Photodecomposition of the carbamate pesticide carbofuran: Kinetics and the influence of dissolved organic matter

Article abstract of DOI:10.1021/es9802652

This study examined the photodecomposition of carbofuran, a carbamate pesticide with high oral toxicity. Rate constants are measured for the pesticide in aqueous solution and in the presence of various samples of dissolved organic matter (DOM). Kinetic experiments are monitored with HPLC, while reaction products are determined using HPLC, GC-MS, and ¹H NMR: mechanisms are proposed for the first three steps of the reaction. It was found that the photodecomposition proceeds via first-order reaction kinetics and that the presence of various DOM samples inhibits the photolysis reaction of carbofuran. This phenomenon can be correlated to the magnitude of the binding interaction between carbofuran and DOM. Finally, techniques such as GC-MS and ¹H NMR are used to identify the photodecomposition products. The first three steps of the reaction are defined. In the first step of the reaction, the carbamate group is cleaved from the molecule. The furan moiety is opened in the second step producing a substituted catechol with a tert-butyl alcohol group as the substituent at the number three carbon. This molecule then undergoes a dehydration reaction to form an alkene side group from the tert-butyl alcohol side group.

Full text of DOI:10.1021/es9802652

Environ. Sci. Technol. 1999, 33, 874-881
Photochemical reactions are especially important in the
Photodecomposition of the
Carbamate Pesticide Carbofuran:
Kinetics and the Influence of
Dissolved Organic Matter
realm of wastewater treatment as a technique used to remove
harmful chemicals from the waste stream. (5). To effectively
employ photolysis as a water treatment technique, the
reaction of the chemical of interest must be thoroughly
studied including the subsequent reactions of intermediates.
Ollis et al. (5) point out that, since complete removal of
harmful chemicals is the goal in wastewater treatment,
demonstration of the formation and elimination of inter-
mediates is required to show that complete remediation has
been achieved. Hence, both kinetic and mechanistic data
are needed to fully understand the photochemical reaction.
Dissolved organic matter (DOM), ubiquitously present in
soils and surface waters, also plays an important role in regard
to pesticide decomposition. Binding interactions to DOM
have been shown to slow the rate of microbial degradation
of pesticides (6). DOM is the primary light-absorbing species
in surface water and, as a result, plays an important role in
photochemical reactions taking place in surface water.
Additionally, reports have found DOM can either enhance
J O H N B A C H M A N † A N D
H O W A R D H . P A T T E R S O N *
Department of Chemistry, University of Maine,
Orono, Maine 04669
This study examined the photodecomposition of carbofuran,
a carbamate pesticide with high oral toxicity. Rate
constants are measured for the pesticide in aqueous
solution and in the presence of various samples of dissolved
organic matter (DOM). Kinetic experiments are monitored
with HPLC, while reaction products are determined
(
7) or inhibit the rate of photolysis (8). Photochemical
decomposition of pesticides represents an important trans-
formation pathway that can occur in surface waters and in
soil.
Carbofuran is a broad spectrum insecticide with a high
oral toxicity. The oral LD50 for carbofuran is 11 mg/ kg body
weight in rats (9). As a comparison, the LD50 for parathion
1
using HPLC, GC-MS, and H NMR: mechanisms are
proposed for the first three steps of the reaction. It was
found that the photodecomposition proceeds via first-order
reaction kinetics and that the presence of various DOM
samples inhibits the photolysis reaction of carbofuran. This
phenomenon can be correlated to the magnitude of the
binding interaction between carbofuran and DOM. Finally,
(
an extremely toxic organophosphorus pesticide) is 8 mg/ kg
whereas the LD50 for atrazine (a triazine herbicide of low to
moderate toxicity) is 1300 mg/ kg (10). Carbofuran is used in
the cultivation of corn, rice, cotton, and other crops. and its
mode of action is cholinesterase inhibition (11). A compara-
tive examination of the kinetics of the photolysis of the three
carbamate pesticides aldicarb, carbaryl, and carbofuran has
been published (12), and the relative rates (from fastest to
slowest) can be listed as aldicarb, carbaryl, and carbofuran.
However, individual rate constants have not been determined
for the various pesticides. Some work has been performed
on the photochemical degradation of carbofuran. Chiron et
al. have recently published a report on carbofuran degrada-
tion by natural sunlight and a xenon arc lamp (13). Two
decomposition products were identified by LC-MS, 3-hy-
droxy-7-carbofuranphenol as a product of hydrolysis and
2-hydroxy-3-(2-m ethylprop-1-enyl)-phenyl-N-m ethylcar-
bamate as a rearrangement product. Raha and Das (14)
examined the photodecomposition of 1000 ppm solutions
of carbofuran in various solvents. They reported two products
for aqueous solutions of carbofuran that had been irradiated
by natural sunlight for 5 days, 2,3-dihydro-2,2-dimethyl-
benzofuran-4,7-diol and 2,3-dihydro-2,2-dimethyl benzo-
furan-7-yl N-methylcarbamate.
