Photochemistry of imidacloprid in model systems

Article abstract of DOI:10.1021/jf801251u

The photochemical behavior of the neonicotinoid insecticide imidacloprid was studied with regard to different chemical environments. Different model solvents simulated the structure moieties mainly occurring in waxes and cutin of the plant cuticle. Cycloh

Full text of DOI:10.1021/jf801251u

J. Agric. Food Chem. 2008, 56, 8023–8029 8023
Photochemistry of Imidacloprid in Model Systems
NICOLE SCHIPPERS AND WOLFGANG SCHWACK*
Institute of Food Chemistry, University of Hohenheim, Garbenstrasse 28, 70593 Stuttgart, Germany
The photochemical behavior of the neonicotinoid insecticide imidacloprid was studied with regard to
different chemical environments. Different model solvents simulated the structure moieties mainly
occurring in waxes and cutin of the plant cuticle. Cyclohexane and cyclohexene substituted saturated
and unsaturated hydrocarbon chains, whereas ethanol and 2-propanol were models for primary and
secondary alcohol groups of cuticular components. After 5 h of irradiation, imidacloprid was completely
degraded in all solvents. With 88-96 mol% 1-[(6-chloropyridin-3-yl)methyl]imidazolidin-2-imine was
formed as the main product, whereas 1-[(6-chloropyridin-3-yl)methyl]imidazolidin-2-one was identified
as minor product in the range 4-6 mol%. By contrast, besides the photoproducts formed in organic
solvents, irradiation of the solid imidacloprid on a glass surface delivered a complex variety of
unidentified photoproducts. The nucleophilic addition reaction of the main photoproduct, 1-[(6-
chloropyridin-3-yl)methyl]imidazolidin-2-imine, with both cyclohexene oxide and methyl 9,10-epoxys-
tearate as model compounds indicates that epoxidized cutin acids are possible reaction partners for
the formation of plant cuticle bound residues of imidacloprid, which could explain the reported findings
of nonextractable residues of imidacloprid in plants.
KEYWORDS: Imidacloprid; photodegradation; bound residues
INTRODUCTION
is still unknown. Scholz and Reinhard (10) performed photolysis
studies of imidacloprid on leaves of tomato plants and described
the cyclic guanidine (5), the cyclic urea (2), the nitroso (6,
Scheme 2), and a hydroxy derivative as main products and up
to 4.4% of nonextractable residues.
The neonicotinoids, with imidacloprid (1-[(6-chloropyridin-
3-yl)methyl]-N-nitroimidazolidin-2-imine, 1, Scheme 1) as the
most prominent representative, are a rather new class of
insecticides, which differ in their mode of action from earlier
pesticides like organophosphates, carbamates, organochlorines,
or pyrethroids. They target at the nicotinic acetylcholine
receptor, acting as agonists of this receptor (1). Because of high
efficacy against sucking and some biting insects, systemic
properties, and low mammalian toxicity, 1 is widely used in
plant protection and veterinary medicine (2).
Photolysis of imidacloprid in water was studied by different
groups (3-7), which resulted in rapid degradation and a variety
of reaction products (Scheme 1). All groups identified 1-[(6-
chloropyridin-3-yl)methyl]imidazolidin-2-one (2) as the main
photoproduct besides 6-chloronicotinic acid (3) and 6-chloroni-
cotinyl-aldehyde (4).
After spray application, there are different possible degrada-
tion pathways on plant surfaces such as hydrolysis, oxidation,
or photolysis. Photoinduced reactions with constituents of the
plant surface, the cuticle, resulting in the formation of bound
residues, are also possible (8). In the 2002 evaluations of the
Joint FAO/WHO Meeting on Pesticide Residues (9), nonex-
tractable residues in tomatoes, with a range 0.2-0.8%, were
reported after spray application of imidacloprid to tomato fruits
on tomato plants. But, the formation pathway of these residues
In this study, we examined the photoreactivity of 1 in different
model solvents, which were introduced by Schwack et
al. (11-13). Waxes and cutin as constituents of the plant cuticle
provide a wide range of possible reaction partners for applied
pesticides or their photoproducts. Cutin is a completely insoluble
polyester consisting of interesterified saturated, unsaturated, and
epoxydized hydroxy fatty acids, which is associated by intra-
and epicuticular lipids called waxes extractable by organic
solvents (14). The models cyclohexane and cyclohexene sub-
stitute saturated and unsaturated hydrocarbons as they occur
(e.g., as alkanes and alkenes, fatty acids, triterpenes, and sterols)
in the plant cuticle. Cutin acids, free fatty alcohols, or sterols
contain primary and secondary alcohol groups, whose influence
was studied by irradiations in ethanol and 2-propanol. In former
studies, most pesticides showed different degradation rates and
different product spectra depending on the solvent used (11-13).
Among the products, photoaddition products with alcohols and
olefins were detected that deliver reaction mechanisms for the
formation of covalently bound residues with constituents of the
plant cuticle. As imidacloprid, parathion also contains a nitro
group, which was first photoreduced to nitrosoparathion and
finally to aminoparathion, both of which are rather reactive
toward plant cuticle constituents (15-17). Thus, we conducted
photolysis studies using our well-established model systems to
* To whom correspondence should be addressed. Phone: ++49-
711-4592-3978. Fax: ++49-711-4592-4096. E-mail: wschwack@uni-
hohenheim.de.
