Development of HPLC/ESI-MS and HPLC/1H NMR methods for the identification of photocatalytic degradation products of iodosulfuron

Article abstract of DOI:10.1021/ac051836t

In the present study, HPLC/ESI-MS and stopped-flow HPLC/¹H NMR methods were developed and applied to separate and characterize the byproducts arising from TiO₂-catalyzed photodegradation of the herbicide iodosulfuron methyl ester (IOME) in aqueous solution under UV irradiation. Prior to identification, irradiated solutions of IOME (200 and 1000 mg·L ^-1) were concentrated by solid-phase extraction using two cartridges: Isolute C₁₈ and Isolute ENV⁺. Analytical separation was achieved on a C₁₈ reversed-phase column with ACN/H₂O (HPLC/MS) or ACN/D₂O (HPLC/NMR) as mobile phase and a linear gradient with a chromatographic run time of 35 min. The combination of UV and MS data allowed the structural elucidation of more than 20 degradation products, whereas ¹H NMR data permitted an unequivocal confirmation of the identities of major products and the differentiation of several positional isomers, in particular, the hydroxylation isomers. The obtained results permitted us to propose a possible degradation scheme and to put in evidence the presence of privileged sites for the attack of OH radicals. This work shows, for the first time, the application of combined HPLC with UV, MS, and NMR detection for complete structural elucidation of photocatalytic degradation products, and it will be of particular value in studies on the elimination of pollutants in aqueous solutions by photocatalysis.

Full text of DOI:10.1021/ac051836t

Anal. Chem. 2006, 78, 2957-2966
Development of HPLC/ESI-MS and HPLC/1H NMR
Methods for the Identification of Photocatalytic
Degradation Products of Iodosulfuron
Mohamad Sleiman,*,† Corinne Ferronato,† Bernard Fenet,‡ Robert Baudot, Farouk Jaber,§ and
Jean-Marc Chovelon†
Laboratoire d’Application de la Chimie a` l’Environnement, UMR 5634, Universite´ Claude Bernard Lyon 1,
43 Boulevard du 11 Novembre 1918, F-69622 Villeurbanne Cedex, France, Centre Commun de RMN UCB-CPE,
F-69622 Villeurbanne Cedex, France, Service Central d’Analyse, USR 059 CNRS, Echangeur de Solaize,
B.P. 22, 69390 Vernaison, France, and Laboratoire des Pesticides et des Micro-Polluants Organiques, Commission
Libanaise de l’Energie Atomique, B.P. 11-8281, Riad El Solh, 1107 2260 Beirut, Lebanon
(TiO2), to produce strongly oxidizing species, usually hydroxyl
radicals (OH•) which are able to destroy a great variety of toxic
organic compounds. Ideally, the end products are carbon dioxide,
water, and inorganic ions.
Generally, in this process, various byproducts are formed en
route to complete mineralization. Some of them can be toxic, and
in some cases, more persistent than the parent compound.5,6
Therefore, a careful identification of these compounds is essential.
Furthermore, this could be very useful to understand and interpret
the degradation mechanism.
This task represents a particular analytical challenge because
the majority of byproducts formed are new chemical entities, for
which standards are not available. Moreover, several positional
isomers are typically produced during the process. Thus, analytical
methods that combine high separation efficiency with a maximum
of structural information are required. GC/MS and LC/UV/MS
techniques are the most frequently used.7-9
For byproducts of high to medium volatility which are
thermally stable, GC/MS under electron impact (EI) conditions
is the method of choice because it combines the high separation
efficiency of the GC with the structural specificity of MS, and
extensive MS libraries are available for the identification of already
known analytes.
In the present study, HPLC/ESI-MS and stopped-flow
HPLC/1H NMR methods were developed and applied to
separate and characterize the byproducts arising from
TiO2-catalyzed photodegradation of the herbicide iodosul-
furon methyl ester (IOME) in aqueous solution under UV
irradiation. Prior to identification, irradiated solutions of
IOME (200 and 1000 mg‚L-1) were concentrated by
solid-phase extraction using two cartridges: Isolute C18
and Isolute ENV+. Analytical separation was achieved on
a C18 reversed-phase column with ACN/H2O (HPLC/MS)
or ACN/D2O (HPLC/NMR) as mobile phase and a linear
gradient with a chromatographic run time of 35 min. The
combination of UV and MS data allowed the structural
elucidation of more than 20 degradation products, whereas
1H NMR data permitted an unequivocal confirmation of
the identities of major products and the differentiation of
several positional isomers, in particular, the hydroxylation
isomers. The obtained results permitted us to propose a
possible degradation scheme and to put in evidence the
presence of privileged sites for the attack of OH radicals.
This work shows, for the first time, the application of
combined HPLC with UV, MS, and NMR detection for
complete structural elucidation of photocatalytic degrada-
tion products, and it will be of particular value in studies
on the elimination of pollutants in aqueous solutions by
photocatalysis.
