Photo-degradation of trans-caffeic acid in aqueous solution and influence of complexation by metal ions

Article abstract of DOI:10.1016/j.jphotochem.2013.05.004

The photo-degradation of metal complexes of caffeic acid was compared to the photo-degradation of free caffeic acid by using UV-vis spectroscopy and HPLC-ESI-mass spectrometry. This article reports first the determination of the products that are formed from the photo-degradation of trans-caffeic acid in aqueous solution and the investigation of the mechanism by a kinetic approach. The good fit between the model and the experimental concentration profiles confirms the photo-isomerization route of the molecule to cis-caffeic acid which then undergoes a cyclization to form the esculetin photo-product. In addition, it reveals, for the first time, another route of major importance leading to the product vinylcatechol. The presence of oxygen leads to an increase of the photo-isomerization rate. Then we report that metallic cations such as Al(III), Pb(II) and Cu(II) can influence the rate and mechanism of caffeic acid photodegradation. Al(III) ions slow down the photo-degradation whereas Pb(II) and Cu(II) ions have a promoter effect on the production of esculetin. In all cases, the photo-isomerization is reduced by the presence of metal ions and the formation of vinylcatechol does not occur.

Full text of DOI:10.1016/j.jphotochem.2013.05.004

Journal of Photochemistry and Photobiology A: Chemistry 265 (2013) 10–19
Contents lists available at SciVerse ScienceDirect
Journal of Photochemistry and Photobiology A:
Chemistry
journal homepage: www.elsevier.com/locate/jphotochem
Photo-degradation of trans-caffeic acid in aqueous solution and influence of
complexation by metal ions
Annaïg Le Persona,∗, Anne-Sophie Lacosteb, Jean-Paul Cornarda
a Laboratoire de Spectrochimie Infrarouge et Raman (LASIR), CNRS UMR 8516, Université Lille 1, Sciences et Technologies, Bâtiment C5, Cité scientifique,
59655 Villeneuve d’Ascq Cedex, France
b Miniaturisation pour l’Analyse, la Synthèse et la Protéomique (MSAP), CNRS USR 3290, Université Lille 1, Sciences et Technologies, Cité scientifique, Bâtiment C4,
59655 Villeneuve d’Ascq Cedex, France
a r t i c l e i n f o
a b s t r a c t
Article history:
The photo-degradation of metal complexes of caffeic acid was compared to the photo-degradation of free
caffeic acid by using UV–vis spectroscopy and HPLC–ESI-mass spectrometry. This article reports first the
determination of the products that are formed from the photo-degradation of trans-caffeic acid in aqueous
solution and the investigation of the mechanism by a kinetic approach. The good fit between the model
and the experimental concentration profiles confirms the photo-isomerization route of the molecule to
cis-caffeic acid which then undergoes a cyclization to form the esculetin photo-product. In addition, it
reveals, for the first time, another route of major importance leading to the product vinylcatechol. The
presence of oxygen leads to an increase of the photo-isomerization rate. Then we report that metallic
cations such as Al(III), Pb(II) and Cu(II) can influence the rate and mechanism of caffeic acid photo-
degradation. Al(III) ions slow down the photo-degradation whereas Pb(II) and Cu(II) ions have a promoter
effect on the production of esculetin. In all cases, the photo-isomerization is reduced by the presence of
metal ions and the formation of vinylcatechol does not occur.
Received 13 November 2012
Received in revised form 12 April 2013
Accepted 11 May 2013
Available online 27 May 2013
Keywords:
Caffeic acid
Humic substances
Photo-degradation
Metal complex
Mechanism
Kinetics
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
This molecule also displays antioxidant properties [7] and is used
as photo-protective substance in preparations and sun-creams [8].
The organic matter of soils is formed from the degradation of
plants and animals and is mainly composed of humic substances.
These are aromatic and aliphatic macromolecules constituting a
heterogeneous and complex system [1,2] and playing an impor-
tant role in the retention and transport of many metal ions in
soils [3]. Most of metal ions are known to be toxic, in particu-
lar if they are in free form, and their accumulation is responsible
for many environmental concerns [4]. In general, their toxicity
and availability diminish with their retention in soils [5]. To reach
a better understanding of the interactions between metal ions
and humic substances (HS), the approach chosen for few years
in our research group consists in studying model systems of HS,
because of the complexity and the poly-functionality of these
substances. One possible model molecule is caffeic acid (trans-3-
(3,4-dihydroxyphenyl)prop-2-enoic acid), naturally omnipresent
in plant kingdom. When plants degradation occurs, this compound
is released and constitutes a precursor of HS [6]. Indeed, this
molecule exhibits two competing metal coordination sites, the car-
boxylic and the catechol groups that are typical functions in HS.
The complexation of metal ions such as Al(III) and Pb(II) by
caffeic acid has already been investigated by a double approach
combining spectroscopic measurements and quantum chemical
calculations [9–13]. The structures and stoichiometries of the
complexes and their equilibrium constants were obtained. The
complexation in solution of Al(III) [10] shows that the 1:1 complex,
formed via the catecholate group, largely predominates at pH = 5
for low amount of metal. Small quantities of metal ions also allow
the formation of a minor 2:1 binuclear complex, for which the two
sites (catechol and carboxylic functions) are involved. At pH = 6.5,
only the catecholate site coordinates to Al(III) leading to the for-
mation of 1:2 and 1:3 complexes. Thus, it was concluded that the
catechol group presents a chelating power which is higher than
that of carboxylic or carboxylate functions toward Al(III) ions. The
trend is different with Pb(II) since these ions coordinate first with
the carboxylate function [12].