10.1021/jf801251u CCC: $40.75  2008 American Chemical Society
Published on Web 08/09/2008

8024 J. Agric. Food Chem., Vol. 56, No. 17, 2008
Schippers and Schwack
Scheme 1. Structural Formula of Imidacloprid (1) and Its Photoproducts in Water Described in the Literature (3-7)
purchased from Fluka (Buchs, Switzerland). Water was purified by a
Synergy ultrapure water system (Millipore GmbH, Schwalbach,
Germany).
Scheme 2. Proposed Mechanism of the Photolysis of Imidacloprid (1)
Dissolved in Organic Solvents
Imidacloprid was precipitated from Confidor (200 g/L, Bayer,
Leverkusen, Germany) by addition of water. After filtration and washing
with water, the crude product was recrystallized twice from methanol
yielding colorless crystals with a melting point of 144 °C (143-144
°C (18)). The determined log ε at 269 nm (methanol) of 4.12 matched
the respective value of the commercial standard substance (99.9%,
Pestanal, Riedel de Haen, Seelze, Germany). LC/MS (ESI+): m/z 294
([M + K]+, 4%), 278 ([M + Na]+, 5%), 256 ([M + H]+, 100%), 210
([M + H - NO2]+•, 23%), 209 ([M + H - HNO2]+, 18%), 175 ([M
+ H - NO2 - Cl]+, 14%).
General techniques. Thin-layer chromatography (TLC) was per-
formed on silica gel 60 TLC aluminum sheets (20 cm × 20 cm, F254
,
Merck, Darmstadt, Germany) with ethyl acetate/methanol/toluene/acetic
acid (98%) (80/60/20/1) as mobile phase.
High-performance liquid chromatography (HPLC) was carried out
on a HP 1100 system (Agilent, Waldbronn, Germany) equipped with
a degasser, an autosampler, a quarternary pump, and a diode array
detector (DAD). For separation at 25 °C, a reversed-phase column
Eurospher 100-C18 (5 µm, 250 mm × 3 mm, Knauer, Berlin, Germany)
with a guard column (C18, 5 µm, 5 mm × 3 mm) and a gradient system
with methanol/phosphate buffer (20 mM, pH 4.0, 2% methanol) was
used. The methanol content of the gradient at a flow rate of 0.5 mL/
min was as follows: 0 min, 5%; 0-2 min, 5-25%; 2-7 min, 25-40%;
7-13 min, 40-50%; 13-18 min, 50%; 18-20 min, 50-100%; 20-30
min, 100%; 30-35 min, 100-5%. An aliquot of 10 µL was injected.
DAD detection was performed at screening wavelengths of 235, 270,
and 320 nm; for quantification, 270 nm was used for all compounds.
Data acquisition and processing were carried out with HP ChemStation
software (rev. A.06.01).
study and understand the general photochemistry of imidacloprid
to be expected on fruit and leaf surfaces.
The high-performance liquid chromatography-mass spectrometry
(LC/MS) system consisted of an HPLC system HP 1100 as described
above and a VG platform II quadrupole mass spectrometer (Micromass,
Manchester, U.K.) equipped with an electrospray interface (ESI). Apart
from the solvents (ammonium formate buffer (10 mM, pH 4.0) and
MATERIALS AND METHODS
Reagents. All solvents used were of analytical grade (Merck,
Darmstadt, Germany) and distilled before use. Cyclohexene was distilled
over P2O5 (Roth, Karlsruhe, Germany). Cyclohexene oxide was

Photochemistry of Imidacloprid in Model Systems
J. Agric. Food Chem., Vol. 56, No. 17, 2008 8025
methanol) used for LC separation, all LC parameters remained the same
as for HPLC/DAD analysis. MS parameters were as follows: ESI+;
source temperature, 120 °C; capillary 3.5 kV; HV lens 0.5 kV; cone
ramp 20-60 V. MS was operated in full-scan mode (m/z 80-1200).
For data acquisition and processing, MassLynx 3.2 software was
employed.
(br, s), 2900 (m), 1663 (s), 1618 (m), 1564 (s), 1452 (m), 1383 (m),
1358 (m), 1306 (w), 1285 (m), 1244 (m), 1209 (m), 1139 (m), 1110
(s), 1097 (s), 1025 (m), 966 (w), 931 (w), 848 (m), 804 (m), 738 (s),
697 (m), 674 (s). UV (methanol): λmax ) 207 (log ε ) 4.30), 268 nm
(log ε ) 3.45).
1-[(6-Chloropyridin-3-yl)methyl]imidazolidin-2-one (2) was synthe-
sized by reaction of 200 mg (0.78 mmol) of imidacloprid with 20 mL
of an aqueous potassium hydroxide solution (40 g/L) for 1 h at 100 °C
according to Rouchaud et al. (20) (yield: 58.6 mg, 35% of theory).
