Water concentration dependent photochemistry of ketoprofen in aqueous solutions

Article abstract of DOI:10.1039/b919330h

Ketoprofen is an important photosensitive drug molecule that has received much attention for the study of its photochemistry in different solvents. In this paper, nanosecond time-resolved resonance Raman spectroscopy was utilized to investigate the photochemistry of ketoprofen in aqueous solutions with varying water concentrations. The rate constants and reaction mechanism of ketoprofen are strongly dependent on the concentration of the solvent. In neat acetonitrile and solvents with low concentrations of water (like water-acetonitrile ≤ 1:1, v/v), ketoprofen exhibits mostly benzophenone-like photochemistry to generate a triplet state which in turn produces a ketyl radical-like species by a hydrogen abstraction reaction. However, in solvents with very high concentrations of water (such as water-acetonitrile ≥ 9:1, v/v), the triplet state ketoprofen is observed first and then undergoes a prompt decarboxylation process to form a triplet protonated biradical carbanion species. For solvents with moderate higher water concentrations (such as between 50% and 90% water by volume), the hydrogen abstraction and decarboxylation processes are two competitive pathways with different rate constants. The triplet state of ketoprofen will simultaneously produce a ketyl radical species and a triplet protonated biradical carbanion species with the amount of each species dependent on the water concentration. the Owner Societies.

Full text of DOI:10.1039/b919330h

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/pccp | Physical Chemistry Chemical Physics
Water concentration dependent photochemistry of ketoprofen in aqueous
solutionsw
Ming-De Li, Yong Du, Yung Ping Chuang, Jiadan Xue and David Lee Phillips*
Received 16th September 2009, Accepted 9th February 2010
First published as an Advance Article on the web 16th March 2010
DOI: 10.1039/b919330h
Ketoprofen is an important photosensitive drug molecule that has received much attention for the
study of its photochemistry in different solvents. In this paper, nanosecond time-resolved
resonance Raman spectroscopy was utilized to investigate the photochemistry of ketoprofen in
aqueous solutions with varying water concentrations. The rate constants and reaction mechanism
of ketoprofen are strongly dependent on the concentration of the solvent. In neat acetonitrile and
solvents with low concentrations of water (like water–acetonitrile r 1 : 1, v/v), ketoprofen exhibits
mostly benzophenone-like photochemistry to generate a triplet state which in turn produces a
ketyl radical-like species by a hydrogen abstraction reaction. However, in solvents with very high
concentrations of water (such as water–acetonitrile Z 9 : 1, v/v), the triplet state ketoprofen is
observed first and then undergoes a prompt decarboxylation process to form a triplet protonated
biradical carbanion species. For solvents with moderate higher water concentrations (such as
between 50% and 90% water by volume), the hydrogen abstraction and decarboxylation
processes are two competitive pathways with different rate constants. The triplet state of
ketoprofen will simultaneously produce a ketyl radical species and a triplet protonated biradical
carbanion species with the amount of each species dependent on the water concentration.
buffer, while the decay of the carbanion in the nanosecond
Introduction
time-scale displayed a clear pH-dependence. The main reason
Ketoprofen (KP) is a common nonsterioidal anti-inflammatory
this affected the characteristic of the carbanion is that after the
drug (NSAID) that is widely used for moderate acute and
decay of the carbanion it will form a long-lived species (triplet
chronic painful conditions to improve the tolerability profile
protonated biradical carbanion species) at low pH values
of NSAIDs. However, KP can induce photosensitivity and
and finally lead to the direct formation of the final product
phototoxicity due to its specific structure and electronic
(3-ethylbenzophenone) at higher pH values. Recently, nano-
features.1,2 A large number of studies have been performed
second time-resolved resonance Raman (ns-TR3) and density
in recent years on both the phototoxic and photoallergic
functional theory (DFT) calculations were used to investigate
the photochemistry of (S)-ketoprofen in different solvents.6 In
reactions induced by KP so as to estimate the safety of KP
and also to investigate the mechanism(s) of the photosensitivity
aqueous solution, KP exhibits a pH dependent photochemistry
reactions when KP is exposed to UV light.3,4 These results
in a 1 : 1 water–acetonitrile (volume ratio) solution. In acidic
show that the photochemistry of KP in aqueous solution is
aqueous and neutral solution, the triplet state ketoprofen
strongly influenced by its acid–base chemistry.5 The neutral
(3[KP]) is formed through intersystem crossing (ISC) from
3
the first excited singlet state (S1), and then the [KP] species
form of ketoprofen that exists in the absence of a base exhibits
typical benzophenone-like photochemistry,5,6 while the
produces an arylphenylketyl (ArPK) radical via a hydrogen
deprotonated ketoprofen that exists in the presence of a base
undergoes an efficient photodecarboxylation process.7
abstraction reaction with the surrounding water molecules. In
A
contrast, a prompt decarboxylation process will take place
from the triplet state deprotonated ketoprofen (3[KPÀ1]) and
then involves the formation of the triplet protonated biradical
carbanion species in a 1 : 1 water–acetonitrile alkaline and 7 : 3
PBS–acetonitrile aqueous solutions.6 Therefore the photo-
chemistry of KP appears determined by the precursors, the
number of investigations have been done to attempt to better
elucidate the mechanism(s) that cause KP’s phototoxicity and
to explore the acid–base influence on the photochemistry of
KP in aqueous solution.8–11 KP was studied in water and in
phosphate buffer solutions in the 5.7–11.0 pH range by laser-
induced optoacoustic spectroscopy12 and these studies showed
that the decarboxylation from the excited singlet deprotonated
ketoprofen did not depend on pH or on the presence of a
3
3
neutral form of [KP] or the deprotonated [KPÀ1].