For polar thermally labile byproducts, such as those resulting
from the photocatalytic degradation of sulfonylurea or phenylurea
compounds, HPLC/UV-DAD and HPLC/MS are the techniques
commonly used.10,11 However, although these techniques are
ideally suited to the characterization of known or suspected
In the past decades, heterogeneous photocatalysis has proved
to be one of the most promising methods for the elimination of
environmental pollutants.1-4 This technique is based upon the use
of UV-irradiated semiconductors, generally titanium dioxide
(3) Blake, D. M. Bibliography of Work on the Photocatalytic Removal of Hazardous
Compounds from Water and Air; NREL/TP-430-22197; National Renewable
Energy Laboratory: Golden, CO, 1997 and 1999.
(4) Herrmann, J. M. Catal. Today 1999, 53, 115-129.
(5) Parra, S.; Sarria, V.; Malato, S.; Pe´ringer, P.; Pulgarin, C. Appl. Catal., B
2000, 27, 153-168.
* Correponding author. Tel: (33) 4 72 43 11 50. Fax: (33) 4 72 44 84 38.
E-mail: mohamad.sleiman@univ-lyon1.fr.
† Universite´ Claude Bernard Lyon 1.
(6) Pramauro, E.; Bianco Prevot, A.; Vincenti, M.; Brizzolesi, G. Environ. Sci.
‡ Centre Commun de RMN UCB-CPE.
Technol. 1997, 37, 3126-3131.
USR 059 CNRS.
(7) Koulombos, V. N.; Tsipi, D. F.; Hiskia, A. E.; Nikolic, D.; van Breemen, R.
B. J. Am. Soc. Mass Spectrom. 2003, 14, 803-817.
§ Commission Libanaise de l’Energie Atomique.
(1) Schiavello, M., Ed. Photocatalysis and Environment: Trends and Applications;
NATO ASI Series C; Vol. 238; Kluwer Academic Publishers: London, 1987.
(2) Ollis, D. F.; Al-Ekabi, H., Eds. Photocatalytic Purification and Treatment of
Water and Air, Elsevier: Amsterdam, 1993.
(8) Amorisco, A.; Losito, F.; Palmisano, F.; Zambonin, P. G. Rapid Commun.
Mass Spectrom. 2005, 19, 1507-1516.
(9) Calza, P.; Baudino, S.; Aigotti, R.; Baiocchi, C.; Pelizzetti, E. J. Chromatogr.,
A 2003, 984, 59-66.
10.1021/ac051836t CCC: $33.50 © 2006 American Chemical Society
Published on Web 03/29/2006
Analytical Chemistry, Vol. 78, No. 9, May 1, 2006 2957

byproducts, the structural elucidation of unknowns as well as the
differentiation of the positional isomers remains difficult, even if
MS/MS or MSn experiments is included.12 As a result, the use of
complementary techniques is necessary to get a full and unam-
biguous identification of byproducts.
Chart 1. Iodosulfuron Methyl Ester Chemical
Structure
The HPLC/NMR technique is one of the most powerful
methods for the structural elucidation of unknown compounds
and the differentiation of structural isomers. It complements the
HPLC/UV/MS methods well: whereas HPLC/MS provides
important information on the molecular weight and the presence
of certain functional groups, HPLC/NMR gives structural informa-
tion on the constitution and the isomeric substitution of unknown
molecules. Moreover, as a result of the very narrow resonance
signals in modern high-field NMR, the resolution of information
is also improved considerably and is comparable to that of GC/
MS.13
The recent progress in pulse field gradients and solvent
suppression, the improvements in probe technology, and the
construction of high-field magnets have increased the sensitivity
of the method and given a new stimulus to this technique, which
has emerged since the mid 1990s as a very efficient tool for the
on-line identification of organic molecules.
Nowadays, directly coupled HPLC/1H NMR spectroscopy is
a commercially available technique that has been applied suc-
cessfully in many cases in analytical chemistry, in particular to
the identification of drug metabolites, natural products, and peptide
libraries.14-20 Furthermore, in the past few years, several studies
have illustrated the capability of HPLC/NMR for the analysis of
complex environmental samples, including the quantification of
pollutants and their degradation products over an environmentally
relevant range of concentrations.21-26
In this study, we present, for the first time, the application of
combined HPLC/UV/MS and HPLC/1H NMR on-line coupling
methods for an unequivocal identification of the byproducts
resulting from the photocatalytic degradation of iodosulfuron (see
structure in Chart 1), a new member of the sulfonylurea group of
herbicides which has recently entered our environment.
EXPERIMENTAL SECTION
Chemicals, Reagents, and Standards. Iodosulfuron methyl
ester (IOME) (methyl-4-iodo-2-[({[(4-methoxy-6-methyl-1,3,5-tri-
azin-2-yl)-amino]carbonyl }amino)sulfonyl]-benzoate (>98%) was
provided by Bayer Crop Science (Lyon, France). The following
reference compounds for confirmation of the identified compounds
were purchased from Aldrich (Steinheim, Germany): cyanuric
acid [2,4,6-trihydroxy-1,3,5-triazine] (98%), ammelide [2-amino-4,6-
dihydroxy-1,3,5-triazine] (97%) and s-triazine. For solubility and
stability reasons,27 stock solutions of IOME (200 and 1000 mg‚L-1
)
were prepared in ultrapure water buffered with KH2PO4/K2HPO4
at pH 6.5 and stored at 4 °C in the dark. The photocatalyst was
TiO2, Degussa P-25, mainly anatase, with a specific area of 50 m2
g-1 and a mean particle size of 30 nm.
