Efficient photodecarboxylation of aroyl-substituted phenylacetic acids in aqueous solution: A general photochemical reaction

Article abstract of DOI:10.1039/b006724p

Photolysis (254-350 nm) of a variety of aroyl-substituted phenylacetic acids and p-acetylphenylacetic acid in aqueous solution at pH > pK(a) resulted in efficient photodecarboxylation (Φ = 0.2-0.7), to give in most cases a single product arising via the corresponding arylmethyl carbanion, indicating that photodecarboxylation is an efficient and general reaction for these types of compounds.

Full text of DOI:10.1039/b006724p

Efficient photodecarboxylation of aroyl-substituted phenylacetic acids in
aqueous solution: a general photochemical reaction
Musheng Xu and Peter Wan*
Department of Chemistry, Box 3065, University of Victoria, Victoria, British Columbia, Canada V8W 3V6.
E-mail: pwan@uvic.ca
Received (in Corvallis, OR, USA) 15th August 2000, Accepted 18th September 2000
First published as an Advance Article on the web
Photolysis (254–350 nm) of a variety of aroyl-substituted
phenylacetic acids and p-acetylphenylacetic acid in aqueous
solution at pH > pKa resulted in efficient photodecarboxyla-
tion (F = 0.2–0.7), to give in most cases a single product
arising via the corresponding arylmethyl carbanion, indicat-
ing that photodecarboxylation is an efficient and general
reaction for these types of compounds.
The photodecarboxylation of organic carboxylates and car-
boxylic acids continues to be an area of active research
interest.1–3 This is an example of a simple photochemical
phores, these compounds might be expected to show reactivity
extrusion reaction that is comparable to the well-known
photodecarbonylation of ketones. However, mechanistically,
photodecarboxylation of carboxylic acids is more complex. The
mechanism in which carbon dioxide is eliminated from
photoexcited (organic) carboxylic acids has been classified into
several types with the simple heterolytic and homolytic
mechanisms being the most readily visualized.1 Some recent
papers dealing with photodecarboxylation of some non-
steroidal anti-inflammatory drugs (NSAIDs) suggest that new
insights and potentially new reactive systems await discov-
ery.2,3a–e A particular example that attracted our attention is the
highly efficiently (F = 0.75) photodecarboxylation of ketopro-
fen [1, eqn. (1)] in aqueous solution orginally reported by
as well. Finally, we have changed the nature of the ‘leaving
group’ in 6. Here, the leaving group is formally formaldehyde,
which is much less labile than CO2. If 6 reacts in an analogous
manner (i.e. deformylation to give the corresponding carban-
ion), then the reaction is indeed much more general than ever
suspected. In this report, we show that all of 2–6 react in the
manner anticipated (with varying degrees of efficiency, all
generally high) demonstrating that the photodecarboxylation of
aroyl and acetoyl-substituted phenylacetic acids is a general
reaction. Moreover, less labile leaving groups such as formal-
dehyde may also be used.
Carboxylic acids 2–5 were readily made by conversion of the
corresponding commercially available methyl-substituted de-
rivative to the a-bromo compound (NBS–CCl4), followed by