1
techniques such as GC-MS and H NMR are used to
identify the photodecomposition products. The first three
steps of the reaction are defined. In the first step of the
reaction, the carbamate group is cleaved from the
molecule. The furan moiety is opened in the second step
producing a substituted catechol with a tert-butyl alcohol
group as the substituent at the number three carbon.
This molecule then undergoes a dehydration reaction to
form an alkene side group from the tert-butyl alcohol side
group.
Introduction
The use of pesticides is an integral part of world food
production as illustrated by the fact that more than 2.5 million
tons of these anthropogenic chemicals were applied to soil
and foliage in 1996 (1). Recent findings that cite the presence
of pesticides in drinking water supplies illustrate the fact
that some fraction of pesticides applied to a particular
location can be transported off site and into surface waters
This paper examines both kinetic and mechanistic aspects
of the photochemical degradation of carbofuran. In addition,
we examine the role of humic materials (DOM) in regulating
this photolysis reaction using DOM from several different
sources. The goals of this research are (i) to determine the
photodecomposition kinetics of carbofuran, (ii) to determine
the effect of dissolved organic matter on reaction rate, and
(
2). Specifically, the carbamate pesticide carbofuran has been
found in surface water that serves as a drinking water source
for Sacramento, CA (3). Given the potential human and
wildlife health risks associated with toxic pesticides in surface
waters (3, 4), it is important to determine the probability that
a certain chemical will persist in the environment by
examining the various reactions of the molecule. Transfor-
mations such as microbial, chemical, and photochemical
degradation are recognized as the significant processes in
determining the fate and transport of pesticides.
(
iii) to determine the photodecomposition reaction products.
Experimental Section
Materials. Carbofuran (2,3-dihydro-2,2-dimethylbenzofuran-
7
-ol N-methylcarbamate) and 2,3-dihydro-2,2-dimethylben-
*
Corresponding author phone: (207)581-1178; fax: (207)581-1191;
zofuran-7-ol were purchased from ChemService and were
used as received (99% purity). Phosphomolybdic acid solution
was used to visualize TLC experiments. A 20% (w/ v) solution
e-mail: howardp@maine.maine.edu.
†
Present address: Great Lakes Environmental Center, 739 Hastings
St., Traverse City, MI 49686.
8
7 4
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 6, 1999
10.1021/es9802652 CCC: $18.00

1999 Am erican Chem ical Society
Published on Web 01/30/1999

of phosphomolybdic acid was purchased from Aldrich
Chemical Co. It was further diluted to 10% (w/ v), and 5%
quantified in terms of carbon content with a calibrated total
organic carbon analyzer (O.I. Corp. model 700).
(
v/ v) concentrated sulfuric acid was added prior to use. The
Photochem ical Reaction Cell. All irradiations were
performed with an EF-260C UV lamp from Spectronics
Corporation (Westbury, NY). The lamp emits a narrow band
of radiation at 254 nm. The irradiation experiments did not
employ additional filters or monochromators. The light
source contains two 6-W tubes that are 8 cm long. Typical
solvents dichloromethane, ethyl acetate, and hexane were
purchased from EM Science. HPLC-grade methanol and
acetonitrile from EM Science were used for HPLC experi-
ments. All H NMR experiments were conducted in deuter-
ated chloroform purchased from Aldrich Chemical Co. The
1
2
CDCl
3
was supplied with 0.03% tetramethylsilane as an
output is 810 µW/ cm . Samples were irradiated in quartz
tubes (1 mm thick). Tube dimensions were 12.5 mm i.d. and
internal reference. All solvents were used as received.
1
00 mm length. A maximum of seven tubes could be
Collection and Preparation of DOM Sam ples. One soil
sample and one water sample were collected for use in the
photochemical experiments. The organic surface horizon was
collected from a stand of red pine in Orono, ME (November
irradiated at the same time. The distance from the light source
to the middle of the sample tube was 3.8 cm. Irradiation
samples reached a maximum temperature of 28 °C in
approximately 30 min. Solutions were not agitated or mixed
during exposure.
The photochemical reaction cell was characterized using
a potassium ferrioxylate actinometer as described in Photo-
chemistry by Calvert and Pitts (15). The calculated value of
1
995). Soil cores, approximately 8 cm deep and 5 cm in
diameter, were collected from random sites within the stand
of trees. The forest soil, litter, and humus was dark in color
and included fallen pine needles. The subsamples were
combined and stored at 4 °C prior to extraction and filtration.
4
-1
-1
the extinction coefficient was 1.18 × 10 L mol cm , which
An additional sample of organic matter was collected from
Caribou Bog in Orono, ME. This wetland site is described as
a sphagnum bog. To collect water from the bog, the vegetation
was pushed aside and the void was allowed to fill with water.