In an attempt to establish the nature of the carbanion’s
precursor, Scaiano and coworkers have conducted studies on
the photochemistry of KP using transient absorption and
quenching experiments in organic solvent–water mixtures.5
They proposed that the same rate constant for the decay of
transient species and spectral characteristics were obtained in
various water–acetonitrile mixtures in the absence of base.
Department of Chemistry, The University of Hong Kong, Pokfulam
Road, Hong Kong, P. R. China. E-mail: phillips@hkucc.hku.hk;
Fax: (852)-2857-1586; Tel: (852)-2859-2160
w Electronic supplementary information (ESI) available: Raman
spectra; nS-TR3 spectra. See DOI: 10.1039/b919330h
ꢀc
This journal is the Owner Societies 2010
4800 | Phys. Chem. Chem. Phys., 2010, 12, 4800–4808

View Article Online
However, the rate constant of the decay of the carbanion
showed an interesting dependence on the mole fraction of
water.5 This result appeared to be consistent with later ns-TR3
experiments that showed the triplet protonated biradical
carbanion decays faster in a 0.5 mM NaOH aqueous solution
than in a PBS solution.6
PBS–MeCN (90% PBS) were used as solvents and unless
specified, all of the mixed solvent ratios are of volume ratio.
PBS was prepared by mixing 2.314 g KH2PO4, 7.384 g
Na2HPO4 in a 1000 ml solution with a final pH value of 7.4.
The ns-TR3 experiments were done using an experimental
apparatus and methods discussed in detail previously,21,22
and only a brief description will be given here. The 266 nm
pump laser pulse generated from the fourth harmonic of a
Nd:YAG nanosecond pulsed laser and a 319.9 nm probe laser
pulse generated from the third anti-Stokes hydrogen Raman
shift laser line from the second harmonic were employed in the
ns-TR3 experiments. The energy for the pump and probe
pulses were in the range of 2.5–3.5 mJ with a 10 Hz repetition
rate. The two Nd:YAG lasers were synchronized electronically
by a pulse delay generator to control the time delay of the
pump and probe lasers with the time delay between the
laser pulses monitored by a fast photodiode and 500 MHz
oscilloscope. The time resolution for the ns-TR3 experiments
was approximately 10 ns. The pump and probe laser beams
were lightly focused onto the flowing sampling system and the
Raman light was collected using reflective optics into a
spectrometer whose grating dispersed the light onto a liquid
nitrogen cooled CCD detector. The Raman signal was
accumulated for 30 s by the CCD before reading out to an
interfaced PC computer and 10–20 scans of the signal were
added together to get a resonance Raman spectrum. The
ns-TR3 spectra presented here were obtained by the subtraction
of a resonance Raman spectrum with negative time delay of
À100 ns (probe-before-pump spectrum) from the resonance
Raman spectrum with a positive time delay (pump–probe
spectrum) and the Raman shifts were calibrated by the known
MeCN solvent’s Raman bands with an estimated accuracy of
5 cmÀ1. A Lorentzian function was applied to integrate the
relevant Raman bands to determine the decay and growth
kinetics of the species observed in the experiments (details for
how the kinetics of the intermediate was obtained are shown in
the ESIw).
It is difficult to clearly distinguish various intermediates that
have similar electronic absorption spectra and time-scales in
time-resolved transient absorption experiments due to the
spectral overlap of the absorption spectra on similar time
scales for the different transient species that may be present.13–17
In these cases it is useful to employ time-resolved vibrational
spectroscopy techniques that can more clearly distinguish the
various transient species present by their vibrational frequencies
that contain structural information and can serve as a
fingerprint for each species. In this work, we have used time-
resolved resonance Raman (TR3) spectroscopy, a type of
time-resolved vibrational spectroscopic technique, which is a
powerful tool to obtain valuable information about the identity,
structure and properties of chemical reaction intermediates.18–20
We have recently utilized ns-TR3 spectroscopy to investigate
the photochemistry of benzophenone in aqueous solutions and
KP in different solvents6,13 and these studies provided new
insight into the identity, structure and properties of several
intermediates associated with the hydrogen abstraction and
decarboxylation reactions observed in these two molecular
systems.
In this article, the photochemistry of KP in aqueous
solutions with varying water–acetonitrile ratios were studied
using nanosecond time-resolved resonance Raman spectro-
scopy (ns-TR3). The ns-TR3 spectra provide vibrational
spectroscopic information that can be used to help elucidate
the photochemical mechanism(s) and solvent effects on the
intermediates involved in the reactions. The experimental
results indicate that the reaction pathways for KP are strongly
dependent on the solvent ratio of water and acetonitrile in the
solvent system. In aqueous solutions with less than 50% water,
the 3[KP] is mainly observed and then forms an arylphenylketyl
(ArPK) radical via a hydrogen abstraction reaction with
water. However, in aqueous solutions with 90% or greater
Solvent effects were considered by UB3LYP/6-311G**
optimizations of the gas-phase stationary points using a simple
SCRF method based on a polarizable continuum model
(PCM).23 All of the density functional theory calculations
presented in this work made use of the Gaussian 03 program
suite24 operated on the High Performance Computing cluster
(HPCPOWER) at the University of Hong Kong.
water, the triplet state mostly promptly experiences
a
decarboxylation process to form a triplet biradical carbanion
species. Both hydrogen abstraction and decarboxylation occur
and take place to an appreciable extent in aqueous solutions
with water concentrations between 50% and 90% water.
