Mechanistic pathways of the photolysis of paracetamol in aqueous solution: An example of photo-Fries rearrangement

Article abstract of DOI:10.1039/c2pp25341k

The mechanism of the photolysis of N-(4-hydroxyphenyl)ethanamide (paracetamol, PA), a widely prescribed analgesic and antipyretic drug, has been investigated in the absence and in the presence of oxygen. Identification of products and kinetic analyses were performed by liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) and by ultra-performance liquid chromatography with a diode array detector (UPLC-PDA). The results show that, under irradiation at 254 nm and independently of the presence of oxygen, the predominant reaction pathway is a photo-Fries rearrangement (PFR), yielding the PA isomer 2′-amino-5′-hydroxyacetophenone (PAI). This reaction occurs from the singlet excited state of the molecule and involves the migration of the acetyl group onto the aromatic ring in the ortho-position to the amine moiety. The formation of 4-aminophenol (4-AP) was observed as a minor competitive pathway. The quantum yield of PA consumption (Φ_-PA) was determined to be 1.0(±0.1) × 10^-3 by chemical actinometry. As its concentration increases, the PFR product (PAI) competes with PA for light absorption and undergoes, in the presence of oxygen, a photooxygenation process leading to the formation of a peroxyester.

Full text of DOI:10.1039/c2pp25341k

Photochemical &
Photobiological Sciences
View Article Online
View Journal | View Issue
PAPER
Mechanistic pathways of the photolysis of paracetamol
in aqueous solution: an example of photo-Fries
rearrangement†
Cite this: Photochem. Photobiol. Sci., 2013,
12, 527
Marion Martignac,a Esther Oliveros,a Marie-Thérèse Maurette,a Catherine Claparolsb,c
and Florence Benoit-Marquié*a
The mechanism of the photolysis of N-(4-hydroxyphenyl)ethanamide (paracetamol, PA), a widely pre-
scribed analgesic and antipyretic drug, has been investigated in the absence and in the presence of
oxygen. Identification of products and kinetic analyses were performed by liquid chromatography
coupled to tandem mass spectrometry (LC-MS/MS) and by ultra-performance liquid chromatography
with a diode array detector (UPLC-PDA). The results show that, under irradiation at 254 nm and indepen-
dently of the presence of oxygen, the predominant reaction pathway is a photo-Fries rearrangement
(PFR), yielding the PA isomer 2’-amino-5’-hydroxyacetophenone (PAI). This reaction occurs from the
singlet excited state of the molecule and involves the migration of the acetyl group onto the aromatic
ring in the ortho-position to the amine moiety. The formation of 4-aminophenol (4-AP) was observed as
a minor competitive pathway. The quantum yield of PA consumption (Φ−PA) was determined to be
1.0( 0.1) × 10−3 by chemical actinometry. As its concentration increases, the PFR product (PAI) competes
with PA for light absorption and undergoes, in the presence of oxygen, a photooxygenation process
leading to the formation of a peroxyester.
Received 11th October 2012,
Accepted 16th November 2012
DOI: 10.1039/c2pp25341k
www.rsc.org/pps
achieves only their partial removal before release of the treated
water into the environment.5
1 Introduction
Detection of pharmaceuticals in the environment was pub-
In the last few decades, increasingly stringent regulations
lished for the first time in 1977.1 Since then, through the for the protection of human health and the preservation of the
development of more advanced analytical techniques such as environment have led to the development of Advanced
chromatography coupled to tandem mass spectrometry, the Oxidation Processes (AOPs) that have improved the efficiency of
persistence of a variety of drugs at very low concentrations oxidative degradation of toxic and non-biodegradable organic
(ng L−1) in surface, ground and drinking water could be compounds. Among the AOPs, photochemical technologies
demonstrated.2,3 This type of “Persistent Organic Pollutant based on the production of the hydroxyl radical, a powerful
(POP)” originates mainly from human and veterinary use as, oxidizing intermediate, have proven most effective for the treat-
after consumption, parent compounds are excreted unaltered ment of polluted water from diverse origins and containing a
and/or as metabolites in both urine and faeces.4 Their pres- variety of organic pollutants, among them pharmaceuticals.6–8
ence in water bodies has been principally related to the fact
N-(4-Hydroxyphenyl)acetamide, commonly called acetami-
that biological oxidation in Sewage Treatment Plants (STPs) nophen or paracetamol (PA), is an analgesic and antipyretic
widely used in human and veterinary medical treatments. The
top three drugs consumed in France contain paracetamol.9
It has been reported to be present at concentrations up to
aLaboratoire des Interactions Moléculaires et Réactivité Chimique et Photochimique
10 μg L−1 in natural waters in the USA, and even higher than
(IMRCP), UMR CNRS 5623, Université de Toulouse III (Paul Sabatier, UPS), 118,
65 μg L−1 in the Tyne River in the UK.10,11 Several studies have
route de Narbonne, F-31062 Toulouse Cedex 9, France.