Humic substances are also sensitive to solar light, leading to a
potential photo-degradation, as previously reported in the litera-
ture [14–17]. However little is known about the photochemistry
of their association with metal ions. To our knowledge, we can
only report few studies devoted to the photo-degradation of metal
complexes of organic molecules [18–20]. This paper focuses on the
photo-degradation of caffeic acid and on the influence of metal
∗
Corresponding author. Tel.: +33 3 20 43 49 14.
E-mail address: annaig.le-person@univ-lille1.fr (A. Le Person).
1010-6030/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jphotochem.2013.05.004

A. Le Person et al. / Journal of Photochemistry and Photobiology A: Chemistry 265 (2013) 10–19
11
ions on this process. Previous studies already suggested that solar
or UV irradiation could modify the behavior of this molecule in
the environment [21,22]. Depending on the experimental condi-
tions, different products and mechanisms were proposed for the
photo-degradation of caffeic acid but they are not all in accor-
dance. On the one hand, most of the authors [23–28] reported
cis-caffeic acid and esculetin (6,7-dihydroxy-2H-1-benzopyran-2-
one) species as products of the degradation. Despite they suggest
different intermediates, they agree about a mechanism based on
the initial isomerization of trans-caffeic to cis-caffeic acid which
then undergoes a cyclization to form esculetin. Nevertheless, this
mechanism was not confirmed by a kinetic study. On the other
hand, some authors [29,30] provided evidence of the formation
of various compounds such as 3,4-dihydroxybenzaldehyde, 3,4-
dihydroxybenzoic acid (or protocatechuic acid), maleic acid and
oxalic acid that do not support the previous mechanism.
In that context, the aim of this work is first to assess the products
that are formed from the photo-degradation of caffeic acid in water
without any added substance. Based on these products, a mecha-
nism and a kinetic model will be proposed. A good fit between the
calculated (with the kinetic model) and the experimental concen-
tration profiles will allow the validation of this mechanism.
Borges and Pinto [26] showed previously the accelerator role of
oxygen for the caffeic acid photo-degradation in ethanol. Conse-
quently, we propose to check this parameter in our experimental
conditions. Then the influence of the coordination to metal ions
such as Al(III), Pb(II) and Cu(II) on the photo-degradation of caffeic
acid is investigated. Both electronic spectroscopies and HPLC–ESI-
MS analyses are used to reach all these objectives.
1 cm path length. Fluorescence spectra were acquired using a Fluo-
rolog (Jobin-Yvon) spectrofluorimeter with slit width varying from
2 to 4 nm. This apparatus is equipped with multichannel detector
that allows the limitation of the recording times and consequently
of the illumination times. Three-dimensional spectra were also
acquired to obtain excitation–emission matrix (EEM) plots, where
the excitation wavelengths are plotted on the y-axis, the emission
wavelengths on the x-axis and the third dimension represents the
relative intensity.
HPLC–ESI-MS analyses were performed on a triple quadrupole
mass spectrometer (Quattro II Micromass-Waters) equipped with
an electrospray ionization source (ESI) coupled with high pressure
liquid chromatography (HPLC) system (HP1100 Agilent). Separa-
tions were achieved on a 250 mm × 2.0 mm i.d. column packed with
5 ␮m Kromasil C18 stationary phase (Interchim) protected by a
10 mm × 2.0 mm C18 precolumn (Interchim) and heated to 25 ◦C.
The flow rate was set at 150 ␮L/min and the injected volume was
20 ␮L. The mobile phase was a mixture of solvent A (0.5% formic
acid in water) and solvent B (acetonitrile). The proportion of solvent
B was increased linearly from 10% to 43% in 25 min, then 43% to 90%
in 5 min. After each injection the column was allowed to reequili-
brate with 10% solvent B for 13 min. The column eluent was first
directed to a UV detector set at 280 nm and then without splitting
to the electrospray interface. The mass spectrometer was operated
in negative ion mode. For phenolic compounds, this mode leads to
a better sensitivity and lower background noise than the positive
mode [31,32]. The source parameters were the following: capillary
voltage of – 3 kV, cone voltage of – 25 V, source temperature of
120 ◦C. Nitrogen was used as nebulization and drying gas at flow
rates of 10–15 and 250–300 L/h respectively. Data were acquired
in full scan MS mode over the range m/z 220–500 in 2 s.
2. Methods
2.1. Chemicals
2.4. Calculations
Trans-caffeic acid ((E)-3-(3,4-dihydroxyphenyl)prop-2-enoic
acid) was obtained from Sigma Aldrich (99%) and used as received
without any purification. Complexes were formed by mixing caf-
feic acid and hexahydrated aluminum chloride (AlCl3(H2O)6), lead
chloride (PbCl2) or copper chloride (CuCl2) in deionized water at a
given molar ratio (R = [metal]/[ligand]). Initial concentrations of caf-
feic acid before irradiation and/or complexation were in the range
of 5.0–5.6 × 10−5 M and the pH before irradiation was fixed at 5 or
6.5, depending on the experiments.
Calculations were performed at the density functional level
of theory with the PBEO global hybrid functional [33,34], using
the Gaussian (G09) program package [35]. Geometry optimiza-
tions were carried out without any symmetry constraints using the
6–311++G(d,p) basis set. Vibrational frequency calculation was per-
formed to ensure that the optimized structure corresponds to an
energy minimum. The low-lying excited states were treated within
the adiabatic approximation of time dependent density functional
theory (TD-DFT) [36] with the PBEO hybrid functional. Vertical exci-
tation energies were computed for the first 50 singlet excited states,
in order to estimate the UV–vis spectra of the molecule. As it is well
known that UV–vis spectra are very sensitive to the solvent effects,
these latter were introduced by the SCRF method, via the polar-
ized continuum model (PCM) [37,38] implemented in the Gaussian
program.
