The role of carbonate radical in limiting the persistence of sulfur-containing chemicals in sunlit natural waters

Article abstract of DOI:10.1016/S0045-6535(00)00042-4

Carbonate radical (·CO₃/^-) is a selective oxidant that may be important in limiting the persistence of a number of sulfur-containing compounds in sunlit natural waters. Thioanisole, dibenzothiophene (DBT), and fenthion were selected

Full text of DOI:10.1016/S0045-6535(00)00042-4

Chemosphere 41 (2000) 1775±1782
The role of carbonate radical in limiting the persistence of
sulfur-containing chemicals in sunlit natural waters
Jiping Huang, Scott A. Mabury *
Department of Chemistry, University of Toronto, Toronto, Ont., Canada M5S 3H6
Received 5 November 1999; accepted 31 January 2000
Abstract
�
Carbonate radical (áCO ) is a selective oxidant that may be important in limiting the persistence of a number of
sulfur-containing compounds in sunlit natural waters. Thioanisole, dibenzothiophene (DBT), and fenthion were
3
�
selected to investigate the degradation pathway initiated by áCO ; electron-rich sulfur compounds are particularly
3
�
reactive towards the áCO . Using HPLC, GC, GC±MS and LC±MS for structural con®rmation, the major
photodegradation products of thioanisole and DBT were the corresponding sulfoxides. The sulfoxide products were
3
�
further oxidized through reaction with áCO3 to the corresponding sulfone derivatives. Fenthion showed a similar
pathway with appearance of fenthion sulfoxide as the major product. The proposed mechanism involves abstraction of
an electron on sulfur to form a radical cation, which is then oxidized by dissolved oxygen. Each of the sulfur probes
were further investigated in a sunlight simulator under varying matrix conditions. The highest rate constants occurred
�
in the áCO3 matrix, and the lowest occurred in a matrix of dissolved organic carbon (DOC) and bicarbonate. In
synthetic and natural ®eld water, thioanisole photodegraded faster than under direct photolysis, with half-lives of 75.1
and 85.8 min, respectively. Fenthion photodegraded more rapidly than thioanisole. DBT photodegraded rapidly in a
�
áCO matrix with a half-life of 24.8 min, while the half-life of direct photolysis was 350 min. Photodegradation products
3
�
of each compound were also investigated. Ultimately, áCO3 was found to contribute toward the photodegradation of
sulfur-containing compounds in natural waters. Ó 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Photolysis; Thioanisole; Fenthion; Dibenzothiophene
1
. Introduction
lysis in sunlight of a speci®c chemical require informa-
tion regarding the absorption spectrum of the chemical,
the quantum yield of the process (/, the eciency), and
the light intensity as a function of wavelength, latitude,
season, and time of day (Zepp and Cline, 1977). Indirect
photolysis involving photochemically produced reactive
chemical transients is also an important fate for these
The persistence of chemical pollutants in natural
waters is determined by the ability of such waters to
cleanse themselves of xenobiotics. Sunlight-induced di-
rect and indirect photoreactions provide an important
sink for chemical pollutants in such environments; direct
photolysis contributes only partially to these sunlight-
induced reactions. The rate constants for direct photo-
pollutants (Zepp, 1991). Such transients include
�
alkylperoxy, hydroxyl, and carbonate radicals (áCO ),
3
singlet oxygen, aqueous electrons, and triplet states. The
nature and steady-state concentration of these transients
strongly depend upon the composition of the water,
particularly on the concentrations of dissolved organic
matter, nitrate, nitrite, and trace metals. The rate of in-
direct photolysis can be predicted from the second-order
*
Corresponding author. Tel.: +1-416-978-1780; fax: +1-416-
9
78-3596.
E-mail address: smabury@alchemy.chem.utoronto.ca
S.A. Mabury).
(
0045-6535/00/$ - see front matter Ó 2000 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 4 5 - 6 5 3 5 ( 0 0 ) 0 0 0 4 2 - 4

1
776
J. Huang, S.A. Mabury / Chemosphere 41 (2000) 1775±1782
rate constant for its reaction with the transient and the
steady-state concentration of the transient (Zepp, 1991).
under various environmental conditions have previously
been investigated (Lartigues and Garrigues, 1995).