For the preparation of 1-[(6-chloropyridin-3-yl)methyl]imidazolidin-
2-one oxohydrazone (6), 100 mg (0.39 mmol) of imidacloprid was
reduced with 150 mg of iron powder (pure, Sigma-Aldrich, Munich,
Germany) in 50 mL of an aqueous ammonium chloride solution (11
mM). After stirring for 2 h at room temperature, the reaction batch
was filtered, and water was evaporated at reduced pressure (30 mbar,
40 °C). The residue was purified by HSCCC. The fraction of 185-225
mL was collected, the organic solvents were evaporated in vacuo, and
the remaining aqueous solution was lyophilized to yield 28 mg of yellow
crystals (30% of theory; melting point 146 °C under decomposition).
LC/MS (ESI+): m/z 262 ([M + Na]+, 10%), 278 ([M + K]+, 6%),
240 ([M + H]+, 100%), 210 ([M + H - NO]+•, 45%), 209 ([M + H
- HNO]+, 12%), 175 ([M + H - NO - Cl]+, 6%). UV (methanol):
λmax ) 212 (log ε ) 3.99), 269 nm (log ε ) 4.28).
Methyl 9,10-epoxystearate was synthesized by the reaction of 1.0 g
(3.2 mmol) of methyl oleate (Sigma-Aldrich, Munich, Germany) with
3 mL of peracetic acid (32% in dilute acetic acid, Sigma-Aldrich) in 5
mL of diethyl ether for 20 h at room temperature. The reaction batch
was poured in 20 mL of a saturated potassium carbonate solution and
extracted three times with 10 mL of diethyl ether. The organic extract
was dried over anhydrous sodium sulfate (Merck, Darmstadt, Germany),
and the solvent was evaporated resulting in 532 mg of a colorless oil
(50,4% of theory).
For kinetic studies, a solution of 10 mg (0.048 mmol) of 1-[(6-
chloropyridin-3-yl)methyl]imidazolidin-2-imine (5) and 500 µL (4.95
mmol) of cyclohexene oxide in 2 mL of methanol was stirred at ambient
temperature in the dark for 14 days. Samples of 10 µL were taken,
diluted with 990 µL of methanol, and analyzed by HPLC. In a
preparative scale, 101 mg (0.48 mmol) of 5 was dissolved in 20 mL of
methanol, 5 mL (49.5 mmol) of cyclohexene oxide were added, and
the reaction batch was stirred at 50 °C for 3 days. The mixture was
concentrated under vacuum to 5 mL, and 5 mL of ammonium formate
buffer (10 mM, pH 4.0) was added. Separation by preparative HPLC
yielded two fractions at 9.0 min (16, 78 mg, 52.8% of theory) and
10.5 min (17, 5.6 mg, 3.8% of theory).
2-({1-[(6-Chloropyridin-3-yl)methyl]imidazolidin-2-ylidene}amino)-
cyclohexanol (16); LC/MS (ESI+): m/z 309 [M + H]+; accurate mass
calculated for C15H22N4OCl ([M + H]+) 309.1482, found 309.1472.
1H NMR (d4-methanol): δ (ppm) ) 8.39 (m, 2H), 7.84 (m, 2H), 7.54
(m, 2H), 4.74-4.60 (m, 4H), 3.77-3.55 (m, 8H), 3.55-3.43 (m, 2H),
3.32-3.21 (m, 2H), 2.16-1.90 (m, 4H), 1.90-1.74 (m, 4H), 1.32-1.24
(m, 8H). 13C NMR (d4-methanol): δ (ppm) ) 158.9, 158.1, 151.0,
148.9, 148.9, 139.2, 139.2, 130.6, 130.5, 124.8, 75.0, 73.0, 69.1, 60.2,
60.0, 47.8, 45.8, 45.7, 45.2, 41.0, 34.4, 33.0, 31.4, 28.2, 24.6, 24.5,
24.2, 24.1, 24.0. UV (methanol/phosphate buffer, from DAD spectra):
λmax ) 214, 269 nm.
An LCT Premier XE LC/TOF-MS system (Waters, Manchester,
U.K.) was used for the determination of accurate masses.
Preparative HPLC was carried out on a system consisting of two
pumps (Kronlab, Sinsheim, Germany), an A 0293 variable-wavelength
monitor (Knauer, Berlin, Germany), a C-R3A integrator (Shimadzu,
Duisburg, Germany), and a YMC HPLC column (guard column 50
mm × 20 mm; column, 250 mm × 20 mm; YMC-Pack ODS-A, 10
µm); a gradient system consisting of ammonium formate buffer (10
mM, pH 4.0) and methanol as eluents (flow rate ) 20 mL/min), with
gradient as above, was used. Pump flow and gradient were controlled
by Prepcon software (SCPA GmbH, Weyhe-Leeste, Germany).
High-speed countercurrent chromatography (HSCCC) was performed
with a model CCC-1000 (Pharma-Tech Research Corp., Baltimore,
Maryland) connected to a A 0293 variable wavelength monitor
(detection at 270 nm, Knauer, Berlin, Germany) and a multirange BD
8 recorder (Kipp & Zonen, Delft, Netherlands). The two-phase solvent
system consisted of water/ethyl acetate/n-hexane/1-butanol (10/8/1/0.5;
v/v/v/v) and was pumped into the column with a Knauer HPLC pump
64 equipped with a preparative pump head. The lower phase was used
as mobile phase and was introduced through the head toward the tail
with a flow rate of 1.5 mL/min. The revolution speed was set at 1000
rpm. Fractions of 5 mL were collected with a Retriever 500 (ISCO,
Lincoln, Nebraska).
