the transforma-
tion being a photochemical reaction that is mediated by
natural dissolved organic matter (DOM) (8). This reaction is
the source of photomirex found in Lake Ontario water,
sediments, and biota (9).
Experimental Section
Materials. Humic acid was obtained from Aldrich Chemical
Company as the sodium salt. Mirex (100%) was obtained
from the U.S. EPA Pesticides and Industrial Chemical
Repositories, and photomirex (97%) was from the Environ-
mental Health Directorate of Health and Welfare Canada.
Additives to humic acid solution were oxygen (UHP), argon
(UHP), sodium or potassium nitrate (Fisher, certified),
hydrochloric acid (EM, reagent), sodium hydroxide (Fisher,
reagent), 2-chloroethanol (Kodak), methanol (Mallinckrodt
Ultimar), and 1-pentanol (Aldrich). Additives to solutions of
N,N-dimethylaniline (Aldrich) included nitrous oxide (Scott
Specialty Gases, 10% in N2) and trichloroacetic acid (Fisher,
certified). Reference compounds used in relative rate experi-
ments were chlorobenzene (Aldrich), lindane (Ultra Scientific,
99%+), and ortho- and meta-dichlorobenzene (Aldrich).
Internal standards were octachlorostyrene and PCB congeners
14, 65, and 209 (99%+), all obtained from Ultra Scientific. All
aqueous solutions were prepared using water purified with
a Millipore four-bowl standard system (Milli-Q water), and
all extraction solvents (hexane, isooctane) were pesticide grade
(J. T. Baker Resi-Analyzed).
Photochemical activity is observed both in native waters
containing DOM and in aqueous solutions of isolated humic
materials (10-13). These reactions include the formation of
excited states capable of energy (10) and electron transfer
(14) as well as the generation of short-lived reactants such as
* Corresponding author present address: Department of Chem-
istry, Western Michigan University, Kalamazoo, MI 49008; tele-
phone: (616)387-2878; fax: (616)387-2909; e-mail address: Burnss@
WMU.edu.
† Present address: Instituto de Quimica, Universidade de Sao Paulo,
Sao Paulo, Brazil.
Mechanistic Experim ents in Hum ic Acid (HA) Solution.
HA solution preparation and DOC measurements are de-
9
S0013-936X(96)00581-0 CCC: $14.00
1997 Am erican Chem ical Society
VOL. 31, NO. 5, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 3 6 5

scribed elsewhere (8). Unless otherwise specified, all HA
solutions were ∼pH 6. Irradiations were performed using an
Osram 150W/ 1 Xenon arc lamp in an Oriel uniform beam
illuminator housing (Model 6148) powered by an Oriel
universal power supply (Model 6240). Reaction vessels were
25 × 150 mm Kimax borosilicate culture tubes having Teflon-
lined caps (Kimble 45066A-25150). The wavelength cutoff of
the tubes was ∼290 nm. HA solutions were amended with
mirex in methanol (10-µL aliquots, final mirex concentration
∼100 ng/ L) and allowed to equilibrate 24 h prior to use.
Aliquots (40 mL) of these solutions were amended with O2,
Ar, NO3-, 2-chloroethanol, methanol, or 1-pentanol after
transfer to culture tubes. Gases (10 min bubbling at 0.1 L/ min)
and NO3- were added immediately prior to irradiation while
the remaining compounds were added 24 h prior to irradia-
tion. Irradiations were performed with the culture tubes
placed in an inverted position directly against the condenser
assembly of the lamp housing. Unless otherwise specified,
solution temperatures during these and other irradiations
were ∼25 °C.
culture tubes, and extracted as previously described for HA
solutions.
Mirex and lindane relative rate experiments not involving
DMA were performed in culture tubes with HA solution
prepared as previously described. Culture tubes were ir-
radiated in a photooxidation apparatus using an Oriel 350-W
Hg arc lamp (Model 6286) powered by an Illumination
Industries 350-W power supply. The distance between lamp
and sample was 2 cm.
Quantum Yield. The wavelength dependence for the
reaction of mirex in air-equilibrated Aldrich HA solution
(10 mg/ L DOC) was investigated over the region of 290-700
nm. Preliminary experiments performed using long wave
pass filters indicated that the reaction did occur at wavelengths
greater than 400 nm but not 500 nm. However, at wavelengths
>310 nm, reaction of mirex was too slow to be measured
when irradiations (<15 h) were performed using light from
the Xe lamp that had passed through a high-intensity
monochromator (19.2 nm band width, Bausch & Lomb Model
338679). A single reaction quantum yield was determined in
a time series experiment performed at 300 nm using the
monochromator and Xe lamp. The reaction vessel was a
cylindrical quartz cell having a 10-cm path length and total
volume of 20 mL. Solution temperature during irradiation
was 22 °C. Light flux in the cell was measured using
ferrioxalate actinometry (27). The preparation and extraction
of the HA solution were performed as previously described,
and each time series sample was matched with a dark control.
Kinetics. The transformation of mirex to photomirex in
HA solution obeys rate laws for a consecutive, pseudo-first-
order reaction (7, 8). This reaction can be followed using
photomirex (P) to mirex (M) concentration ratios and the
integrated rate equation (7)
Mirex and photomirex were extracted from HA solution
following a reported amended method (8) of Driscoll et al.
(25). Analyses in these and all subsequent experiments were
performed using methods and gas chromatography instru-
mentation described elsewhere (8).