Fenthion degraded much faster under sunlight condi-
tions than in the dark, with the following half-lives at
22°C: 42 d in darkness and 2 d under sunlight for river
water at pH 7.3; 26 d in darkness and 5 d under sunlight
for sea water at pH 8.1. In estuarine waters, the half-
lives of fenthion and its major transformation product,
fenthion sulfoxide were reported to be 4.6 and 6.9 d,
respectively (Lacorte et al., 1995). DBT is found in fossil
fuels, the fuel value of which is partly lowered due to the
high organic sulfur content which, upon combustion,
can release sulfur dioxide into the atmosphere, causing
acid rain. A desulfurization process for DBT through a
combination of photochemical reaction and liquid±
liquid extraction has been investigated (Hirai et al., 1997).
During the desulfurization of DBT by Rhodococcus sp.
strain IGTS8, the formation of key metabolites, in-
cluding dibenz[c,e][1,2] oxathiin 6-oxide (sultine) and
dibenz[c,e][1,2] oxathiin 6,6-dioxide (sultone) were in-
vestigated through extensive GC/FTIR/MS analysis
(Olson et al., 1993). The metabolites from the biode-
sulfurization of DBT in bitumen were also reported by
using SPME/GC±MS (Macpherson et al., 1998).
�
áCO3 is a sunlight-generated secondary radical pro-
duced from the reaction of hydroxyl radical áꢀOH† with
either carbonate or bicarbonate ions. In sunlit natural
waters, the áOH is formed primarily through the pho-
tolysis of nitrate (Hagg and Hoign ꢀe , 1985; Zepp et al.,
1
tolysis of H (Larson and Zepp, 1988), nitrite (Za®-
987); hydroxyl can also be formed through the pho-
O
2
2
riou, 1974), and dissolved organic carbon (DOC) (Mill
et al., 1980) although these are generally considered
�
minor pathways. áCO3 is a more selective oxidant than
áOH which leads to relatively higher steady-state con-
�
centrations (Zepp, 1991); the concentration of áCO will
3
be a function of hydroxyl production rates, and the
relative proportion of scavenging by DOC or carbonate/
bicarbonate.
Larson and Zepp (1988) investigated the reaction
�
pathway of áCO3 with a number of aromatic amine
compounds and concluded that, particularly for elec-
�
tron-rich systems, áCO3 reacts rapidly by electron
�
transfer. The second-order rate constant of áCO3 was
high for indole and its derivatives such as tryptophan
8
� 1 � 1
(
k > 10 M
s
) (Chen and Ho€man, 1974). As a se-
�
�
lective oxidant, áCO3 may play an important role in
oxidizing sulfur-containing compounds with a rate
Our hypothesis was that áCO3 would contribute to
the natural cleansing of these model pollutants from
sunlit natural waters. Our objectives were to identify the
major photodegradation products of these selected ar-
omatic sulfur-containing compounds and measure their
reactivities under varying natural and synthetic ®eld
water (SFW) conditions to reveal the relative impor-
6
7
� 1 � 1
constant at 10 to 10
1
M
973), thus contributing to their transformation. Ulti-
s (Chen and Ho€man,
�
mately, áCO3 may also a€ect the cycling of naturally
occurring sulfur compounds in aqueous systems.
Therefore, it is important both to measure the reactivity
�
�
and to understand the pathways of a áCO reaction with
sulfur-containing compounds.
tance of áCO3 to limiting their persistence.
3
The compounds selected for this investigation were
thioanisole, fenthion, and dibenzothiophene (DBT)
2. Materials and methods
(
Fig. 1). Thioanisole was selected to identify the reaction
�
pathways of áCO3 towards an aromatic sulfur com-
pound with structural similarities to an organophosph-
orous insecticide, fenthion (Cabras et al., 1991). Once
used extensively as an e€ective insecticide, fenthion was
classi®ed by the US EPA as a restricted use pesticide
2.1. Chemicals
Thioanisole, methyl phenyl sulfoxide, DBT and dib-
enzothiophene sulfone were obtained from Aldrich
(Oakville, ON), as well as sodium nitrate, sodium car-
bonate and sodium bicarbonate. Fenthion was obtained
from Chem Service (West Chester, PA). The interme-
diate photodegradation products such as fenthion sulf-
one were synthesized from slightly modi®ed published
methods (Cabras et al., 1991). An aqueous 0.15 M so-
(
RUP) on the basis of its toxicity (Kamrin, 1997). Five
metabolites of fenthion have been isolated in animals
and plants (Cabras et al., 1993). The fenthion sulfoxide
and fenthion sulfone were observed as fenthion photo-
lysis products on fruit surfaces (Minelli et al., 1996).