Ultraviolet spectra were measured from 190 to 400 nm with a Cary
1E spectrophotometer (Varian, Darmstadt, Germany) using quartz glass
cuvettes; scan rate: 60 nm/min; ave time: 0.1 s; data intervall: 0.1 nm.
1H and 13C nuclear magnetic resonance (NMR) spectra were obtained
using a Unity Inova 300 spectrometer (Varian, Darmstadt, Germany)
at 298 K at 300 and 75 MHz (nominal frequency), respectively, in
dimethyl-d6-sulfoxide (d6-DMSO) or d4-methanol. Chemical shifts are
given in δ (ppm) relative to trimethylsilane (TMS).
An infrared attenuated total reflectance (IR-ATR) spectrum was
recorded on a Fourier transform infrared (FT-IR) spectrometer Nicolet
Avatar 320 ESP (Thermo Electron, Dreieich, Germany).
Melting points were determined on a digital melting point apparatus
model 8100 (Electrothermal, Southend-on-Sea, U.K.) and are not
corrected.
For irradiation experiments in solutions, a metal halogen lamp (SOL
500, Dr. K. Ho¨nle GmbH, Martinsried, Germany; 120 000 lx, 900
W/m2) with a glass filter WG 295 (Schott Glaswerke, Mainz, Germany,
λ > 280 nm) was used. Irradiation on glass surfaces was performed
under a Suntest CPS + (Heraeus-Industrietechnik, Kleinostheim,
Germany) with a xenon lamp, UV filters (λ > 290 nm) and air cooling;
standard black temperature: 35 °C; irradiance: 250 W/m2.
Syntheses. Synthesis of 1-[(6-chloropyridin-3-yl)methyl]imidazoli-
din-2-imine (5) was performed following a modified procedure of
Hayakawa et al. (19). In a 100 mL screw-capped bottle, 400 mg of
imidacloprid (1.56 mmol) and 800 mg of stannous(II) chloride dihydrate
(3.54 mmol, analytical grade, Merck, Darmstadt, Germany) in 75 mL
of formic acid (98-100%, analytical grade, Merck, Darmstadt) were
heated at 100 °C for 4 h. After cooling, the reaction batch was
transferred into a 250 mL round-bottomed flask; 20 g of silica gel 60
(0.063-0.200 nm, for column chromatography, Merck, Darmstadt) was
added, and formic acid was distilled off on a rotary evaporator under
reduced pressure (40 mbar, 40 °C). The residue was suspended in 40
mL of ethyl acetate and transferred into a glass column (3 cm × 40
cm); the product was eluted with ethyl acetate/methanol (1/1, v/v).
Fractions of 10 mL were analyzed by TLC, and the ones containing 5
(Rf value ) 0.05) were combined. The residue obtained after evapora-
tion was recrystallized from ethanol to yield 180 mg of colorless crystals
(55% of theory; melting point 250 °C under decomposition). LC/MS
2-({1-[(6-chloropyridin-3-yl)methyl]imidazolidin-2-ylidene}amino)-
cyclohexanone (17); LC/MS (ESI+): m/z 307 [M + H]+; accurate mass
calculated for C15H20N4OCl ([M + H]+) 307.1326, found 307.1321.
1H NMR (d4-methanol): δ (ppm) ) 8.36 (m, 1H), 7.80 (m, 1H), 7.52
(m, 1H), 4.67-4.56 (m, 2H), 3.75-3.56 (m, 4H), 3.89 (m, 1H),
2.08-1.95 (m, 4H), 1.88-1.78 (m, 2H), 1.61-1.42 (m, 2H); 13C NMR
(d4-methanol): δ (ppm) ) 155.6, 151.2, 149.0, 139.3, 130.4, 124.8,
81.4, 51.3, 46.6, 44.6, 45.2, 39.0, 36.6, 27.0, 18.3. UV (methanol/
phosphate buffer, from DAD spectra): λmax ) 213, 269 nm.
Methyl-9,10-epoxystearate (105 mg, 0.336 mmol) and 5 (3 mg, 0.014
mmol) were dissolved in 5 mL of methanol and heated under reflux
for 24 h. The reaction mixture (10 µL) was directly injected into the
LC/MS system.
1
(ESI+): m/z 211 ([M + H]+). H NMR (d6-DMSO): δ (ppm) ) 8.39
(m, 1H), 7.81 (m, 1H), 7.56 (m, 1H), 4.56 (s, 2H), 3.34-3.53 (m, 4H).
13C NMR (DMSO-d6): δ (ppm) ) 160.3, 150.5, 150.0, 140.1, 131.6,
125.1, 47.7, 45.0, 41.1. IR (ATR, cm-1) 3347 (w), 3130 (sh), 3037
Irradiation experiments. Imidacloprid (15 mg, 0.058 mmol) were
dissolved in 50 mL of methanol, ethanol, 2-propanol, cyclohexane/
ethanol, or cyclohexene/ethanol (20 vol% ethanol to improve solubility).

