Transient intermediates in the laser flash photolysis of ketoprofen in aqueous solutions: Unusual photochemistry for the benzophenone chromophore

Article abstract of DOI:10.1021/ja970818t

The transient intermediates the nanosecond laser flash photolysis of ketoprofen, an arylpropionic acid, show the formation of a carbanion in aqueous solutions at pH 7.1. This carbanion incorporates spectroscopic properties from both a ketyl radical anion and a benzylic radical. The ketoprofen carboxylate undergoes biphotonic photoionization, a process that contributes less than 10% to its photodecomposition- and leads to a benzylic-type radical after decarboxylation with a rate constant ≤1 x 10⁷ s^-1. On the other hand, the carbanion forms monophotonically and the unsuccessful attempts to sensitize the formation of the ketoprofen triplet excited state in aqueous solutions suggest that the carbanion precursor is either an excited singlet state or an extremely short-lived triplet. In organic solvents of lower polarity, the excited triplet state is readily detectable.

Full text of DOI:10.1021/ja970818t

11066
J. Am. Chem. Soc. 1997, 119, 11066-11070
Transient Intermediates in the Laser Flash Photolysis of
Ketoprofen in Aqueous Solutions: Unusual Photochemistry for
the Benzophenone Chromophore
L. J. Mart´ınez and J. C. Scaiano*
Contribution from the Department of Chemistry, UniVersity of Ottawa, 10 Marie Curie,
Ottawa, Ontario K1N 6N5, Canada
ReceiVed March 13, 1997X
Abstract: The transient intermediates in the nanosecond laser flash photolysis of ketoprofen, an aryl propionic acid,
show the formation of a carbanion in aqueous solutions at pH 7.1. This carbanion incorporates spectroscopic properties
from both a ketyl radical anion and a benzylic radical. The ketoprofen carboxylate undergoes biphotonic
photoionization, a process that contributes less than 10% to its photodecomposition and leads to a benzylic-type
radical after decarboxylation with a rate constant g1 × 107 s-1
. On the other hand, the carbanion forms
monophotonically and the unsuccessful attempts to sensitize the formation of the ketoprofen triplet excited state in
aqueous solutions suggest that the carbanion precursor is either an excited singlet state or an extremely short-lived
triplet. In organic solvents of lower polarity, the excited triplet state is readily detectable.
Introduction
the transient intermediates produced after excitation of KP in
solution have been reported to support either mechanism.18
The photochemistry of aryl ketones in solution and micelles
has been extensively studied.1-6 Aryl ketones find a variety of
biochemical applications as site-specific reagents for probing
proteins and nucleic acids.7 They also have pharmacological
applications as drugs or sunscreening agents.8-11 Benzophenone
(BP) is probably the best known aryl ketone whose photochem-
istry is characterized by an efficient intersystem crossing that
populates an n,π* triplet state.2,5,12 In aqueous solutions,13
benzophenone undergoes biphotonic photoionization and triplet-
triplet annihilation involving electron transfer to yield a radical
pair.
O
COOH
KP
We have employed nanosecond laser flash photolysis tech-
niques to elucidate the photodecarboxylation mechanism of the
KP carboxylate in aqueous buffered solutions.
Experimental Section
Ketoprofen (KP) is a non-steroidal anti-inflammatory agent
whose chemical structure is basically that of a substituted
benzophenone. Product studies of the UV irradiation of KP in
aqueous solutions14,15 point to an efficient photodecarboxylation
process of the KP carboxylate (pKa ) 4.7). Two mechanisms
have been proposed to account for the photodecarboxylation
process (i.e. ionic vs radical).16,17 Nevertheless, no studies on
Materials. Ketoprofen (2-[3-benzoylphenyl]propionic acid), potas-
sium peroxydisulfate (K2S2O8), sodium phosphate monobasic, sodium
phosphate dibasic, citric acid, and cysteine were purchased from Sigma
Chemical Co. and used as received. Sorbic acid (2,4-hexadienoic acid)
and 4-methoxyacetophenone were from Aldrich Chemical Company
Inc. 1-azaxanthone was from Lancaster, and phenol was purchased
from Fisher Scientific. All solvents were HPLC grade and were used
without further purification. Water was purified through a Millipore
MilliQ system.