Acetonitrile (quality HPLC grade and quality NMR), methanol
(HPLC grade), and deuterium oxide (99.9%) were purchased from
SDS (Peypin, France). Water was obtained from a Millipore
Waters Milli-Q water purification system (Molsheim, France).
ortho-Phosphoric acid (85%) and formic acid (MS grade, 99%
purity) were from Aldrich (Steinheim, Germany). Other reagents
were of at least analytical grade.
Irradiation Procedure. The irradiation experiments were
carried out in an open Pyrex glass cell (cutoff at 295 nm, 4.0-cm
diameter, 2.3-cm height), containing 25 mL of the aqueous
suspension of IOME and TiO2 powder. The light source was a
HPK 125 W Philips mercury lamp, cooled with water circulation.
For all experiments, the cell temperature during irradiation was
adjusted to 20 °C, the suspensions were magnetically stirred, and
(10) Malato, S.; Caceres, J.; Fernandez-Alba, A. R.; Piedra, L.; Hernando, A. D.;
Aguera, A.; Vial, J. Environ. Sci. Technol. 2003, 37, 2516-2524.
(11) Vulliet, E.; Emmelin, C.; Chovelon, J. M.; Guillard, C.; Herrmann, J. M. Appl.
Catal., B 2002, 38, 127-137.
(12) Rafqah, S.; Wong-Wah-Chung, P.; Aamili, A.; Sarakha, M. J. Mol. Catal. A:
Chem. 2005, 237, 50-59.
(13) Preiss, A.; Levsen, K.; Humpfer, E.; Spraul, M. Fresenius’ J. Anal. Chem.
1996, 356, 445-451.
the concentration of TiO2 was set to 2.5 g‚L-1
.
(14) Albert, K. J. Chromatogr., A 1999, 856, 199-211.
(15) Lindon, J. C.; Nicholson, J. K.; Wilson, I. D. J. Chromatogr., B 2000, 748,
233-258.
During the irradiation, aliquots of the aqueous suspension were
collected at predefined times and filtered through 0.45-µm PVDF
filters (Millipore) to remove TiO2 particles.
(16) Sandvoss, M.; Weltring, A.; Preiss, A.; Levsen, K.; Wuensch, G. J. Chro-
matogr., A 2001, 917, 75-86.
Solid-Phase Extraction (SPE). To develop a sample extrac-
tion method adapted to a maximum number of degradation
products which may have very different chemical properties and
polarities (and a priori unknown), two types of SPE cartridges
from International Sorbent Technology (IST, Cambridge, UK)
were used: (a) Isolute C18 (500 mg/6 mL), to extract IOME and
its “first generation products” and (b) Isolute ENV+ (200 mg/6
mL), a hydroxylated, highly cross-linked polystyrene/divinyl
benzene copolymer (St-DVB) to concentrate more polar products
as triazine derivatives.
(17) Peng, S. X. Biomed. Chromatogr. 2000, 14, 430-441.
(18) Bobzin, S. C.; Yang, S. T.; Kasten, T. P. J. Chromatogr., B 2000, 748, 259-
267.
(19) Wolfender, J. L.; Ndjoko, K.; Hostettmann, K. Phytochem. Anal. 2001, 12,
2-22.
(20) Wilson, I. D.; Lindon, J. C.; Nicholson, J. K. Anal. Chem. 2000, 72, 534-
542.
(21) Levsen, K.; Preiss, A.; Godejohann, M. Trends Anal. Chem. 2000, 19, 27-
48.
(22) Benfenati, E.; Pierucci, P.; Fanelli, R.; Preiss, A.; Godejohann, M.; Astratov,
M.; Levsen, K.; Barcelo´, D. J. Chromatogr., A 1999, 831, 243-256.
(23) Preiss, A.; Sa¨nger, U.; Karfich, N.; Levsen, K.; Mu¨gge, C. Anal. Chem. 2000,
72, 992-998.
(24) Godejohann, M.; Preiss, A.; Mu¨gge, C.; Wu¨nsch, G. Anal. Chem. 1997,
69, 3832-3837.
(25) Godejohann, M.; Astratov, M.; Preiss, A.; Levsen, K.; Mu¨gge, C. Anal. Chem.
1998, 70, 4104-4110.
(26) Godejohann, M.; Preiss, A.; Mu¨gge, C. Anal. Chem. 1998, 70, 590-595.
The efficiency of the preconcentration of each of both used
cartridges has been studied for different conditions (pH of the
(27) Brigante, M.; Emmelin, C.; Previtera, L.; Baudot, R.; Chovelon, J. M. J. Agric.
Food Chem. 2005, 53, 5347-5352.