Cleaned and sterilized 1-L nalgene bottles were submerged
in the water and allowed to fill. The water was dark in color
and turbid. A total of 2 L was collected and stored at 4 °C
prior to filtration.
4
is in good agreement with the published value of 1.10 × 10
-
1
-1
L mol cm . Lamp flux was determined by irradiating the
potassium ferrioxylate reactant for 12 s and determining the
number of Fe2 molecules produced. The average value of
+
1
6
16
the flux is 2.664 × 10 photons/ s, with a range of 2.77 × 10
1
6
to 2.58 × 10 photons/ s.
All solutions for the photolysis experiments were prepared
immediately prior to irradiation. Solutions were made in an
appropriately sized volumetric flask and then immediately
transferred to a nalgene bottle. Long periods of storage in
glass were avoided to prevent loss of carbofuran due to
molecules adhering to the glass walls of the flask. Exactly 4
mL of the solution to be irradiated was pipetted into the
reaction tubes. Typically, six reaction tubes were irradiated
simultaneously. At various times, two reaction tubes were
collected for each data point. The contents of the two tubes
were stored in a small (20 mL) nalgene bottle until HPLC
testing could be started. Again, the solution was stored in
nalgene to avoid loss of carbofuran due to adhesion to glass.
Quantification by HPLC was performed as soon as possible
following the completion of the irradiation experiment.
Product Separation. To identify the decomposition
products, a suitable quantity of each compound had to be
generated and isolated. To maintain consistency to the
kinetics experiments, a series of irradiations were conducted
with 30 ppm aqueous solutions of carbofuran. Following
irradiation, the aqueous solution was saturated with sodium
chloride to improve extraction efficiency. The reaction
products were extracted with five 15-mL portions of ethyl
acetate. Ethyl acetate was removed by rotary evaporation
until the sample volume was approximately 1 mL.
A humic acid and a fulvic acid sample used in the
photochemical experiments were purchased from Aldrich
Chemical Co. and provided by Dr. Christopher Cronan from
the Program in Ecology and Environmental Science, Uni-
versity of Maine, respectively. The Aldrich humic acid sample
was determined to be 40.7% C. The fulvic acid sample was
isolated from a southern hardwood forest soil, and carbon
content of this sample was 47.2% C. Both the humic and
fulvic acids were received as brown solid material that was
stored in a desiccator prior to use in the photolysis experi-
ments.
To separate the organic matter from the pine forest soil
sample, approximately 1.5 kg of soil was placed in a 2-L
nalgene bottle and 1500 mL of distilled, deionized water was
added. The sample was extracted for 24 h at room temper-
ature with periodic shaking. Following the extraction, the
liquid from the pine forest soil extract was darkly colored
and turbid.
The remaining steps of sample processing apply to both
the pine forest soil sample and the water sample collected
from the bog. The soil extract and the water sample were
passed through a series of filters to remove clay particles,
plant debris, and other unwanted particulates. First, the
samples were passed without vacuum through Whatman 41
ashless filter paper. The resulting filtrant was passed through
a Gelman A/ E glass fiber filter with a pore size of 1 µm with
the aid of vacuum. The final filtration was conducted with
a Gelman A/ E glass fiber filter with a pore size of 0.45 µm,
again, using vacuum filtration. At this point the samples were
no longer turbid. The bog sample and the coniferous sample
were both deep amber in color.
The product mixture was analyzed with HPLC and GC-
MS. Both techniques indicated that the mixture contained
seven compounds. These products were separated via
preparative thin-layer chromatography (TLC) using plates
purchased from VWR Scientific (Merck, Item 13894). The
stationary phase was 0.50 mm layer of 60 mesh silica,
containing no fluorescent indicator, and coated onto a 20 ×
2
0 cm glass plate. The mobile phase consisted of 3:1 hexane:
Following filtration, the samples were passed through a
cation exchange column. Columns with inside diameters of
ethyl acetate. As many as three developments of the TLC
plate were required to achieve adequate separation. Products
were visualized with an alcoholic phosphomolybdic acid
solution. The indicator was applied to both vertical edges of
the TLC plate and dried with a heat gun.
Upon visualization, the silica from the appropriate region
of the plate was removed with a razor blade. The compound
was resolvated by placing the silica in 100 mL of ethyl acetate
and stirring for 10 min. The silica was removed with vacuum
filtration. Again, the excess solvent was removed via rotary
evaporation.
4
.5 mm were packed with Rexyn 101H beads. The packing
material had previously been Soxhlet extracted with methanol
to remove residual carbon. The pumping sequence through
the column was as follows: 1 M HCl for 15 min to condition
the packing material, distilled/ deionized (dd) H O for 30 min,
2
DOM sample. Approximately 200 mL of DOM sample was
collected at a time. Following sample collection, the columns
were flushed with water (10-15 min) and the HCl/ H
2
O
conditioning step was repeated. Finally, the samples were
VOL. 33, NO. 6, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8 7 5

FIGURE 1. Rate plots for the irradiation of carbofuran in the presence of various concentrations of DOM.