Degradation kinetics of fenthion in di€erent waters
lution of KMnO was added to a solution of fenthion in
4
Fig. 1. Chemical structures of selected compounds.

J. Huang, S.A. Mabury / Chemosphere 41 (2000) 1775±1782
1777
acetic acid. The reaction mixture was then agitated for
2 h, diluted with water and extracted with chloroform.
pressor. The column was an IonPac AS 14 4 mm and
IonPac AG 14 4 mm (Dionex). The gradient was initi-
ated and held for 6 min at 4% 50 mM borate and 96%
deionized water followed by a linear increase to 30:70
(borate:deionized water) over 18 min; the instrument
was allowed to re-equilibrate for 5 min at initial condi-
tions.
1
Using silica gel column chromatography with 60% ethyl
acetate and 40% hexane as the mix solvent, the crude
fenthion sulfone residue was puri®ed to yield a white
crystalline solid; methyl phenyl sulfone was synthesized
through similar procedures. Fenthion sulfoxide and
DBT-sulfoxide were synthesized, respectively, from
fenthion and DBT in methanol by reaction with
2.4. Analysis
equimolar
2 2 2
H O /SeO . The reaction mixture was
cleaned up using silica gel column chromatography.
DOC was obtained by photobleaching Aldrich humic
acid in direct sunlight for 2 weeks. The mixture was
diluted with HPLC grade water to a ®nal concentration
of 50 mg carbon/l as determined by a DOC analyzer (OI
Corporation Model 1010 TOC analyzer); ®nal absor-
bance of the solutions was 0.43 in a 1 cm path length at
Photodegradation products were extracted (ꢁ50 ml)
and concentrated by solid-phase extraction (500 mg C18
SPE, Supelco), and spike recovery of each analyte was
above 80%. The concentrated samples were then ana-
lyzed through reverse phase HPLC (Waters 600S) with a
996 PDA detector (C18 column; 250 mm X ID 4.6 mm)
and GC using a Perkin-Elmer Autosystem XL GC with
an FPD detector analysis. The GC column was equip-
ped with Supelco MDN-5 column (0:25 mm  30 m;
370 nm. The stock solution was kept in the refrigerator
with no headspace in the container.
0
.25 lm ®lm thickness) (split ratio: 5:1) and injector
2
.2. Photolysis
temperature of 250°C; initial temperature was 50°C for 1
min, followed by a 10°C/min ramp to 200°C, with a ®nal
1 min at 200°C. Hydrogen was used as carrier gas with a
¯ow rate of 5 ml/min. The ¯ow rates of air and hydrogen
were 90 and 75 ml/min, respectively. Structural con®r-
mation was obtained by GC±MS (TurboMass) or LC±
MS (PE Sciex API 3) in comparison with authentic
standards. Kinetic analysis of each probe during pho-
tolysis was obtained through direct injection onto the
HPLC-PDA under isocratic conditions with acetonitri-
le±water (varying ratios) at a ¯ow rate of 1.0 ml/min; all
injections were conducted in triplicate.
Irradiation of the sample solutions in quartz test
tubes was carried out in a sunlight simulator (Atlas
�
Suntest CPS). áCO3 was generated from the photolysis
of 0.1 M Na
1
2 3 2 2
CO and 3 mM H O (Larson and Zepp,
988) in order to determine reaction pathways. SFWs
with varying concentrations of nitrate, DOC, and bi-
carbonate were then run at ®xed positions in the pho-
toreactor (Table 1); the sunlight simulator was set at a
2
maximum irradiance level of 765 W/m . Dark controls
were run excluding light, while direct photolysis was run
in distilled/deionized HPLC grade water; all experiments
were run in duplicate.