proposed to use AOPs for the oxidative degradation of this
compound.12–15 Among the processes investigated, the combi-
nation of hydrogen peroxide with ultraviolet radiation (H2O2/
UV process)6,16 has received considerable attention and the
formation and fate of by-products have been investigated.13,14
However, to the best of our knowledge, the photolysis of PA
E-mail: florence@chimie.ups-tlse.fr
bCNRS, LCC (Laboratoire de Chimie de Coordination), 205 route de Narbonne,
BP 44099, F-31077 Toulouse Cedex 4, France
cService commun de spectrométrie de masse, Université de Toulouse, UPS, INPT,
F-31077 Toulouse Cedex 4, France
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c2pp25341k
This journal is © The Royal Society of Chemistry and Owner Societies 2013
Photochem. Photobiol. Sci., 2013, 12, 527–535 | 527

View Article Online
Paper
Photochemical & Photobiological Sciences
has not been studied, although this compound competes for the incident photons were absorbed by H2O2 ([H2O2]0 = 3.18 ×
light absorption at the wavelength generally used for the pho- 10−3 mol L−1, [A-AP]0 = 1.55 × 10−4 mol L−1). Conversion was
tolysis of H2O2 (254 nm, emission from commercial Hg arc kept low enough, so that further oxidation was minimized. Sol-
lamps). Moreover, UV irradiation with so-called germicidal utions were continuously bubbled with oxygen, air or argon
light sources (low pressure Hg lamps) is also used for disinfec- and the experiments were performed at a temperature of
tion of drinking water, as UV-C radiation (wavelength shorter 20( 1) °C. Samples (800 μL) were withdrawn from the reactor
than 280 nm) is harmful for microorganisms and may kill at regular intervals and analyzed by UV-Visible spectropho-
pathogens, such as bacteria and viruses.17,18 Therefore, it is of tometry, UPLC and liquid chromatography coupled with mass
particular interest to identify the intermediates formed during spectrometry (LC-MS/MS, § 2.4).
PA photolysis, as this process would also occur during photo-
chemical processes involving UV-C irradiation.
2.3 Actinometry and quantum yield determination
In the present work, we have investigated the mechanistic
pathways of the photolysis of PA at 254 nm in Ar- and in O2-
saturated aqueous solutions. Identification of products and
kinetic analyses have been performed using a combination of
analytical techniques: spectrophotometry, liquid chromato-
graphy coupled to tandem mass spectrometry (LC-MS/MS) and
ultra-performance liquid chromatography with a diode array
detector (UPLC-PDA). The main product of PA photolysis
has also been identified by hydroxylation of a commercial
compound using the UV/H2O2 process.
Uridine (1-β-D-ribofuranosyluracil) was used as an actino-
meter.19,20 The quantum yield of uridine photolysis at 254 nm
was taken as Φuridine = 0.018.21 The molar absorption coeffi-
cients of uridine (determined in this work) at 254 nm and
262 nm are 8520 L mol−1 cm−1 and 9970 L mol−1 cm−1
respectively (5% experimental error). To ensure total absorp-
tion of the radiation by the actinometric solution at 254 nm on
1 cm, a 2.37 × 10−4 mol L−1 solution of uridine was used. The
incident photon flux (P0) defined as the number of photons
entering the solution per unit of time and unit volume was
calculated from eqn (1). A value of 1.5 × 10−5 einstein L−1 s−1
(incident photon flux emitted by the lamp: 4.5 × 10−6 einstein
s−1) was obtained.
2 Experimental
2.1 Chemicals
ꢀd½uridineꢁ=dt ¼ Φuridine P0 ð1 ꢀ 10ꢀA
Þ
ð1Þ
uridine
N-(4-Hydroxyphenyl)ethanamide or paracetamol (PA, 98%),
4-aminophenol (4-AP, 98%), 2′-aminoacetophenone (2A-AP,
98%), 3′-aminoacetophenone (3A-AP, 98%), 1-β-D-ribofuranosy-
luracil or uridine (99%) and H2O2 (35% w/w in H2O) were pur-
chased from Sigma Aldrich and used without further
purification. LC-MS Chromasolv grade methanol, acetonitrile,
trifluoroacetic acid (TFA) and formic acid were purchased from
Fluka Analytical (Sigma Aldrich). The working solutions (pH =
5.5) were prepared with ultrapure water (Milli Q Reagent Water
System apparatus).
where −d[uridine]/dt is the initial rate of uridine disappear-
ance (mol L−1 −1), Φuridine is the quantum yield of uridine
s
photolysis at 254 nm, P0 is the incident photon flux (einstein
L−1 s−1) and Auridine is the absorbance of uridine at 262 nm.
The value of the quantum yield of PA consumption (Φ−PA
at low conversion (≤30%) was calculated using eqn (2).
)
Φ
¼ ðd½PAꢁ=dtÞ0=Pa
ð2Þ
ꢀPA
where (d[PA]/dt)0 is the initial rate of PA disappearance
(mol L−1 s−1) and Pa is the photon flux absorbed by PA
(einstein L−1 −1). The initial rate of PA disappearance was
s
2.2 Photochemical reactors
calculated to be 1.5 ( 0.1) × 10−8 mol L−1 s−1 ([PA]0 = 2.6 × 10−4
mol L−1).
The experiments were performed in an annular photochemical
reactor. The laboratory batch reactor had a capacity of 300 mL
(DEMA, Mangels, Bornheim-Roisdorf, Germany, internal
radius: 3.6 cm; external radius: 5.5 cm; height: 27 cm). It was
2.4 Analytical methods
equipped with a low pressure Hg lamp positioned in a quartz UV/Visible absorption spectra were registered on an HP-8452A
well in the axis of the reactor and emitting at 254 nm (electri- diode array single beam spectrophotometer from Hewlett
cal power: 5 W; radiant power: 1.2 W; UV-Consulting Peschl, Packard. Quartz cells of 1 cm or 1 mm optical path length were
Mainz, Germany). The initial concentration of paracetamol employed.
was 2.65 × 10−4 mol L−1
.