2958 Analytical Chemistry, Vol. 78, No. 9, May 1, 2006

sample; nature, composition, and volume of eluting solution). The
best extraction conditions were conditioning of the sorbents with
6 mL of MeOH followed by 6 mL of pure water. The sorbents
were not allowed to dry, and subsequently irradiated samples of
IOME (200 and 1000 mg‚L-1) of 20 mL volume were passed
through the cartridges at a rate of ∼3 mL‚min-1 using a Varian
vacuum manifold (Sigma-Aldrich). For the ENV+ cartridge, the
pH of the samples was adjusted to between 2.5 and 2.7 using 40%
orthophosphoric acid. The analytes were eluted using 1 mL of
region. The analytical column and gradient were the same as used
for the HPLC/MS experiments.
Prior to data acquisition, solvent-signal suppression was
achieved using water suppression enhanced through T1 effects
(WET) solvent suppression, which included standard 4-fold
selective excitation with 20-ms seduced-1 pulses flanked by 1-ms
sine-shaped gradient pulses, with GARP 13C decoupling during
WET dephasing and acquisition to suppress 13C satellites of ACN.28
FIDs were collected into 16-K data points (spectral width of 10 000
Hz) with a flip angle of 90° and a relaxation delay of 3.0 s, resulting
in a total repetition time of 4.1 s. Depending on the concentration
of the analyte in the sample, the total number of scans varied
between 64 and 256. After acquisition, data were multiplied with
an exponential function f2 leading to a line broadening of 0.3 Hz
(LB ) 0.3 Hz) prior to Fourier transformation. 1H NMR chemical
shift values were referenced to acetonitrile, which was set to 2.0
ppm.
MeOH (2 × 0.5 mL) at a rate of ∼1 mL.min-1
.
HPLC Separations. Optimization of HPLC separations were
carried out on a 150 × 3 mm i.d., Interchim HDO C18 (Montluc¸on,
France) reversed-phase column (3 µm, 128 Å) using a Shimadzu
VP series HPLC system (Kyoto, Japan) consisting of a LC-10AT
binary pump, a SPD-M10A DAD, and Shimadzu Class-VP software
(version 5.0). The mobile phase was a mixture of acetonitrile (A)
and water acidified with formic acid to pH 2.8 (B). The flow rate
was set to 0.4 mL‚min-1, and injection volumes were 20 µL. The
optimized linear gradient was programmed as follows: 0 min, 5%
A; 30 min, 70% A; 33 min, 70% A; and 35 min, 100% A.
HPLC/MS Experiments. The HPLC/MS analysis was per-
formed on a HP 1100 (Agilent Technologies, Waldbronn, Ger-
many) LC system (Hewlett-Packard GmbH) consisting of quater-
nary pump, degasser, autosampler, column oven, and DAD UV
detector, interfaced with a HP 1100 MSD quadrupole mass
spectrometer equipped with an electrospray ionization source. The
chromatographic conditions were similar to those of the HPLC
separations (see HPLC Separations).
Safety Considerations. Iodosulfuron, ammelide and s-triazine
are toxic compounds and should, therefore, be handled with
special care. UV light used for irradiation is harmful to the eyes,
and a UV mask must be used for protection.
RESULTS AND DISCUSSION
SPE Preconcentration. After optimization of the conditions,
experiments were performed in triplicate. Recoveries were deter-
mined by comparing the analytes’ UV responses before and after
preconcentration, taking into account the preconcentration factor,
which is theoretically equal to 20.
Using the C18 cartridge, good recoveries (58-83%, RSD < 9%)
were obtained for IOME and its degradation products 11-21
(see Table 1), whereas for the highly polar products (1-10), such
as triazine derivatives, the preconcentration efficiency was unsat-
isfactory, with poor recoveries ranging from 8 to 21% (RSD < 13%).
As a consequence, the C18 cartridge was not suitable for the
preconcentration of the polar products.
Using the ENV+ cartridge, extraction efficiency was much
better for the highly polar products (1-10), with satisfactory
recoveries between 38 and 54% (RSD < 8%) except for cyanuric
acid, which had a moderate recovery of 32 ( 5%. For the less
polar products, such 11-21, ENV+ allowed us to obtain accept-
able recoveries, but lower than with the C18 (42-61%, RSD < 11%).
Study of recoveries obtained with the two tested cartridges
(C18 and ENV+) shows well that the use of both cartridges was
appropriate to make an efficient preconcentration of the degrada-
tion products of IOME to detect and identify them further by
HPLC/MS and HPLC/NMR.
HPLC/UV/ESI-MS. The characterization of byproducts was
first achieved by HPLC/UV/ESI-MS. As described in the Experi-
mental Section, the MS spectra were acquired in both positive
and negative ion modes to confirm the structural assignment of
the quasimolecular ion, [M + H]+ in positive ion mode by the
corresponding molecular anion [M - H]-. By comparing the two
ionization modes, the first was found to be more sensitive and
suitable for the majority of the analyzed compounds. Thus, all
the detected byproducts form [M + H]+ ions, but only some of
these, that is, 4, 6-8, and 14-21, form [M - H]- ions, as well
In preliminary experiments, an optimization of the ESI interface
and ion optics parameters was accomplished to maximize the
ionization current signal of the analytes. For this purpose, a 200
mg‚L-1 IOME irradiated solution was infused (flow rate 20
µL‚min-1) into the ESI interface using the syringe pump incor-
porated into the mass spectrometer.