Reactions in N,N-Dim ethylaniline (DMA) Solution. Ir-
radiations were performed using the lamp system described
in the previous section. Reaction vessels were 110 × 285 mm
(2 L) Wheaton-33 (borosilicate glass) roll bottles (wavelength
cutoff ∼295 nm). Roll bottle caps were customized with
stainless steel disks having a through port (1/ 8-in. SS tubing)
that was used both for sample withdrawal and gas introduc-
tion.
DMA solutions (∼0.001 M) were prepared in roll bottles.
DMA and analytes (mirex, chlorobenzene, lindane) dissolved
in methanol were added to 1.7 L of water, and the solution
was stirred in the dark with a glass-sealed magnetic stirbar
for 24 h prior to use. Final analyte concentrations were 50-
100 ng/ L. Following the method of Kohler et al. (26), all DMA
solutions were amended with methanol (final concentration
0.05 M) to suppress side reactions. Trichloroacetate and
2-chloroethanol were added to solutions at the same time as
DMA. Gases (Ar, 10% N2O in N2) were added immediately
prior to irradiation. Solutions in roll bottles were irradiated
with stirring flush against the condenser assembly of the lamp
housing. All experiments were replicated and matched with
dark and light (irradiated solutions containing no DMA)
controls.
kM′
M
P
P/ M )
(et(k ′-k ′) - 1)
(1)
kM′ - kP′
where t is time and k
M′ and kP′ are the respective reaction rate
constants for mirex and photomirex. The alternate equation
(7)
P
M
ln 1 +
) kM′t
(2)
(
)
applies when transformation of mirex to photomirex is <25%
(8). Equation 1 was used to calculate kM′ in time series
experiments, and eq 2 was used in single point irradiations.
Results
Samples (40-50 mL) were withdrawn from roll bottles
with a glass syringe and transferred to culture tubes containing
0.4 g of NaCl (Fisher certified). NaCl was heated to 600 °C
overnight in a muffle furnace prior to use and was added to
culture tubes to promote extraction of analytes. After sample
transfer, the syringe was rinsed with 2 mL of methanol, and
the methanol was added to the sample. Analytes were
extracted from solution into hexane or isooctane (2-4 mL)
by liquid-liquid extraction performed in the culture tubes.
Hum ic Acid (HA) Solutions. The HA-mediated dechlori-
nation of mirex follows apparent first-order kinetics according
to
d[M]/ dt ) -k′[M]
(3)
where [M] is the concentration of mirex and k′ is the apparent
first-order rate constant for the reaction of mirex. For reaction
with a single transient reactant, this apparent rate constant
is defined by (12)
Relative Rate Experim ents in HA Solution. Most relative
rate experiments were performed using roll bottles as reaction
vessels. The light source was a Hanovia 450-W medium
pressure Hg lamp housed in a reflector (Ace Glass 7883-02)
and powered by an Ace Glass cased power supply (7830-60).
HA solutions in roll bottles were amended as needed with
mirex, lindane, DMA, and ortho- and meta-dichlorobenzene
24 h prior to irradiation. All solutions were bubbled with Ar
immediately prior to irradiation. Solutions were irradiated
with stirring ∼20 cm in front of the lamp and were kept cool
during irradiation with both a fan and a water-driven turbine
stirrer (Fisherbrand 14511210). Samples (40 mL) were
withdrawn from solutions with a glass syringe, transferred to
k′ ) kM[X]ss
(4)
where [X]ss is the steady-state concentration of the unknown
DOM-photogenerated transient reactant X and kM is the
second-order rate constant for the reaction of mirex with X.
Previously, we reported that k′ in Aldrich HA solution
decreased when select compounds were added to HA solution
to scavenge X and thus reduce its steady-state concentration
(8). These additives include oxygen (O2), hydrogen ion (H+),
nitrate (NO3-), and 2-chloroethanol but not methanol or
1-pentanol (Figure 2).
9
1 3 6 6 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 5, 1997

FIGURE 3. Effect of pH on the apparent rate constant (k′) for the
indirect photolysis of mirex in humic acid solution (33 mg/L DOC).
FIGURE 2. Effect of oxygenation (O SAT’D), pH, nitrate (NO -),
2
3
2-chloroethanol (ClEtOH), methanol (MeOH), and pentanol (PeOH)
on the apparent rate constant (k′) for the indirect photolysis of mirex
in humic acid solution. Values of k′ are reported relative to values
in control solutions (ko′). Error bars represent standard deviation.
The increase in k′ at high pH may be significant because
hydroxide (HO-) and NO3- quench electron transfer reactions
with equal facility (38, 39). Quenching of electron transfer
reactions by these and similar inorganic anions is thought to
occur through electron exchange reactions that are initiated
when an anion donates an electron to the ground-state “hole”
in the excited electron donor (35, 38, 39). Supporting this
mechanism, the quenching capabilities of these anions are
correlated to the anion oxidation potentials rather than
reduction potentials (35, 38, 39). The respective aqueous
phase oxidation potentials of HO- and NO3- are reported to
be 2.29 and 2.3 eV vs the standard hydrogen electrode (40).