8026 J. Agric. Food Chem., Vol. 56, No. 17, 2008
Schippers and Schwack
Table 1. Formation of the Photoproducts
1-[(6-Chloropyridin-3-yl)methyl]imidazolidin-2-imine (5) and
1-[(6-Chloropyridin-3-yl)methyl]imidazolidin-2-one (2) during UV Irradiation
of Imidacloprid (A, 7 h) and N-Nitroso-imidacloprid (B, 30 min) in Different
Model Solvents
A
B
solvent
methanol
ethanol
mol% 5a
mol% 2a
mol% 5a
mol% 2a
87.9 ( 1.2 3.6 ( 0.2
96.2 ( 1.7 5.8 ( 0.4
60.8 ( 5.2 6.2 ( 1.1
76.6 ( 0.7 6.4 ( 0.4
2-propanol
91.2 ( 8.2 6.6 ( 0.8 100.7 ( 8.2 <2 mol%b
cyclohexene/ethanol 95.4 ( 6.4 6.1 ( 0.2
(80/20, v/v)
cyclohexane/ethanol 66.7 ( 4.3 6.5 ( 1.4
(80/20, v/v)
86.4 ( 5.4 2.5 ( 1.1
94.0 ( 5.1 2.4 ( 0.5
Figure 1. Photodegradation of imidacloprid (1) (300 mg/L) in different
model solvents (b, methanol; 2, ethanol; 9, 2-propanol; 0, cyclohexane/
ethanol (80/20, v/v); O, cyclohexene/ethanol (80/20, v/v)).
a n g 3. b Below limit of quantification.
The solutions were purged with nitrogen for 2 min. Irradiation was
carried out in water-cooled 40 mL quartz cuvettes (60 mm × 35 mm)
with Teflon caps while stirring for up to 7 h using a metal halogen
lamp. For HPLC analysis and measurement of UV spectra, samples of
0.5 mL were taken in adequate intervals and evaporated under a stream
of nitrogen; the residue was dissolved in 5 mL of methanol.
Following the same procedure, methanolic solutions of 1-[(6-
chloropyridin-3-yl)methyl]imidazolidin-2-imine (5) (300 mg/L) were
irradiated.
Solutions of 1-[(6-chloropyridin-3-yl)methyl]imidazolidin-2-one oxo-
hydrazone (6) in the above-mentioned solvents (300 mg/L) were
irradiated in water-cooled quartz glass test tubes (12 mm × 100 mm)
for up to 20 min. Samples of 100 µL were taken, diluted with 900 µL
of methanol, and analyzed by HPLC.
For the experiments with solid imidacloprid, 500 µL of a methanolic
solution (2 g/L) was pipettied onto Petri dishes (diameter 9 cm, without
cover). The solvent was evaporated at ambient temperature releasing
1 mg of imidacloprid for each irradiation assay. Irradiations were
performed for 24 h in a suntest CPS+ sunlight simulator. For sample
preparation, the Petri dishes were rinsed with 10 mL of methanol, and
the resulting solutions were subjected to HPLC analysis.
Figure 2. UV spectra of imidacloprid solutions during irradiation (300
mg/L, methanol) for up to 300 min (isosbestic point at 227 nm).
moiety but no change of the chloropyridine ring, where
dechlorination could be expected (23).
In addition to HPLC measurements, UV spectra of irradiation
assays in methanol were recorded. The curves crossed in one
point at 227 nm (Figure 2). This isosbestic point could be seen
as a hint that the reaction is either first order or that an
intermediate product possesses a very short lifetime or a low
absorbance in the wavelength range recorded (24). Thus, the
formation of an intermediate with certain photostability seemed
unlikely.
A possible reaction mechanism was delivered by Chow et
al. (25) who carried out photolysis studies with N-nitrodialky-
lnitramines in different solvents such as methanol, n-hexane,
and acetonitrile in the presence and absence of cyclohexene.
These irradiation experiments resulted in the corresponding
amino, formamido, or nitroso compounds in different ratios
depending on the solvent and the reaction conditions. As in the
present study, photoaddition products of N-nitrodialkylnitra-
mines and cyclohexene were not obtained. The photolysis is
explained by Chow et al. (25) by a radical mechanism where
in an initial step homolytic cleavage of the N-N bond generates
nitrogen dioxide and the aminyl radical, which could undergo
further reactions. N-N bond cleavage can also be assumed
during the photolysis of 1 resulting in 5 by hydrogen transfer
from the organic solvents (Scheme 2). Such a mechanism was
also proposed by Moza et al. (4) for photolysis of 1 in water
but with a subsequent hydrolysis of 5 leading to the formation
of 2.
RESULTS AND DISCUSSION
Photodegradation of Imidacloprid. Irradiation of imida-
cloprid (1) in the presence of methanol, ethanol, 2-propanol,
cyclohexene, or cyclohexane resulted in nearly the same
degradation curves (Figure 1). After an irradiation time of 5 h,
1 was completely degraded in all solvents, whereas respective
solutions kept in the dark remained unchanged during the same
period of time. Previous studies on the photochemistry of
pesticides (e.g., parathion (11), folpet (21), or vinclozolin (12))
strongly showed different degradation rates and product distri-
butions depending on the solvent system used. Therefore, it was
rather surprising that the photodegradation rate of 1 was not
influenced by the solvent environment.