Chromatographic analyses of the reaction systems were per-
For the UV/H2O2 experiments, a given volume of an formed using an ACQUITY UPLC from Waters with a PDA
aqueous solution of H2O2 at the required concentration was detector, and an Acquity UPLC BEH C18 column (2.1 ×
introduced into the reactor containing the solution of 2A-AP or 100 mm, 1.7 μm, Waters). The optimized mobile phase for PA
3A-AP just before starting the irradiation. 2A-AP, 3A-AP and consisted of a mixture of two solutions: 98% of a 0.1% TFA
H2O2 absorb at 254 nm, the molar absorption coefficients of aqueous solution (A) and 2% of acetonitrile containing 0.1%
2A-AP and 3A-AP (determined in this work) being much higher TFA (B). Runs were carried out in isocratic mode at a flow rate
(6930 L mol−1 cm−1 and 7290 L mol−1 cm−1, respectively, 5% of 0.3 mL min−1. In the case of aminoacetophenones, the flow
experimental error) than that of H2O2 (18.7 L mol−1 cm−1).16 rate was the same, but the eluent consisted of a mixture of
The amount of H2O2 used in this work was such that 5% of A and B at a ratio of 98/2 for only 2.5 min, the A/B ratio was
528 | Photochem. Photobiol. Sci., 2013, 12, 527–535
This journal is © The Royal Society of Chemistry and Owner Societies 2013

View Article Online
Photochemical & Photobiological Sciences
Paper
then changed to 65/35 within 1 min, maintained at this value
for 2.5 min, changed back to 98/2 within 0.5 min and kept con-
stant until the end of the run. The detection range of the PDA
detector was set between 190 nm and 500 nm. Samples of 3 μL
were injected into the UPLC system. Data acquisition was
handled by the Empower2 software.
Preparative chromatography was performed using an Autop-
urif Waters system with a PDA 2998 detector, an injector collec-
tor 1767 and a pump 2545. Beforehand, the 300 mL solution
was lyophilized (Lyoph Alpha 1–2, Fisher Bioblock Scientific)
and redissolved in 1.5 mL of methanol. Chromatographic con-
Table 1 Optimized parameters for the ionization of paracetamol (PA) and its
photolysis products (PAI, 4-AP and PAIO)
PA and PAIa
PAIOa
4-AP
Transition MRM
152/110
57
5
184/166
81
6
110/65
51
10
Declustering potential (V)
Entrance potential (V)
Collision energy (V)
21
11
28
a PAI: 2′-amino-5′-hydroxyacetophenone (2A5H-AP); PAIO: peroxyester
(Scheme 3).
ditions were optimized on an analytical column (XBridge C18
,
4.6 × 150 mm, 5 μm, Waters); the flow rate was 1.2 mL min−1
in isocratic mode with 95% of A and 5% of B. These conditions
were then transposed on a preparative column (XBridge C18,
19 × 150 mm, 5 μm, Waters), the flow rate was 20 mL min−1
,
the eluent consisted of a mixture of A and B at a ratio of 95/5
during 7 min, the A/B ratio was then changed to 100% of B
within 1 min, maintained at this value for 3 min, changed
back to 95/5 within 1 min and kept constant until the end of
the run. The detection range of the PDA detector was set
between 200 nm and 400 nm. Samples of 500 μL were injected
into the preparative system. Data acquisition was handled by
the MassLynx and FractionLynx softwares. 21 fractions were
collected and checked on the ACQUITY UPLC system with the
same method described above for the analysis of PA. The
solvent from the various fractions was evaporated by Speed-
Vacuum.
LC-electrospray ionization (ESI) – MS/MS analyses were
carried out on an instrument consisting of an Agilent 1100 LC
(Waters) equipped with a UV/visible detector and a Q-trap
Applied Biosystems mass spectrometer with an ESI source in
positive mode. Data acquisition was handled by the Analyst
1.4.2 software. A Symmetry C18 column (4.6 × 75 mm, 3.5 μm,
Waters) was used for HPLC analyses with a 0.1% formic acid
aqueous solution as the eluent (C) or pure methanol (eluent
D) at a flow rate of 0.3 mL min−1. The mobile phase gradient
was changed as follows. Elution was started with a volumetric
ratio C/D of 90/10, changed to 20/80 after 5 min and to 90/10
at 5.5 min and held for 9 min. The UV/visible detector was set
at 254 nm. The mass spectrometer was operated in a positive
ionization mode. The source temperature was set at 400 °C
and the ionization source at 5 kV. MS analyses were performed
using extracted ion current (XIC) and multiple reaction moni-
toring (MRM) of specific product and parent ions (tandem
mass spectrometry). The parameters for the MRM mode were
optimized for each compound (Table 1).
Fig. 1 Evolution of the absorption spectrum of an Ar-saturated solution of PA
during 3 hours of irradiation at 254 nm ([PA]0 = 2.65 × 10−4 mol L−1); PA absorp-
tion spectrum in bold; arrows indicate increase or decrease of corresponding
absorption bands.
cm−1, respectively (experimental error of 5%). Irradiation of an
Ar-saturated solution at 254 nm induced significant changes
in the absorption spectrum (Fig. 1). Evolution of the latter
during irradiation shows two distinct isosbestic points at
208 nm and 240 nm, suggesting that PA is transformed into
one main product under these conditions. The solution
changed from colorless to yellow due to a new absorption
band centered at 380 nm and extending into the visible spec-
tral range.
The disappearance of PA during irradiation was followed by
liquid chromatography coupled to mass spectrometry (LC-MS/
MS), using the multiple reaction monitoring (MRM) mode.
The concentration of PA was calculated from the peak area of
the MRM transition 152/110, between the parent ion at 152
(MH+ of PA) and the fragment ion at 110 (loss of the acetyl
group).
3 Results and discussion
3.1 Photolysis of paracetamol in the absence of oxygen
The major product formed during the photolysis of PA in
an Ar-saturated solution had the same MRM transition 152/
110 as PA, but a different retention time on the chromatogram.