The optimized electrospray parameters were spray voltage, 4
kV; cone voltage, 80 V; drying gas (nitrogen), 12 L‚min-1
;
nebulizer pressure, 55 psi; and source temperature, 350 °C. Mass
spectra were detected in both positive and negative ion modes in
a mass range of m/z 80-1000.
HPLC/NMR Stopped-Flow Experiments. Experiments were
conducted on an Agilent 1100 series HPLC system consisting of
a solvent degasser (model G1322A), a quaternary pump, a manual
injector (model 7725i) from Rheodyne (Rohnert Park, U.S.A.) and
a VWD (variable wavelength detector). A BSFU (Bruker stop flow
unit) was used to interface the HPLC to a DRX 500-MHz NMR
spectrometer from Bruker (Rheinstetten, Germany). The NMR
1
spectrometer was equipped with a H-{13C} inverse flow probe
(3-mm i.d. of measuring cell, with a detection volume of 60 µL).
The chromatographic equipment was controlled by HyStar NT
version 2.3 software (Bruker, Germany). Stopped-flow NMR
spectra were acquired using XWINNMR software (version 3.5,
Bruker, Rheinstetten, Germany).
The chromatographic separation was performed using ACN
and (D2O + H3PO4, pH 2.8) as mobile phase. D2O was used for
the deuterium lock to ensure stable measurement conditions. In
addition, formic acid used in the HPLC/MS experiments was
replaced by ortho-phosphoric acid to avoid superposition of the
signals of the analytes and formic acid in the aromatic spectral
(28) Smallcombe, S. H.; Patt, S. L.; Keifer, P. A. J. Magn. Reson. A 1995, 117,
295-303.
Analytical Chemistry, Vol. 78, No. 9, May 1, 2006 2959

Table 1. UV and MS Data for the Structural Characterization of IOME and Its Photocatalytic Transformation
Products
a Rt: Retention time.
(see Table 1). This could be explained by the presence of nitrogen
atom(s), having an important proton affinity (basic character), in
the majority of the byproducts.
products are eluted before IOME, which is normal because the
photocatalytic process involves the formation of smaller and more
polar products when compared to the parent compound; however,
some compounds are eluted after IOME, and thus, they are more
hydrophobic than IOME. Mass spectra of some of these products
showed very high masses (m/z > 600), indicating that they
probably result from reactions of dimerization. This hypothesis
was supported by the absence of these compounds when the
degradation experiment was achieved at low levels of concentra-
tions of IOME (<100 mg‚L-1). Structures of these later-eluted
compounds were not determined and will not be discussed here.
In contrast, structures of the majority of compounds eluted before
Figure 1 depicts the UV trace recorded at 234 nm and ESI-
MS full-scan (positive mode) chromatograms, obtained for a
mixture of three aqueous solutions of IOME (200 mg‚L-1
)
irradiated successively for 15 min, 1 h, and 3 h. On the basis of
a preliminary monitoring by HPLC/DAD of the degradation
products during 5 h of irradiation, this mixture represents as much
as possible the different byproducts formed during the degrada-
tion. As can be seen in Figure 1, many compounds in addition to
IOME are present in the analyzed sample. The majority of these
2960 Analytical Chemistry, Vol. 78, No. 9, May 1, 2006

Figure 1. HPLC/UV at 234 nm (A) and HPLC/ESI-MS in positive mode (B) chromatograms of a mixture of three aqueous solutions of IOME
(200 mg‚L-1) irradiated successively for 15 min, 1 h, and 3 h. For chromatographic conditions, refer to hte Experimental Section. Peak numbers
correspond to compound numbers in Table 1.
30 min (Figure 1) were tentatively identified by interpretation of
their UV and MS spectra. A summary of the UV and MS data
(absorption maximums, quasimolecular ions, and the most im-
portant fragment ions), along with the proposed structures for
the byproducts, is reported in Table 1.
Chart 2. Chemical Structures of the Main
Fragment Ions, m/z 127, 141, 153 and 167,
Observed in MS Spectra of IOME Byproducts
The identified byproducts can be classified in three principal
groups according to their structures:
(a) Byproducts 1-8. These early-eluting, highly polar com-
pounds showed UV spectra (not showed) typical of a triazine ring
with two maximums between 195 and 200 and 220-230 nm (Table
1). No absorption band was observed above 270 nm, suggesting
the absence of a benzene ring in their structures. This conclusion
was corroborated by the mass spectrometric data which revealed
quasimolecular (127-264) and fragment ions (127, 153, 141, 167)
compatible with structures of triazine derivatives (see Table 1).