3
. Results and discussions
2
.3. Field water analysis
3
.1. Reaction pathways
A Lake Huron water sample served as our reference
eld water. DOC was determined as previously de-
®
Fenthion and thioanisole both react rapidly with
�
7
scribed while nitrate and bicarbonate ions were analyzed
via ion-exchange chromatography with a Perkin-Elmer
series 200 IC pump and an Alltech 550 conductivity
detector, equipped with an ERIS 1000 HP Autosup-
áCO with a second-order rate constant of 2:0 Â 10 and
3
7
� 1
� 1
s , respectively, which we previously
2:3 Â 10
M
measured using a new competition kinetic system
(Huang and Mabury, 2000). Methyl phenyl sulfoxide was
Table 1
Synthetic ®eld waters and photolysis solutions
Matrix
pH
6.50
Components
DI water
Pure DI water
�
High áCO3
11.05
8.30
8.30
8.28
6.75
8.10
8.30
0.04 M Na
2
CO
3
+ 1.5 mM H
500 ppm HCO + 50 ppm NO
2
O
2
�
�
Biocarbonate radical
Synthetic ®eld water
DOC/bicarbonate
DOC/nitrate
3
3
�
�
5 mg/l DOC + 500 ppm HCO + 50 ppm NO
5 mg/l DOC + 500 ppm HCO
3
3
�
3
�
5 mg/l DOC + 50 ppm NO3
2.6mg/l DOC + 105 ppm HCO + 6.2 ppm NO
�
�
Filtered Lake Huron water
Dark control matrix
3
3
�
�
500 ppm HCO + 50 ppm NO
3
3

1
778
J. Huang, S.A. Mabury / Chemosphere 41 (2000) 1775±1782
�
identi®ed by reverse-phase HPLC-PDA as the major
�
at m/z 79.6 of SO3 were observed, which were the same
as observed for a standard injection.
product after the irradiation of thioanisole with áCO3
�
(
Fig. 2); the UV spectrum and retention time of the
Fenthion irradiated with áCO3 for 15 min resulted in
unknown peak matched those for an authentic standard
of methyl phenyl sulfoxide. The reacted sample was
extracted and concentrated through C-18 solid phase
extraction (SPE) cartridge and was subsequently run on
a GC-FPD detector to show clearly the presence of the
sulfoxide product; ®nal structural con®rmation was
many observed peaks in the HPLC chromatogram.
Fenthion sulfoxide was the major product, which was
similar to previous reports of photolysis in environ-
mental matrices (Cabras et al., 1991; Lacorte et al.,
�
1995). Further reaction with áCO3 yielded fenthion
sulfone; the presence of both compounds was con®rmed
‡
provided by GC±MS with m/z 140 (M , 100), 125 (95),
9
using UV and GC±MS spectra. Elucidation of the re-
�
7 (40) and 77 (53). Methyl phenyl sulfoxide was run
action pathways of fenthion with áCO3 were compli-
under the same conditions and yielded slow conversion
to the sulfone (Fig. 2), which was con®rmed by HPLC-
PDA, GC±FPD and GC±MS. Direct injection of thio-
anisole photolysate into an LC±MS (negative ion mode)
tentatively identi®ed benzenesulfonic acid as a minor
product. A molecular ion peak at m/z 157 and fragment
cated, due to the likely hydrolysis of the photo-oxidized
products. In a previous work we showed each of the
metabolites of fenthion to hydrolytically degrade faster
than the parent compound (Huang and Mabury, 1999).
�
Similarly, DBT irradiated with áCO3 for 15 min
yielded DBT-sulfoxide, identi®ed through HPLC-PDA
�
Fig. 2. Thioanisole photoreaction with áCO3 for 1 h; the sample was concentrated through SPE and separated by HPLC (60%
�
acetonitrile and 40% water as mobile phase; ¯ow rate ˆ 1.0 ml/min) (A). Methyl phenyl sulfoxide photoreaction with áCO3 for 1.2 h;
the sample was concentrated through SPE and separated through HPLC (40% acetonitrile and 60% water as mobile phase; ¯ow
rate ˆ 1.0 ml/min) (B).

J. Huang, S.A. Mabury / Chemosphere 41 (2000) 1775±1782
1779
�
and structurally con®rmed through GC±MS, as the
major product; DBT-sulfone was con®rmed as the
radical. The second step involves O transfer from a
�
second áCO to the sulfur cation radical. It was shown
that áCO could react with other radicals via the transfer
of an O ion, and O transfer to molecules with closed
3
�
�
product of áCO3 reaction with DBT-sulfoxide. Similar
3
�
�
metabolites were also reported in the biodesulfurization
of DBT in bitumen (Macpherson et al., 1998). All pri-
mary reaction pathways of the selected compounds with
electron shells has not previously been observed (Lilie
�
et al., 1978). The reaction of áCO3 probably occurred
through both mechanisms. However, under realistic
�
the áCO3 are shown in Fig. 3.
natural water conditions, the steady-state concentration
�
�
of áCO3 would be too low for O transfer to be signif-
icant in comparison to dissolved oxygen.