Nanosecond Laser Flash Photolysis. Ketoprofen samples at
concentrations ranging from 0.1 to 0.3 mM were prepared in 10 mM
phosphate buffer at pH 7.1. All the transient spectra and kinetics were
recorded by employing a flow system with a 7 × 7 mm2 Suprasil quartz
cell with a 2 mL capacity. The laser flash photolysis system has been
previously described.19 Briefly, samples were excited with a Lumonics
EX-530 laser with a Xe/HCl/Ne mixture generating pulses at 308 nm
of ∼6 ns and e60 mJ/pulse. For the sensitization experiments we
employed either the third harmonic from a Surelite Nd/YAG laser (355
nm, 25 mJ per 8 ns pulse) when exciting 1-azaxanthone or a Molectron
UV-24 nitrogen laser for exciting the p-methoxyacetophenone at 337
X Abstract published in AdVance ACS Abstracts, October 15, 1997.
(1) Scaiano, J. C. J. Photochem. 1973/74, 2, 81.
(2) Turro, N. J. Modern Molecular Photochemistry; Benjamin/Cummings
Publishing Co.: Menlo Park, 1978; p 628.
(3) Scaiano, J. C.; Selwyn, J. C. Can. J. Chem. 1981, 59, 2368.
(4) Scaiano, J. C.; Abuin, E. B.; Stewart, L. C. J. Am. Chem. Soc. 1982,
104, 5673.
(5) Bensasson, R. V.; Gramain, J.-C. J. Chem. Soc., Faraday Trans. 1
1980, 76, 1801.
(6) Beckett, A.; Porter, G. Trans. Faraday Soc. 1963, 59, 2051.
(7) Dorma´n, G.; Prestwich, G. D. Biochemistry 1994, 33, 5661.
(8) Schallreuter, K. U.; Wood, J. M.; Farwell, D. W.; Moore, J.; Edwards,
H. G. M. J. InVest. Dermatol. 1996, 106, 583.
(9) Silva, R.; Almeida, L. M. S.; Brandao, F. M. Contact Dermatitis
1995, 32, 176.
(10) Alomar, A.; Cerda, M. T. Contact Dermatitis 1989, 20, 74.
(11) Szczurko, C.; Dompmartin, A.; Michel, M.; Moreau, A.; Leroy, D.
Photodermatol., Photoimmunol. Photomed. 1994, 10, 144.
(12) Bell, J. A.; Linschitz, H. J. Am. Chem. Soc. 1963, 85, 528.
(13) Elisei, F.; Favaro, G.; Go¨rner, H. J. Photochem. Photobiol. A: Chem.
1991, 59, 243.
(14) Bosca´, F.; Miranda, M. A.; Carganico, G.; Mauleo´n, D. Photochem.
Photobiol. 1994, 60, 96.
(15) Constanzo, L. L.; De Guidi, G.; Condorelli, G.; Cambria, A.; Fama,
M. Photochem. Photobiol. 1989, 50, 359.
(16) Budac, D.; Wan, P. J. Photochem. Photobiol. A: Chem 1992, 67,
135.
(17) Studies in cellular systems suggest free radical involvement:
Chignell, C. F.; Sik, R. H. Photochem. Photobiol. 1995, 62, 205.
(18) While this manuscript was being reviewed we became aware of
similar studies conducted by S. Monti (Monti, S.; Sortino, S.; De Guidi,
G.; Marconi, G. J. Chem. Soc., Faraday Trans. Submitted for publication).
(19) Other aspects of the system are similar to those described earlier:
Scaiano, J. C. J. Am. Chem. Soc. 1980, 102, 7747. Scaiano, J. C.; Tanner,
M.; Weir, D. J. Am. Chem. Soc. 1985, 107, 4396.