Instrum entation. The technique of synchronous scan
fluorescence spectroscopy (SSFS) involves scanning both the
excitation and emission monochromators over a specified
range of wavelengths. In typical emission fluorescence
spectroscopy, the excitation monochromator is fixed at a
particular wavelength while the emission monochromator
is scanned. In SSFS, a constant wavelength difference (∆λ)
is maintained between the excitation and emission mono-
chromators as they scan the specified spectral range. SSFS
measurements were collected on a model QM-1 fluorescence
spectrometer from Photon Technologies International (PTI),
Inc. The instrument is equipped with a 75-W xenon arc lamp.
Excitation slit widths were typically set at 1 nm with the
emission slits set at 10 nm. A ∆λ value of 27 nm was used
for all SSFS measurements. Quartz fluorimetry cuvettes with
a path length of 10 mm were used for all fluorescence and
absorption spectroscopy measurements.
The mobile phase used for all experiments was 45.45% water,
45.45% methanol, and 9.10% acetonitrile. The detector was
set to monitor 276 and 216 nm. Authentic samples of
carbofuran and 2,3-dihydro-2,2-dimethylbenzofuran-7-ol
were analyzed to determine separation parameters and
retention times. Calibration standards of the two compounds
were prepared for both injection volumes. Calibration curves
were linear and had excellent correlation coefficients (car-
bofuran, 20 µL injection, r ) 0.9999; carbofuran, 95 µL
injection, r ) 0.9983; 2,3-dihydro-2,2-dimethylbenzofuran-
7-ol, 20 µL injection, r ) 0.9987; and 2,3-dihydro-2,2-
dimethylbenzofuran-7-ol, 95 µL injection, r ) 0.9894).
Proton NMR spectra were collected on a 300 MHz Varian
Gemini XL 300. Peaks were referenced to an internal TMS
standard. Sample concentrations were approximately 5 mg/
mL.
GC-MS measurements were made on a Hewlett-Packard
890 gas chromatograph with Hewlett-Packard MSD 5970
Results and Discussion
5
serving as the detector. Two chromatography columns were
Kinetic Experim ents. Carbofuran was irradiated in aqueous
solution and in the presence of DOM; samples for quanti-
tation of carbofuran were collected over a period of ap-
proximately 2 h. The concentration of the pesticide was
determined by HPLC using a comparison of peak area to a
standard curve. The data were plotted as the natural log of
carbofuran concentration (ln[cf]) versus irradiation time. All
experiments produced linear plots of ln[cf] vs t, indicating
that the photochemical degradation reaction is first order in
carbofuran. The first-order rate constants for 3 and 0.5 ppm
used in the analyses, specifically a 30 m × 0.25 mm i.d. DB-
5
ms column from J&W Scientific and a 30 m × 0.25 mm 50%
phenyl silicone column from Quadrex. Components of the
various samples were separated using the following param-
eters: injector temperature set to 100 °C, the initial oven
temperature of 100 °C was held for 3 min, the temperature
was then ramped to 180 °C at a rate of 25 °C/ min. The oven
was held at 180 °C for 1 min and then ramped to a temperature
of 250 °C at a rate of 10 °C/ min and held at this temperature
for 5 min. The detector temperature was set at 320 °C, and
the injector was set at 100 °C. Helium was used as the carrier
gas with a flow rate of 50 mL/ min. Prior to injection, samples
were filtered through a Gelman 2 mm Acrodisc syringe filter
-
1
carbofuran solutions are 0.76 and 0.69 h , respectively.
Experiments were conducted with 3 ppm initial carbo-
furan concentration and various concentrations of the four
DOM samples. Again, all experiments produced linear first-
order rate plots, allowing for the calculation of the rate
constants. In all cases the presence of DOM slowed the rate
of photolysis. For example, experiments conducted with 3
ppm carbofuran and various concentrations of the bog DOM
sample produced the following rate constants: 0.62, 0.51,
(
0.2 µm pore size) to remove small particles that could
obstruct flow through the column.
HPLC was used to quantify the samples from the
photolysis experiments. Samples were analyzed on a Hewlett-
Packard model 1050 high-performance liquid chromato-
graph. The detector was a UV/ vis diode array detector, model
-
1
0.45, 0.32, and 0.23 h for solutions with 5, 8, 15, 25, and 30
ppm DOM, respectively. Figure 1 displays the rate data from
these experiments. One can see that increasing the DOM
concentration decreases the slope of the rate plots, indicating
a decreasing reaction rate. The data from all the experiments
with other DOM samples are summarized in Tables 1-3.