As the main photoproduct, 1-[(6-chloropyridin-3-yl)meth-
yl]imidazolidin-2-imine (5) could be isolated by preparative
1
HPLC and was identified by LC/MS and H- and 13C NMR
spectroscopy. It was formed up to 96 mol% of the degraded
imidacloprid (Table 1). Additionally, 1-[(6-chloropyridin-3-
yl)methyl]imidazolidin-2-one (2), which is already known as a
hydrolysis product especially under alkaline conditions (22), was
quantified as a minor photoproduct (Table 1).
Contrarily to the photochemical behavior of imidacloprid in
water (3-7), the photolysis in organic solvents is obviously less
complex, yielding 5 as main product that is not further
photodegraded. However, irradiations in both water (3-7) and
organic solvents only showed reactions at the nitroguanidine
To elucidate the reaction pathway of the formation of 2 and
to test if 2 is a photoproduct of 5, we irradiated solutions of 5

Photochemistry of Imidacloprid in Model Systems
J. Agric. Food Chem., Vol. 56, No. 17, 2008 8027
Figure 4. Extended section of an HPLC chromatogram (270 nm) of the
reaction batch of 1-[(6-chloropyridin-3-yl)methyl]imidazolidin-2-imine (2, 5
mg/mL) and cyclohexene oxide (0.25 mg/mL) in methanol, yielding the
addition products 16 and 17.
spectra as 5 were quantified at 270 nm and yielded 36 mol%
calculated and expressed as 5.
The irradiation assays were also analyzed by LC/MS, and
some compounds could be tentatively attributed to known
photoproducts. Both peaks eluting at 12.1 and 13.6 min delivered
protonated molecules of m/z 226 corresponding to the reported
oxidation products 13, 14, or 15 (Scheme 1) found after
irradiation of 1 in water (3, 4, 7). An [M + H]+ at m/z 243
under the peak at 7.5 min can be assigned to 12, another reported
photooxidation product (7). Cleavage of the imidazolidine ring
of 1 leads to the photoproducts 9 or 11, both with [M + H]+ at
m/z 171, appearing in the LC/MS chromatogram at 12.4 and
12.9 min. The protonated molecule of m/z 184, detected at the
retention time of 8.1 min, fits with compound 8 or 10, both
reported as photoproducts of 1 (4, 7). Only for the protonated
molecule at m/z 242, no photoproduct described in literature
could be assigned. A possible photoproduct yielding an [M +
H]+ at m/z 242 is the hydroxylamino derivative of 1 formed by
photoreduction of the nitro group. Summarizing, irradiation of
solid 1 delivered a totally different and more complex product
spectrum than irradiation in organic solvents.
Reaction of Photoproduct 5 with Epoxides. It is known
that amines react with epoxides by opening the oxirane ring
resulting in ꢀ-amino alcohols (27). After spiking tomato cuticles
with aminoparathion as a well-known photoproduct of parathion
and incubation in the dark, it could be shown that aminopar-
athion was covalently bound to the fruit cuticle (28). The
formation of nonextractable residues of aminoparathion was
explained by reactions with epoxides occurring in cutin acids
(14). In model studies with cyclohexene oxide and 9,10-
epoxystearic acid, the respective reactivity of aminoparathion
could be proved (28). Since during the present studies the imine
5 was found to be an important photoproduct of 1, the reactivity
of 5 toward epoxides was checked. During incubation of 5 and
cyclohexene oxide in methanol at ambient temperature in the
dark, 35% of 5 had disappeared after 14 days, and two products
could be detected by HPLC whereas the peak of compound 16
was slightly split up (Figure 4). LC/MS experiments revealed
the expected protonated molecule at m/z 309 for the main
product 16 throughout the double peak, whereas the minor
product 17 offered an [M + H]+ at m/z 307. Both products
showed the typical isotope pattern of singly chlorinated com-
pounds. As aminolysis of cyclohexene oxide should only deliver
the corresponding anti-amino alcohol (29), splitting up of the
main peak can only be attributed to the formation of E/Z isomers
Figure 3. Extended section of HPLC chromatograms (270 nm) of a
solution of imidacloprid (1) in cyclohexane/ethanol (300 mg/L) after 7 h
of irradiation (A) and an irradiation assay of solid 1 after 24 h (B). 5:
1-[(6-chloropyridin-3-yl)methyl]imidazolidin-2-imine; 2: (1-[(6-chloropyridin-
3-yl)methyl]imidazolidin-2-one.
in methanol. After 5 h of irradiation, 5 was not degraded at all,
and 2 could not be detected. Consequently, 5 is to be termed as
photostable under the chosen irradiation conditions.
A possible intermediate in the formation of 2 is N-nitrosoimi-
dacloprid (6). Nitrosamines are described as main photoproducts
of nitramines (25), and during irradiation of 1 on leaves of
tomato plants, Scholz and Reinhard (10) observed 6 as one of
four main photoproducts. But, during the irradiations of 1 in
our solvent systems, 6 could not be identified as photoproduct.