The same MRM transition means that this product is a con-
figuration isomer of PA (PAI). The minor product of PA photo-
lysis was identified by UPLC-PDA and mass spectrometry as
The absorption spectrum of paracetamol (PA) in aqueous sol-
ution shows two absorption bands centered at 196 nm and
244 nm (Fig. 1), with molar absorption coefficients (deter-
mined in this work) of 13 660 L mol−1 cm−1 and 9680 L mol−1
This journal is © The Royal Society of Chemistry and Owner Societies 2013
Photochem. Photobiol. Sci., 2013, 12, 527–535 | 529

View Article Online
Paper
Photochemical & Photobiological Sciences
4-aminophenol (4-AP) by comparison to a commercial stan- absorption spectrum after 7 hours, in contrast to Ar-saturated
dard. Fig. 2 represents the evolution of the concentrations of solutions (Fig. 3b and S1†). This result suggests that, in
PA, its isomer (PAI) and 4-AP during irradiation at 254 nm in the presence of O2, the reaction proceeds to yield second-
Ar-saturated solutions. Under these conditions, the isomer ary products. This was confirmed by LC-MS/MS measure-
concentration increased during irradiation time. After 5 hours ments, showing that the PAI concentration reached
a
of irradiation, the conversion rate of PA was approx. 80% with maximum value after 3 hours of irradiation, and then
formation of 73% of PAI and 6% of 4-AP.
decreased (Fig. 4).
The product of PAI oxidation (PAIO) was analyzed by mass
spectrometry. Its molecular mass (183 g mol−1) is equal to that
of PA (and PAI) (151 g mol−1) plus 32 g mol−1. The concen-
tration of PAIO was followed by measuring the peak area of the
MRM transition 184/166, between the parent ion at 184 (MH+
of PAIO) and the fragment ion at 166 (loss of H2O). Fig. 4 con-
firms that PAI underwent an oxygenation process in a second-
ary reaction step under UV irradiation in the presence of O2:
the PAIO formation was shifted at longer times than that of
3.2 Photolysis of paracetamol in the presence of oxygen
When the PA aqueous solution was saturated with O2 and irra-
diated, changes in the absorption spectrum were similar to
those in an Ar-saturated solution (identical isosbestic points)
up to 3 hours of irradiation at 254 nm. PA disappearance fol-
lowed the same kinetics within experimental error (Fig. 3a).
Under prolonged irradiation, O2-saturated solutions of PA
became light brown and no distinct band was present in the
Fig. 4 Evolution of the relative concentrations of PA (triangles), of its isomer
PAI (circles) and of the PAI oxygenation product (squares) during irradiation at
254 nm in an O2-saturated aqueous solution; [PA]0 = 2.65 × 10−4 mol L−1; analy-
sis by LC-MS/MS in the MRM mode: transitions 152/110 (PA and PAI) and 184/
166 (PAIO).
Fig. 2 Consumption of PA (triangles) and formation of its isomer PAI (dots)
and of 4-AP (diamonds) during irradiation at 254 nm in an Ar-saturated
aqueous solution; [PA]0 = 2.65 × 10−4 mol L−1; analysis by LC-MS/MS in the
MRM mode: transitions 152/110 (PA and PAI) and 110/65 (4-AP).
Fig. 3 (a) Evolution of the relative concentration of PA in Ar-saturated (filled triangles) and in O2-saturated (open circles) solutions during irradiation at 254 nm. (b)
Comparison of the absorption spectra of PA after 7 hours of irradiation in Ar-saturated (full line) and O2-saturated (dotted line) solutions; [PA]0 = 2.65 × 10−4 mol
L
−1; analysis by LC-MS/MS in the MRM mode for the transition 152/110.
530 | Photochem. Photobiol. Sci., 2013, 12, 527–535
This journal is © The Royal Society of Chemistry and Owner Societies 2013

View Article Online
Photochemical & Photobiological Sciences
Paper
Fig. 5 (a) Evolution of the concentration of 4-AP in Ar-saturated (filled diamonds) and in O2-saturated (open diamonds) solutions of PA under irradiation at
254 nm. (b) XIC (extracted ion chromatogram) of 110 (MH+ of 4-AP, retention time: 2.4 min) after 7 hours of PA irradiation in Ar- and O2-saturated aqueous sol-
utions; [PA]0 = 2.65 × 10−4 mol L−1; analysis by LC-MS/MS in the MRM mode for the transition 110/65.
PAI and reached its maximum when the PAI concentration had
significantly decreased (Fig. 4).
Formation of 4-AP was also observed in O2-saturated
solutions. The kinetics of its formation at the beginning of
the irradiation time were the same in Ar-saturated and O2-
saturated solutions (Fig. 5a). After 7 hours of irradiation at
254 nm, 4-AP was less concentrated in the presence of O2 than
in its absence (Fig. 5b); a maximum concentration of 4-AP of
6.2 × 10−6 mol L−1 (2.3% of [PA]0) was achieved after 3 hours of
irradiation, then the 4-AP concentration decreased. This com-
pound is known for its poor stability in the presence of oxygen
and undergoes an autooxidation process in the dark, in par-
ticular at alkaline pH.22,23 A product with a molecular mass of
107 g mol−1 was detected by mass spectrometry (ESI, positive
ionization, MH+ = 108); such a mass, corresponding to that of
4-AP minus 2, may be attributed to the quinonoid structure
Scheme 1 Photo-Fries rearrangement (PFR): (a) products formed by photolysis
formed by autooxidation of 4-AP.23 This reaction may explain
of catechol monoacetate,24 (b) of acetanilide34 and (c) potential product of PA
the yellow-brownish color of the solution at long irradiation
photolysis.
times.