1
1
040A. Separations were made on a Columbus C-18 column,
50 mm × 4.6 mm, 5µm particle size. Samples were injected
by a Hewlett-Packard model 1050 autosampler. Injection size
was either 20 or 95 µL depending on initial carbofuran
concentration. Flow rate was 1 mL/ min for all experiments.
8
7 6
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 6, 1999

TABLE 1. Rate Constants, 3 ppm Carbofuran with Various DOM
Concentrations
TABLE 3. Rate Constants, 3 ppm Carbofuran with Various
Concentrations of Southern Hardwood Fulvic Acid (shfa)
DOM concn
DOM concn
DOM sample
5 ppm
8 ppm
15 ppm
25 ppm
30 ppm
DOM sample
5 ppm
7.2 ppm
15 ppm
-
1
1
-1
-1
-1
-1
0.48 h-1
0.42 h-1
0.26 h-1
bog
pine
0.62 h
0.58 h
0.51 h
0.51 h
0.45 h
0.40 h
0.32 h
0.25 h
0.23 h
0.24 h
shfa
-
-1
-1
-1
-1
was used in the fluorescence quenching experiments. As the
amount of free carbofuran in solution is decreased due to
binding to DOM, the fluorescence intensity is also decreased.
The magnitude of fluorescence intensity reduction is de-
termined by the DOM concentration and the binding
constant. The Stern-Volmer equation defines the fluores-
cence quenching/ binding relationship:
TABLE 2. Rate Constants, 3 ppm Carbofuran with Various
Concentrations of Aldrich Humic Acid (aha)
DOM concn
DOM sample
3.4 ppm
10.2 ppm
20.4 ppm
1
0.27 h-1
0.16 h-1
aha
0.47 h-
F/ F ) 1 + k [DOM]
(3)
0
b
The rate constant data clearly show that the presence of
DOM slows the rate of carbofuran photolysis. The magnitude
of rate inhibition is different for each of the four DOM
samples. For a given DOM concentration, the bog DOM
sample produces the least amount of rate reduction followed
by the pine sample, the southern hardwood fulvic acid, and,
finally, the Aldrich humic acid sample. The rate effects are
not linearly related to DOM concentration.
It was postulated that the decrease in the rate of
photodegradation could be due either to DOM competing
with carbofuran for the available photons or to binding
between DOM and carbofuran. Competing photochemical
reactions were initially investigated as the cause of the
measured rate effect. It is clear from the spectroscopic data
that the DOM does react in the presence of light (Figure 2).
Additionally, it has been reported elsewhere that DOM can
react photochemically (16). However, the amount of light
needed for the carbofuran reaction is small relative to the
total flux from the lamp. For example, at the highest DOM
concentration used (30 ppm), it can be determined through
photometry experiments that DOM is absorbing 58.3% of
the total flux while a 3 ppm solution of carbofuran absorbs
where F is the fluorescence intensity of carbofuran/ DOM
solution, F is the fluorescence intensity of carbofuran, k is
the binding constant, and [DOM] is the DOM concentration.
Fluorescence data were collected on carbofuran solutions
with various DOM concentrations such that binding con-
0
b
stants could be determined. Specifically, the Caribou Bog
4
DOM sample had a binding constant (k
b
) of 1.75 × 10 L/ kg.
The binding constants for the red pine, southern hardwood
4
4
fulvic acid, and Aldrich humic acid were 2.2 × 10 , 4.10 × 10 ,
4
and 4.26 × 10 L/ kg, respectively. It can be noted that the
ranking of DOM samples from least to most, in terms of rate
effects (bog, pine, southern hardwood fulvic acid, Aldrich
humic acid), is the same as the ranking of binding constants
(bog, pine, southern hardwood fulvic acid, Aldrich humic
acid).
To further illustrate that the photolysis rate effects are a
result of binding interactions, one can examine the rate
constant data as related to the percent of bound carbofuran
in solution. The percentage of bound pesticide is calculated
using the binding constant and the DOM concentration:
1
8.7% of the total flux. Thus, in the presence of DOM there
%
bound ) (k [DOM])/ (1 + k [DOM]) × 100 (4)
b
b
remains a sufficient amount of photons in the system for the
carbofuran photolysis. Given the magnitude of the lamp flux
and the concentrations of the experimental solutions, it is
not possible to account for the total of rate decrease with a
competing photochemical pathways model. One of the
shortcomings of this model is that it can only account for the
absorption due to carbofuran and DOM but ignores the third
species present, the bound form of carbofuran.