reaction with NaCN–EtOH and then hydrolysis using either
conc. HCl or aqueous H2SO4. All of the acids were purified by
repeated crystallization from aqueous EtOH to achieve > 98%
purity.† Keto-alcohol 6 was readily prepared by BH3–THF
reduction of 2.
Photolysis of 2–4 in 1+1 H2O–CH3CN (50 mL total volume;
water portion set to pH 7 with dilute NaOH after adding
compound to ensure complete dissociation of the acid;
≈ 1023 M substrate; Rayonet photochemical reactor; 254, 300
or 350 nm lamps; quartz vessel; solution cooled with a cold-
finger to ≈ 15 °C; purged continuously with argon; 5–20 min)
gave cleanly the corresponding photodecarboxylation product
[e.g., eqn. (2)] in high yield (up to 100% on sufficiently long
(1)
Costanzo et al.2 At least two groups3a–e have studied the
mechanism of the photodecarboxylation. The latest report we
are aware of is from Scaiano and co-workers3a who showed that
the reaction proceeds from the singlet excited state of the
carboxylate form and that water is not essential (but probably
helpful). Singlet state reactivity is unusual for benzophenone
photochemistry. The nature of the primary photochemical step
remains an issue. It could be either (a) a simple heterolytic
cleavage mechanism to generate the carbanion directly or (b) a
mesolytic mechanism in which there is an electron transfer from
the carboxylate oxygen to the ketone to generate a biradical
anion, followed by bond cleavage to give the carbanion.
Ketoprofen (1) is an aroyl-substituted phenylacetic acid.
Compounds of this general structural type are readily available
by routine synthesis. The high (and clean) reactivity of 1
suggested to us that other aroyl-substituted phenylacetic acids
might react in a similar manner and that a general type of
efficient photodecarboxylation awaits discovery. The availabil-
ity of compounds of similar structural type that also react will
certainly help in the elucidation of the reaction mechanism. We
have now made five analogs of 1, viz., 2–6. Acids 2 and 3 are
simple analogs of ketoprofen (1). If the photodecarboxylation of
1 is respresentative of a general process, compounds 2 and 3
should react, neglecting of course any dramatic substituent
effects on excited state reactivity that we have not taken into
account. Compounds 4 and 5 have a different carbonyl
chromophore (anthraquinone and acetophenone, respectively).
If the reaction is general for other types of carbonyl chromo-
(2)
photolysis time). No reaction was observed when the solution
was kept in the dark. Quantum yields for photodecarboxylation
were estimated using the reaction of ketoprofen (1) as a
1
secondary standard (F = 0.75, in neat water, pH 7)2 and H
NMR to monitor conversions. These experiments gave high
quantum yields for the photodecarboxylations for 2–4 (in 1+1
H2O-CH3CN, pH 7): 2 (F = 0.66); 3 (F = 0.62); 4 (F = 0.22).
Photolysis of 2–5 in neat CH3CN under photolysis times used
above gave no more than 5% reaction and the compounds could
be essentially recovered unchanged.
It’s clear from these results that (efficient) photodecarboxyla-
tion is a general photoreaction for aroyl-substituted phenyl-
DOI: 10.1039/b006724p
Chem. Commun., 2000, 2147–2148
This journal is © The Royal Society of Chemistry 2000
2147