Results and discussion
Experimental and computational methods
KP is a benzophenone derivative and hence exhibits benzo-
phenone-like photochemistry in acidic and neutral solution.
The triplet state benzophenone (3[BP]) (without a carboxyl
group when compared with KP) will abstract hydrogen from a
water molecule in neutral solution (1 : 1 H2O–CH3CN) to form
a ketyl radical species, but no ketyl radical intermediate is
(S)-Ketoprofen (KP) was obtained from Aldrich with 99%
purity and used as received. Spectroscopic grade acetonitrile
(MeCN), potassium dihydrogen phosphate, anhydrous
sodium phosphate dibasic of reagent grade from Aldrich,
and deionized water were used for preparing sample solutions.
Experimental sample solutions of KP with a concentration of
1 mM in neat MeCN and mixed aqueous solvents were used in
the ns-TR3 experiments. Seven sample solutions including neat
MeCN, 1 : 9 H2O–MeCN (10% water), 3 : 7 H2O–MeCN
(30% water), 1 : 1 H2O–MeCN (50% water), 7 : 3 H2O–MeCN
(70% water), 9 : 1 H2O–MeCN (90% water) and 9 : 1
3
observed in CH3CN solution.17 This indicates that the [BP],
and very likely KP as well, do not abstract hydrogen easily
from CH3CN but from a water molecule instead. This may be
due to the hydrogen bond attraction in water moves the
molecules closer together and forms an electrostatic-type
3
attraction between them. Therefore, the propensity for [BP]
ꢀc
This journal is the Owner Societies 2010
Phys. Chem. Chem. Phys., 2010, 12, 4800–4808 | 4801

View Article Online
3
and [KP] to abstract hydrogen from H2O over CH3CN (the
latter having an apparently weaker C–H bond) may be due to
the initial hydrogen bonding of the O–H to the carbonyl of the
benzophenone moiety. This might be sufficient to reduce the
activation barrier such that abstraction occurs via a pathway
that seems intrinsically less likely based on examination of the
thermochemisty (that is, the relative bond strengths) that does
not consider explicit hydrogen bonding effects on the reaction.
Besides, the evidence for hydrogen abstraction from water by
a triplet of benzophenone was previously found in water
solution.25 The initial very fast photophysical processes
produce the triplet state ketoprofen (3[KP]) that can react
with a nearby molecule to form a ArPK radical via a hydrogen
abstraction reaction in a 1 : 1 H2O–MeCN neutral solution
and also in an acidic solution and this reaction has been
observed by ns-TR3 experiments employing a 266 nm pump
laser and 319.9 nm probe laser excitation wavelengths.6 The
growth time constant of the ArPK radical was determined to
be 100 ns in a 1 : 1 H2O–MeCN solution and 204 ns in a 1 : 1
D2O–MeCN solution from ns-TR3 experiments. A typical
primary kinetics isotope effect of kD/kH = 2 was observed.6
3
This demonstrates that the [KP] species reacts with the poor
hydrogen donor water molecules by a hydrogen abstraction
process to form the ketyl radical species. In order to learn
more about how water molecules are involved in the photo-
chemistry of KP, ns-TR3 experiments were carried out in
H2O–MeCN solutions containing 10%, 30%, 50%, 70%,
90% (vol%) water.
Fig. 1 Ns-TR3 spectra of 1 mM KP in 1 : 9 H2O–MeCN solution
obtained using a 266 nm pump excitation wavelength and a 319.9 nm
probe wavelength at various time delays indicated next to the spectra.
The asterisk (*) marks regions affected by solvent subtraction artifacts
and/or stray light.
Triplet state hydrogen abstraction—generation of the ArPK
radical
Fig. 1 shows ns-TR3 spectra of 1 mM KP obtained in a 10%
water–MeCN solution. Two intermediates can be observed
over the 10 ns to 1 ms time-scale. The first intermediate has
characteristic Raman bands at 966, 1173, 1121 and 1538 cmÀ1
and is assigned to the 3[KP] species because the resonance
Raman spectrum of this species displays excellent agreement
3
with that of an authentic [KP] spectrum which was obtained
in neat acetonitrile solvent (the comparison of these ns-TR3
spectra are shown in Fig. 2).6 The integrated areas of the
3
characteristic Raman band at 1538 cmÀ1 of the [KP] species
as a function of the time delay can be fit satisfactorily by a
single exponential function and gave a time constant for the
decay of ca. 119 ns (see Fig. 1Sw). As the 3[KP] decays, a small
shoulder peak (at 1582 cmÀ1) belonging to the second species
gradually increases and it appears that the 3[KP] is the
precursor of the second intermediate species. The second
species is detected at 10 ns time delay, which appears at an
earlier time than in neat acetonitrile where the 1582 cmÀ1
Raman band appears at 30 ns and this indirectly suggests that
water molecules are involved in the decay of the 3[KP]
intermediate.