Alternatively, a second model to describe our photo-
chemical results has been considered in which binding
interactions partition a significant quantity of the pesticide
molecules into a bound form. When carbofuran and DOM
are in solution together, a certain fraction of the pesticide
is bound to the DOM. It is believed that binding occurs
predominantly through a hydrophobic partitioning mech-
anism, with polar interactions playing a much smaller role
For example, at a concentration of 15 ppm, the bog DOM
sample binds 21.3% of the total carbofuran concentration,
while the Aldrich humic acid (15 ppm) binds 39.0% of the
total carbofuran in solution. The rate constants for the same
two DOM samples at 15 ppm are 0.45 and 0.27 h for the
bog and Aldrich humic acid, respectively. It can be noted
that an approximately 2-fold increase in the amount of
carbofuran bound in solution (20.3% vs 39.0%) is reflected
in an approximately 2-fold decrease in the rate constant (0.45
-
1
-
1
and 0.27 h ).
To test if all of the experimental data fit this model, a plot
of rate constant versus percent of carbofuran bound to DOM
was prepared (Figure 3). When the data are plotted in this
fashion, the four DOM samples produce essentially equivalent
straight lines. Therefore, the photolysis rate constant is
inversely proportional to the amount of bound pesticide.
Although four different DOM samples were used in these
experiments, the amount of rate reduction can be related
(
17, 18). Representing the bound carbofuran as cf-DOM, the
following equilibrium can be written:
cf + DOM T cf-DOM
(1)
(2)
(
inversely) to the binding ability of the DOM, regardless of
kb ) [cf-DOM]/ [cf][DOM]
which particular DOM sample is considered. Also, the data
in Figure 3 indicate that a single mechanism is responsible
for the binding in all four DOM samples. As mentioned
previously, it is believed that the binding mechanism is
hydrophobic partitioning. Therefore, we conclude that all of
the experimental data can be described by a model in which
the magnitude of the DOM-pesticide binding is inversely
proportional to the photolysis rate reduction.
b
The binding constant, k , can be determined by a
fluorescence quenching method. Carbofuran lends itself to
this method since it has a strong fluorescence peak especially
when measured by SSFS. It was determined experimentally
that carbofuran has a maximum fluorescence intensity when
spectra are collected with a ∆λ of 27 nm; thus, this parameter
VOL. 33, NO. 6, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8 7 7

FIGURE 2. Synchronous scan fluorescence spectra (∆λ ) 27 nm, optical path length ) 10 mm) of southern hardwood forest DOM, initial
and irradiated for 2 h.
FIGURE 3. Rate constants for the photolysis of carbofuran/DOM solutions versus the percent of bound carbofuran in the corresponding
solution.
It is clear that the magnitude of the DOM-carbofuran
binding controls the magnitude of the photolysis rate effects.
However, a physical picture is needed to fully understand
the rate effects of DOM. The hydrophobic binding mechanism
draws the carbofuran into an aggregate of humic molecules.
This arrangement allows carbofuran to transfer the excess
energy resulting from the absorption of a photon to the
surrounding DOM molecules. Examination of the absorption
spectrum of a mixture of carbofuran and DOM shows that
carbofuran is still able to absorb photons despite being bound
to DOM. Absorption spectroscopy (Figure 4) shows that the
absorbance at 276 nm of a 3 ppm carbofuran solution is
rates indicate that the energy absorbed by the carbofuran in
the mixture is released via an alternative process such as
energy transfer to DOM.
Identification of Interm ediates and Products. The first
step of the photochemical reaction of carbofuran in water
is similar to its base hydrolysis reaction. Namely, the
carbamate group is cleaved from the molecule to form 2,3-
dihydro-2,2-dimethylbenzofuran-7-ol and carbamic acid.
This step was confirmed by multiple experimental tech-
niques. Peaks in both the GC-MS and HPLC chromatograms
matched the respective retention times for an authentic
sample of 2,3-dihydro-2,2-dimethylbenzofuran-7-ol. The fact
that the retention times from the two chomatographic
techniques match the retention times for the standard
compounds is strong evidence that the identity of the
decomposition product is 2,3-dihydro-2,2-dimethylbenzo-
0
.392, a 30 ppm bog DOM solution is 0.358, and a mixture
of 3 ppm carbofuran and 30 ppm bog DOM is 0.760 or roughly
the sum of the two individual absorbances. However,
fluorescence quenching and reduced photochemical reaction
8
7 8
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 6, 1999

FIGURE 4. Absorption spectra of carbofuran, bog DOM, and carbofuran/DOM mixture (optical path length ) 10 mm).
SCHEME 1. Photodegradation of carbofuran, steps 1 and 2.
both of which are gases at room temperature. The formation
of the two gaseous products is a driving force for this reaction.
Analysis of the product mixture by GC-MS and HPLC
shows a total of seven peaks including the parent molecule
and 2,3-dihydro-2,2-dimethylbenzofuran-7-ol (Figure 5).
Having analyzed the mixture with GC-MS, the molecular
weights of the reaction products were obtained using GC-
MS and are summarized in Table 4.