Therefore, 6 was synthesized, and irradiation experiments were
performed to determine the photoreactivity of this possible
photodegradation intermediate. Irradiations were carried out in
methanol, ethanol, 2-propanol, cyclohexane, and cyclohexene
under the same conditions used for 1. Photodegradation of 6
took place very fast with a half-life of about 3 min in all
solvents. Control samples stored in the dark for the same period
of time showed no degradation at all. Hence, 6 is rather short-
living under the chosen irradiation conditions. As in the case
of imidacloprid, 5 was formed as main product in amounts
between 61 and 100 mol% (Table 1), and 2 was found as minor
product. Loss of nitrogen can directly lead to 2 and serve as an
explanation for its formation pathway (Scheme 2). Other
photoproducts as the corresponding formamide as reported for
irradiated N-nitrosopiperidine (26) were not detected. Therefore,
during irradiation of 1, possibly formed 6 will immediately
degrade resulting in 2 and 5 and cannot be detected as
intermediate.
After spray application of imidacloprid in the field and
evaporation of the solvents, a solid residue will first remain on
fruit and leaf surfaces. Therefore, further irradiation experiments
were carried out with solid 1 on glass as a comparatively inert
surface to compare the behavior of solid and dissolved imida-
cloprid. After 24 h of irradiation, 1 was completely degraded.
Contrary to irradiation in solution, the formation of 2 and 5
occurred in rather low amounts (3.3 and 8.1 mol%, respectively),
but a variety of additional photoproducts was detected by HPLC
(Figure 3). All unknown photoproducts having similar UV

8028 J. Agric. Food Chem., Vol. 56, No. 17, 2008
Schippers and Schwack
ACKNOWLEDGMENT
Scheme 3. Reaction of 1-[(6-Chloropyridin-3-yl)methyl]imidazolidin-2-imine
(5) with Cyclohexene Oxide
We thank S. Mika (Institute of Chemistry, University of
Hohenheim) for recording the NMR spectra and Waters
(Manchester, U.K.) for providing the accurate masses by means
of LC/TOF-MS.
LITERATURE CITED
(1) Casida, J. E.; Quistad, G. B. Why insecticides are more toxic to
insects than people: the unique toxicology of insects. J. Pestic.
Sci. 2004, 29, 81–86.
(2) Imidacloprid. In The Pesticide Manual, Incorporating the Agro-
chemical Handbook, 10th ed.; Tomlin, C., Ed.; British Crop
Protection Council: Hampshire, U.K., 1994; pp 591-593.
(3) Wamhoff, H.; Schneider, V. Photodegradation of imidacloprid.
J. Agric. Food Chem. 1999, 47, 1730–1734.
(4) Moza, P. N.; Justert, K.; Feicht, E.; Kettrup, A. Photolysis of
imidacloprid in aqueous solution. Chemosphere 1997, 36, 497–
502.
(5) Aguera, A.; Almansa, E.; Malato, S.; Maldonado, M. I.; Fernan-
dez-Alba, A. R. Evaluation of photocatalytic degradation of
imidacloprid in industrial water by GC-MS and LC-MS. Analusis
1998, 26, 245–251.
(6) Kole, R. K.; Banerjee, H.; Bhattacharyya, A.; Chowdhury, A.;
Adityachaudhury, N. Phototransformation of some pesticides.
J. Indian Chem. Soc. 1999, 76, 595–600.
(7) Imidacloprid. In Metabolic Pathways of Agrochemicals, Part 2:
Insecticides and Fungicides; Roberts, T. R., Hutson, D. H., Eds.;
The Royal Society of Chemistry Information Services: Cambridge,
U.K., 1999; pp 111-120.
(8) Schmidt, B. Non-extractable residues of pesticides and xenobiotics
in plantssa review. Recent Res. DeV. Agric. Food Chem. 1999,
3, 329–354.
(9) Pesticide Residues in Food-EValuations 2002, Joint FAO/WHO
Meeting on Pesticide Residues; World Health Organisation, Food
and Agriculture Organization of the United Nations: Geneva,
Switzerland, 2002; Vol. 2, pp 687-1007.
(10) Scholz, K.; Reinhard, F. Photolysis of imidacloprid (NTN 33893) on
the leaf surface of tomato plants. Pestic. Sci. 1999, 55, 652–654.
(11) Schwack, W. Photoreduction of parathion-ethyl. Toxicol. EnViron.
Chem. 1987, 14, 63–72.
(12) Schwack, W.; Walker, F.; Bourgeois, B. Fungicides and photo-
chemistry: photodegradation of the dicarboximide fungicide
vinclozolin. J. Agric. Food Chem. 1995, 43, 3088–3092.
(13) Sinderhauf, K.; Schwack, W. Photolysis experiments on phosmet,
an organophosphorus insecticide. J. Agric. Food Chem. 2003, 51,
5990–5995.
at the carbon nitrogen double bond. Although this should also
hold true for the minor product 17, it could not be observed by
HPLC.
From a preparative assay, 16 was isolated by preparative
HPLC in an amount of about 50%. The determined exact
mass [M + H]+ of 309.1472 confirmed the molecular formula
of C15H21N4OCl corresponding to an [1 + 1] addition product
of 5 with cyclohexene oxide (Scheme 3). The NMR spectra of
the mixture clearly showed both signals of aromatic protons at
8.39, 7.84, and 7.54 ppm belonging to the pyridine part of the
molecule and signals of methylene protons in the area between
1.24 and 2.16 ppm and 3.21 and 3.55 ppm, which can be
attributed to the protons of the cyclohexane ring. The exact mass
of the minor product 17 of 307.1321 ([M + H]+) and the quite
similar NMR spectra as obtained for 16 support 17 as an
oxidation product of 16 (Scheme 3).