2.2. Irradiation setup
The solutions of trans-caffeic acid or metal-caffeic acid com-
plexes were placed into a cell (Suprasil Quartz) also aimed at
spectroscopic measurements and were kept under stirring upon
irradiation. The main source used for irradiation is a 200 W
mercury–xenon lamp (LC8 Hamamatsu) connected with a light
guide and directed toward the cell. Band-pass filters were used
in order to get a light beam centered at 312 10 nm and with a
∼10 ␮W power at the sample. The cell was irradiated directly in
the UV–vis spectrometer. A more powerful lamp (1000 W Xenon
from Oriel) equipped with a monochromator (wavelength fixed at
312 nm) was also used for one experiment to get a faster kinetic.
HPLC–ESI-MS analyses (see next section) were performed by samp-
ling a small volume (100 ␮L) of the solution directly inside the
cell.
3. Results and discussion
3.1. Photo-degradation of trans-caffeic acid followed by
electronic spectroscopies
3.1.1. Absorption spectra
The photo-degradation of trans-caffeic acid was followed by
UV–vis absorption spectroscopy. As seen in Fig. 1a, trans-caffeic
acid at pH = 6.5 absorbs light below 350 nm and presents two
absorption bands centered at 285 and 312 nm. Upon irradiation, the
band intensity decrease clearly shows that caffeic acid is degraded.
This consumption is also accompanied by a hypsochromic shift.
The UV spectrum after 196 min of irradiation exhibits an absorption
band at 280 nm, higher than the small contribution around 312 nm.
In parallel to this feature, the baseline in the 350–450 nm range is
2.3. Instrumentation
UV–vis spectra were recorded using a double-beam spectrom-
eter (Cary 100-Varian), in the 200–700 nm region with cells of

12
A. Le Person et al. / Journal of Photochemistry and Photobiology A: Chemistry 265 (2013) 10–19
0,8
0,7
0,6
0,5
0,4
0,3
0,2
0,1
0
t=0
pH=9.8
385nm
(b)
(a)
285
1
0,8
0,6
0,4
0,2
0
pH=8.8
312
2'
pH=8
5'
36'
pH=7.5
92'
pH=7
1000 W source
385
196'
280
pH=5
410
pH=4
360
250
300
350
400
450
260
310
460
Wavelength (nm)
Wavelength (nm)
Fig. 1. (a) UV–vis spectra (pH = 6.5) obtained for various times of irradiation (0, 2, 5, 36, 92 and 196 min) of trans-caffeic acid with the Hg–Xe lamp (initial concentration of
5.0 × 10−5 M) and UV–vis spectrum obtained after a longer time of irradiation of trans-caffeic acid from the 1000 W Xe light source; (b) UV–vis spectra of pure solutions of
esculetin in water (concentration of 5.7 × 10−5 M) at various pH.
not equal to zero and a zoom on the spectral part shows a very weak
absorbance. For this reason, the solution has been irradiated with a
much more intense source (1000 W) and an absorption band with
a maximum around 385 nm (Fig. 1a) is clearly observed. Based on
the UV–vis spectra of pure solutions of esculetin acquired at differ-
ent pH (Fig. 1b), this absorption can be attributed to this product at
pH ∼8, indicating an increase of the pH upon photo-degradation. A
one unity pH increase was also directly measured in the solution.
This increase in absorption in the visible range was also observed
by Cilliers and Singleton [39] for the oxidation of caffeic acid but the
product was not identified. Thus our spectral data let us suggest that
esculetin formation is due to trans-caffeic acid photo-degradation.
Unfortunately, it is not possible to quantify the concentrations of
esculetin in our irradiation conditions by UV–vis spectroscopy. Fur-
thermore, the determination of cis-caffeic acid concentrations from
these spectra was not possible for two main reasons: (i) due to
very close structures, trans and cis caffeic acids have very similar
absorption spectra so it is hard to discriminate them, (ii) the band
observed at 280 nm could be partially assigned to this isomer but
also to esculetin and to other eventual photo-products.
fluorescence emission features of cis- and trans-caffeic acid are too
close to be discriminated, as previously observed for absorption.
Another fluorescence band also appears at 280/310 nm (56 min
of irradiation). This feature could be assigned to protocatechuic
acid because the matrix obtained from pure protocatechuic acid is
closed to this broad band (Fig. 2). However the fluorescence of this
compound does not wholly reproduce the band shape observed
in the photo-degradation process. The presence of vinylcatechol
product could be another hypothesis. Unfortunately, no experi-
mental proof of its presence could be given since this compound
is not available from chemical suppliers. Nevertheless, we have
been able to calculate the absorption and fluorescence spectra of
this molecule by TD-DFT/PBEO/6-311+G(d,p)/PCM. The theoreti-
cal wavelengths of vinylcatechol are found to be 225, 253 and
284 nm for the absorption electronic transitions and 340 nm for the
fluorescence emission. The spectroscopic properties of this com-
pound are thus also close to the broad band observed from the
photo-degradation of caffeic acid. Both protocatechuic acid and
vinylcatechol were identified by using mass spectrometry (see Sec-
tion 2.2).
3.1.2. Fluorescence measurements
3.1.2.2. Fluorescence spectra. Emission fluorescence spectra (Fig. 3)
were recorded by exciting the reactant mixture at the maximum
absorption wavelengths of the different species (312, 385 and
280 nm) and for different times of irradiation. By exciting at 312 nm,
we clearly show the intensity decrease of trans-isomer fluorescence
band at 415 nm with the irradiation time (Fig. 3a). Concomitantly,
a fluorescence band at 468 nm (Fig. 3a) due to esculetin formation
grows, despite the low absorption of this product at 312 nm. This
puts forward the high fluorescence quantum yield of esculetin. As
expected, the fluorescence band observed at 468 nm is enhanced by
exciting at the maximum absorption band of esculetin (385 nm), as
seen in Fig. 3b. According to the data previously obtained from exci-
tation emission matrixes, the emission spectra recorded with an
excitation at 280 nm (Fig. 3c) exhibit a fluorescence band around
310 nm which was previously suggested to be due to the forma-
tion of protocatechuic acid and/or vinylcatechol. The continuous
increase of this fluorescence with irradiation time confirms that
it cannot be attributed to cis-caffeic acid. This isomer can still
not be detected here. Profiles of fluorescence intensities (Fig. 3d)
were extracted at 415, 468 and 310 nm from the three emis-
sion fluorescence spectra. The fluorescence intensity evolutions
are directly linked to the concentrations of the different species.