The identities of byproducts 1, 2, and 6 were confirmed by
direct comparison of their UV, MS, and chromatographic data with
those of reference compounds cyanuric acid, ammelide, and
s-triazine (AMMT), respectively. These compounds were previ-
ously reported in several studies on the photocatalytic degradation
of sulfonylureas.11,29
The remaining byproducts were not commercially available,
and therefore, a confirmation of their identities was relatively
difficult. However, byproducts 4, 7, and 8 exhibit some charac-
teristic fragment ions in positive ion ESI-MS, providing further
structural information. Thus, the MS spectrum of byproduct 4
showed two main fragment ions at m/z 127 and 153. The first is
interpreted as arising from cleavage of the bond between the
carbonyl group and the N-atom substituted with the triazinic
group, whereas the second is due to the cleavage of the carbonyl
C-NH2 bond. By analogy, formation of fragment ions at m/z 141
and 167 in MS spectra of byproducts 7 and 8 can be rationalized.
The structures of the identified fragments (m/z 127, 141, 153, 167)
are showed in Chart 2. It should be pointed out that this type of
fragmentation was already observed for sulfonylurea products by
Marek and Koskinen.30
(b) Byproducts 9, 10. UV spectra of these two products were
typical of a benzene ring with two bands between 240 and 250
(band I) and between 285 and 290 nm (band II). This, along with
the [M + H]+ quasimolecular ions and [M + Na]+ adducts
detected in their MS spectra suggests that these two byproducts
correspond to the structures proposed in Table 1. Unfortunately,
the MS data did not provide more detailed structural information
to determine the position of hydroxyl group on the benzene ring,
so that several isomers can be hypothesized for these compounds.
(29) Maurino, V.; Minero, C.; Pelizzetti, E.; Vincenti, M. Colloids Surf., A 1999,
151, 329-338.
(30) Marek, L.; Koskinen, W. J. Agric. Food Chem. 1996, 44, 3878-3881.
Analytical Chemistry, Vol. 78, No. 9, May 1, 2006 2961

Figure 2. HPLC/UV chromatogram (234 nm) as obtained from the Stopped-flow LC/NMR run of a concentrated solution of IOME 1000 mg‚L-1
.
(c) Byproducts 11-21. These products were characterized by
high retention times (>15 min, Figure 1), indicating that they are
less polar than those of the first two groups (a, b), their structures
are likely to contain the two moieties (benzene and triazine rings).
This was verified by comparing the UV spectra of these com-
pounds with that of IOME with two characteristic bands between
225 and 235 nm (band I) and 280-300 nm (band II), indicating
the presence of the two rings. Furthermore, MS spectra of these
products revealed, as already observed for byproducts of group
(a), characteristic fragment ions at m/z 127, 153 (byproducts 11-
13, 17) or at m/z 141, 167 (14-16, 18-21) which proves that
they contain the triazine ring. On the other hand, these fragment
ions permitted knowing whether a byproduct contains a triazinic
ring substituted with OH and CH3 groups (m/z 127 and 153) or
with OCH3 and CH3 (m/z 141 and 167). As a consequence, we
have distinguished the two isomers 17 and 18 having the same
mass ([M + H]+ at m/z 494) considering that compound 17 gave
the two fragment ions at m/z 127 and 153, whereas the MS
spectrum of byproduct 18 showed the two other fragment ions
at m/z 141 and 167.
In contrast, the information provided by mass spectrometry
was not sufficient to make the differentiation between the
positional isomers 14, 15 (two isomers) and 19-21 (three
isomers), resulting from the attack of OH• on the benzene ring.
Therefore, HPLC/1H NMR coupling was used to determine the
substitution pattern of the aromatic ring for these isomers and
also to verify the structure assignment by HPLC/UV/MS of those
compounds for which reference compounds were not available,
in particular, the group of byproducts 11-21.
By comparing the chromatogram displayed in Figure 2 with
that showed in Figure 1, it can be seen that the use of D2O as the
mobile phase and the replacement of formic acid by phosphoric
acid had no significant effect on the elution order and the retention
times of the main byproducts. Therefore, a correlation between
the HPLC/MS peaks in Figure 1 and the HPLC/NMR peaks in
Figure 3 was exploited to identify as many IOME byproducts as
possible. As a result, the structures of six major byproducts, 8,
14 (isomer of 15), 16, 17 (isomer of 18), 19 and 21 (two mono-
hydroxylation isomers), were unambiguously identified. Figure
1
3 shows relevant regions of the corresponding H NMR spectra
1
and that of IOME, obtained in the stopped-flow mode. The H
NMR signal assignments are summarized in Table 2.
According to Figure 3, spectrum A corresponding to the parent
compound (IOME) is divided in two sections, an aliphatic region
(2-4 ppm) displaying the protons of the methyl groups (a-c)
and an aromatic region (7-8.5 ppm) showing the signals of a 1,2,4
trisubstituted aromatic system: a doublet with a small meta
coupling constant for H3, a doublet of doublet with a small meta
and a large ortho coupling constant for H5, and a doublet with a
large ortho coupling constant for H6. Signals of NH protons are
missing because of their exchange with deuterium in the D2O
matrix. Note that the signal at 3.3 ppm corresponds to methanol,
which is the solvent in which the injected samples were dissolved
after preconcentration.