Fig. 3 shows the ns-TR3 spectra of 1 mM KP obtained in
30% and 50% water–MeCN solutions and these spectra are
very similar to those shown in Fig. 1 except for the kinetics of
the transient species. The 3[KP] intermediate is observed in the
range of the time resolution of the ns-TR3 experiments and
the Raman band (1582 cmÀ1) for the second species is also
Fig. 2 Comparison of the experimental resonance Raman spectrum
of KP in a 1 : 9 H2O–MeCN solution acquired at a time delay of 5 ns
(top) and the authentic 3[KP] spectrum in neat acetonitrile solvent
acquired at a time delay of 5 ns (bottom). The asterisk (*) marks
regions affected by solvent subtraction artifacts and/or stray light.
observed at 10 ns time delay. The intensity of the second
species gradually becomes larger as the concentration of water
increases. The ns-TR3 spectra obtained at time delays later
than 400 ns belong to the second species and the resonance
Raman spectrum of the second species in 30% and 50%
aqueous solution is almost identical to that of an authentic
ArPK radical which is obtained in neat isopropanol solvent
and detected with the same probe wavelength as used here.6
Hence the second species observed on the nanosecond time
ꢀc
This journal is the Owner Societies 2010
4802 | Phys. Chem. Chem. Phys., 2010, 12, 4800–4808

View Article Online
scale in 30% and 50% water solutions shown in Fig. 3 is
reasonably assigned to the ArPK radical. Although the ArPK
radical is typically generated in H-donor solvents such as
alcohols and amines, the observation of the ArPK radical
here in 10%, 30% and 50% water–MeCN solutions is
consistent with previous results that reported the observation
of the ketyl radical species.6,13 This observation provides
valuable evidence for a hydrogen abstraction reaction with
decay time constant of ca. 107 ns and the time dependence of
the 1582 cmÀ1 Raman band integrated areas for the ArPK
radical were fit by a two-exponential function with a ca. 109 ns
growth time constant and a ca. 230 ns decay time constant. In
the 50% water solution, the time dependence of the 1538 cmÀ1
Raman band integrated areas for the 3[KP] intermediate were
fit by a single-exponential function with a decay time constant
of ca. 97 ns and the time dependence of the 1582 cmÀ1 Raman
band areas for the ArPK radical were fit by a two-exponential
function with a ca. 118 ns growth time constant and a ca.
910 ns decay time constant. Examination of the ns-TR3
spectra in the 10%, 30%, 50% water–MeCN solutions shows
3
water molecules by the [KP] intermediate.
In order to determine the kinetics of those intermediates, the
contribution of other species generated at post time delays
need to be removed. The characteristic 1538 and 1582 cmÀ1
3
3
Raman bands in Fig. 3 which belong to the [KP] and ArPK
that the time constant for the decay of the [KP] intermediate
radical respectively were fit with a Lorentzian function to
obtain the time dependence of these two species and the
kinetics determined from this analysis are displayed in
Fig. 3. In the 30% water–MeCN solution, the time dependence
gradually decreases as the water concentration increases and
this suggests that the reaction rate of the hydrogen abstraction
reaction for the 3[KP] species is significantly affected by the
surrounding water molecules. The time constant for the
growth of the ArPK radical is somewhat longer than the time
constant for the decay of the 3[KP] intermediate suggests that
3
of the 1538 cmÀ1 Raman band integrated areas for the [KP]
intermediate were fit by a single-exponential function with a
Fig. 3 Ns-TR3 spectra of 1 mM KP in 3 : 7 (A, left) and 1 : 1 (B, right) H2O–MeCN solution obtained using 266 nm pump excitation wavelength
and 319.9 nm probe wavelength at various time delays indicated next to the spectra and below are its corresponding kinetics. The asterisk (*) marks
regions affected by solvent subtraction artifacts and/or stray light. In 30% water solution, the time constant for the decay of the 3[KP] (squares) is
ca. 107 ns and the time constant for the growth and decay of the ArPK radical (triangles) are ca. 109 ns and ca. 230 ns respectively. In 50% water
solution, the time constant for the decay of the 3[KP] (squares) is ca. 97 ns and the time constants for the growth and decay of the ArPK radical
(triangles) are ca. 118 ns and ca. 970 ns respectively. The inset displays an expanded view of the kinetics for the early 200 ns time delay.
ꢀc
This journal is the Owner Societies 2010
Phys. Chem. Chem. Phys., 2010, 12, 4800–4808 | 4803

View Article Online
a small portion of the 3[KP] species may be quenched by
oxygen under the open air conditions of the experiments. The
time constant for the decay of the ArPK radical gradually
increases as the water concentration increases. This seems
mainly attributed to there being more ArPK radicals generated
as the water concentration increases. Therefore, in the 50%
water solution, the ArPK radical is still observed at a 1.5 ms
time delay.
As the ArPK radical decays, a new species is observed after
the decay of the ArPK radical which may form from the
self-coupling reaction between ArPK radicals or the cross-
coupling reaction between ArPK radicals and water molecules.
Different concentrations of KP and pump power conditions
experiments were carried out to determine the decay time
constants of the ArPK radical. These results make it clear
that the kinetics for the decay of ArPK radical does not
depend on the concentration of KP and the pump power
conditions. This implies that the third species does not arise
from the self-coupling reaction of the ArPK radical. On the
contrary, it is more likely to couple with the surrounding water
molecules which are present in a large excess. Therefore, the
third species is tentatively assigned to (2S)-2-[3-(hydroxyl-(4-
hydroxycyclohexa-2,5-dienylidene)-methyl)-phenyl]-propanoic
acid (p-OH).6
Fig. 4 Ns-TR3 spectra of 1 mM KP in a 7 : 3 H2O–MeCN solution
obtained using a 266 nm pump excitation wavelength and a 319.9 nm
probe wavelength at various time delays indicated next to the spectra.