S0002-7863(97)00818-4 CCC: $14.00 © 1997 American Chemical Society

Absolute Rate Constants for Carbanions in Solution
J. Am. Chem. Soc., Vol. 119, No. 45, 1997 11067
Figure 2. Transient intermediates in the photolysis of KP in aceto-
nitrile, 0.72 µs after the pulse (O), methanol, 0.13 µs after the pulse
(b), and cyclohexane, 0.32 µs after the pulse (]).
Figure 1. Transient absorption spectra obtained from the 308 nm laser
flash photolysis of KP in aqueous buffered solutions at pH 7.1: (A)
under N2-saturated (]) and O2-saturated (2) conditions; (B) under N2O-
saturated conditions (O). All spectra were recorded 72 ns after the laser
pulse.
nm. The signals from the monochromator/photomultiplier system were
initially captured by a Tektronix 2440 digitizer and transferred to a
PowerMacintosh computer that controlled the experiment with software
developed in the LabVIEW 3.1.1 environment from National Instru-
ments. Samples were purged with the appropriate gas (i.e. N2, O2, or
N2O) for 30 min before and during the acquisition of the transient
spectrum.
Figure 3. Power dependence for the formation of solvated electrons
(4) and the 600-nm transient (1) observed in aqueous solutions.
methanol, we also detected the contribution from ketyl radical
absorptions to the transient spectra.22
To determine if the photoionization of KP is a monophotonic
or biphotonic process, the optical density at 720 nm was
monitored as a function of laser energy between 0.5 and 4.5
Product Studies. Photolysis of ketoprofen solutions at a concentra-
tion of 0.5 mM in H2O and D2O was done by using as the excitation
source the Lumonics EX-530 308-nm laser. Samples received 100
pulses with an average power of 4 mJ/pulse measured at the sample
holder. The solvent was removed under vacuum, and the resulting
mixture was analyzed with a Fisons GC/MS 8060 8000 Series equipped
with a J&W DB5 30 m × 0.32 mm column.
mJ. Figure 3 shows that the e- data are best-fitted by a
aq
second-order polynomial, whereas for the signal at 600 nm a
linear dependence was obtained. For the 600-nm transient, the
linear fit goes through the origin, a requisite for monophotonic
processes.23 Furthermore, a log-log plot (not shown) yields a
slope of 0.93 for the 600-nm transient, while a slope of 2.05
was calculated for the e-aq data. Therefore, the KP carboxylate
undergoes biphotonic photoionization in aqueous solutions as
reported for benzophenone and other carboxylic acid deriva-
tives.24,25 As has been previously suggested for benzophenone,25
the extinction coefficient of the KP triplet at the excitation
wavelength (i.e. 308 nm) might be higher than that for the
ground state KP at the same wavelength, thus increasing the
probability of absorption of a second photon by the KP triplet
(reaction 1).
Results and Discussion
The transient absorption spectra obtained from the 308 nm
laser photolysis of 0.1 mM KP in aqueous solutions at pH 7.1
(Figure 1) show absorptions from 280 to 380 nm and from 500
to 800 nm. The broad band in the visible region has contribu-
tions from two species that decay almost simultaneously
following first-order kinetics (k720 nm ) 6.1 × 106 s-1; k600 nm
) 5.1 × 106 s-1). The effect of O2 and N2O on these signals
(Figure 1) establishes the contribution from the solvated electron
(e-aq) and shows clearly the spectrum of another species
absorbing at 600 nm. This result indicates that the KP
carboxylate photoionizes in aqueous solutions. From the optical
density immediately after the laser pulse at 720 nm, and knowing
3KP* f KP+ + e-
(1)
aq
The absorption for the benzophenone radical cation in
cyclohexane has a maximum at 850 nm.26 No such signal was
the extinction coefficient of e- 20 we estimate, at most, an 8%
,
aq
contribution from photoionization to the KP photodegradation.
In contrast, in solvents of lower polarity such as acetonitrile,
methanol, or cyclohexane (Figure 2), photoionization is not
observed, but the characteristic triplet-triplet benzophenone-
like absorption21 is readily detectable. In cyclohexane and
(22) The differences observed in organic solvents could be related to
the protonation state of this propionic acid. This is currently under
investigation.