In comparison, it can be noted that aliphatic regions of 1H NMR
spectra B, C, and D are almost identical to that of IOME, revealing
the presence of a methyl group around 2.5 ppm (a), a methyl ester
group around 3.9 ppm (c), and a methoxy group around 4.0 ppm
(b). On the other hand, the aromatic parts of spectra B-D show
only two signals integrated for 1 H each, indicating the substitution
of one of the three aromatic protons by a hydroxyl group (OH).
This finding is supported by the chemical shift values of the
aromatic protons and by the MS data.
Stopped-Flow HPLC/1H NMR. Since the sensitivity of the
NMR detection is low compared to MS, more concentrated
samples of IOME (1000 mg‚L-1) were irradiated and, later on,
preconcentrated by SPE (as described in Experimental Section).
Moreover, volumes of 100 µL were injected instead of 20 µL.
The UV chromatogram at 234 nm as obtained in the stop-flow
Byproduct 21 (spectrum C) exhibits two ortho-coupling
doublets centered at 8.17 and 6.95 ppm (J ∼ 8.0 Hz), which are
assigned to H5 and H6, respectively. Thus, the proton H3 was
substituted by a hydroxyl group (OH). In contrast, the aromatic
pattern of byproducts 14 (isomer of 15) and 19 (isomer of 21)
reveal two singlets (see spectra B and D) corresponding to the
noncoupled protons H3 and H6. The OH group is thereby located
in position 5 (C5) on the aromatic ring (Table 2).
mode for a mixture of three solutions of IOME (1000 mg‚L-1
)
irradiated for 45, 150, and 480 min is depicted in Figure 2. These
irradiation times permitted us to obtain a mixture that contains
the different byproducts at their maximum levels of concentra-
tions. Noting that the degradation of IOME (1000 mg‚L-1) was
completely achieved within ∼160 min, as compared to 60 min for
the solution of IOME at 200 mg‚L-1, which is normal because
the degradation of higher concentrations need more UV irradiation
time.
Unfortunately, byproducts 15 (isomer of 14) and 20 (isomer
of 19 and 21) were not identified because of their low concentra-
2962 Analytical Chemistry, Vol. 78, No. 9, May 1, 2006

Figure 3. Stopped-flow 1H NMR spectra of IOME (A) and its byproducts 19 (B), 21 (C), 14 (D), 16 and 17 (E), and 8 (F). An asterisk (*)
denotes residual solvent signals (ACN, 1.93 ppm; MeOH, 3.30 ppm; HOD ∼ 4.5 ppm).
1
tions and their overlapping with other chromatographic peaks.
Nevertheless, the structure of isomer 20 could be assigned taking
into account that only one free site (position 6) remains for the
OH function to be on the aromatic ring, since the two others,
positions 5 and 3, were already attributed to isomers 19 and 21,
respectively.
A closer look at spectrum E, extracted at ∼21 min, shows two
groups of signals probably resulting from a mixture of two
byproducts. Indeed, the chromatographic peak at ∼21 min (see
Figure 3) exhibits a left threshold which indicates a coelution
process. Comparison with the separation profile in the HPLC/
UV/MS experiments (Figure 1) suggests that the two coeluted
byproducts are likely to be 16 (minor) and 17 (major). This
suggestion and the structures of both byproducts were confirmed
by the interpretation of the H NMR signals of spectrum E: in
the aliphatic region, the two intense singlets are attributed to the
methyl group (a) and the methyl ester group (c) of byproduct
17, whereas the three remaining singlets of weak intensities
correspond to the methyl group (a), the methyl ester group (c),
and the methoxy group (b) of byproduct 16 (Table 2). On the
other hand, in the aromatic region, each product exhibits three
signals (two doublets of doublets and one doublet) assigned to
the aromatic protons H3 (d), H5 (dd), and H6 (dd) (see Table 2).
Finally, contrary to the preceding spectra A-E, spectrum F
does not show any signals in the aromatic part, indicating the
absence of a benzene ring. The aliphatic region displays only two
singlets, which might be assigned to the methyl group (a) and
the methoxy group (b). This information is not sufficient to deduce
Analytical Chemistry, Vol. 78, No. 9, May 1, 2006 2963

Table 2. 1H NMR Data for IOME and Its Byproducts Identified by Stopped-Flow HPLC/NMR
a Rt: Retention time. b s, singlet; d, doublet; dd, doublet of doublet.
the complete structure of the compound because each one of
indicates that C5 on the benzene ring of IOME is privileged for
byproducts 6, 7, and 8 has both methyl groups a and b and can
thereby give almost the same H NMR spectrum. However, due
the attack of OH radicals. This result might be explained by the
mesomeric effect of the benzene ring substituents (-COOMe,
-SO2, and -I). The groups -COOMe and -SO2 decrease the
electron density in ortho and para positions (C3, C5, and C6) with
respect to them, whereas the iodide atom increases the electron
density in ortho and para positions (C3 and C5) with respect to it.
Therefore, sites C3 and C5 are more nucleophilic than site C6,
which explains the formation of byproducts 19 and 21 at higher
levels of concentration than that of byproduct 20 (not identified).