The asterisk (*) marks regions affected by solvent subtraction artifacts
and/or stray light.
Triplet state hydrogen abstraction and triplet state
decarboxylation reactions-simultaneously produce the ArPK
radical and the triplet protonated biradical carbanion species
The rate constants exhibit a water concentration dependence
in solutions with less than 50% water present. In addition, the
kinetics of the decarboxylation reaction have been examined
due to the strong dependence of the rate constant on
the polarity of the solvent.26 The polar protic solvent was
proposed to be the donor of a proton or hydrogen to aid the
formation of an intramolecular proton transfer as the initial
step to induce the photodecarboxylation.27 Thus a further
increase in the concentration of water may lead KP to exhibit
an obvious decarboxylation process. Fig. 4 displays the
ns-TR3 spectra of 1 mM KP obtained in a 7 : 3 H2O–MeCN
aqueous solution. The Raman bands at 966, 1173, 1121 and
1538 cmÀ1 can be observed at early time delays and these four
Raman bands are assigned to the 3[KP] species because the
spectra of this species are essentially identical to that of the
3[KP] intermediate obtained in neat acetonitrile. Examination
of the ns-TR3 spectra in Fig. 4 shows that the resonance
Raman signals of the intermediates are greatly enhanced and
an obvious spectral change is present so that four new bands
appear at 966, 1024, 1282 and 1519 cmÀ1. Fig. 5 shows a
comparison of the experimental ns-TR3 spectrum of KP in a
1 : 1 water–MeCN solution acquired at a time delay of 800 ns
(top) and the experimental ns-TR3 spectrum of KP in a 7 : 3
water–MeCN solution acquired at a time delay of 800 ns
(bottom). Our previous work demonstrated that a triplet
protonated biradical carbanion species was produced through
a decarboxylation process in a PBS solution when 266 nm is
used as the pump laser. Fig. 6 shows a comparison of the
experimental resonance Raman spectrum obtained in a 7 : 3
water–MeCN solution acquired at a time delay of 800 ns and
the spectrum of the triplet protonated biradical carbanion
obtained in a 9 : 1 PBS–MeCN solvent acquired at a time delay
of 800 ns. The Raman characteristic feature at 1519 cmÀ1 of
the triplet protonated biradical carbanion species can be seen
to emerge in the ns-TR3 spectra obtained in the 70% neutral
aqueous solution demonstrates that this new species is the
triplet protonated biradical carbanion species. The overlapping
of the resonance Raman spectra for the triplet protonated
biradical carbanion and ArPK radical leads to an up-shift of
two Raman bands (987 cmÀ1 and 1177 cmÀ1). Therefore, two
intermediates are clearly produced in the same range of time
delays and this is easily observed in the ns-TR3 spectra.
Previous time-resolved transient absorption studies had
difficulties in clearly distinguishing these intermediates due to
their overlapping broad absorption spectra.13 However, time-
resolved resonance Raman can readily overcome this problem
because each intermediate has its own fingerprint vibrational
frequencies with higher resolution than their absorption
spectra.
In order to remove the contributions from other intermediates
(including the 3[KP] intermediate and other species) generated
at post time delays, the Raman spectrum of the ArPK radical
was extracted by subtracting the spectrum of the 3[KP]
obtained in neat acetonitrile and the spectrum of the triplet
protonated biradical carbanion obtained in the PBS solution
from that acquired at a different time delay in the 70% water
solution shown in Fig. 4. Fig. 2Sw shows the extracted ns-TR3
spectra of the ArPK radical. The characteristic Raman bands
3
at 1538 and 1582 cmÀ1 in Fig. 4 belonging to the [KP] and
ꢀc
This journal is the Owner Societies 2010
4804 | Phys. Chem. Chem. Phys., 2010, 12, 4800–4808

View Article Online
Fig. 5 Comparison of the experimental ns-TR3 spectrum of KP in a
1 : 1 H2O–MeCN solution acquired at a time delay of 800 ns (top) and
the experimental ns-TR3 spectrum of KP in a 7 : 3 H2O–MeCN
solution acquired at a time delay of 800 ns (bottom). The asterisk
(*) marks regions affected by solvent subtraction artifacts and/or stray
light.
Fig.
7
The time dependence of the 1538 cmÀ1 Raman band
integrated areas of 3[KP] (closed squares) in a 7 : 3 water–MeCN
solution were fit by a single-exponential function with a decay time
constant of ca. 46 ns and the time dependence of the 1582 cmÀ1
Raman band areas of the ArPK radical (closed triangles) were fit by a
two-exponential function with a ca. 220 ns growth time constant and a
ca. 1120 ns decay time constant. The inset displays an expanded view
of the kinetics for the 3[KP] intermediate. The data displayed here were
derived from the ns-TR3 spectra shown in Fig. 4.
giving rise to the triplet protonated biradical carbanion
at higher water concentrations, it is possible that the 3[KP]
species decays and then generates the triplet protonated
biradical carbanion.
There appear to be three pathways for the decay of the
3[KP] intermediate: one is quenching by oxygen dissolved in
the water solution with a quenched rate constant k1; the
second is the formation of the ArPK radical by a hydrogen
abstraction reaction with water molecules with a reaction rate
constant k2 (220 nsÀ1); the third produces the triplet proto-
nated biradical carbanion species through a decarboxylation
process with a reaction rate constant k3. The observed rate
constant (46 nsÀ1) for the decay of the 3[KP] in the 70% water
solution is the sum of these three reaction pathways. Therefore,
Fig. 6 Comparison of the experimental ns-TR3 spectrum of KP in a
7 : 3 H2O–MeCN solution acquired at a time delay of 800 ns (top) and
the spectrum of the triplet protonated biradical carbanion obtained in
a 9 : 1 PBS–MeCN solution acquired at a time delay of 800 ns
(bottom). The asterisk (*) marks regions affected by solvent subtraction
artifacts and/or stray light.