3.3 Identification of the configuration isomer formed during
photolysis of paracetamol
cleavage of the C–O bond in the α-position to the carbonyl
group (α-cleavage, Norrish type I reaction)28,29 and migration of
Taking into account the structure of PA and the absence of the the acyl group to the ortho- and/or para-positions on the aro-
O2 effect on the formation of its configuration isomer, the matic ring (not meta due to electronic factors).30 Although there
hypothesis of a photo-Fries rearrangement (PFR) occurring are much less literature reports on PFR of aromatic amides (ani-
from the singlet excited state of PA may be formulated. This lides) than on analogous esters,27 a few early studies on acetani-
type of rearrangement was first observed by Anderson and lide have shown that, similarly to phenyl acetate, acetanilide
Reese in 1960,24,25 when they reported that 2-hydroxyphenylace- may undergo PFR with cleavage of the C–N bond and migration
tate rearranged to 2,3- and 3,4-dihydroxyacetophenone upon UV of the acetyl group.31–36 A mixture of ortho and para isomers (2-
irradiation in ethanol (Scheme 1a). Similarly, Kobsa26 showed amino and 4-aminoacetophenone) was obtained (Scheme 1b).34
that the photolysis of p-(tert-butyl) phenyl esters of aromatic
In the case of PA, due to the presence of a hydroxyl substitu-
carboxylic acids gave the corresponding o-hydroxyphenyl ent in the para position of the amide moiety, a PFR should
ketones. Such rearrangements were later investigated for a large lead to the formation of only one rearrangement product:
variety of phenyl esters in organic solvents.27 They imply the 2′-amino-5′-hydroxyacetophenone (2A5H-AP) (Scheme 1c). To
This journal is © The Royal Society of Chemistry and Owner Societies 2013
Photochem. Photobiol. Sci., 2013, 12, 527–535 | 531

View Article Online
Paper
Photochemical & Photobiological Sciences
the best of our knowledge, this compound is not commercially The latter are efficiently trapped by O2 and the peroxy radicals
available for comparison with PAI. Nevertheless, the corre- thus formed undergo re-aromatization to hydroxylated deriva-
sponding non-hydroxylated derivative, 2′-aminoacetophenone tives by elimination of HO2˙ (hydroperoxyl radical) (Scheme 2).
(2A-AP), could be purchased. 2A5H-AP might be obtained from
2A-AP by hydroxylation of the aromatic ring in the para posi- ents on 2A-AP, electrophilic addition of HO˙ on the ortho and
tion to the amino group. para positions relative to –NH2 should be favored (activation by
Taking into account the electronic effects of the substitu-
Hydroxylation of 2A-AP was performed in aqueous solution –NH2 of these positions that are also less deactivated by the
using the UV/H2O2 process (irradiation at 254 nm) (§ 2.2). In electron attracting effect of –COCH3),38 thus yielding 2A5H-AP
this process, photolysis of H2O2 results in the homolytic clea- as one of the main products.
vage of the oxygen–oxygen bond, producing hydroxyl radicals
In fact, PAI and the main product obtained by the hydroxy-
(HO˙) with a quantum yield of 1.16 The HO˙ radical is a highly lation of 2A-AP were found to be the same compound. This
reactive species, and the first step of its reaction with aromatic conclusion was supported by the results of LC-MS/MS and
compounds is generally electrophilic addition to the aromatic UPLC-PDA analyses (same retention time, same mass spec-
ring yielding intermediate hydroxycyclohexadienyl radicals.37 trum and same UV-visible absorption spectrum) (Fig. 6). The
only possible common product resulting from the PFR of PA
with migration of the acetyl group to the ortho-position on the
aromatic ring (Scheme 1c) and from the hydroxylation of
2A-AP on the para-position to the –NH2 (Scheme 2) is 2A5H-AP
(=PAI). The difference in the absorption spectra observed in
Fig. 3b (PAI, main product of PA photolysis under Ar) and in
Fig. 6 (chromatographic peak of PAI) is due to a pH difference:
water at natural pH in the former case and elution solvent
(pH approx. 2) in the latter case. A control experiment using
preparative HPLC (§ 2.4) confirmed this pH effect on the
absorption spectrum (Fig. S2†).
It could be confirmed that PFR of PA involving migration of
the acetyl group to the meta position on the aromatic ring did
not occur. Indeed, hydroxylation of 3′-aminoacetophenone
yielded a mixture of hydroxylated products (because of oppo-
site electronic effects of the two substituents, –NH2 and
–COCH3) but none of these compounds could be identified as
a product of PA photolysis.
It should be noted that the same secondary product as
found during PA photolysis in the presence of O2 (PAIO) was
Scheme 2 Hydroxylation of 2’-aminoacetophenone (2A-AP) by addition of
HO˙ in the para position to the amino group in the presence of oxygen.
Fig. 6 UPLC-PDA chromatograms (left) and absorption spectra corresponding to the different peaks observed (right): (a) chromatogram obtained after photolysis
of PA ([PA]0 = 2.65 × 10−4 mol L−1) during 3 h; (b) chromatogram obtained after hydroxylation of 2’-aminoacetophenone ([2A-AP]0 = 1.55 × 10−4 mol L−1) by the
H2O2/UV process during 15 min; O2-saturated aqueous solutions; irradiation at 254 nm.
532 | Photochem. Photobiol. Sci., 2013, 12, 527–535
This journal is © The Royal Society of Chemistry and Owner Societies 2013

View Article Online
Photochemical & Photobiological Sciences
Paper
observed during the hydroxylation of 2A-AP to 2A5H-AP (=PAI),
thus confirming that PAIO comes from the oxygenation of PAI
during the PFR of PA in the presence of O2 (Fig. 6).