From these data, we identified the compounds at retention
times of 10.5 and 6.6 min as carbofuran and 2,3-dihydro-
2
,2-dimethylbenzofuran-7-ol, respectively. These assign-
ments are based on a comparison of retention times and
mass spectra to standard molecules. The compounds ap-
pearing at the other retention times required further char-
acterization before a confident identification could be made.
The compounds were isolated using preparative TLC, and
their structure was determined by comparing 1H NMR
analysis with the mass spectral data.
1
The H NMR spectra for the compound at retention time
8
.9 min (GC-MS) contains a multiplet at δ ) 6.8 ppm, a
singlet at δ ) 2.8 ppm, and a singlet at δ ) 1.3 ppm. The
integration of the peaks are approximately 3, 2, and 6,
respectively. The multiplet at 6.8 ppm is assigned to three
aromatic protons, the singlet at 2.8 ppm indicates two
methylene protons, and the singlet at 1.3 ppm indicates six
methyl protons.
furan-7-ol. It is possible that a different molecule could elute
at the same retention time. However, the mass spectra of the
product found in the irradiated solution is identical to the
mass spectra for the standard, confirming the identification.
The mechanism of the first three steps of the photodeg-
radation of carbofuran is shown in Scheme 1. To produce
The major ions in the mass spectra are as follows; m/z
+
)182 (M
), 167 (M - CH ), 164, M - H O) and, 149 (M -
3
2
2CH ). Other significant ions are m/z ) 91 corresponding to
the benzyl ion and m/z ) 77 corresponding to C H . From
5
3
+
6
the available data, a chemical structure of the compound
was determined as shown:
2
,3-dihydro-2,2-dimethylbenzofuran-7-ol from carbofuran,
the C-O bond of the carbamate group must be broken. The
lamp used in the irradiation experiments produced a narrow
line at 254 nm that corresponds to an energy of 470.9 kJ/ mol.
Carbon-oxygen single bonds require between 334.4 and
4
18.0 kJ/ mol to cleave the bond; thus, sufficient energy is
available in this system to break the bond. Cleavage of the
C-O bond forms a phenolic anion and a cation on a basic
nitrogen in the cleaved carbamate group (19). The phenolic
anion can abstract a proton from the solvent (water) to form
the intermediate, 2,3-dihydro-2,2-dimethylbenzofuran-7-ol.
The cation can abstract an OH-, also from the water, to form
carbamic acid, which is known to be unstable (20). Carbamic
acid rapidly degrades to methylamine and carbon dioxide,
Having identified the reaction product, a reasonable
mechanism must be outlined for the second step of the
photolysis. As illustrated in the reaction scheme, the starting
point of this reaction is 2,3-dihydro-2,2-dimethylbenzofuran-
7-ol in which the C-O bond of the ether linkage in the furan
moiety is broken. Opening of the furan group creates a
negative charge that can be distributed over the aromatic
VOL. 33, NO. 6, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8 7 9

FIGURE 5. GC (A) and HPLC (B) chromatograms of the photolysis reaction mixture following 1 h of irradiation.
collected was quite small. However, aromatic protons can
TABLE 4. GC-MS Data for Reaction Mixture
be deciphered at δ ) 6.6-6.9 ppm. Also, three singlets appear
at the following δ values: 1.25, 1.75, and 3.35 ppm. Also,
there is a peak at 5.35 ppm, indicative of alkene protons. Due
to the low resolution, the spectrum was not integrated. Using
the available chromatographic and spectral data the following
structure is assigned to the product of the third step of the
photolysis of carbofuran. This compound results from the
dehydration of the product of step two of the photolysis:
retention
time
mass of predominant ions
in fragmentation pattern
mol wt
1
0.5
221
182
164
166
164
148
126
221, 164, 149, 131, 91, 77
182, 167, 164, 149, 91, 77
164, 149, 131, 91, 77
148, 133, 108, 91, 77
164, 149, 131, 91, 77
148, 133, 105, 91, 77
126, 97, 55
8
7
7
6
4
3
.9
.7
.4
.6
.9
.7
ring and a positive charge that can be placed on the carbon
at position 2 to form a tertiary carbocation. A water molecule
can then add to the carbocation, with a charged species
remaining at that position. The negative charge can then be
Chromatographic data indicate the presence of three
additional products; however, the evidence is insufficient to
propose structures for the remaining compounds. The
mechanism results can be summarized as follows. Carbofuran
decomposes via direct photolysis. In the first step of the
reaction, the carbamate group is cleaved forming carbamic
acid and 2,3-dihydro-2,2-dimethylbenzofuran-7-ol. Carbamic
acid readily decomposes to carbon dioxide and methylamine.
The phenolic product formed in the first step is also
photochemically reactive. This molecule decomposes upon
irradiation into a substituted catechol. The ether linkage of
the furan group is cleaved forming a tertiary alcohol side
chain attached to catechol. This molecule also reacts photo-
chemically, converting the alcoholic side chain to an alkene
via a dehydration reaction.
released from the ring and deprotonate the H
tertiary alcohol.