Because NO3- and HO- have opposite effects on the reaction
of mirex in HA solution, these potentials mean that NO3- is
likely to be interfering in the reaction by some mechanism
other than quenching of excited electron donors. Consistent
with this, fluorescence quenching commonly has been used
O2 is known or thought to be involved in the production
and decay of various DOM-generated reactants. For example,
O2 can scavenge triplet energy to form singlet oxygen (10)
and can scavenge hydrated electron (11) and reductants (28)
to form superoxide anion (12, 28). Figure 2 shows that the
mirex apparent reaction rate constant (k′) decreased 40% in
O2 rich HA solution (33 mg/ L DOC) relative to matching
solution depleted in O2. This precludes transient reactants
such as superoxide anion or singlet oxygen from causing the
reaction of mirex because O2 is required for their production
(10, 28). At the same time, this decrease implicates reactants
for which O2 is a scavenger, such as carbon-centered radicals
(20), electron donors (28), or hydrated electron (11). Alter-
natively, O2 could decrease k′ by quenching excited states
that are precursors to the reactant responsible for mirex
transformation. However, involvement of energy transfer in
the reaction is not likely because, even in systems with high
enough energy to access the excited states of chlorocarbons
and chlorinated aliphatic hydrocarbons, quenching of excited
singlet and triplet states typically occurs through electron
transfer (29, 30).
-
to detect quenching of electron transfer reactions by NO3
-
(35, 38), but NO3 does not measurably decrease the
fluorescence intensity of humic materials (34).
The decrease in k′ at low pH is potentially significant
because H+, NO3-, and O2 all are known to react with the
hydrated electron (eaq-) at diffusion-controlled rates (38),
suggesting that eaq- could participate in the transformation
of mirex to photomirex in HA solution. In support of this,
2-chloroethanol decreased k′ in O2-depleted HA solution (10
NO3- is a photochemical source of hydroxyl radical (31),
is a scavenger of hydrated electron (32-34), and quenches
some electron transfer reactions through what is thought to
be an electron-exchange reaction initiated by NO3- on contact
with excited electron donors (35). As reported previously, k′
decreased with increasing NO3- concentration in O2-depleted
HA solution (33 mg/ L DOC) (8). Figure 2 shows that at 0.2
M NO3-, k′ decreased by ∼60%. This decrease in k′ precludes
the involvement of the hydroxyl radical as significant to the
reaction of mirex. This was confirmed using Fenton’s reagent
following the method of Draper and Crosby (36), which did
not cause an observable reaction of mirex.
-
mg/ L DOC) (8). 2-Chloroethanol has been used as an eaq
probe in many systems, including HA solution (11), and has
a reduction potential greater than that of water (42). At 0.15
M 2-chloroethanol, k′ decreased by >90% (Figure 2). Metha-
nol and pentanol had no effect on k′ (Figure 2), indicating
that 2-chloroethanol decreased k′ not as an alcohol but as a
chlorinated compound. This last result tends to preclude
the involvement of carbon-centered radicals as significant to
the reaction of mirex because carbon-centered radicals
generally preferentially abstract hydrogen over chlorine (43).
The effect of pH on the reaction of mirex in O2-depleted
HA solution (33 mg/ L DOC) is shown in Figure 3. Between
pH 4 and pH 9, k′ was relatively constant. However, k′
decreased at low pH and increased at high pH. These trends
are not likely to be due to pH-induced changes in solution
absorbance because, while k′ does not increase significantly
between pH 6 and pH 9 (95% CI), the absorptivity of the HAs
at any wavelength between 290 and 700 nm more than
doubles when the pH is increased over the same range.
These trends in k′ also cannot be due to pH-induced changes
in mirex HA-water partitioning because the reaction of mirex
in HA solution is limited to mirex molecules that are bound
to HA molecules (8) and hydrophobic binding to HAs
decreases with increasing pH (37).
Previous work has shown that mirex can be dechlorinated
through electron transfer both from ground-state porphyrins
(24) and in a charge-transfer complex with triethylamine (23).
The dominant product of those reactions is 10-monohy-
dromirex (Figure 1), with photomirex formed as only a minor
product (23). The formation of 10-monohydromirex as the
dominant product of electron transfer reactions is somewhat
expected because these reactions have been shown to be
largely controlled by redox potentials (44), and geminal
dihalides typically are more easily reduced than isolated
halides (45, 46). Nevertheless, the formation of photomirex
as the product of the DOM-mediated reaction of mirex could
be expected of an electron transfer reaction from DOM to
-
mirex. In contrast, the reactivity of mirex toward eaq and
9
VOL. 31, NO. 5, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 3 6 7

-
appears to react with eaq to form photomirex, the same
product formed in HA solution.
The second-order rate constant for the reaction of mirex
with eaq- was determined in DMA solution using the method
of relative rates with CB as the original reference compound.
CB was chosen because its rate constant for reaction with
eaq- is known (5 × 108 M-1 s-1; 48, 49) and is typical of rate
constants for chlorinated, aliphatic hydrocarbons (108-109
M-1 s-1; 41). However, because of the low solubility of mirex
(∼100 ng/ L; 3) and the low relative response factor of CB in
GC-ECD analysis, CB could not be detected when added to
DMA solution at concentrations equal to that of mirex.
Consequently, lindane was chosen as an intermediate refer-
ence compound for the relative rate experiments. Lindane
has a GC-ECD response factor similar to mirex and, like
mirex, is a chlorinated, aliphatic hydrocarbon. Lindane also
is more soluble in water (∼10 ppm; 51) than mirex, and its
rate constant with eaq- therefore could be determined using
CB as the reference compound.
The relative rate method was tested using meta- and ortho-
dichlorobenzene. The compounds have a relative reaction
rate (meta-/ ortho-) with eaq- of 1.1 (48), the same ratio found
in this study. No loss of lindane occurred in light or dark
controls, and quenching of the lindane reaction by 2-chlo-
roethanol, trichloroacetate, and N2O was similar to that of
mirex (Figure 4c). On average, lindane reacted 1.25 times
faster than CB in DMA solution, and mirex reacted 1.44 times
faster than lindane. Given the rate constant for reaction of
CB with eaq-, these results translate into second-order rate
constants for reaction with eaq- of 8.71 ( 0.11 × 108 M-1 s-1
for mirex and 6.05 ( 0.77 × 108 M-1 s-1 for lindane.