3
.2. Reaction mechanism
The most reactive site of the selected aromatic com-
�
pounds when reacting with áCO3 occurred at the sulfur
atom. The primary reaction pathway of the compounds
selected indicated primary conversion to the corre-
sponding sulfoxide followed by slow production of the
sulfone. A study of aquatic photodegradation of al-
bendazole under sunlight showed similar metabolites of
albendazole sulfoxide and sulfone (Weerasinghe et al.,
3.3. Persistence in di€erent matrices
During a cloudless summer noon hour, surface wa-
2
ters receive approximately 1 kW/m of sunlight. The
photoreactor simulated sunlight wavelengths with a
2
®xed irradiance level at 765 W/m . A suite of SFWs
(Table 1) was used to delineate more clearly the putative
�
1
992). Draper and Crosby (1984) investigated solar
role of áCO3 contributing to the natural cleansing of
photo-oxidation of sulfur-containing pesticides in dilute
hydrogen peroxide. Electron transfer reactions were
proposed to initiate sulfur oxidation observed in the
thiocarbamates molinate and thiobencarb with sulfoxide
formation resulting from radical coupling of dissolved
oxygen in water. This mechanism may partially explain
sulfur-containing chemical pollutants in ®eld waters.
The amount of DOC was about 5 mg/ml, which is a
reasonable range for most ®eld waters. DOC can not
only contribute to the production of transients in nat-
ural waters including áOH, but can also scavenge the
4
áOH with a second-order rate constant of 2:5 Â 10 (l/mg
�
results in this investigation with áCO3 as the initiator of
DOCs) (Larson and Zepp, 1988). The rate constants and
half-lives of selected compounds in di€erent matrices are
shown in Table 2; the rate constants for duplicate runs
had relative errors less than 5%.
the electron transfer that ultimately yields the corre-
sponding sulfoxide and sulfone. Under direct photolysis
conditions in DI water, relatively small amounts of
sulfoxide and sulfone were produced, which presumably
occurred through a similar mechanism.
The pseudo-®rst-order photoreaction of thioanisole
in the di€erent waters is shown in Fig. 4. In the dark
control solution at pH 11.0, there was no apparent
change of concentration over the time of the experiment.
In contrast with direct photolysis in DI water, the most
rapid photoreaction occurred in the matrix with a
�
Under high áCO3 conditions, sulfoxide formation
could also proceed through O transfer from a second
�
�
áCO . Initial electron transfer from the sulfur atom to
3
�
áCO , as described above, occurs to form the cation
3
�
Fig. 3. Primary reaction pathways towards áCO .
3

1
780
J. Huang, S.A. Mabury / Chemosphere 41 (2000) 1775±1782
Table 2
Photoreaction rate constants and half-lives of selected compounds at di€erent matrices
Reaction matrix
Thioanisole
Fenthion
DBT
�
1
� 1
� 1
k (min )
k (min
)
t1=2 (min)
k (min
)
t1=2 (min)
t
1=2 (min)
�
� 2
� 2
� 2
� 2
� 2
� 2
� 2
� 2
� 2
� 3
� 3
� 3
� 3
� 3
� 3
High áCO3
2:02 Â 10
9:46 Â 10
9:18 Â 10
6:84 Â 10
9:23 Â 10
8:08 Â 10
34.4
73.2
75.5
3:84 Â 10
2:76 Â 10
2:26 Â 10
1:95 Â 10
2:22 Â 10
2:63 Â 10
2:48 Â 10
18.0
25.1
30.7
35.5
31.2
26.3
27.9
2:80 Â 10
2:76 Â 10
2:60 Â 10
1:92 Â 10
2:71 Â 10
2:47 Â 10
1:98 Â 10
24.8
�
�
�
�
�
3
3
3
3
3
3
Bicarbonate radical
Synthetic ®eld water
DOC + bicarbonate
DOC + nitrate
251
267
361
256
281
350
101
75.1
Filtered ®eld water
DI water
85.8
96.9
7:15 Â 10�
relatively high concentration of Na
O
2
CO
3
(0.04 M) and
(1.5 mM) owing to the higher production of hy-
similar (half-life ˆ 101 min) to direct photolysis in DI
water. The rate in DOC/nitrate was the same as in SFW,
indicating an increased áOH availability for reaction
with thioanisole and reduced bicarbonate scavenging.