(23) Lachish, U.; Shafferman, A.; Stein, G. J. Chem. Phys. 1976, 64,
4205.
(24) Sauberlich, J.; Brede, O.; Beckert, D. J. Phys. Chem. 1996, 100,
18101.
(25) Sauberlich, J.; Beckert, D. J. Phys. Chem. 1995, 99, 12520.
(26) Baral-Tosh, S.; Chattopadhyay, S. K.; Das, P. K. J. Phys. Chem.
1984, 88, 1404.
(20) Hug, G. L. Optical Spectra of Nonmetallic Inorganic Transient
Species in Aqueous Solution; National Bureau of Standards: Washington,
1981; NSRDS-NBS 69, p 160.
(21) Carmichael, I.; Hug, G. L. J. Phys. Chem. Ref. Data 1986, 15, 1.

11068 J. Am. Chem. Soc., Vol. 119, No. 45, 1997
Mart´ınez and Scaiano
observed for ketoprofen. If the radical formed from the
photoionization of KP was similar to BP•+, it would rapidly
react with water yielding a KP-OH• adduct (i.e. cyclohexadienyl
radical). However, the electron is most likely lost from the
carboxylate moiety followed by rapid decarboxylation. In N2O-
saturated solutions we detected the absorption of the KP-OH•
radical formed from the reaction with the hydroxy radical
generated from reaction 2.
N2O + e-aq f N2 + HO• + HO-
(2)
No absorption near 400 nm was detected under N2-saturated
conditions that would confirm the formation of a KP radical of
the cyclohexadienyl type.
An experiment using the photolysis of potassium peroxy-
disulfate (K2S2O8) to generate the KP radical formed after loss
of an electron showed the formation of a transient absorbing
with maxima at 330 and 360 nm. This transient is assigned to
the benzylic radical generated from the decarboxylation of the
initial aryloxyl KP radical.27,28
Figure 4. Transient absorption of the 600-nm transient as a function
of pH (right) and rate constant of decay of the 600-nm transient as a
function of pH (left).
lower than that for benzophenone,32 as expected for a negatively
charged species. The reaction of KP with e- could lead to
aq
hν
the formation of KP radical anions (reaction 3). Due to the
–•
2–
2SO4
S2O8
O
COO•
CH3
KP + e-aq f KP-
(3)
2–
SO4–• + KP
SO4
+
nature of the KP chromophore, we expect these radical anions
to have absorptions with maxima at 340 and 600 nm just as the
ketyl radical anions of benzophenones.33 The radical anion
formed would be rapidly protonated at pH 7.1 to yield a ketyl
radical absorbing in the 530-nm region.34 However, this process
is not occurring under the conditions employed in our experi-
ments for three reasons. First, the absorption at 600 nm is still
present when solutions where bubbled with N2O, an efficient
–CO2
O
CH3
e- scavenger. Second, the transient absorbing at 600 nm is
aq
From the growth observed at either 330 or 360 nm, a rate
constant >1.3 × 107 s-1 was estimated for the decarboxylation
process.29-31 The decay of this radical followed second-order
kinetics (k ) 2.1 × 106 M-1 s-1), thus implying that dimer-
ization through a radical recombination process is their principal
mode of decay. The direct excitation of KP under N2-saturated
conditions shows a band with maximum at a 330 nm and a
small shoulder with a maximum at 360 nm. Thus, the KP
radical generated after photoionization and decarboxylation is
a benzylic-type radical.
produced after absorption of one photon (Vide supra). If this
transient was formed by reaction 3, a biphotonic dependence
would have been expected. Third, the formation of ketyl
radicals was negligible at the low initial KP concentrations
employed. Therefore, we rule out significant contributions to
the 600-nm band from reaction 3.
We also monitored the pH dependence for the decay of the
600-nm transient in an attempt to determine its pKa (Figure 4).
At pH values close or below the ground state pKa of KP other
species contribute to the absorption at 600-nm, such as the KP
triplet (Vide infra). In addition, the nature of the absorbing
species changes from the carboxylate to the neutral acid, thus
only a pH-dependence half-wave curve could be obtained.