The rapid decrease of the fluorescence intensity of trans-caffeic
Fluorescence spectroscopy is a more sensitive technique than
UV–vis spectroscopy. For this reason, fluorescence measurements
were performed in order to highlight the different species involved
in the mechanism of trans-caffeic acid photo-degradation.
3.1.2.1. Excitation emission matrix. Three-dimensional spectra also
called excitation emission matrix (EEM) were acquired by varying
the excitation and the emission wavelengths. Before photo-
degradation (t = 0), a fluorescence band detected at 312/415 nm
(ꢀex/ꢀem) is characteristic of trans-caffeic acid (Fig. 2). After
6 min of irradiation, the additional fluorescence which appears at
385/468 nm is attributed to esculetin, with respect to the spectral
data obtained by UV–vis absorption. Then, the emission of trans-
caffeic acid continues to decrease while that of esculetin increases
from 6 to 180 min. By comparison with the UV spectra where the
absorbance related to the formation of esculetin is relatively low,
the quantum yield of esculetin fluorescence is assumed to be high,
since it was detected in the first minutes of photo-degradation. On
the contrary, cis-caffeic acid which is known to be formed from
trans-caffeic acid in the first step of the mechanism could not be
observed on the EEM since no other fluorescence band appeared
in the first minutes of photo-degradation. One can conclude that

A. Le Person et al. / Journal of Photochemistry and Photobiology A: Chemistry 265 (2013) 10–19
13
Fig. 2. Excitation emission matrix (EEM) obtained at different times of irradiation (0, 6, 16, 56 and 80 min) of trans-caffeic acid (initial concentration of 5.0 × 10−5 M) and
EEM of pure protocatechuic acid (concentration of 2.0 × 10−5 M).
observed in the first 2 min can be connected to the fast conversion
to the cis-isomer. At 200 min, trans-caffeic acid is almost completely
consumed whereas the concentration of esculetin reaches a max-
imum. On the contrary, the emission at 310 nm does not show
the same behavior since it seems to reach a plateau value after
250 min. Consequently, the formation pathways of the different
observed photo-products are not linked and then protocatechuic
acid and/or vinylcatechol are not formed from esculetin degrada-
tion, as expected.
Vinylcatechol is also proposed since a peak was detected at its cor-
responding M = 136 g mol−1. The concentration of protocatechuic
acid cannot be obtained because it remains too low during the
degradation process to be quantified. The concentration of vinyl-
catechol cannot be determined as well since this compound is not
commercially available. Nevertheless, it can be noted that their
concentrations both increase upon photo-degradation.
The concentration profiles of trans-caffeic acid, cis-caffeic acid
and esculetin and the estimated or calculated concentration pro-
files of protocatechuic acid and vinylcatechol are reported in Fig. 5.
These data were compared to the model mechanism presented
in Fig. 6. This model takes into account (i) the reversible trans/cis
3.2. Kinetic investigation of the photo-degradation by mass
spectrometry
isomerization characterized by the kinetic parameters k1 and k
,
−1
(ii) the formation of esculetin from cis-isomer with a rate con-
stant k2 and (iii) the two other routes leading to vinylcatechol
and protocatechuic from trans-caffeic acid and characterized by
two independent rate constants k1ꢀ and k1ꢀꢀ, respectively. Although
the concentration of vinylcatechol remains unknown all along the
degradation process, the areas of the peaks related to this product
are available and proportional to the concentrations. The k1ꢀ rate
constant was then adjusted in order to get the same trend between
the experimental areas profile and the calculated concentration
profile, i.e. in considering a constant ratio between the experi-
mental and calculated profiles. Then a scale factor was applied to
the experimental data to obtain the real concentrations of vinyl-
catechol. Furthermore, the concentrations of protocatechuic acid
are also not precisely determined but they are below the quan-
tification limit of 1 × 10−6 mol L−1. This limit was obtained from
HPLC–ESI-MS measurements of protocatechuic acid standard and
High performance liquid chromatography coupled to mass spec-
trometry was used to separate the different species involved in
the mechanism, in order to get information about the kinetics
and the mechanism of the photo-degradation. The trans-caffeic
acid solution was analyzed by HPLC–ESI-MS for different irradi-
ation times with the mercury–xenon lamp. The initial extracted
ion chromatogram (Fig. 4) obtained in the absence of irradia-
tion underlines the main occurrence of trans-isomer of caffeic
acid and a very small contribution detected at the same mass
(M = 180 g mol−1) attributed to cis-isomer. Upon irradiation, trans-
caffeic acid is consumed while cis-caffeic acid (M = 180 g mol−1) is
formed and reaches a maximum concentration after ∼5 min under
these experimental conditions. Then, its concentration decreases
and is followed by the formation of esculetin (M = 178 g mol−1
)
which is detected after 20 min of irradiation. The presence of pro-
tocatechuic acid was confirmed here from the standard compound.