Fenthion and DBT show similar reactivities in dif-
ferent matrices (Table 2). The most rapid photoreaction
was in the sodium carbonate with hydrogen peroxide
matrix, followed by the sodium bicarbonate with nitrate
ion matrix. The slowest photoreaction occurred in the
matrix of DOC with sodium bicarbonate. The photolysis
rate constants for fenthion were higher than thioanisole
H
2
2
�
droxyl and thus higher áCO produced via the scavenging
by carbonate (Chen and Ho€man, 1974). The rate con-
3
�
3
� 1
stants ranged from 7:15 Â 10 min in DI water to
�
2
� 1
�
2
:02 Â 10 min in a high áCO matrix, which corre-
3
spond to half-lives of 96.9 and 34.4 min, respectively.
More importantly the degradation rates under realistic
�
áCO3 concentrations were faster than direct photolysis
(
Fig. 4). In the ®ltered Lake Huron water at pH 8.1, the
concentration of nitrate and bicarbonate was 6.2 and
05 ppm, respectively, while the amount of DOC was
�
2
� 1
1
and ranged from 2:48 Â 10 min in DI water to
�
2
� 1
�
about 2.5 mg/l. These values were lower than in SFW,
especially the amount of nitrate ion. The SFW and Lake
Huron water yielded half-lives of 75.5 and 85.8 min,
respectively. Enhanced degradation in these solutions
3:84 Â 10 min in a high áCO matrix; corresponding
3
to half-lives from 27.9 to 18.0 min, respectively. Due to
the relatively high direct photolysis rate in DI water for
fenthion, the di€erences in rate constants between vari-
ous matrices were small. The direct photolysis rate (half-
life ˆ 27.9 min) of fenthion was even slightly higher than
the rate in SFW (half-life ˆ 30.7 min) and in the DOC
�
suggests that the áCO3 could contribute to the cleansing
of this compound from sunlit natural waters. Photolysis
in the solution containing only DOC/bicarbonate was
Fig. 4. Photoreaction kinetics of thioanisole under di€erent reaction matrices.

J. Huang, S.A. Mabury / Chemosphere 41 (2000) 1775±1782
1781
plus nitrate matrix (half-life ˆ 31.2 min), while the
slowest rate (half-life ˆ 35.5 min) was in the DOC/
bicarbonate ion matrix. Photodegradation of fenthion in
three matrices containing DOC all showed slower pho-
tolysis than under direct photolysis. These results sug-
gest DOC's primary role in fenthion photolysis was to
reduce the overall light intensity available for production
of aqueous reagents and also to reduce the inherent di-
rect photolysis pathway (Torrents et al., 1997). The ®l-
tered Lake Huron water and the bicarbonate radical
solution yielded rate constants marginally faster than
direct photolysis (26.3 and 25.1 min half-lives, respec-
tively). In comparison to direct photolysis, the hydroly-
the duration of these experiments (>2 h). No methyl
phenyl sulfone was detected as a degradation product of
thioanisole in di€erent matrices due to the lower reac-
�
tivity of methyl phenyl sulfoxide toward the áCO3
(Huang and Mabury, 2000).
The primary photodegradation product for fenthion
was fenthion sulfoxide and 3-methyl 4-methylthiophe-
nol. In addition to direct oxidation on the sulfur atom,
as observed with thioanisole, direct photolysis breaking
of the P±O bond to form the 3-methyl 4-methylthi-
ophenol also occurred; as indicated above, hydrolysis
would be too slow to yield the amount of phenol ob-
served. Both mechanisms could have contributed to a
more rapid photolysis of fenthion in comparison with
thioanisole. Moreover, no fenthion sulfone was detected
as a degradation product, presumably for similar rea-
sons as for thioanisole.
�
6
� 1
sis rate constant of fenthion at 25°C was 8.7 ´ 10 min
at pH 9 (Huang and Mabury, 1999), which is much
slower than that for the photolysis rate observed here;
similar conclusions for fenthion sulfoxide hold since the
�
5
� 1
hydrolysis rate is 1.4 ´ 10 min at pH 9. Therefore,
photolysis of fenthion should contribute signi®cantly to
its primary degradation in sunlit ®eld waters. The role of
Contrary to our expectations, DBT-sulfoxide was not
observed as a degradation product in the photolysis of
DBT in the SFW and related solutions. Under direct
photolysis in DI water, DBT-sulfoxide degraded faster,
�
indirect photolysis via áCO3 reaction would be small.
with a rate constant of 5:0 Â 10� min in comparison
2
� 1
DBT had
2
a
high photolysis rate constant of
�
�
� 1
� 3
� 1
2
:80 Â 10 min in the high áCO3 matrix (half-life of
with DBT with a rate constant of 2:0 Â 10 min . Ap-
proximately 5% of the DBT-sulfoxide was converted into
DBT. It was also reported that photolysis of DBT-sulf-
oxide in organic solvents resulted in the formation of
DBT and oxidized solvent (Gregory et al., 1997). These
two observations may explain why a sucient amount of
DBT-sulfoxide was not found in the DBT photolysis.