(14) Kolattukudy, P. E. Polyesters in higher plants. AdV. Biochem. Eng.
Biotechnol. 2001, 71, 1–49.
(15) Schynowski, F.; Schwack, W. Photochemistry of parathion on
plant surfaces: relationship between photodecomposition and
iodine number of the plant cuticle. Chemosphere 1996, 33, 2255–
2262.
(16) Wettach, J. W.; Rung, B.; Schwack, W. Detection of photochemi-
cally induced cuticle-bound residues of parathion by immunoassay.
Food Agric. Immunol. 2002, 14, 5–13.
(17) Rung, B.; Schwack, W. Aminoparathion: a highly reactive
metabolite of parathion. 1. Reactions with polyphenols and
polyphenol oxidase. J. Agric. Food Chem. 2005, 53, 9140–9145.
(18) Kagabu, S.; Yokoyama, K.; Iwaya, K.; Tanaka, M. Chloronicotinyl
insecticides. Part X. Imidacloprid and related compounds: structure
and water solubility of N-alkyl derivatives of imidacloprid. Biosci.
Biotechnol. Biochem. 1998, 62, 1216–1224.
(19) Hayakawa, T.; Fujiwara, Y.; Noguchi, J. A new method of
reducing nitroarginine-peptide into arginine-peptide, with reference
to the synthesis of poly(L-arginine hydrochloride). Bull. Chem.
Soc. Jpn. 1967, 40, 1205–1208.
The same reactivity of 5 was found in the presence of 9,10-
epoxystearic acid occurring as 18-hydroxy-9,10-epoxystearic
acid in several plant cuticles (14). LC/MS screening of incuba-
tion experiments with 5 and methyl 9,10-epoxystearate afforded
products with [M + H]+ at m/z 523 and 521 proving the
formation of an addition product that partly underwent oxidation
during incubation.
Concludingly, imidacloprid is rather susceptible to photo-
degradation independent upon the chemical environment, result-
ing in 5 as main product in organic solvents but without the
formation of photoaddition products with solvent molecules.
However, photoproduct 5 is quite reactive toward epoxidized
cutin acids supporting a possible pathway for the formation of
nonextractable residues of imidacloprid in plants, which were
described in the literature. Therefore, further research is needed
to transfer the results from model systems to plant surfaces.
(20) Rouchaud, J.; Gustin, F.; Wauters, A. Imidacloprid insecticide
soil metabolism in sugar beet field corps. Bull. EnViron. Contam.
Toxicol. 1996, 56, 29–36.

Photochemistry of Imidacloprid in Model Systems
J. Agric. Food Chem., Vol. 56, No. 17, 2008 8029
(21) Schwack, W. Photoinduzierte additionen von pesticiden an
biomoleku¨le, III. Modellreaktionen von folpet mit cyclohexen.
Z. Lebensm.-Unters. Forsch. 1990, 190, 420–424.
(27) Swern, D.; Findley, T. W. Chemistry of epoxy compounds. XIV.
Reaction of cis-9,10-epoxystearic acid with ammonia and amines.
J. Am. Chem. Soc. 1952, 74, 6139–6141.
(22) Zheng, W.; Liu, W. Kinetics and mechanism of the hydrolysis of
imidacloprid. Pestic. Sci. 1999, 55, 482–485.
(28) Rung, B. Bildung und Analytik gebundener Ru¨cksta¨nde des
Insektizids Parathion in pflanzlichen Lebensmitteln. Ph.D. Thesis,
University of Hohenheim, Stuttgart, Germany, 2003.
(29) Robinson, M. W. C.; Timms, D. A.; Williams, S. M.; Graham,
A. E. A mild and efficient synthesis of ꢀ-amino alcohols from
epoxides using a mesoporous aluminosilicate catalyst. Tetrahedron
Lett. 2007, 48, 6249–6251.
(23) Ohkura, K.; Seki, K.; Terashima, M.; Kanaoka, Y. Direct
photolysis of halopyridines in solutions; generation of the 2-pyr-
idyl cation. Chem. Pharm. Bull. 1991, 39, 3168–3169.
(24) Bertlett, B. S.; Levine, R. L.; Stadtman, E. R. Use of isosbestic
point wavelength shifts to estimate the fraction of a precursor
that is converted to a given product. Anal. Biochem. 2000, 287,
329–333.
(25) Chow, Y. L.; Richard, H.; Snyder, R. W.; Lockhart, R. W. Generation
of aminyl and aminium radicals by photolysis of N-nitrodialkylamines
in solution. Can. J. Chem. 1979, 57, 2936–2943.
Received for review April 21, 2008. Revised manuscript received July
8, 2008. Accepted July 8, 2008.
(26) Michejda, C. J.; Rydstrom, T. Photochemistry of N-nitrosamines
in neutral media. IARC Sci. Publ. 1984, 57, 365–369.
JF801251U