3.4 Mechanistic pathways of the photolysis of paracetamol
Mechanism in the absence of oxygen. We have shown that
the excitation of PA at 254 nm in de-aerated aqueous solution
leads to a photo-Fries rearrangement with formation of
2′-amino-5-hydroxyacetophenone (2A5H-AP) as the main product
(§ 3.3). The lack of sensitivity of PA disappearance to the pres-
ence of oxygen (Fig. 3a) confirms that the reaction occurs from
the singlet excited state of the molecule, as is most often
observed in PFRs.
The value of the quantum yield of PA consumption (Φ−PA
)
,
up to 30% conversion was calculated to be 1.0 ( 0.1) × 10−3
from the rate of PA disappearance (Fig. 2 and 3a) and the
photon flux absorbed by PA determined by actinometry (§ 2.3).
This low value is in agreement with earlier measurements for
acetanilide showing that the quantum yield for the PFR with
ortho migration of the acetyl group was 0.004 in water, 4 times
lower than in ethanol and more than 15 times lower than in
cyclohexane.34 The OH substituent in the para position in PA
further decreases its reactivity compared to acetanilide. These
results confirm the much lower photoreactivity of anilides
compared to that of analogous phenyl esters.
The accepted PFR mechanism for aromatic esters involves
the homolytic scission of the O–CO bond to give a radical pair
(phenoxy and acyl radicals) enclosed in a solvent cage, fol-
lowed by an intramolecular recombination with migration of
the acyl group to form intermediate cyclohexadienones that
tautomerize to the final products.27 A similar mechanism may
explain the formation of 2A5H-AP by PFR of PA. The homolytic
cleavage of the N–CO bond leads to the formation of a cyclo-
hexadieneimine intermediate I that undergoes a 1,3-H shift to
yield PAI (Scheme 3).
Scheme 3 Mechanistic pathways of PA photolysis at 254 nm in the absence
and presence of oxygen.
The formation of 4-aminophenol (4-AP) represents a minor
competitive pathway. The origin of the corresponding phenol
observed in the case of phenyl esters was attributed to the out-
of-cage escape of the phenoxy radical to form phenol by hydro-
gen abstraction from the organic solvent.27 However, in the
case of PA in aqueous solution, hydrogen abstraction from
water by the Ph–(H)N˙ radical is unlikely but an in-cage redox
reaction (H-transfer) between the two radicals could explain
the formation of 4-AP with low yield (Scheme 3). The reported
yield of aniline (18%) in the case of the photolysis of acetani-
lide in absolute ethanol in the absence of oxygen supports this
interpretation.32
the carbonyl group (Norrish type I cleavage) has been reported
to occur from n,π* or π,π* excited states of aromatic (and ali-
phatic) ketones.28,29,39 PAI being an acetophenone derivative,
cleavage of the C–CH3 bond, weaker than the C–Ph bond, is
expected.40 In aqueous solution (no source of hydrogen atoms)
and in the absence of oxygen, the radical pair recombines and
no product formation is observed. However, in the presence of
oxygen, the radical(s) may be trapped by oxygen in the solvent
cage and subsequent recombination may lead to the formation
of a peroxyester (PAIO, molecular mass of 183 g mol−1
(Scheme 3).
)
Mechanism in the presence of oxygen. As already discussed
(§ 3.2), the presence of oxygen does not affect the PFR of PA to
yield PAI (2A5H-AP), but PAI undergoes an oxygenation
4 Conclusion
process leading to the formation of PAIO (Fig. 4). Since PAI We have investigated the photolysis of paracetamol (PA) at
absorbs at the irradiation wavelength (254 nm) as does PA 254 nm in aqueous solution and we have elucidated the main
(Fig. 1), the fraction of photons absorbed by PAI increases as mechanistic pathways following excitation of PA in the pres-
its concentration increases and that of PA decreases. There- ence and in the absence of oxygen (Scheme 3). Independently
fore, PAI itself may undergo a photochemical transformation. of the presence of O2, the major reaction product is a PA
The homolytic cleavage of the C–C bonds in the α-position to isomer (PAI), identified as 2′-amino-5′-hydroxyacetophenone
This journal is © The Royal Society of Chemistry and Owner Societies 2013
Photochem. Photobiol. Sci., 2013, 12, 527–535 | 533

View Article Online
Paper
Photochemical & Photobiological Sciences
(2A5H-AP) by combining different analytical techniques. This
compound results from a photo-Fries rearrangement (PFR),
involving cleavage of the C–N bond in the α-position to the car-
bonyl group (α-cleavage, Norrish type I reaction) and migration
of the acetyl group to the ortho-position relative to the amino
group on the aromatic ring. This reaction proceeds with a low
quantum yield, as expected from a few earlier studies on the
PFR of acetanilide.
and salicylic acid in sewage water effluent, Life Sci., 1977,
20(2), 337–341.
2 T. A. Ternes, Occurrence of drugs in German sewage
treatment plants and rivers, Water Res., 1998, 32(11),
3245–3260.
3 A. Togola and H. Budzinski, Multi-residue analysis of
pharmaceutical compounds in aqueous samples, J. Chro-
matogr., A, 2008, 1177(1), 150–158.
In a minor pathway, 4-aminophenol (4-AP) was formed with
a yield more than 10 times lower than the PFR product (PAI),
most probably through a cage disproportionation of the two
radicals resulting from the initial cleavage of the C–N bond. In
the presence of oxygen, 4-AP underwent an autooxidation
reaction.