14
A. Le Person et al. / Journal of Photochemistry and Photobiology A: Chemistry 265 (2013) 10–19
Fig. 3. Emission fluorescence spectra obtained at different times of irradiation of trans-caffeic acid from 0 to 5 h by using the spectrofluorimeter as the source of irradiation
(initial concentration of 5.6 × 10−5 M). (a) Excitation at 312 nm and slit width of 4 nm; (b) excitation at 385 nm and slit width of 2 nm; (c) excitation at 280 nm and slit width
of 2 nm; (d) profiles of fluorescence obtained from the emission spectra a, b and c.
corresponds to the concentration for which the ratio between
the height of the corresponding peak and the noise is 10. An
upper limit of the k1ꢀꢀ rate constant was thus estimated in order
to get a maximum concentration of this product below this limit.
With the hypothesis of pseudo-first order reactions, the simulated
data fit well with experimental ones (Fig. 5) by optimizing the
kinetic parameters to the following values: k1 = 3.8 0.1 × 10−3 s−1
,
k
−1 = 7.0 0.1 × 10−3 s−1, k2 = 1.5 0.1 × 10−5 s−1, k1ꢀ = 1.5 0.1 ×
10−4
s
−1 and k1ꢀꢀ = 2.5 0.1 × 10−6
s
−1. The pseudo-first order rates
are in disagreement with the work of Carlotti et al. [21] who
observed a degradation of caffeic acid in aqueous solutions upon
UV irradiation following a pseudo-zero order kinetic. It can then be
noticed that the sum of the concentrations of the species present
at a given time are roughly equal to the initial concentration of
caffeic acid (5 × 10−5 mol L−1). This means that possible other prod-
ucts that could not be detected might have lower concentrations.
The values of rate constants depend on experimental conditions
and notably on irradiation wavelength range. The most important
results of this study rely on the mechanistic aspect of the caffeic acid
photo-degradation. We show that the photo-isomerization and the
intramolecular cyclization really occur but are not the major routes,
Fig. 4. Extracted ion chromatograms obtained before irradiation and at 196 min of
irradiation of trans-caffeic acid (initial concentration of 5.0 × 10−5 M) by using the
mercury–xenon lamp.

A. Le Person et al. / Journal of Photochemistry and Photobiology A: Chemistry 265 (2013) 10–19
15
5
4
3
2
1
0
trans- caffeic acid
vinylcatechol
cis- caffeic acid
esculeꢀn
protocatechuic acid
0 5 20 36
92
Irradiaꢀon ꢀme (min)
196
Fig. 7. Comparison between the compositions of the oxygen containing solution
and the deoxygenated solution (trans-caffeic acid, cis-caffeic acid and esculetin con-
centrations in arbitrary units obtained from extracted ion chromatograms).
Fig. 5. Experimental (HPLC) and calculated or estimated concentration profiles of
trans-caffeic acid, cis-caffeic acid, esculetin, vinylcatechol and protocatechuic acid
at different times of irradiation (0, 2, 5, 20, 36, 92 and 196 min) by using the Hg–Xe
lamp.
using two cells: one containing air and the other deoxygenated
by degassing 30 min with a slow Ar stream before irradiation. The
concentrations of trans-caffeic acid and cis-caffeic acid were then
compared for similar irradiation times (5 and 70 min), as seen in
Fig. 7. The decrease of trans to cis conversion in the presence of
Ar compared to oxygen is clear. One can thus conclude that the
photo-isomerization rate is higher in the presence of air. Borges
and Pinto [26] showed that the production of esculetin decreases in
deoxygenated solutions. This formation that occurs in the absence
of oxygen was explained by an auto-oxidation as suggested by these
authors. In our experiments, the concentrations of esculetin are
similar in both conditions but the very low content of this prod-
uct for the studied irradiation time does not permit to conclude.
Nevertheless, the results reported by Borges and Pinto [26] are
also consistent with an increase of the photo-isomerization. We
underline the fact that the concentrations of vinycatechol and pro-
tocatechuic acid are equal in both conditions.
since the formation of vinylcatechol seems to be important. So far,
this latter compound was never detected as a product of caffeic
acid photo-degradation. We can just mention that Grimes et al. [29]
observed an increase of the total amount of carbon dioxide upon
photo-degradation. This could be linked to the loss of the caffeic
acid carboxylic group, thus leading to the formation of vinylcate-
chol species. It was also shown that this compound can be formed
by thermal decarboxylation of caffeic acid [40–42] and it was also
detected in roasted coffee [43]. Finally, the pathway leading to the
protocatechuic acid product, also observed by Grimes et al. [29] and
Amat et al. [30], is minor.
3.3. Influence of oxygen on the photo-degradation of trans-caffeic
acid
3.4. Influence of the complexation on the photo-degradation
The results presented above were obtained from solutions con-
taining air. Complementary experiments were carried out to test
the real influence of oxygen in our experimental conditions by
Photo-degradation of metal complexes of caffeic acid was com-
pared to the photo-degradation of free caffeic acid using the
Fig. 6. Proposed mechanism of the photo-degradation of trans-caffeic acid.

16
A. Le Person et al. / Journal of Photochemistry and Photobiology A: Chemistry 265 (2013) 10–19
acid since the photo-degradation is prior to analysis. The compar-
ison with the photo-degradation of free caffeic acid is based on
HPLC–ESI-MS and UV–vis results. Vinylcatechol and protocatechuic
acid were not detected upon irradiation of metal complexes.