Nevertheless, our results based on signal amplitude suggest that
the pKa of this transient must be around 6.9, while the kinetics
indicate a pKa between 4.5 and 7.0, indicating that its acidity
constant is about two orders of magnitude lower than for the
benzophenone ketyl radical. The presence of a propionic acid
side chain at the meta position should not affect dramatically
the properties of a ketyl radical anion derived from the
benzophenone-like chromophore since delocalization of the
electron density through resonance cannot occur. If a biradical-
type structure dominated, one would expect the ketyl moiety
to behave essentially the same way as in a monoradical.35-37
The decay of the e-aq depends on the initial KP concentration
leading to a rate constant of 2.1 × 1010 M-1 s-1 for electron
trapping by KP. This second-order rate constant is slightly
(27) Chatgilialoglu, C. In Handbook of Organic Photochemistry; Scaiano,
J. C., Ed.; CRC Press: Boca Raton, FL, 1989; Vol. II, p 3.
(28) The transient absorption spectra of several benzoyloxyl radicals show
broad absorptions covering the visible range. In our case, the quenching
and decrease of the initial absorption in the visible region when solutions
are gassed with N2O or O2 discard the possibility that we are observing
similar aryloxyl radicals from KP. Since the lifetime of the transient
absorption remaining at 600 nm goes from 5.1 × 106 s-1 under N2 to 1.7
× 107 s-1 under O2 we believe that decarboxylation following the direct
excitation of KP occurs within the duration of the laser pulse. (a) Wang, J.;
Itoh, H.; Tsuchiya, M.; Tokumaru, K.; Sakuragi, H. Tetrahedron 1995, 51,
11967. (b) Wang, J.; Tateno, T.; Sakuragi, H.; Tokumaru, K. Tetrahedron
1995, 92, 53.
(29) The rate of oxidation of ketoprofen with the sulfate radical is quite
fast (k g 2 × 109 M-1 s-1). Therefore, we can only provide a lower limit
for the decarboxylation process.
(30) The estimate is in line with similar lower limits obtained for other
acyl radicals leading to benzylic radical centers: Steenken, S.; Warren, C.
J.; Gilbert, B. C. J. Chem. Soc., Perkin Trans. 2 1990, 335.
(32) Ross, A. B.; Mallard, W. G.; Helman, W. P.; Bielski, B. H. J.;
Buxton, G. V.; Cabelli, D. A.; Greenstock, C. L.; Huie, R. E.; Neta, P.
NDRL/NIST Solution Kinetics Database: Ver. 1, 1.0 ed.; NIST, U.S.
Department of Commerce: Gaithersburg, MD, 1992.
(33) Hayon, E.; Ibata, T.; Lichtin, N. N.; Simic, M. J. Phys. Chem. 1972,
76, 2072.
(34) The pKa of the benzophenone ketyl radical anion has been reported
to be 9.2 (see ref 28).
(35) Small, R. D.; Scaiano, J. C. J. Phys. Chem. 1977, 81, 828.
(31) For phenylacetoxyl radical the rate constant for decarboxylation has
been estimated as 5 × 109 s-1 (Hilborn, J. W.; Pincock, J. A. J. Am. Chem.
Soc. 1991, 113, 2683). The only absolute measurement corresponds to the
radical derived from 9-methylfluorene-9-carboxylate with a lifetime of ca.
55 ps: Falvey, D. E.; Schuster, G. B. J. Am. Chem. Soc. 1986, 108, 7419.

Absolute Rate Constants for Carbanions in Solution
J. Am. Chem. Soc., Vol. 119, No. 45, 1997 11069
Scheme 1
Figure 5. Transient intermediates in the 308 nm laser flash photolysis
of KP in aqueous solutions at pH 7.1 under N2O-saturated conditions
(to remove contributions from the solvated electron) (0) and under
N2-saturated conditions with 1 mM sorbic acid (b). Spectra were
recorded 64 ns after the laser pulse.