Relative Rate in HA solution. As was found with mirex,
the reaction of lindane in HA solution was quenched by
2-chloroethanol (0.05 M) but was unaffected by methanol
(up to 0.1 M; 52). Furthermore, when the concentration of
lindane in HA solution was increased to 8 mg/ L, near its
solubility (∼10 mg/ L; 51), the mirex k′ decreased by 43% (8).
These results suggest that lindane and mirex are competing
for the same transient reactant in HA solution. However,
while mirex reacted only 1.44 times faster than lindane in
DMA solution, it reacted 31 ( 2.4 times faster in HA solution
(10 mg/ L DOC) (Figure 5a). If the reactions of mirex and
lindane were occurring in a single-phase or homogeneous
solution, then the relative rate results would rule out eaq- as
the sole transient reactant responsible for the reactions.
However, HA solution is not homogeneous, at least not
to molecules capable of associating with or partitioning to
HAs. In the case of mirex, reaction is limited to mirex
molecules that are bound to HA molecules (8). This indicates
that the transient reactant responsible for the dechlorination
occurs mostly in the HA phase. The DOC-water partition
coefficient (KDOC) of mirex in Aldrich HA solution is 106.26 (8),
meaning that 95% of mirex molecules are bound in HA
solution containing 10 mg/ L DOC. In contrast, approximately
2% of the lindane molecules would be expected to be bound
in this solution (Table 1). As a consequence, even if lindane
FIGURE 4. (a) Disappearance of mirex and appearance of photomirex
in irradiated DMA solution. (b) Effect of 0.05 M 2-chloroethanol
(2-ClEtOH), 0.005 M N2O, and 0.05 M trichloroacetate on the apparent
rate constant (k′) for transformation of mirex (M) to photomirex (P)
in irradiated DMA solution. (c) Effect of 0.05 M 2-chloroethanol
(2-ClEtOH), 0.005 M N2O, and 0.05 M trichloroacetate (Cl3COO-) on
the apparent rate constant (k′) for the reaction of lindane in irradiated
(λ g 290 nm) DMA solution.
-
the product or products of the reaction of mirex with eaq
have not been reported.
-
Dim ethylaniline Solution. To test whether eaq could
transform mirex to photomirex, the reaction of mirex was
studied in irradiated (λ g 290 nm) aqueous solutions of N,N-
-
dimethylaniline (DMA), which are known to produce eaq
(26, 47). Figure 4a shows the disappearance of mirex and the
appearance of photomirex in DMA solution. Loss of mirex
or formation of photomirex was not detected in dark or light
controls. As with the reaction of mirex in HA solution, the
reaction of mirex in DMA solution obeyed kinetics for a
consecutive, pseudo-first-order reaction.
Figure 4b shows that the reaction of mirex in DMA solution
was nearly quenched by 0.05 M 2-chloroethanol and was
completely quenched by 0.05 M trichloroacetate (pH 8) and
-
∼0.005 M N2O, both of which are also eaq scavengers (11,
-
26). These results are expected if mirex reacts with eaq in
DMA solution. Using chlorobenzene (CB, 50 µM) as a
-
and mirex were reacting only with eaq in the HA solution,
molecular probe for eaq- (48, 49), the measured steady-state
the relative reaction rate of the two compounds in that
solution would be expected to differ from the relative rate in
homogeneous DMA solution.
-
-
concentration of eaq ([eaq ]ss) in DMA solution was 5.4 ×
10-14 M. Assuming that measured residual oxygen (8.3 ×
-
10-5 M) controls [eaq
]
ss
in DMA solution and given a rate
For a compound that reacts only with eaq- in HA solution,
if the intra- and extrahumic reaction rate constants are the
same, the effect of binding on apparent reaction rate constant
(k′) can be represented by (8)
-
constant of 2 × 1010 M-1 s-1 for the reaction of O2 with eaq
(41), the calculated production rate of eaq- in these solutions
was 8.9 × 10-8 M s-1. Given a rate constant for reaction of
-
2-chloroethanol with eaq of 2 × 108 M-1 s-1 (50), 0.05 M
-
2-chloroethanol should decrease [eaq ]ss, and thus k′, by 86%,
k′ ) k[eaq-]dfd + k([eaq-]b/ [DOC])fb
(5)
in good agreement with observation (85%). Furthermore,
given the rate constants for the reaction of trichloroacetate
and N2O with eaq- (8.5 × 109 and 9.1 × 109 M-1 s-1, respectively;
where k is the rate constant for reaction of the compound
41), 0.05 M trichloroacetate and ∼0.005 M N2O should
with eaq-; [eaq-]d and [eaq-]b are the respective steady-state
-
-
completely quench (>97%) the reaction of mirex with eaq
.
concentrations of eaq in the dissolved (d) and bound (b)
This is also in agreement with observation. Mirex therefore
phases of HA solution, both expressed as molar concentrations
9
1 3 6 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 5, 1997

lindane molecules were in the bound phase of HA solution
-
and therefore had access to eaq
. This value is in good
agreement with expectation (∼2%, Table 1). Similar good
agreement between the expected and the observed binding
of lindane in HA solution was reported previously and was
based on the ability of lindane to quench the intrahumic
reaction of mirex (8). Therefore, the relative rate results are
quantitatively consistent with eaq- as the species responsible
for the reaction of mirex in HA solution.