3.4.1. Influence of Pb(II) ions
Caffeic acid is known to complex Pb(II) ions [12]. These ions
first coordinate with the carboxylate function to form a complex
of 1:1 stoichiometry. Then both carboxylate and catechol func-
tions are naturally involved in the 2:1 complex [13]. However,
the complexation of caffeic acid is not complete for molar ratio
below 3, preventing the precipitation of lead hydroxide. UV–vis
spectra of Pb(II)/caffeic acid solution were recorded during the
photo-degradation at pH = 6.5 (Fig. 9a), in the same conditions as
the photo-degradation of the free molecule (Fig. 1a). The UV absorp-
tion bands observed in the 260–360 nm spectral range before
irradiation are linked to the simultaneous presence of free and
complexed (1:1 and 2:1) forms of caffeic acid for a molar ratio of
R = 2 [13]. Thus the decrease of the absorption bands upon irra-
diation is due to the photo-degradation of the free molecule and
of the two possible complexed forms. Despite the apparent diffi-
culty to distinguish between the three processes, it appears that
this decrease is accompanied by the formation of a neat absorption
band at 390 nm which can be attributed to the complex formed
between esculetin and Pb(II) ions (Fig. 9b). The intensity of this
band is much higher than the absorbance measured in the same
conditions for the photo-degradation of the free form (Fig. 1a),
showing that the formation of esculetin is enhanced by the pres-
ence of Pb(II) ions. These results are confirmed by HPLC–ESI-MS
measurements. Indeed, the esculetin concentration reached in the
Fig. 8. Extracted ion chromatograms obtained with free trans-caffeic acid (a) and
with Pb(II) ions complexed by caffeic acid before irradiation (b) and after 60 min of
irradiation (c).
same experimental conditions. Samples were first analyzed before
degradation: the extracted ion chromatograms obtained from a
metal–caffeic acid complex are similar to those obtained from free
caffeic acid, in terms of retention time (17.7 min) and peak area,
showing that complexes dissociate in the HPLC device (Fig. 8).
Therefore, the peaks observed at this retention time are due to the
total contribution of free and complexed caffeic acid. HPLC–ESI-MS
can still be used for the photochemical study of complexed caffeic
Fig. 9. (a) UV–vis spectra (pH = 6.5) obtained at different times of irradiation (0–180 min) of Pb(II)/caffeic acid complexes (molar ratio of R = 2); (b) UV–vis spectra (pH = 6.5)
obtained from the complexation of Pb(II) ions by esculetin (molar ratio R = Pb(II)]/[esculetin] from 0 to 3); (c) photo-degradation (pH = 6.5) of free caffeic acid compared to
the photo-degradation of Pb(II)/caffeic acid complexes (molar ratio of R = 2) for the same time of irradiation (60 and 180 min) obtained from extracted ion chromatograms.
The y-axis represents the areas of the peaks related to trans-caffeic acid, cis-caffeic acid and esculetin.

A. Le Person et al. / Journal of Photochemistry and Photobiology A: Chemistry 265 (2013) 10–19
17
Fig. 10. UV–vis spectra (pH = 5) obtained at different times of irradiation (0–180 min) of (a) free caffeic acid (b) Al(III)/caffeic acid complexes (molar ratio of R = 10); (c)
UV–vis spectra (pH = 5) obtained from the complexation of Al(III) ions by esculetin (molar ratio R = [Al(III)]/[esculetin] from 0 to 2); (d) photo-degradation (pH = 5) of free
caffeic acid compared to the photo-degradation of Al(III)/caffeic acid complexes (molar ratio of R = 10) for the same time of irradiation (180 min) obtained from extracted ion
chromatograms. The y-axis represents the areas of the peaks related to trans-caffeic acid, cis-caffeic acid and esculetin.
presence of the complex is higher than with the free molecule,
for a similar irradiation time (Fig. 9c). As this photo-product is
formed from cis-caffeic acid, this partly explains that the concen-
tration of the latter isomer is lower in the presence of Pb(II) than in
the absence of these ions (Fig. 9c). One can conclude that the rate
constant k2, related to the formation of esculetin from cis-isomer,
increases in the presence of Pb(II) ions (Fig. 6). As for trans-caffeic
acid, its concentration is higher with metal ions at 60 min of irra-
diation but is lower at 180 min. It implies that the isomerization
reaction is also modified by the complexation process. For low
times of irradiation such as 60 min, the decrease of the conver-
sion from the trans form to the cis form in the presence of Pb(II)
acid complex was followed by UV–vis spectroscopy (Fig. 10b) and
compared to the photo-degradation of the free molecule at pH = 5
(Fig. 10a). The UV–vis spectrum (Fig. 10b) recorded before irradi-
ation in the presence of Al(III) ions presents an absorption band
at 345 nm which fits well with the maxima at 335 and 347 nm of
the 1:1 and 2:1 complexes, respectively [10]. Upon irradiation, the
decrease of this band is slow (less than 8% after 180 min) and no shift
or new absorption band is observed. Moreover, since no absorption
band at 350 or 390 nm, corresponding to free or complexed forms
of esculetin (Fig. 10c), respectively, is detected, we can assume that
the presence of Al(III) ions does not enhance the photo-degradation.
The decrease of the photo-degradation of caffeic acid in the pres-
ence of Al(III) ions is confirmed by the HPLC–ESI-MS results. The
areas of the peaks obtained from the extracted ion chromatograms
and related to the concentrations of trans-caffeic acid, cis-caffeic
acid and esculetin are presented on Fig. 10d. Three observations can
be deduced from this figure: (i) the degradation rate of trans-caffeic
acid is much higher in the absence (∼40%) than in the presence (less
than 2%) of Al(III), after the same irradiation time (180 min). (ii)
Conversion to cis-isomer is also hampered by the presence of Al(III)
ions. (iii) The photo-product (esculetin) is observed by irradiation
of free caffeic acid but not by irradiation of the complexes. Thus it is
clear that the presence of Al(III) ions inhibits the photo-degradation
of the molecule. This behavior can be explained by the preferential
fixation site of Al(III) ions on the catechol function of caffeic acid,
this function being also involved in the photo-degradation mech-
anism [23,27,28]. If this latter is blocked by Al(III) ions, the best
stability of the molecule toward photo-degradation can be easily
understood. It is interesting to note that the isomerization is also
inhibited by the presence of this metal ion coordinated to catechol
site.
ions can account for a decrease of the k1/k ratio, where k1 and
−1
k
−1
are the rate constants corresponding to the isomerization pro-
cess. In contrast, the k2 rate constant enhancement explains that
the amount of caffeic acid that reacted becomes higher after a
given irradiation time. Thus we confirm the promoter effect of
Pb(II) ions on the caffeic acid photo-degradation leading to the
formation of esculetin. This effect is probably linked to the coor-
dination of Pb(II) ions on the carboxylate function of caffeic acid
[13].