Caldwell et al.37 have determined the pKa for the biradical from
γ-phenylbutyrophenone as 10.2 in water; the value is directly
relevant to the KP case since the biradical also has a ketyl and
a benzylic site. Clearly the 600-nm transient from KP is not a
typical biradical with a ketyl radical anion site.
transfer from the 3KP* had occurred.40,41 This mechanism
would explain the lack of singlet oxygen sensitization reported
by Constanzo and co-workers.15 The efficient electron transfer
from the KP triplet state in polar environments would decrease
the amount of KP triplets available for energy transfer reactions
to oxygen.
The product studies reported by Bosca and co-workers14 in
aqueous solutions showed an almost quantitative conversion of
the ketoprofen carboxylate to the 3-ethylbenzophenone under
anaerobic conditions. Additional evidence for an ionic pathway
in the photodecomposition of KP comes from deuterium
incorporation at the benzilic position of the major photoproduct
(i.e. 3-ethylbenzophenone). We performed an experiment using
the 308 nm laser excitation of ketoprofen in aqueous solutions
and in D2O at pD > pKa. Incorporation of deuterium was
quantitative, as revealed by mass spectrometric analysis of the
product (data not shown).
The time-resolved results mentioned above and published
product studies14,15 are consistent with the assignment of the
transient intermediate absorbing with maxima at 330 and 600
nm to a carbanion that incorporates spectroscopic features
reminiscent of a ketyl radical anion and a benzylic radical (see
Scheme 1).42 This carbanion is quenched by oxygen, cysteine
(k ) 3.0 × 109 M-1 s-1), and phenol (1.04 × 108 M-1 s-1).
Only with cysteine were we able to detect absorptions with
maxima at 330 and 545 nm, indicative of the generation of ketyl
radicals. This is not surprising due to the known H-atom
donating ability of cysteine, and it exemplifies the biradicaloid
nature of this intermediate (Vide infra).43 The reaction with
phenol does not involve electron or hydrogen atom transfer since
we did not detect the formation of phenoxy radicals even at the
highest concentrations of phenol employed.
In solvents other than water, the KP triplet (3KP*) absorbs
with maxima at 320 and 520 nm. A large red shift (i.e., 80 nm
for the longer wavelength maximum) for the triplet in a polar
solvent like water would not be expected as the T-T absorption
maximum for benzophenone is not solvent dependent.21 An
experiment with sorbic acid (a conjugated diene and excellent
triplet quencher and e- scavenger) showed no effect on the
aq
transient absorption at 600 nm (Figure 5), indicating that it does
3
not correspond to KP*; further, its precursor is either not a
triplet or a remarkably short lived triplet.
Attempts to sensitize triplet formation in water with p-
methoxyacetophenone as a sensitizer lead to signals that could
be identified neither as triplet KP nor as the 600-nm transient
mentioned above, in spite of extensive (>90%; k ) 2.4 × 109
M-1 s-1 38
triplet quenching by KP. In contrast, sensitization
)
in acetonitrile leads to the readily detectable KP triplet. The
absence of the 600-nm signal in water is not particularly
significant, since the highest concentration of KP that could be
employed was limited by its competitive absorption of the
excitation beam. The shortest triplet lifetime for p-methoxy-
acetophenone achieved under these conditions was ca. 200 ns,
i.e. longer than the 600-nm signal from KP.39 Similarly, in
another experiment where we used 1-azaxanthone to sensitize
the formation of the KP triplet, we were unable to detect any
transient that would correspond to 3KP* even at KP concentra-
tions that would quench g75% of the 1-azaxanthone triplets
and lead to 1-azaxanthone triplet lifetimes of 250 ns. Interest-
ingly, we detected the formation of the 1-azaxanthone ketyl
radical anion (k g 2.7 × 107 s-1), thus suggesting that electron
Budac and Wan16 proposed that intramolecular electron
transfer in the triplet state could lead to the formation of a
carbanion having as precursor a biradical-like species. We
(36) Small, R. D., Jr.; Scaiano, J. C. J. Phys. Chem. 1977, 81, 2126.
(37) Caldwell, R. A.; Dhawan, S. N.; D. E., M. J. Am. Chem. Soc. 1985,
107, 5163.