3
the growth time constant for the [KP] photodegradation by
the decarboxylation reaction in the 70% water solution is ca.
102 ns (obtained using 46 nsÀ1 = 220 nsÀ1 + 136 nsÀ1 + k2).
3
Oxygen will quench the [KP] with a ca. 136 ns time constant
ArPK radical respectively were fit with a Lorentzian function
to obtain the time dependence of these two species and the
kinetics obtained from this analysis are displayed in Fig. 7.
The integrated areas of the characteristic 1538 cmÀ1 Raman
in neat acetonitrile. Since the solubility of oxygen in the 70%
water solution is smaller than that of in neat acetonitrile, the
quenching rate k1 should be smaller than 136 nsÀ1. Thus, the
reaction rate constant k2 for the decarboxylation process is
likely larger than 102 nsÀ1. In other words, the growth time
constant for the 3[KP] photodegradation by the decarboxylation
reaction in the 70% water solution is smaller than 102 ns.
It was difficult to obtain the decay time constant for the
triplet protonated biradical carbanion due to overlapping of
the resonance Raman spectra. The observation of the triplet
protonated biradical carbanion in the 70% water solution
indicates the decarboxylation process becomes more favored
at higher concentrations of water in aqueous solutions. The
ns-TR3 spectra obtained in 90% water solution are similar to
those obtained in 70% water solution and also to those obtained
in 90% PBS solution (shown in Fig. 3S). Examination of Fig. 8
3
band for the [KP] intermediate were fit by a single-exponential
function with a time constant of ca. 46 ns. The integrated areas
of the characteristic 1582 cmÀ1 Raman band for the ArPK
radical were fit by a two-exponential function with a ca. 220 ns
growth time constant and a ca. 1120 ns decay time constant.
As mentioned above, the 3[KP] species is the precursor for the
formation of the ArPK radical. The time constant for the
decay of the 3[KP] (46 ns) is noticeably smaller than the growth
of the ArPK radical (220 ns), within the uncertainty of the
measurements, and this indicates that another pathway for the
3
decay of the [KP] species exists in the 70% water solution.
Although we cannot exclude that there may be other pathways
ꢀc
This journal is the Owner Societies 2010
Phys. Chem. Chem. Phys., 2010, 12, 4800–4808 | 4805

View Article Online
Fig.
9
The time dependence of the 1578 cmÀ1 Raman band
integrated areas of the triplet protonated biradical carbanion species
were fit by a two-exponential function with a ca. 1547 ns decay time
constant in 9 : 1 H2O–MeCN solution (closed squares).
most biological reactions take place in water solution
environments and our results suggest that the triplet protonated
biradical carbanion is very likely responsible for much of the
photosensitization and phototoxicity of KP. Scaiano and
coworkers tried to extend the life-time of the carbanion to
explore the carbanion’s reactivity and our results appear
consistent with theirs.28
Fig. 8 Ns-TR3 spectra of 1 mM KP in a 9 : 1 H2O–MeCN solution
obtained using a 266 nm pump excitation wavelength and a 319.9 nm
probe wavelength at various time delays indicated next to the
spectra are shown. The asterisk (*) marks regions affected by solvent
subtraction artifacts and/or stray light.
Proposed photochemical mechanism of KP in aqueous solutions
with varying concentrations of water
The propionic acid chain substitution in KP results in a
ground state acid–base equilibrium with a pKa = 4.7 and this
indicates that the main component in a neutral aqueous
solution is the neutral KP.29 Thus, the neutral KP is excited
and undergoes efficient photophysical processes to produce the
3[KP] which appears to be a predominant intermediate formed
in a neutral aqueous solution. According to our experimental
results, the 3[KP] species exhibits a water concentration
dependent photochemistry. Scheme 1 shows a proposed
photochemical mechanism for KP in water solutions with
varying concentrations of water. In water solutions with lower
water concentrations (like those with less than 50% water
concentration), the photolysis of KP leads to formation of the
first excited singlet state (S1) and that then forms the 3[KP]
species through a highly efficient intersystem crossing (ISC)
process.6,30 This triplet state easily undergoes mainly photo-
reduction by hydrogen abstraction from water molecules and
forms the ArPK radical with different rate constants depending
on the water concentration in the aqueous solutions. For
aqueous solutions with a modest to high water concentration
shows that the Raman band of 1519 cmÀ1 is obviously
enhanced compared with the one obtained in 70% water
solution at the same time delay. This indicates that a noticeable
amount of 3[KP] will experience the decarboxylation to
generate the triplet protonated biradical carbanion.