The photo-Fries rearrangement of PA to yield PAI is not
affected by the presence of O2. However, in O2-saturated sol-
utions, the formation of a secondary product (PAIO) resulting
from the oxygenation of PAI was observed under prolonged
irradiation. This result was explained by the photochemical
reactivity of PAI. Indeed, as its concentration increases, PAI
competes more and more efficiently with PA for light absorp-
tion and, as an acetophenone derivative, may undergo cleavage
of the C–O bond in the α-position to the carbonyl group
(α-cleavage). In the absence of O2, the radicals formed recom-
bine and PAI is stable. However, in O2-saturated solutions, the
radical(s) may be trapped by O2, leading to the formation of a
peroxyester (PAIO).
4 T. Heberer, U. Dünnbier, C. Reilich and H. J. Stan, Detec-
tion of drugs and drug metabolites in groundwater
samples of a drinking water treatment plant, Fresenius
Environ. Bull., 1997, 6, 438–443.
5 A. Piram, A. Salvador, J.-Y. Gauvrit, P. Lanteri and R. Faure,
Development and optimisation of a single extraction pro-
cedure for the LC/MS/MS analysis of two pharmaceutical
classes residues in sewage treatment plant, Talanta, 2008,
74(5), 1463–1475.
6 O. Legrini, E. Oliveros and A. M. Braun, Photochemical
processes for water treatment, Chem. Rev., 1993, 93(2),
671–698.
7 C. von Sonntag, Advanced oxidation processes:
mechanistic aspects, Water Sci. Technol., 2008, 58(5),
1015–1021.
8 P. R. Gogate and A. B. Pandit, A review of imperative tech-
nologies for wastewater treatment I: oxidation technologies
at ambient conditions, Adv. Environ. Res., 2004, 8(3–4),
501–551.
It should be noted that, although the quantum yield of PA
photolysis is low, both reaction products (major and minor)
are aromatic amines (anilines). These compounds are known
to be toxic, in particular for fish41 and also for humans,42 and
9 AFSSAPS (Agence Française de Sécurité Sanitaire des pro-
duits de Santé) expert’s report: Analyse des ventes de médi-
caments aux officines et aux hôpitaux en France
1999–2009, Juillet 2011, 2011.
UV-C (254 nm) disinfection under prolonged irradiation in the 10 D. W. Kolpin, E. T. Furlong, M. T. Meyer, E. M. Thurman,
presence of PA may induce their formation. More generally,
this raises the question of the use of 254 nm irradiation alone
for disinfection of water containing organic aromatic residues
that absorb at this wavelength (such as drugs or pesticides).
S. D. Zaugg, L. B. Barber and H. T. Buxton, Pharmaceuti-
cals, hormones, and other organic wastewater contaminants
in U.S. streams, 1999–2000 – a national reconnaissance,
Environ. Sci. Technol., 2002, 36(6), 1202–1211.
Under these conditions, the use of an adequate amount of 11 P. H. Roberts and K. V. Thomas, The occurrence of selected
additive absorbing at 254 nm (e.g. H2O2, O3, TiO2) for produ-
cing hydroxyl radicals and ensuring pollutant mineralization
to CO2, H2O and inorganic acids would be required.
pharmaceuticals in wastewater effluent and surface waters
of the lower Tyne catchment, Sci. Total Environ., 2006,
356(1–3), 143–153.
12 X. Zhang, F. Wu, X.-W. Wu, P. Chen and N. Deny, Photode-
gradation of acetaminophen in TiO2 suspended solution,
J. Hazard. Mater., 2008, 157(2–3), 300–307.
Acknowledgements
13 D. Vogna, R. Marotta, A. Napolitano and M. d’Ischia,
Advanced oxidation chemistry of paracetamol. UV/H2O2-
induced hydroxylation/degradation pathways and 15N-
aided inventory of nitrogenous breakdown products, J. Org.
Chem., 2002, 67(17), 6143–6151.
14 R. Andreozzi, V. Caprio, R. Marotta and D. Vogna, Paraceta-
mol oxidation from aqueous solutions by means of
ozonation and H2O2/UV system, Water Res., 2003, 37(5),
993–1004.
M. M. thanks the CNRS, the “Conseil Régional Midi-Pyrénées”
and the “PME” Loïra (Derichebourg Aqua) for a doctoral grant.
The authors also thank Chantal Zedde (research engineer at
the “Service commun de HPLC de l’Institut Chimie de Tou-
louse” (FR2599), Université Paul Sabatier, Toulouse, France)
for her valuable technical support and helpful discussions of
UPLC experiments.
15 L. Yang, L. E. Yu and M. B. Ray, Photocatalytic oxidation
of paracetamol: dominant reactant intermediates and
reaction mechanisms, Environ. Sci. Technol., 2009, 43(2),
460–465.
References
1 C. Hignite and D. L. Aznaroff, Drugs and drug metabolites
as environmental contaminants: chlorophenoxyisobutyrate
534 | Photochem. Photobiol. Sci., 2013, 12, 527–535
This journal is © The Royal Society of Chemistry and Owner Societies 2013

View Article Online
Photochemical & Photobiological Sciences
Paper
16 J. R. Bolton, S. R. Cater, G. R. Helz and R. G. Zepp, 29 J. C. Scaiano, K. G. Stamplecoskie and G. L. Hallett-Tapley,
Homogeneous photodegradation of pollutants in contami-
nated waters, in Aquatic and Surface Photochemistry, ed.
D. G. Crosby, Lewis Publishers, Boca Raton, FL, 1994,
pp. 467–490.
Photochemical Norrish type I reaction as a tool for metal
nanoparticule synthesis: importance of proton coupled
electron transfer, Chem. Commun., 2012, 48(40), 4798–4808.
30 T. J. Stone and W. A. Waters, Aryloxy radicals. III. Electron
spin resonance of radicals from some substituted resorci-
nols, J. Chem. Soc., 1964, 4302–4307. and references
therein.