Furthermore, no DBT-sulfone was detected which may
2
4.8 min), which was much higher than the direct pho-
�
3
� 1
tolysis rate constant of 1:98 Â 10 min in DI water
half-life of 350 min) (Table 2). The rate constant in the
high carbonate matrix was also much higher than the
(
�
3
rate constant in the bicarbonate matrix (2:76 Â 10
�
1
min ), which indicated that the higher steady-state
�
concentration of áCO3 formed in the system could en-
hance the photolysis rate dramatically. The lowest rate
indicate its rapid photolysis or an alternate pathway of
�
constant for DBT of 1:92 Â 10� min was found in the
DOC and bicarbonate ion matrix as well. However, the
measured half-life value of DBT in DI photolysis was
much faster than the reference value (94 h) reported by
Mill et al. (1981), presumably due to di€erent wave-
length and light intensities.
3
� 1
DBT-sulfoxide degradation. In contrast, the ``high áCO ''
solution did yield small amounts of DBT-sulfoxide but
these were observed only after concentrating the sample.
3
4. Conclusions
In general, the absorption of DOC gradually de-
creases with increasing wavelength, with essentially no
absorption occurring above 550 nm. Most of the solar
energy absorbed by DOC is between 300 and 500 nm. As
the concentration of DOC increases, the integrated ab-
sorption also increases. As a result, DOC could shield the
light intensity to lower the photolysis rate for the tested
compounds and provide some indication that secondary
DOC radicals (alkyl peroxy radicals) are not contribut-
ing much if any to the indirect photolysis pathway.
The primary reaction pathways of the selected sulfur-
�
containing compounds with áCO3 were through the
sulfoxide, followed by the corresponding sulfone prod-
ucts. Three selected compounds were investigated under
a photoreactor in di€erent reaction matrices. Among all
matrices, the most rapid photodegradation occurred in a
carbonate radical matrix and bicarbonate radical matrix
in comparison with direct photolysis. Because of the
relatively high photodegradation rate in deionized wa-
�
ter, the contribution of indirect photolysis through áCO3
was of intermediate importance.
3
.4. Photodegradation products
The primary photodegradation product for thioani-
sole in the di€erent matrices was methyl phenyl sulfox-
ide. Approximately 30% of the thioanisole was
Acknowledgements
�
converted to methyl phenyl sulfoxide in the high áCO3
matrix, while only 6% thioanisole was in DI water.
Methyl phenyl sulfoxide was relatively stable in DI for
This research project was supported by the Natural
Science and Engineering Research Council of Canada
(NSERC) and DuPont Canada. We appreciate the

1
782
J. Huang, S.A. Mabury / Chemosphere 41 (2000) 1775±1782
contribution made by Perkin-Elmer toward the creation
of the Analytical Lab for Environmental Science Re-
search and Training (ANALEST). This facility was used
to a signi®cant extent throughout the course of this
investigation.
Lartigues, B.S., Garrigues, P.P., 1995. Degradation kinetics of
organophosphorous and organonitrogen pesticides in dif-
ferent waters under various environmental conditions.
Environ. Sci. Technol. 29, 1246±1254.
�
Lilie, J., Hanrahan, R.J., Hengkein, A., 1978. O transfer
reactions of the carbonate radical anion. Radiat. Phys.
Chem. 11, 225±227.
Macpherson, T., Greer, C.W., Zhou, E., Jones, A.M., Wisse,
G., Lau, P.C.K., Sankey, B., Grossman, M.J., Hawari, J.,
References
1998. Application of SPME/GC±MS to characterize me-
Cabras, P., Garau, V.L., Melis, M., Pirisi, F.M., Spanedda, L.,
1
tabolites in the biodesulfurization of organosulfur model
compounds in bitumen. Environ. Sci. Technol. 32, 421±
426.
993. Persistence and fate of fenthion in olives and olive
products. J. Agric. Food Chem. 41, 2431±2433.
Cabras, P., Plumitallo, A., Spanedda, L., 1991. High-perfor-
mance liquid chromatographic separation of fenthion and
its metabolites. J. Chromatogr. 540, 406±410.