Results obtained using ortho-dichlorobenzene (o-DCB)
and meta-dichlorobenzene (m-DCB) also are consistent with
eaq-. The DCBs have second-order rate constants for reaction
-
with eaq nearly an order of magnitude greater than that of
mirex (48) but, like lindane, have little affinity for HAs (Table
1). The relative rate of reaction of m-DCB to o-DCB in HA
solution (1.73 ( 0.96) was not significantly different from the
ratio of their rate constants (1.1; 48), but mirex reacted 2-3.5
times faster than the DCBs (Figure 5b). Furthermore, at the
DCB concentrations (∼12 µg/ L) needed to detect them
simultaneously with mirex, the DCBs had the effect of
decreasing the mirex k′ by 73%. As with lindane, this result
suggests that the DCBs and mirex are competing for the same
-
transient reactant. Correcting for the decrease in [eaq
]
b
caused by the DCBs and assuming reaction of mirex and the
DCBs only with intrahumic eaq-, 3.5% of the DCB molecules
-
in this experiment apparently had access to intrahumic eaq
as compared with a prediction of 2.6% (Table 1).
These results illustrate one of the consequences of the
effect of binding on DOM-mediated photoreactions: for
compounds having different affinities for HAs, competition
and relative rate results are expected to differ in HA solution
vs homogeneous solution. This is true regardless of the
relative contributions of the bound and dissolved phases to
overall reaction rate constants (8). The reaction of mirex in
HA solution represents one extreme, with the reactivity
transforming mirex being restricted to the bound phase. At
this extreme, compounds occurring primarily in the dissolved
phase are expected to have both little access to bound-phase
reactivity and little ability to compete for this reactivity.
Reactivity restricted to the dissolved phase would be expected
to give the opposite result, with bound-phase compounds
being effectively protected from transformation. This is
supported by a comparison of the reactions of lindane and
mirex in HA solution (10 mg/ L DOC) with and without added
DMA (∼0.001 M) (Figure 6). DMA has been used in previous
work as a source of eaq- in HA solution supplementing that
generated by HAs (11). Assuming the KDOC of DMA is similar
to that of other anilines (∼100; 51, 55), DMA has little affinity
for HAs (fb < 0.01). This means that DMA in HA solution
constitutes a dissolved-phase source of eaq-. As Figure 6a
illustrates, the loss of lindane was undetectable (95% CI) in
the HA solution without added DMA even though the reaction
of mirex went to ∼75% completion. With added DMA (Figure
6b), the reaction of lindane was greatly accelerated but no
change could be detected in the reaction of mirex (95% CI).
The relative rate of reaction of mirex to lindane with added
DMA was 1 ( 0.37, compared to 31 in HA solution having no
DMA (Figure 5a). These results clearly indicate that the
additional reactivity provided by DMA is unequally available
to mirex and lindane. Assuming that DMA-photogenerated
eaq- is available only to dissolved-phase compounds, the mirex
k′ in this solution would be predicted to increased by only
1.1%. This prediction agrees with the experimental result
that DMA did not measurably affect the mirex reaction rate.
FIGURE 5. (a) Sample relative rate plot for the reactions of mirex
and lindane in irradiated humic acid solution (10 mg of DOC/L). (b)
Relative rate plot showing the transformation of mirex (M) to
photomirex (P) and the disappearance of ortho- (O-DCB) and meta-
dichlorobenzene (M-DCB) in irradiated humic acid solution (10 mg/L
DOC).
TABLE 1. Comparison of Calculated and Apparent DOC-Bound
Concentration Fractions (fb) in HA Solution (10 mg/L DOC)
a
b
c
log K
calculated fb
apparent fb
DOC
lindane
o-DCB
m -DCB
3.22d
0.017
0.010
0.016
0.044
0.023
0.012
2.99d
3.21d
a
b
DOC-water partition coefficient, L/kg. Calculated using KDOC
c
values and eq 6. Calculated from eq 7 assum ing reaction with bound
phase eaq only. Refs 51 and 54.
-
d
in the aqueous solution; [DOC] is the HA concentration (in
kg/ L); and fd and fb are the respective fractions of the
compound in the dissolved phase and bound phases of HA
solution. The value of fd ) 1 - fb and fb is defined by (53)
K
DOC[DOC]
fb )
(6)
1 + KDOC[DOC]
From eq 5, if eaq- were confined to the bound phase, then the
ratio of the apparent reaction rate constants for two com-
pounds (compounds 1 and 2) with eaq- in HA solution (k′1/
k′2) would be
f
k1
b1
k′1/ k′2 )
(7)
k2 fb
2
where k1/ k2 is the ratio of their rate constants for reaction
-
with eaq and fb / fb is the ratio of the fractions of each
Quantum Yield. Quantum yields for eaq- production from
humic materials in solution are reported to be as high as 0.22
when measured using flash photolysis (56). However, previ-
ous work generally has concluded that eaq- occurs in natural
waters and solutions of humic materials at concentrations
too low to transform pollutants (∼10-17 M; 11). The primary
reason eaq- has been dismissed as an important intermediate
1
2
compound bound. Note that eq 7 applies only when
concentrations of the compounds are too low to significantly
affect [eaq-]b. For mirex and lindane in Aldrich HA solution
having 10 mg/ L DOC, if eaq- occurred only in the bound phase
and mirex and lindane reacted only with eaq-, then a value
of k′1/ k′2 ) 31 would mean that approximately 4% of the
9
VOL. 31, NO. 5, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 3 6 9

300 nm. The light flux (I300) in the solution was 1.1 × 10-9
Einsteins/ L s-1. Using the measured ꢀ300 value of 2.1 × 104
(kg/ L DOC)-1 cm-1, a path length of 10 cm, and the average
apparent kr value of 1.33 × 109 s-1 calculated by Breugem et
al. (34) yields Φ300 ) 0.03 from eqs 5 and 8. This quantum
yield falls within the range of values previously measured
(0.22-0.0017; 11, 56) using flash photolysis (355 nm) in both
natural waters and solutions of humic materials. Therefore,
the calculated quantum yield is consistent with the assump-
tion that the intrahumic dechlorination of mirex is mediated
-
by eaq
.