acetic acids. Quantum yields decreased on lowering the pH‡
which is consistent with the carboxylate form being reactive,
consistent with the latest findings for ketoprofen (1) reported by
Scaiano and co-workers.3a Furthermore, photolysis of 2 and 3 in
D2O–CH3CN (as above) gave exclusively the corresponding a-
deutero-methyl product, consistent with a carbanion inter-
mediate [eqn. (2)]. Interestingly, photolysis of acetophenone
derivative 5 gave two products in a 7+3 mole ratio (F for loss of
substrate ≈ 0.2), respectively [eqn. (3)]: the expected simple
decarboxylation product (p-methylacetophenone) via a simple
phenylmethyl carbanion and a ‘dimeric’ product formally
derivable from the phenylmethyl radical via coupling. Inde-
pendent photolysis of p-methylacetophenone under identical
conditions gave no reaction. Although radical coupling prod-
ucts have not been observed for ketoprofen (1) or from any of
2–4, the fact that it is observed for 5 indicates that a radical-type
mechanism (e.g. initial electron transfer from the carboxylate to
the ketone)3a–c cannot be excluded for the photodecarboxyla-
tion. The reactivity of acetophenone derivative 5 allows for
detailed investigations on this reaction pathway.
which would be required for deformylation via a carbanion
intermediate. These initial results suggest that the aroylphe-
nylmethyl group may be thought of as ‘photolabile’ carbanion
leaving group which is able to induce a variety of benzylic C–C
bond heterolysis in appropriately designed molecular systems.
This is akin to the nitrobenzyl group4 and selected diarylmethyl
systems5 that are known to have this ability.
In summary, we have shown that the photodecarboxylation of
aroyl and acetoyl-substituted phenylacetic acids is a general
reaction, all proceeding with high quantum yields. Less labile
‘leaving groups’ such as formaldehyde may also be used. In this
way, one may visualize the aroyl-substituted phenylmethyl
group to be a photolabile carbanion leaving group. This type of
general reactivity may have application in the design of photo-
labile protecting groups and other photochemical applications
in addition to providing new structural types for the elucidation
of detailed reaction mechanism.
We acknowledge the continued support of the Natural
Sciences and Engineering Research Council (NSERC) of
Canada and the University of Victoria.
Notes and references
† All compounds had satisfactory H NMR (300 MHz) and mass spectral
1
data.
‡ We have preliminary results indicating that the meta isomer 2 becomes
increasingly more reactive at acidities greater than pH 2 (and into the
Hammett acidity region) suggesting that a new mechanism for photo-
decarboxylation is available for this compound in acid involving the acid
form.
(3)
The generally high reactivity towards photodecarboxylation
of the above aroyl (and acetoyl) substituted phenylacetic acids
prompted us to think about the possible driving force for these
reactions. A simple rationalization (assuming S1 reactivity) is
that the excited singlet states of these compounds have highly
polarized p electron densities, shifted towards the keto oxygen.
This would favour loss of CO2 via benzylic C–C bond
heterolysis, and subsequent generation of the carbanion. This
mechanism bypasses the need for an initial electron transfer
from the carboxylate oxygen to the carbonyl group, which
appears to be warranted only for 5. In any event, we wanted to
test the former proposal and have made derivative 6 to test
whether a similar reaction could operate using a much poorer
‘leaving group’, viz., formaldehyde, and a moiety incapable of
efficient electron transfer to the carbonyl group. To our surprise,
photolysis of 6 in 1+1 H2O–CH3CN gave the anticipated
deformylated product [p-methylbenzophenone (7)] although
with yields that are several fold (5–10%) less than observed for
the corresponding acid 2, along with what appears to be major
product(s) derived from photoreduction of the ketone. How-
ever, photolysis at pH 12 resulted in clean deformylation
( > 70% yield) consistent with a hydroxide ion catalyzed process
1 (a) D. Budac and P. Wan, J. Photochem. Photobiol., A, 1992, 67, 135; (b)
S. A Fleming and J. A. Pincock, Molecular and Supramolecular
Photochemistry, Vol. 3, V. Ramamurthy and K. Schanze, eds., M. Dekker
Inc., New York, 1999, p. 211.
2 L. L. Costanzo, D. DeGuidi, G. Conderelli, A. Cambria and M. Fama,
Photochem. Photobiol., 1989, 50, 359.
3 (a) G. Cosa, L. J. Martínez and J. C. Scaiano, Phys. Chem. Chem. Phys.,
1999, 1, 3533; (b) J. L. Martínez and J. C. Scaiano, J. Am. Chem. Soc.,
1997, 119, 11 066; (c) S. Monti, S. Sortino, G. De Guidi and G. Marconi,
J. Chem. Soc., Faraday Trans., 1997, 93, 2269; (d) S. Sortino and J. C.
Scaiano, Photochem. Photobiol., 1999, 69, 167; (e) F. Boscá and M. A.
Miranda, Photochem. Photobiol., 1999, 70, 853; (f) F. L. Cozens, W.
Ortiz and N. P. Schepp, J. Am. Chem. Soc., 1998, 120, 13543; (g) T. M.
Bockman, S. M. Hubig and J. K. Kochi, J. Org. Chem., 1997, 62, 2210;
(h) H. Koshima, E. Hayashi, K. Shirafuji, M. Hamada, D. Matsushige, M.
Miyauchi and T. Matsuura, J. Photochem. Photobiol., A, 1999, 129, 121;
(i) H. Yokoi, T. Nakano, W. Fujita, K. Ishiguro and Y. Sawaki, J. Am.
Chem. Soc., 1998, 120, 12 453; (j) H. Görner and H. J. Kuhn, J. Chem.
Soc., Perkin Trans. 2, 1999, 2671; (k) C. K. Lee and P. Wan,
J. Photochem. Photobiol., A, 1993, 76, 39.
4 P. Wan and S. Muralidharan, J. Am. Chem. Soc., 1988, 110, 4336.
5 E. Krogh and P. Wan, J. Am. Chem. Soc., 1992, 114, 705.
2148
Chem. Commun., 2000, 2147–2148