On the other hand, attack of OH radicals on position 3 is more
difficult than that on position 5 because of the steric hindering
by the -SO2 group. As consequence, the attack of OH radicals,
known as strong electrophilic species, takes place preferentially
at C5 and forms compounds 19.
(iii) The substitution of an iodide atom by hydroxyl groups
forms the byproduct 16, which in turn can be transformed by a
hydroxylation process into the two products 14 and 15. These
compounds could be also formed by substitution of an iodide atom
of the hydroxylation products 19-21 by a hydroxyl group.
(iv) Byproducts 17 and 18 result from an O-demethylation of
the methoxy moiety of the triazine ring and the methyl ester
groups, respectively. Both products can undergo a second O-
demethylation, leading to byproduct 13. Decarboxylation via
1
to the fact that the peaks’ retention times in both chromatograms
(Figures 1 and 2) are almost identical, spectrum F, which is
extracted at ∼10 min corresponds well to byproduct 8 having a
[M + H]+ at m/z 184 in the MS spectrum.
Deductions Concerning the Photocatalytic Degradation of
IOME. On the basis of the foregoing results of the LC/UV/MS
and LC/NMR experiments, a general reaction pathway can be
proposed for the photocatalytic degradation of IOME (Scheme
1). Although this task would be difficult starting only from the
identification of the intermediates, some considerations about the
process can be made:
(i) The first steps of the degradation seem to be hydroxylation
of the benzene ring, O-demethylation of the methoxy and methyl
ester moieties, substitution of the iodide atom by a hydroxyl group,
and cleavages of the sulfonylurea bridge. This means that the
products 3, 4, and 6-21 might be formed during the initial
degradation of IOME (see Scheme 1).
(ii) The hydroxylation process leads to three isomeric products,
19, 20, and 21. By assuming that the response coefficients for
the three isomers in UV or MS detection (Figure 1) are relatively
similar, it is clear that 19 is the major reaction product. This
2964 Analytical Chemistry, Vol. 78, No. 9, May 1, 2006

Scheme 1. Proposition of Photocatalytic Degradation Pathway of IOME
photo-Kolbe reactions31 can be invoked to justify the formation of
compounds 11 and 12 from 13.
(v) The formation of other s-triazinic compounds, such as
products 2 and 5, suggests that the degradation process continues
via a series of reactions, including further cleavages of the
sulfonylurea bridge, O-demethylation of the methoxy moiety of
the triazine, and oxidation of methyl and amino groups leading
to cyanuric acid (1), which cannot undergo further oxidative
attacks. It constitutes a final product, as has been observed in
previous studies.11,28,33
(iv) Cleavages of the sulfonylurea bridge of IOME and its
products 11-21 generate various s-triazinic derivatives (3, 4,
6-8) dominated by the products 6 and 8 resulting from the
rupture of carbon-nitrogen (C-N) and nitrogen-sulfur (N-S)
bonds, respectively. Cleavage of the sulfonylurea bridge also leads
to the formation of benzene derivatives. Only two of them were
detected and identified during the degradation process (byprod-
ucts 9 and 10). This could be attributed to the fact that they are
rapidly transformed by an oxidative opening of the aromatic ring,
giving rise to the formation of smaller and more oxidized
molecules, such as short carboxylic acids (formic, acetic, oxalic,
etc). This behavior has been confirmed in a previous study in
which several carboxylic acids, such as formic, acetic, oxalic and
glycolic acids, have been detected after the aromatic ring open-
ing.11,32
CONCLUSIONS
The combined use of HPLC/UV/MS and HPLC/NMR
under similar chromatographic conditions was demonstrated
to be a very powerful approach for the identification of a
large number of compounds generated by photocatalytic
transformation of IOME, even if no reference compounds are
(32) Pramauro, E.; Vincenti, M.; Augugliaro, V.; Palmisano, L. Environ. Sci.
Technol. 1993, 27, 1790-1795.
(33) Pelizzetti, E.; Maurino, V.; Minero, C.; Carlin, V.; Pramauro, E.; Zerbinati,
(31) Krauetler, B.; Bard, A. J. J. Am. Chem. Soc. 1978, 10, 2239.
O.; Tosato, M. L. Environ. Sci. Technol. 1990, 24, 1559-1565.
Analytical Chemistry, Vol. 78, No. 9, May 1, 2006 2965

commercially available. This approach could be further
applied successfully for the structural elucidation of byproducts
resulting from the treatment of others pollutants by photo-
catalysis.
On the basis of the results of the identification of byproducts,
a general scheme of degradation in which cyanuric acid was the
final organic product has been proposed. Fortunately, this
compound is known to be innocuous, nontoxic, and easily
biodegradable; thus, its formation during the photocatalytic
process does not represent any danger to the environment and
human beings.
Taking into account the results obtained during this study, it
would be interesting to quantify the main byproducts identified
using HPLC/NMR, since at a given concentration, the signal
intensity depends only on the molecular mass and the number of
chemical equivalent protons. In addition, an evaluation of the
toxicity of the different byproducts would be essential for the
continuation of this study.
Received for review October 13, 2005. Accepted March 6,
2006.
AC051836T
2966 Analytical Chemistry, Vol. 78, No. 9, May 1, 2006