An interesting phenomena observed in our study is that the
decay time constants of the triplet protonated biradical
carbanion in the 90% PBS solution (1770 ns, shown in
Fig. 4Sw) and the 90% water solution (1547 ns, shown in
Fig. 9) are noticeably longer than the decay time constant of
triplet protonated biradical carbanion (411 ns) in the 70% PBS
solution.6 This indicates that increasing the concentration of
water can help KP undergo the decarboxylation process to
generate the carbanion species more easily. On the other hand,
this also provides an implication that water can stabilize the
triplet protonated biradical carbanion but the reason for this
stabilization is not clear. Thus the solvent effects on the
carbanion and triplet protonated biradical carbanion were
investigated using the UB3LYP/6-311G** optimized geometries
of the gas-phase stationary points and then using the SCRF
method based on a polarizable continuum model (PCM)
developed by Tomasi and co-workers.23 The calculations were
done for acetonitrile and water solvents and show that water
can noticeably decrease the free energy of the carbanion and
triplet protonated biradical carbanion by about ca. 1.3 and ca.
0.6 kcal respectively and this suggests that the solvent polarity
may have a noticeable impact on the lifetime of the triplet
protonated biradical carbanion. It is important to note that
3
(like solutions with between 50% and 90% water), the [KP]
species appears to lead to appreciable degrees of hydrogen
abstraction and decarboxylation reactions. These two photo-
chemical pathways lead to formation of the ArPK radical and
the triplet protonated biradical carbanion with different
kinetics dependent on the water concentration of the solution.
A further increase in the water concentration (such as aqueous
solutions with 90% and greater concentration of water) leads
to the decarboxylation process becoming more predominant
ꢀc
This journal is the Owner Societies 2010
4806 | Phys. Chem. Chem. Phys., 2010, 12, 4800–4808

View Article Online
to generate a greater degree of the carbanion species. The
detailed mechanism for the decarboxylation of the neutral
triplet state KP is not completely clear and further investigation
on this mechanism will be done in the future. Finally, we note
that the triplet protonated biradical carbanion can undergo
a water-assisted proton shift and radiationless decay to the
singlet ground state 3-ethylbenzophenone product.12,31 As
mentioned above, KP is a weak acid, the dissociated carboxylate
form should be in equilibrium with the acid form for either the
ground state KP or triplet state KP. A small amount of
deprotonated KP will easily undergo decarboxylation and give
arise to the carbanion intermediate (shown in Scheme 1).
Scaiano and coworkers reported that the strong transient
absorption peak of the carbanion intermediate appeared at
650 nm.5,7 In the case of the ns-TR3 experiments performed on
KP reported here, the carbanion intermediate may not have
been observed when 319.9 nm was used as the probe laser
wavelength in the experiments. One of our recent papers6
reported TD-DFT calculations for the carbanion and triplet
protonated biradical carbanion and found that the triplet
protonated biradical carbanion has a higher oscillator strength
near the probe laser wavelength (319.9 nm) while the absorption
spectrum of the carbanion displayed a relatively weak
oscillator strength near the 319.9 nm probe laser wavelength.
Thus, the carbanion intermediate was probably not observed
in ns-TR3 experiments because it had a very small Raman
signal due to its weak resonance enhancement. The carbanion
and its resonant structure will promptly receive a proton and
generate the triplet protonated biradical carbanion species in
aqueous solution.6 The triplet protonated biradical carbanion
transforms into the final product (3-ethylbenzophenone)
presumably through electron transfer.12
study indicates that the photochemistry of KP has substantial
water concentration dependence and suggests that the ratio or
concentrations of the ArPK radical and the triplet protonated
biradical carbanion reactive intermediates will vary with
the hydrophilic versus hydrophobic nature of the micro-
environments present in the different parts of complex bio-
logical systems.
Conclusions
We have explored the photochemistry of KP in aqueous
solutions with varying water concentrations using nanosecond
time-resolved resonance Raman spectroscopy. The ns-TR3
spectra reveal that the 3[KP] species can be observed and then
it reacts to produce mainly an ArPK radical species by
hydrogen abstraction from water molecules in solutions that
have less than 50% water. The rate constant of the hydrogen
abstraction reaction will be affected by the concentration
of water. In solution with water concentration between
50% and 90%, the photoinduced hydrogen abstraction and
decarboxylation outcomes of the 3[KP] are two competitive
processes that lead to the formation of the corresponding
intermediates (the ArPK radical and the triplet protonated
biradical carbanion species). For aqueous solutions with more
3
than 90% water, the [KP] can be observed at an early time
delay (20 ns) and then mainly undergoes decarboxylation to
form the triplet protonated biradical carbanion species. Thus,
the photo-induced decarboxylation of KP in higher water
concentration solution can be (at least in part) a triplet-
mediated mechanism which leads to formation of the triplet
protonated biradical carbanion species. This species has a very
long lifetime in 90% water and 90% PBS solutions (with the
lifetime reaching to the microsecond time scale in some cases)
and is likely an important intermediate that reacts with
biological components in biological systems.
The photochemistry of KP appears to be strongly influenced
by its molecular environment.32,33 When dissolved in a plasma
which is similar to a water predominant environment (like the
90% water solution examined here), our results suggest that
the photodecarboxylation should be the predominant process
and the KP triplet state should be short-lived. In contrast,
when KP is immersed in a protein hydrophobic core which
would be more like an acetonitrile solution with lower
concentrations of water (such as the neat MeCN and 30%
water–MeCN solutions examined here), a higher contribution
of the hydrogen abstraction would be expected. Our current
Acknowledgements
This work was supported by a grant from the Research
Grants Council of Hong Kong (HKU 7035/08P), the award of
a Croucher Foundation Senior Research Fellowship (2006–07)
from the Croucher Foundation and an Outstanding Researcher
Award (2006) from the University of Hong Kong to D. L. P.
Scheme 1 A schematic diagram of the proposed photochemical mechanism of KP in different water concentration solutions is displayed.