2
O to form a
The product appearing at r ) 7.7 min in the GC-MS
t
chromatogram was also isolated and identified. The major
+
ions in the mass spectra include m/ z ) 164 (M ), 149 (M -
+ +
2 6 5 2 6 5
), 131 (149 - H O), 91 (C H CH ), and 77 (C H ). This
CH
3
sample was derivatized with bis-trimethylsilyl acetamide (in
excess), which converts a free alcohol to a trimethylsilyl (TMS)
group. Two compounds are produced in the derivatization.
The molecular weights of the derivatives are 236 and 308, as
determined by GC-MS analysis. The addition of one TMS
group adds 72 mass units to the molecular weight. Since 164
+
72 ) 236 and 164 + (2 × 72) ) 308, the two derivatives
formed correspond to the addition of one and two TMS units,
respectively.
The present work demonstrates that the pesticide carbo-
furan degrades photochemically via a first-order reaction
initially to 2,3-dihydro-2,2-dimethylbenzofuran-7-ol, which
subsequently reacts to form additional products. Two of these
The NMR spectrum of the compound that elutes at 7.7
min was not well resolved since the amount of product
8
8 0
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 6, 1999

additional products have been identified. Also shown in this
work is the fact that the presence of DOM slows the rate of
photolysis of carbofuran. The magnitude of rate inhibition
is directly proportional to the binding capacity of a particular
DOM sample. Biological decomposition is also inhibited in
the presence of DOM (4). Thus, DOM slows two important
transformation pathways of carbofuran, thereby increasing
its persistence.
(10) Hayes, W. J., Jr., Laws, E. R., Jr., Eds. Handbook of Pesticide
Toxicology; Academic Press: New York, 1991; p 1042.
(
(
(
11) Briggs, S. A. Basic Guide to Pesticides; Their Characteristics and
Hazards; Hemisphere Publishing: Washington, DC, 1994; p 201.
12) de Bertrand, N.; Barcelo, D. Anal. Chim. Acta 1991, 254, 235-
2
44.
13) Chiron, S.; Torres, J. A.; Fernandez-Alba, A.; Alpendurada, M.
F.; Barcelo, D. Int. J. Anal. Chem. 1996, 65, 37-52.
(14) Raha, P.; Das, A. K. Chemosphere 1990, 21, 99-106.
(
(
(
15) Calvert, J.; Pitts, D., Jr. Photochemistry; John Wiley and Sons:
New York, 1986; pp 781-789.
16) Bruccoleri, A.; Pant, B. C.; Sharma, D. K.; Langford, C. H. Environ.
Sci. Technol. 1993, 27, 889-894.
17) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M.
Environmental Organic Chemistry; John Wiley and Sons: New
York, 1993: pp 255-341.
Literature Cited
(
1) Brown, L. R.; Flavin, C.; Kane, H. Vital Signs-The Trends That
Are Shaping Our Future.;W. W. Norton: New York, 1996.
2) Koplin, D. W.; Thurman, E. M.; Goolsby, D. A. Environ. Sci.
Technol. 1996, 30, 335-340.
(
(
(
3) Johnson, W. G.; Lavy, T. L. J. Environ. Qual. 1995, 24, 487-493.
4) Nicosia, S.; Carr, N.; Gonzales, D. A.; Orr, M. K. J. Environ. Qual.
(
(
(
18) Xing, B.; Pignatello, J. J.; Gigliotti, B. Environ. Sci. Technol. 1996,
3
0, 2432-2440.
1
991, 20, 532-539.
5) Ollis, D. F.; Pelizzetti, E.; Serpone, N. Environ. Sci. Technol. 1991,
5, 1523-1529.
6) Getzin, L. W. Environ. Entomol. 1973, 2, 461-467.
7) Canonica, S.; Jans, U.; Stemmler, K.; Hoigne, J. Environ. Sci.
Technol. 1995, 29, 1822-1831.
19) Morrison, R. T.; Boyd, R. N. Organic Chemistry, 5th ed.; Allyn
and Bacon: Boston, 1987; pp 202-206, 1041.
20) McMurray, J. Organic Chemistry, 3rd ed.; Brooks/ Cole Publish-
ing: Belmont, CA, 1992; p 1204.
(
2
(
(
(
(
8) Torrents, A.; Anderson, B. G.; Bilboulian, S.; Johnson, W. E.;
Hapeman, C. J. Environ. Sci. Technol. 1997, 31, 1476-1482.
9) Kuhr, R. J.; Dorough H. W. Carbamate Insecticides: Chemistry,
Biochemistry, and Toxicology; CRC Press: Cleveland, 1976; p 8.
Received for review March 16, 1998. Revised manuscript
received November 24, 1998. Accepted December 24, 1998.
ES9802652
VOL. 33, NO. 6, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8 8 1