17 R. L. Wolfe, Ultraviolet disinfection of potable water,
Environ. Sci. Technol., 1990, 24(6), 768–773.
18 C. von Sonntag and H.-P. Schuchmann, UV disinfection of
drinking water and by-product formation – some basic con- 31 D. Elad, Photochemical studies. IV. The photochemical
siderations, J. Water SRT – Aqua, 1992, 41(2), 67–74.
19 H. Görner, Chromophore loss of uracil derivatives and poly-
rearrangement of N-acylanilines, Tetrahedron Lett., 1963
(14), 873–875.
uridylic acid in aqueous solution caused by 248 nm laser 32 D. Elad, D. V. Rao and V. I. Stenberg, Photoanilide
pulses and continuous UV irradiation: mechanism of the rearrangement, J. Org. Chem., 1965, 30(9), 3252–3254.
photohydration of pyrimidines, J. Photochem. Photobiol., B, 33 H. Shizuka and I. Tanaka, Photochemistry of
1991, 10, 91–110.
20 J. Y. Zhang, I. W. Boyd and H. Esrom, UV intensity
measurement for a novel 222 nm excimer lamp using
acetanilide. I. Quantum yields of the rearrangement and
benzene photosensitized reaction, Bull. Chem. Soc. Jpn.,
1968, 41(10), 2343–2349.
chemical actinometer, Appl. Surf. Sci., 1997, 109/110, 34 H. Shizuka and I. Tanaka, Photochemistry of acetanilide.
482–486.
II. The primary processes in the photochemical reaction,
21 H. J. Kuhn, S. E. Braslavsky and R. Schmidt, Chemical acti-
Bull. Chem. Soc. Jpn., 1969, 42(1), 52–57.
nometry (IUPAC technical report), Pure Appl. Chem., 2004, 35 H. Shizuka and I. Tanaka, Photochemistry of acetanilide.
76(12), 2105–2146.
III. The secondary processes in the photochemical reaction,
22 B. N. Chandrashekar, B. E. Kumara Swamy,
Bull. Chem. Soc. Jpn., 1969, 42(1), 57–65.
M. Pandurangachar, T. V. Sathisha and B. S. Sherigara, 36 J. S. Bradshaw, R. D. Knudsen and E. L. Loveridge, Photo-
Electrochemical investigation of 4-aminophenol at CTAB
modified carbon paste electrode: a cyclic voltammetric
technique, Anal. Bioanal. Electrochem., 2011, 3, 227–231.
23 R. Apak, S. D. Çekiç, A. Çetinkaya, H. Filik, M. Hayvah and
E. Kiliç, Selective determination of catechin among pheno-
reactions of 2,4-dimethoxyacetanilide, J. Org. Chem., 1970,
35(4), 1219–1221.
37 C. von Sonntag and H. P. Schuchmann, Peroxyl radicals in
aqueous solutions, in Peroxyl Radicals, ed. Z. B. Alfassi,
John Wiley and Sons, New York, 1997, pp. 173–234.
lic antioxidants with the use of a novel optical fiber reflec- 38 K. P. C. Vollhardt and N. Schore, Electronic Effects of Substi-
tance sensor based on indophenols dye formation on
tuents: Traité de Chimie Organique, DeBoeck University,
nano-sized TiO2, J. Agric. Food Chem., 2012, 60, 2769–2777.
Bruxelles, 2nd edn, 1995.
24 J. C. Anderson and C. B. Reese, Photo-induced Fries 39 M. C. Jimenez, P. Leal, M. A. Miranda and R. Tormos,
rearrangement, Proc. Chem. Soc., London, 1960, 217.
25 J. C. Anderson and C. B. Reese, The photochemical Fries
reaction, J. Chem. Soc., 1963, 1781–1784.
Norrish type I, photoreaction in presence of phenols; an
intramolecular photo-Fries rearrangement, J. Chem. Soc.,
Chem. Commun., 1995 (19), 2009–2010.
26 H. Kobsa, Rearrangement of aromatic esters by ultraviolet 40 H.-Q. Zhao, Y.-S. Cheung, C.-L. Liao, C.-X. Liao, C. Y. Ng
radiation, J. Org. Chem., 1962, 27, 2293–2298.
27 M. A. Miranda and F. Galindo, Photochemistry of organic
molecules in isotropic and anisotropic media, in Molecular
and W.-K. Li, A laser photofragmentation time-of-flight
mass spectrometric study of acetophenone at 193 and
248 nm, J. Phys. Chem., 1997, 107(18), 7230–7241.
and Supramolecular Photochemistry, ed. V. Ramamurthy and 41 S. Marchini, M. D. Hoglund, S. J. Broderius and
K. S. Shanze, 2003, ch. 2, vol. 9, pp. 43–132.
28 N. J. Turro, J. C. Dalton, K. Dawes, G. Farrington,
R. Hautala, D. Morton, M. Niemczyk and N. Schore, Mole-
M. L. Tosato, Comparison of the susceptibility of daphnids
and fish to benzene derivatives, Sci. Total Environ. Suppl.,
1993, 799–808.
cular photochemistry of alkanones in solution. α-cleavage, 42 J. V. Rodricks, Carcinogens, in Calculated Risks: The Toxicity
hydrogen abstraction, cycloaddition, and sensitization reac-
and Human Health Risks of Chemicals in Our Environment,
tions, Acc. Chem. Res., 1972, 5(3), 92–101.
Cambridge Univ. Press, 1992, ch. 7.
This journal is © The Royal Society of Chemistry and Owner Societies 2013
Photochem. Photobiol. Sci., 2013, 12, 527–535 | 535