3.4.2. Influence of Al(III) ions
The complexation of Al(III) ions by caffeic acid was previously
studied [10]. Unlike Pb(II) ions, Al(III) ions are known to coordinate
preferentially to the catechol function. The optimal conditions to
reach a complete complexation of Al(III) ions by caffeic acid con-
sist in working at pH = 5 and with a [Al(III)]/[caffeic acid] molar
ratio (R) higher than 10 [10]. By this way, the concentration of free
caffeic acid can be considered as negligible compared to the concen-
trations of the complexes. The photo-degradation of Al(III)/caffeic

18
A. Le Person et al. / Journal of Photochemistry and Photobiology A: Chemistry 265 (2013) 10–19
Fig. 11. UV–vis spectra (pH = 5) obtained at different times of irradiation (0–180 min) of (a) Cu(II)/caffeic acid complexes (molar ratio of R = 2); (b) UV–vis spectra (pH = 5)
obtained from the complexation of Cu(II) ions by esculetin (molar ratio R = [Cu(II)]/[esculetin] from 0 to 5); (c) photo-degradation (pH = 5) of free caffeic acid compared to
the photo-degradation of Cu(II)/caffeic acid complexes (molar ratio of R = 2) for the same time of irradiation (60 min) obtained from extracted ion chromatograms. The y-axis
represents the areas of the peaks related to trans-caffeic acid, cis-caffeic acid and esculetin.
3.4.3. Influence of Cu(II) ions
is based on the photo-isomerization of the molecule followed by
a cyclization to form esculetin as photo-product. However, this
report reveals the occurrence of another route of major impor-
tance leading to the likely formation of vinylcatechol which was
never observed up to now. As shown previously [29,30], protoca-
techuic acid is also formed but in small amounts. As expected, the
presence of oxygen leads to an increase of the photo-isomerization
rate. This study provides also clear evidence of the influence of com-
plexation to metal ions on caffeic acid photo-degradation. Note that
complexation process with the metal ions was also observed on
esculetin when this photo-product is formed. In all cases, the photo-
isomerization is reduced by the presence of Pb(II), Al(III) and Cu(II)
ions. In addition, the metal type was shown to play a significant
role for the esculetin formation. Indeed, the presence of Al(III) ions
leads to a very photo-stable complex and inhibits the formation
of esculetin, whereas Pb(II) and Cu(II) ions enhance the produc-
tion of this compound. This phenomenon might be explained by
the different coordination sites of the metal ions to the ligand: the
preferential fixation site of Al(III) ions is the catechol function and
since this function is involved in the photo-degradation mecha-
nism, the complexation blocks the reaction. On the contrary, the
coordination of Pb(II) ions on the carboxylate function promotes
the formation of esculetin. Finally, one can conclude that, in soils
polluted by heavy metals, the possibility of formation of metal com-
plexes with humic substances contributes to significantly change
the photo-degradation of these substances.
The study of the photo-degradation of the Cu(II)/caffeic acid
system has been carried out in the same conditions as those used
for Al(III) ions (pH = 5). The photo-degradation of Cu(II)/caffeic acid
complex (Fig. 11a) was compared to the photo-degradation of the
free molecule (Fig. 10a). The absorption spectra of Cu(II)/caffeic acid
complex and free caffeic acid (dotted line in Fig. 11a) are very sim-
ilar. Only a slight shoulder appears at approximately 350 nm. Upon
irradiation, the UV–vis spectra (Fig. 11a) highlight the growth of an
absorption band in the 350–450 nm range, which can be attributed
to the formation of the complexed form of esculetin (Fig. 11b).
Compared to lead ions, cupric ions have a similar and even more
marked influence on the photo-degradation of caffeic acid. Indeed,
HPLC–ESI-MS results (Fig. 11c) show that the formation of the
photo-product is 20 times larger in the presence of Cu(II) ions than
in its absence whereas it was only 4 times larger in the presence of
Pb(II) ions. For an irradiation time of 60 min, the cis-isomer formed
was almost completely consumed in the presence of Cu(II) ions
contrary to the degradation of the free molecule (Fig. 11c). The
consumption of this isomer is more pronounced than what was
observed with Pb(II) solutions.
To our knowledge, the coordination site of Cu(II) ions on caffeic
acid has not been identified yet. However, the similar influences
of Pb(II) and Cu(II) ions on the photo-degradation of caffeic acid
could be explained by a common site of fixation of metal ions on
the molecule such as the carboxylate function.
4. Conclusion
Acknowledgments
This study corroborates the first route of trans-caffeic acid
photo-degradation already proposed by several authors. This one
The Mass Spectrometry facility used in this study was funded by
the European Community (FEDER), the Région Nord-Pas de Calais

A. Le Person et al. / Journal of Photochemistry and Photobiology A: Chemistry 265 (2013) 10–19
19
(France), the CNRS, and the Université Lille 1, Sciences et Tech-
nologies. The authors are also grateful to Dr. A. Moncomble for his
precious help about the quantum chemical calculations.
[26] F. Borges, M. Pinto, Helvetica Chimica Acta 75 (1992) 1061–1068.
[27] G. Pandey, A. Krishna, J.M. Rao, Tetrahedron Letters 27 (34) (1986) 4075–4076.
[28] M. Satô, A. Hiraoka, Chemical and Pharmaceutical Bulletin 33 (3) (1985)
1289–1292.