Mill, T., Hendry, D.G., Richardson, H., 1980. Free-radical
oxidants in natural waters. Science 207, 886±887.
Mill, T., Mabey, W.R., Lan, B.Y., Baraze, A., 1981. Photolysis
of polycyclic aromatic hydrocarbons in water. Chemosphere
10, 1281±1290.
Chen, S.N., Ho€man, M.Z., 1973. Rate constants for the
reaction of the carbonate radical with compounds of
biochemical interest in neutral aqueous solution. Radiat.
Res. 56, 40±47.
Minelli, E.V., Cabras, P., Angioni, A., Garau, V.L., Melis, M.,
Pirisi, F.M., Cabitza, F., Cubeddu, M., 1996. Persistence
and metabolism of fenthion in orange fruit. J. Agric. Food
Chem. 44, 936±939.
Chen, S.N., Ho€man, M.Z., 1974. Reactivity of the carbonate
radical in aqueous solution. Tryptophan and its derivatives.
J. Phys. Chem. 78, 2099±2102.
Olson, E.S., Stanley, D.C., Gallagher, J.R., 1993. Character-
ization of intermediates in the microbial desulfurization of
dibenzothiophene. Energy and Fuels 7, 159±164.
Torrents, A., Anderson, B.G., Bilboulian, S., Johnson, W.E.,
Hapeman, C.J., 1997. Atrizine photolysis: mechanistic
investigations of direct and nitrate-mediated hydroxyl
radical processes and the in¯uence of dissolved organic
carbon from chesapeake bay. Environ. Sci. Technol. 31,
1476±1482.
Draper, W.M., Crosby, D.G., 1984. Solar photoxidation of
pesticides in dilute hydrogen peroxide. J. Agric. Food
Chem. 32, 231±237.
Gregory, D.D., Wan, Z., Jenks, W.S., 1997. Photodeoxygen-
ation of dibenzothiophene sulfoxide: evidence for a unimo-
lecular S±O cleavage mechanism. J. Am. Chem. Soc. 119,
9
4±102.
Hagg, W.R., Hoign ꢀe , J., 1985. Photo-sensitized oxidation in
natural water via OH radicals. Chemosphere 14, 1659±1671.
Hirai, T., Shiraushi, Y., Ogawa, K., Komasawa, I., 1997. E€ect
of photosensitizer and hydrogen peroxide on desulfurization
of light oil by photochemical reaction and liquid±liquid
extraction. Ind. Eng. Chem. Res. 36, 530±533.
Weerasinghe, C.A., Lewis, D.O., Mathews, J.W., Je€coat,
A.R., Troxler, P.M., Wang, R.Y., 1992. Aquatic photode-
gradation of albendazole and its major metabolites. 1.
Photolysis rate and half-life for reaction in a tube. J. Agric.
Food. Chem. 40, 1413±1418.
Huang, J., Mabury, S.A., 1999. Hydrolysis kinetics of fenthion
and its metabolites. Submitted.
Za®riou, O.C., 1974. Sources and reactions of OH and
daughter radicals in seawater. J. Geophys. Res. 79, 4491±
4497.
Huang, J., Mabury, S.A., 2000. A new method for measuring
carbonate radical reactivity towards pesticide. Environ.
Toxicol. Chem., in press.
Zepp, R.G., 1991. Photochemical fate of agrochemicals in
natural waters. Pesticide Chemistry: advances in interna-
tional research, development, and legislation. In: Proceeding
of the Seventh International Congress of Pesticide Chem-
istry (IUPAC) Hamburg 1990. VCH, Weinheim, pp. 329±
345.
Kamrin, M.A., 1997. Pesticide Pro®les: Toxicity, Environman-
tal Impact, and Fate. CRC Lewis, New York.
Lacorte, S., Lartigue, S.B., Garrigues, P., Barcelo, D., 1995.
Degradation of organophosphorous pesticides and their
transformation products in estuarine waters. Environ. Sci.
Technol. 29, 431±438.
Zepp, R.G., Cline, D.M., 1977. Rates of direct photolysis in
aquatic environment. Environ. Sci. Technol. 11, 359±366.
Zepp, R.G., Hoign ꢀe , J., Bader, H., 1987. Nitrate-induced
photooxidation of trace organic chemicals in water. Envi-
ron. Sci. Technol. 21, 443±450.
Larson, R.A., Zepp, R.G., 1988. Reactivity of the carbonate
radical with aniline derivatives. Environ. Toxicol. Chem. 7,
2
65±274.