An intrahumic eaq- quantum yield measured by mirex in
air-equilibrated HA solution (20 °C) would not be expected
to differ significantly from one measured by mirex in deoxy-
genated solution. As previously demonstrated (8), the ability
of scavengers to interfere in intrahumic reactions depends
on the number of scavenger molecules present in individual
HA molecules, with scavenger molecules being distributed
in HA molecules according to a Poisson distribution. Ac-
cording to this distribution, the probability (Prn) of finding
no scavenger molecules in an HA molecule (n ) 0) is
Prn ) e-N
(9)
where N in HA solution is the average number of humic-
associated scavenger molecules per humic molecule. N is
given by (8)
KDOC
HA1 + KDOC[DOC]
N ) 10-3fcMW
[S]
(10)
where fc is the weight fraction of organic carbon in humic
molecules, MWHA is the gram molecular weight of the humic
molecules, [S] is the truly dissolved (not bound) concentration
of the scavenger, and KDOC and [DOC] are expressed in units
of L/ kg and kg/ L, respectively. Assuming fc ) 0.5 (57), MWHA
) 105 g/ mol (58), and KDOC ) 1 for O2 then, given a bulk
solution O2 concentration of 2.84 × 10-4 M (20 °C), eqs 9 and
10 indicate that there is a >98% probability of finding no O2
molecules in an HA molecule (10 mg of DOC/ L). This means
that >98% of all bound mirex molecules would be likely to
FIGURE 6. Relative rate plots showing the disappearance of mirex
and lindane in Aldrich humic acid solution (10 mg of DOC/L) in the
absence (a) and presence (b) of added DMA (0.001 M).
is the contrast between eaq- quantum yields measured using
flash photolysis and those measured using 2-chloroethanol
as a molecular probe for eaq- (11). Using the same solutions,
the range of quantum yields determined by nanosecond flash
photolysis was 0.0017-0.0076 while the range of quantum
yields measured using 2-chloroethanol was <0.8-12 × 10-5
.
-
react with eaq without competition from O2. This result
It was originally hypothesized that this difference might be
explained if some eaq- measured by flash photolysis existed
in a kind of radical cage that collapsed before eaq- could escape
to react with, and thus be measured by, 2-chloroethanol.
Another explanation is that eaq- does not escape the HA matrix
and that 2-chloroethanol, with an estimated KDOC of 1.4 (51,
54), occurred in these solutions almost exclusively in the
illustrates another consequence of the effect of binding on
DOM-mediated photoreactions: because it is possible for a
hydrophilic scavenger to have little access to intrahumic
transient reactants in spite of high bulk solution scavenger
concentrations, it is also possible for transient reactant
production rates and intrahumic steady-state concentrations
to be substantially underestimated by molecular probes, such
as 2-chloroethanol, that have little tendency to partition to
HAs.
-
dissolved phase. The possibility of a purely intrahumic eaq
is supported by the work of Breugem et al. (34), who used
NO3-, Zn2+, and chloroform as molecular probes for eaq- to
estimate the lifetime of eaq- generated in natural waters and
HA solutions under steady-state conditions. Based on an
estimated lifetime of 0.2 ns in HA solution and an average
distance covered during that lifetime of 1.3 nm, the authors
concluded that eaq- was confined to the humic matrix during
its lifetime.
Environmental Significance
All of the experimental results are consistent with the
assumption that eaq- is the transient reactant responsible for
the intrahumic dechlorination of mirex to form photomirex.
The results also eliminate oxidants such as hydroxyl radical
and superoxide as being important to the reaction. However,
they cannot rule out the possibility that some other photo-
generated, humic-associated reductant is actually responsible
for the reaction. Even though the actual reactant has not
been firmly identified, the relative experiments with lindane
Assuming that reaction of mirex in HA solution is mediated
solely by eaq- and is entirely intrahumic, eq 5 can be used to
-
calculate the eaq quantum yield if [eaq-]b is defined as
I300(1 - 10-ꢀ [DOC]l )Φ300
300
-
[eaq-]b )
(8)
and the DCBs indicate that reactivity with eaq combined
kr
with DOM-binding affinity is a potentially useful predictor of
photoreactivity of chlorinated compounds in natural water.
This reaction is potentially very significant for compounds
that are associated with natural organic matter in sunlit water.
This is exemplified by the photomirex to mirex concentration
ratios (P/ M ratios) found in the Lake Ontario system (9). Up
to 65% of the mirex in the Lake Ontario water column is
bound to DOM (4), but >99% of the incident sunlight is
where ꢀ300 is the absorption coefficient for HA at 300 nm [in
units of (kg/ L DOC)-1 cm-1], Φ300 is the quantum yield for
eaq- production at 300 nm, and kr is the apparent rate constant
(s-1) for all processes removing eaq- from the humic matrix.
The apparent rate constant for reaction of mirex was 2.15
× 10-6 s-1 in air-equilibrated HA solution (10 mg/ L DOC) at
9
1 3 7 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 5, 1997

(26) Kohler, G.; Getoff, N.; Rotkiewicz, K.; Grabowski, Z. R. J.