[29] S.M. Grimes, L.K. Mehta, H.C. Ngwang, Journal of Environment Science and
Health, Part A Environmental Science 36 (5) (2001) 599–612.
[30] A.M. Amat, A. Arques, M.A. Miranda, Applied Catalysis B: Environmental 23
(1999) 205–214.
[31] S. Perez-Magarino, I. Revilla, M.L. Gonzalez-SanJose, S. Beltran, Journal of Chro-
matography A 847 (1999) 75–81.
[32] I. Nicoletti, A. De Rossi, G. Giovinazzo, D. Corradini, Journal of Agricultural and
Food Chemistry 55 (2007) 3304–3311.
[33] J.P. Perdew, K. Burke, M. Ernzerhof, Physical Review Letters 77 (1996)
3865–3868.
[34] C. Adamo, V.J. Barone, Journal of Chemical Physics 110 (1999) 6158–6169.
[35] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato,
X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada,
M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery Jr., J.E. Peralta, F. Ogliaro, M.
Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J. Nor-
mand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N.
Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo,
R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W.
Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador,
J.J. Dannenberg, S. Dapprich, A.D. Daniels, Ö. Farkas, J.B. Foresman, J.V. Ortiz, J.
Cioslowski, D.J. Fox, Gaussian 09, Revision B.01, Gaussian, Inc., Wallingford, CT,
2009.
[36] R. Bauernschmitt, R. Ahlrichs, Chemical Physics 256 (1996) 454–464.
[37] M. Cossi, G. Scalmani, N. Rega, V.J. Barone, Journal of Chemical Physics 117
(2002) 43–54.
[38] J. Tomasi, R. Cammi, B. Mennucci, C. Cappelli, S. Corni, Physical Chemistry
Chemical Physics 4 (2002) 5697–5712.
[39] J.J.L. Cilliers, V.L. Singleton, Journal of Agricultural and Food Chemistry 39 (1991)
1298–1303.
[40] P. Terpinc, T. Polak, N. Segatin, A. Hanzlowsky, N.P. Ulrih, H. Abramovic, Food
Chemistry 128 (2011) 62–69.
[41] R. Stadler, H.D. Welti, A. Stämpfli, L. Fay, Journal of Agricultural and Food Chem-
istry 44 (1996) 898–905.
[42] G. Rizzi, L. Boekley, Journal of Agricultural and Food Chemistry 40 (1992)
1666–1670.
[43] O. Frank, S. Blumberg, C. Kunert, G. Zehentbauer, T. Hofmann, Journal of Agri-
cultural and Food Chemistry 55 (2007) 1945–1954.
References
[1] F.J. Stevenson, Humus chemistry: Genesis, Composition Reactions, Wiley, New
York, 1982.
[2] H.R. Schulten, M. Schnitzer, Soil Science 162 (1997) 115–130.
[3] R.L. Wershaw, G.R. Aiken, Humic Substances is Soils, Sediments and Water,
Wiley Interscience, New York, 1995.
[4] E. Baath, Water Air Soil Pollution 47 (1989) 335–379.
[5] P. Doelman, L. Haanstra, Soil Biology and Biochemistry 11 (1979) 475–479.
[6] R.A. Olsen, J.C. Brown, J.H. Bennet, D. Blume, Journal of Plant Nutrition 5 (1982)
433–445.
[7] I. Gülc¸ in, Toxicology 217 (2006) 213–220.
[8] A. Saija, A. Tomaino, D. Trombetta, A. De Pasquale, N. Uccella, T. Barbuzzi, T.
Paolino, F. Bonina, International Journal of Pharmaceutics 199 (2000) 39–47.
[9] C. Lapouge, J.P. Cornard, ChemPhysChem 8 (2007) 473–479.
[10] J.P. Cornard, A. Caudron, J.C. Merlin, Polyhedron 25 (2006) 2215–2222.
[11] J.P. Cornard, C. Lapouge, Journal of Physical Chemistry A 108 (2004) 4470–4478.
[12] J.P. Cornard, C. Lapouge, Chemical Physics Letters 438 (2007) 41–46.
[13] L. Boilet, J.P. Cornard, C. Lapouge, Journal of Physical Chemistry A 109 (2005)
1952–1960.
[14] I.V. Sokolova, O.N. Tchaikovskaya, Atmospheric and Oceanic Optics 19 (2006)
220–222.
[15] D.E. Latch, R. McNeil, Science 311 (2006) 1743–1747.
[16] W.L. Miller, Ecological Studies 133 (1998) 125–143.
[17] R.G. Zepp, G.L. Baughman, P.F. Schlotzhauer, Chemosphere 10 (1981) 109–117.
[18] S. Protti, A. Mezzetti, C. Lapouge, J.P. Cornard, Photochemistry & Photobiological
Sciences 7 (2008) 109–119.
[19] J. Buschmann, S. Canonica, L. Sigg, Environmental Science and Technology 39
(2005) 5335–5341.
[20] Ba L. Tran, Seth M. Cohen, Chemical Communications (2006) 203–205.
[21] M.E. Carlotti, E. Ugazio, S. Sapino, E. Peira, L. Battaglia, R. Cavalli, Journal of
Dispersion Science and Technology 29 (2008) 1435–1444.
[22] F.J. Benitez, J. Beltran-Heredia, J.L. Acero, M.L. Pinilla, Journal of Chemical Tech-
nology and Biotechnology 70 (1997) 253–260.
[23] W.L. Butler, H.W. Siegelman, Nature (London) 183 (1959) 1813–1814.
[24] R.D. Hartley, E.C. Jones, Journal of Chromatography 107 (1975) 213–218.
[25] F. Borges, M. Pinto, Journal of Liquid Chromatography 12 (12) (1989)
2345–2354.