ꢀc
This journal is the Owner Societies 2010
Phys. Chem. Chem. Phys., 2010, 12, 4800–4808 | 4807

View Article Online
22 P. Zhu, S. Y. Ong, P. Y. Chan, K. H. Leung and D. L. Phillips,
J. Am. Chem. Soc., 2001, 123, 2645.
References
1 F. Bosca, M. A. Miranda, G. Carganico and D. Mauleon, Photo-
chem. Photobiol., 1994, 60, 96.
23 E. Cances, B. Mennucci and J. Tomasi, J. Chem. Phys., 1997, 107,
3032.
2 F. Bosca and M. A. Miranda, J. Photochem. Photobiol., B, 1998, 43, 1.
3 H. Bagheri, V. Lhiaubet, J. L. Montastruc and N. Louis-Lalanne,
Drug Saf., 2000, 22, 339.
4 V. Lhiaubet-Vallet and M. A. Miranda, Pure Appl. Chem., 2006,
78, 2277.
5 G. Cosa, L. J. Martinez and J. C. Scaiano, Phys. Chem. Chem.
Phys., 1999, 1, 3533.
6 Y. P. Chuang, J. Xue, Y. Du, M. D. Li, H. Y. An and
D. L. Phillips, J. Phys. Chem. B, 2009, 113, 10530.
7 M. Laferriere, C. N. Sanrame and J. C. Scaiano, Org. Lett., 2004,
6, 873.
8 M. Lukeman and J. C. Scaiano, J. Am. Chem. Soc., 2005, 127, 7698.
9 S. Montis, I. Manet, F. Manoli, R. Morrone, G. Nicolosi and
S. Sortino, Photochem. Photobiol., 2006, 82, 13.
10 H. Suzuki, T. Suzuki, T. Ichimura, K. Ikesue and M. Sakai,
J. Phys. Chem. B, 2007, 111, 3062.
24 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. A. Montgomery, Jr.,
T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam,
S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi,
G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada,
M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida,
T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li,
J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo,
J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. Ochterski, P. Y. Ayala,
K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg,
V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain,
O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford,
J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz,
I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham,
C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill,
B. G. Johnson, W. Chen, M. W. Wong, C. Gonzalez and
J. A. Pople, GAUSSIAN 03 (Revision C.02), Gaussian, Inc.,
Wallingford, CT, 2004.
11 T. Suzuki, T. Okita, Y. Osanai and T. Ichimura, J. Phys. Chem. B,
2008, 112, 15212.
12 B. C. Borsarelli, S. Braslavsky, E. S. Sortino, G. Marconi and
S. Monti, Photochem. Photobiol., 2000, 72, 163.
13 F. Bosca, G. Carganico, J. V. Castell, M. J. Gomez-Lechon,
D. Hernandez, D. Mauleon, L. A. Martinez and M. A. Miranda,
J. Photochem. Photobiol., B, 1995, 31, 133.
25 R. V. Bensasson and J. C. Gramain, J. Chem. Soc., Faraday Trans.
1, 1980, 76, 1801.
26 D. Sicinska, D. G. Truhlar and P. Paneth, J. Am. Chem. Soc.,
2001, 123, 7683.
27 L. Ding, X. Chen and W. H. Fang, Org. Lett., 2009, 11,
1495.
28 G. Cosa, M. Lukeman and J. C. Scaiano, Acc. Chem. Res., 2009,
42, 599.
14 B. Przybilla, U. Schwab-Przybilla, T. Ruzicka and J. Ring, Photo-
dermatology, 1987, 4, 73.
15 B. Ljunggren, Photodermatology, 1985, 2, 3.
16 L. L. Costanzo, G. De Guidi, G. Condorelli, A. Cambria and
M. Fama, Photochem. Photobiol., 1989, 50, 359.
17 Y. Du, J. Xue, M. D. Li and D. L. Phillips, J. Phys. Chem. A, 2009,
113, 3344.
29 L. J. Martinez and J. C. Scaiano, J. Am. Chem. Soc., 1997, 119,
11066.
18 C. Ma, W. M. Kwok, W. S. Chan, Y. Du, J. T. W. Kan, P. H. Toy
and D. L. Phillips, J. Am. Chem. Soc., 2006, 128, 2558.
19 J. Xue, P. Y. Chan, Y. Du, Z. Guo, C. W. Y. Chung, P. H. Toy and
D. L. Phillips, J. Phys. Chem. B, 2007, 111, 12676.
20 Y. Du, X. Guan, W. M. Kwok, L. M. Chu and D. L. Phillips,
J. Phys. Chem. A, 2005, 109, 5872.
30 Y. Du, C. Ma, W. M. Kwok, J. Xue and D. L. Phillips, J. Org.
Chem., 2007, 72, 7148.
31 K. A. K. Musa, J. M. Matxain and L. A. Eriksson, J. Med. Chem.,
2007, 50, 1735.
32 V. Lhiaubet-Vallet, N. Belmadoui, M. J. Climent and
M. A. Miranda, J. Phys. Chem. B, 2007, 111, 8277.
33 S. Moti, I. Manet, F. Manoli and S. Sortino, Photochem. Photo-
biol. Sci, 2007, 6, 462.
21 P. Y. Chan, S. Y. Ong, P. Zhu, C. Zhao and D. L. Phillips, J. Phys.
Chem. A, 2003, 107, 8067.
ꢀc
This journal is the Owner Societies 2010
4808 | Phys. Chem. Chem. Phys., 2010, 12, 4800–4808