Photochem. 1985, 28, 537.
(27) Kuhn, H. J.; Braslavsky, S. E.; Schmidt, R. Pure Appl. Chem. 1989,
61, 187.
captured within only 15% of the lake volume (59, 60). As a
consequence, P/ M ratios in lake surface sediments are
approximately 0.1 (5, 9). This ratio represents a time-
integrated average for the entire lake water column and
indicates that, on a whole lake basis, intrahumic reaction is
relatively unimportant to the fate of mirex. However, P/ M
ratios peak at ∼1 in Lake Ontario surface waters when the
lake is stratified (9). In the St. Lawrence River, which receives
only Lake Ontario surface waters, the average P/ M ratio in
sediments is 0.35 (9). This ratio is similar to those found in
Lake Ontario biota, which include 0.34-0.42 in herring gull
eggs (61, 62), 0.37 in herring gulls (62), and 0.44-0.48 in fish
(63). P/ M ratios in the Lake Ontario food chain consequently
reflect P/ M ratios in Lake Ontario surface waters. On the
basis of these ratios, the intrahumic reaction of mirex is
significant. If this reaction is in fact mediated by eaq-, then
eaq- acquires renewed environmental significance. Further-
more, because chlorinated aromatic hydrocarbons such as
dichlorobenzenes react with eaq- about 20 times faster than
does mirex, intrahumic eaq- could be even more significant
to the environmental fate of hydrophobic halogenated
aromatic hydrocarbons such as PCBs.
(28) Blough, N. V.; Caron, S. Preprints of Papers, 210th National
Meeting of the American Chemical Society, Chicago, IL; Division
of Environmental Chemistry, American Chemical Society: Wash-
ington, DC, 1995; Paper 68.
(29) Buhler, R. E. In The Chemistry of the Carbon-Halogen Bond;
Patai, S., Ed.; Wiley: New York, 1973; Chapter 12.
(30) Gunning, H. E.; Strausz, O. P. In Advances in Photochemistry;
Noyes, W. A., Jr., Hammond, G. S., Pitts, J. N., Jr., Eds.;
Interscience: New York, 1963; Vol. 1, pp 209-274.
(31) Zepp, R. G.; Hoigne, J.; Bader, H. Environ. Sci. Technol. 1987, 21,
443.
(32) Gordon, S.; Hart, E. J.; Matheson, M. S.; Rabani, J.; Thomas, J.
K. Disc. Faraday Soc. 1963, 36, 193.
(33) Rotenberg, A. A.; Lakomov, V. I.; Pleskov, Y. V. J. Electroanal.
Chem. 1970, 403.
(34) Breugem, P.; van Noort, P.; Velberg, S.; Wondergem, E.; Zijlstra,
J. Chemosphere 1986, 15, 717.
(35) Watkins, A. R. J. Phys. Chem. 1973, 77, 1207.
(36) Draper, W. M.; Crosby, D. G. J. Agric. Food Chem. 1981, 29, 699.
(37) Jota, M. A. T.; Hassett, J. P. Environ. Toxicol. Chem. 1991, 10,
483.
(38) Watkins, A. R. J. Phys. Chem. 1974, 78, 2555.
(39) Shizuka, H.; Nakamura, M.; Morita, T. J. Phys. Chem. 1980, 84,
989.
(40) Berdnikov, V. M.; Bazhin, N. M. Russ. J. Phys. Chem. 1970, 44,
395.
(41) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. J.
Phys. Chem. Ref. Data 1988, 17, 513.
(42) Elving, P. J.; Bennett, C. E. J. Am. Chem. Soc. 1954, 76, 4473.
(43) Fossey, J.; Lefort, D.; Sorba, J. Free Radicals in Organic Chemistry;
Wiley: Chichester, 1995.
(44) Michl, J.; Bonacic-Koutecky, V. Electronic Aspects of Organic
Photochemistry; Wiley-Interscience: New York, 1990.
(45) Wawzonek, S. Anal. Chem. 1952, 24, 35.
(46) Fingerling, G.; Hertkorn, N.; Parlar, H. Environ. Sci. Technol.
1996, 30, 2984.
(47) Kohler, G.; Getoff, N. J. Chem. Soc. Faraday Trans. 1 1980, 76,
1576.
(48) Anbar, M.; Hart, E. J. J. Am. Chem. Soc. 1964, 86, 5633.
(49) Lichtscheidl, J.; Getoff, N. Int. J. Radiat. Phys. Chem. 1976, 8,
661.
Acknowledgments
M.V.R. is grateful to CNPq (Conselho Nacional de Pesquisa,
Brasilia, Brazil) for financial support. This work was partially
funded by the U.S. Air Force Office of Scientific Research
(AFOSR) under Grant F49620-94-1-0296.
Literature Cited
(1) Kaiser, K. L. E. Environ. Sci. Technol. 1978, 12, 520.
(2) Smith, J. H.; Mabey, W. R.; Bahonos, N.; Holt, B. R.; Lee, S. S.;
Chou, T. W.; Venberger, D. C.; Mill, T. Environmental pathways
of selected chemicals in fresh water systems: Part II. Laboratory
Studies; Interagency Energy-Environment Research Program
Report EPA-600/ 7-78-074; U.S. Environmental Protection Agency,
Laboratory Office of Research and Development: Athens, GA,
1978.
(3) Yin, C.; Hassett, J. P. Environ. Sci. Technol. 1986, 20, 1213.
(4) Yin, C.; Hassett, J. P. Chemosphere 1989, 19, 1289.