(38) The rate constant is below diffusion control, but somewhat faster
than comparable isoenergetic triplet energy migration as expected for a
slightly exothermic transfer: Encinas, M. V.; Scaiano, J. C. Chem. Phys.
Lett. 1979, 63, 305.
(39) Naturally, the precursor process (sensitization) needs to be faster
than the decay of the intermediate of interest for this species to be detectable.
For the 600-nm transient this would require such high concentrations of
KP that selective excitation of the sensitizer would be impossible. The
experiments do show the absence of KP triplet in the time scale available
through sensitization.
(40) The 1-azaxanthone triplet is a much better electron acceptor than
the electron rich triplet of 4-methoxyacetophenone.
(41) Alternatively, the KP carbanion could be the species involved in
the electron transfer reaction. It is well documented that carbanions that
are not resonance stabilized can transfer an electron to photoexcited aromatic
hydrocarbons. Budac, D.; Wan, P. AdV. Carbohydr. Chem. 1996, 2, 147.
(42) A seminal study of the carbanion derived fron p-nitrophenyl acetate
has been reported: Craig, B. B.; Weiss, R. G.; Atherton, S. J. J. Phys.
Chem. 1987, 91, 5906.
(43) Thiols were among the earliest reported biradical scavengers:
Wagner, P. J.; Kelso, P. A.; Zepp, R. G. J. Am. Chem. Soc. 1972, 94, 7480.

11070 J. Am. Chem. Soc., Vol. 119, No. 45, 1997
Mart´ınez and Scaiano
mophore. Our results are consistent with carbanion formation,
either directly (Scheme 1), or following excited state electron
transfer through a 1,5-biradical intermediate. We presume that
the biradical would be the most stable electronic configuration
in the triplet manifold, while the carbanion structure will be
preferred in the singlet state.
Additional evidence for this mechanism comes from the
studies with the KP ethyl ester (Figure 6). The transient
intermediates observed for the KP ethyl ester corresponded to
the triplet state (λmax at 320 and 520 nm) and to ketyl radicals
(λmax at 330 and 545 nm). Esterification of KP shuts down the
intramolecular electron transfer process in the excited singlet
or triplet state and allows the detection of transients characteristic
of the benzophenone moiety. This is also in agreement with
the results from the photolysis of the KP free acid (pH 3.2)
showing the formation of e-aq, KP triplets, and ketyl radicals.
Figure 6. Transient absorption spectra for the ketoprofen ethyl ester
in a 1:1 MeOH/H2O mixture: (9) 72 ns, (O) 0.32 ms, (]) 0.79 ms,
and (+) 1.47 ms after the pulse.
In conclusion, the ketoprofen carboxylate undergoes bipho-
tonic photoionization. The formation of a carbanion as inter-
mediate has been confirmed, indicating that the photodegrada-
tion of ketoprofen in aqueous solutions proceeds mainly (≈90%)
through an ionic mechanism involving either the singlet state
or a very short lived triplet.
propose that the decay of the KP carbanion is due to protonation
by the solvent (i.e., water) with a lifetime of 150 ns. This
lifetime was independent of buffer concentration at least in the
5-60 mM range. The monophotonic dependence for its
formation and the lack of triplet absorption agree with the
mechanism proposed. The decay kinetics for this carbanion
show a small isotope effect (kH/kD ) 1.2). A small isotope effect
has also been reported for the benzyl carbanion.44 Preliminary
AM1 calculations show that the KP carbanion would have a
biradical-like structure in which part of the electron density is
shared with the carbonyl group in the benzophenone chro-
Acknowledgment. We thank Dr. Colin F. Chignell for
providing the ketoprofen ethyl ester and Mr. William Skene
for performing the AM1 calculations. Thanks are also due to
Dr. S. Monti for sharing with us a preprint of her manuscript.
This work has been supported by the Natural Sciences and
Engineering Research Council of Canada.
(44) Bockrath, B. a. L. M. D. J. Am. Chem. Soc. 1974, 96, 5708.
JA970818T