(5) Oliver, B. G.; Niimi, A. J. Environ. Sci. Technol. 1985, 19, 842.
(6) Mackay, D.; Paterson, S. Environ. Sci. Technol. 1991, 25, 427.
(7) Mudambi, A. R.; Hassett, J. P. Chemosphere 1988, 17, 1133.
(8) Burns, S. E.; Hassett, J. P.; Rossi, M. V. Environ. Sci. Technol.
1996, 30, 2934-2941.
(50) Anbar, M. In Solvated Electron; Hart, E. J., Ed.; ACS Advances in
Chemistry Series 50; American Chemical Society: Washington,
DC, 1965; Chapter 6.
(51) Karickhoff, S. W. Chemosphere 1981, 10, 833.
(52) Burns, S. E. Ph.D. Dissertation, State University of New York
College of Environmental Science and Forestry, 1995.
(53) Carter, C. W.; Suffet, I. H. Environ. Sci. Technol. 1982, 16, 735.
(54) Hansch, C.; Leo, A. Substituent Constants for Correlation Analysis
in Chemistry and Biology; Wiley: New York, 1979.
(55) Briggs, G. G. J. Agric. Food Chem. 1981, 29, 1050.
(56) Bruccoleri, A.; Pant, B. C.; Sharma, D. K.; Langford, C. Environ.
Sci. Technol. 1993, 27, 889.
(9) Mudambi, A. R.; Hassett, J. P.; McDowell, W. H.; Scrudato, R. J.
J. Great Lakes Res. 1992, 18, 405.
(10) Zepp, R. G.; Schlotzhauer, P. F.; Sink, R. M. Environ. Sci. Technol.
1985, 19, 74.
(11) Zepp, R. G.; Braun, A. M.; Hoigne, J.; Leenheer, J. A. Environ. Sci.
Technol. 1987, 21, 485.
(12) Haag, W. R.; Mill, T. In Effects of Solar Ultraviolet Radiation on
Biogeochemical Dynamics in Aquatic Environments; Blough, N.
V., Zepp, R. G., Eds.; Woods Hole Oceanographic Institute
Technical Reptort, WHOI-90-09; Woods Hole Oceanographic
Institute: Woods Hole, MA, 1990; pp 82-88.
(13) Haag, W. R.; Hoigne, J. Environ. Sci. Technol. 1986, 20, 341.
(14) Zepp, R. G.; Ritmiller, L. F. Aquatic Chemistry. In Interfacial and
Interspecies Processes; Huang, C. P., O’Melia, C. R., Morgan, J. J.,
Eds; ACS Advances in Chemistry Series 244; American Chemical
Society: Washington, DC, 1995.
(15) Zhou, X.; Mopper, K. Mar. Chem. 1990, 30, 71.
(16) Cooper, W. J.; Zika, R. G.; Petasne, R. G.; Plane, J. M. C. Environ.
Sci. Technol. 1988, 22, 1156.
(17) Zepp, R. G.; Wolfe, N. L.; Gordon, J. A.; Fincher, R. C. J. Agric.
Food Chem. 1976, 24, 727.
(57) MacCarthy, P.; Malcolm, R. L. In Aquatic Humic Substances;
Suffet, I. H., MacCarthy, P., Eds.; Advances in Chemistry Series
219; American Chemical Society: Washington, DC, 1989; Chapter
4.
(58) Kim, J. S.; Chian, E. S. K.; Saunders, F. M.; Perdue, E. M.; Giabbai,
M. F. In Aquatic Humic Substances; Suffet, I. H., MacCarthy, P.,
Eds.; Advances in Chemistry Series 219; American Chemical
Society: Washington, DC, 1989; Chapter 29.
(59) Beeton, A. M. In Proceedings Fifth Conference on Great Lakes
Research; Publication No. 9; Great Lakes Research Division,
Institute of Science & Technology, University of Michigan: Ann
Arbor, MI, 1962; pp 68-76.
(60) Allen, E. R. Lake Ontario Atlas: Chemistry; New York Sea Grant
Institute: Albany, NY, 1977.
(61) Norstrom, R. J.; Hallett, D. U.; Onuska, F. I.; Comba, M. E. Environ.
Sci. Technol. 1980, 14, 860.
(62) Braune, B. M.; Norstrom, R. J. Environ. Toxicol. Chem. 1989, 8,
957.
(18) Miller, G. C.; Zisook, R.; Zepp, R. J. Agric. Food Chem. 1980, 28,
1053.
(19) Skurlatov, Y. I.; Zepp, R. G.; Baughman, G. L. J. Agric. Food Chem.
1983, 31, 1065.
(63) Oliver, B. G.; Niimi, A. J. Environ. Sci. Technol. 1988, 22, 388.
(20) Blough, N. V. Environ. Sci. Technol. 1987, 22, 77.
(21) Zavitsas, A. A.; Eherenson, S. J. Am. Chem. Soc. 1965, 7, 2841.
(22) Anbar, M.; Hart, E. J. J. Phys. Chem. 1965, 69, 271.
(23) Alley, E. G.; Layton, B. R.; Minyard, J. P., Jr. J. Agric. Food Chem.
1974, 22, 727.
(24) Holmstead, R. L. J. Agric. Food Chem. 1976, 24, 620.
(25) Driscoll, M. S.; Hassett, J. P.; Fish, C. L. Environ. Sci. Technol.
1991, 25, 1432.
Received for review July 3, 1996. Revised manuscript received
January 15, 1997. Accepted January 17, 1997.X
ES960581L
X Abstract published in Advance ACS Abstracts, March 15, 1997.
9
VOL. 31, NO. 5, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 3 7 1