Development of a catalytic aromatic decarboxylation reaction

Article abstract of DOI:10.1021/ol070749f

A palladium-catalyzed aromatic decarboxylation reaction has been developed. With electron-rich aromatic adds, the reaction proceeds efficiently under fairly mild conditions in good yields. The method was useful with complex functionalized substrates containing hindered carboxylic adds.

Full text of DOI:10.1021/ol070749f

ORGANIC
LETTERS
2
007
Vol. 9, No. 13
441-2444
Development of a Catalytic Aromatic
Decarboxylation Reaction
2
Joshua S. Dickstein, Carol A. Mulrooney, Erin M. O’Brien,
Barbara J. Morgan, and Marisa C. Kozlowski*
Department of Chemistry, Roy and Diana Vagelos Laboratories,
UniVersity of PennsylVania, Philadelphia, PennsylVania 19104
marisa@sas.upenn.edu
Received March 27, 2007
ABSTRACT
A palladium-catalyzed aromatic decarboxylation reaction has been developed. With electron-rich aromatic acids, the reaction proceeds efficiently
under fairly mild conditions in good yields. The method was useful with complex functionalized substrates containing hindered carboxylic
acids.
A mild and efficient method for aromatic decarboxylation
has yet to be reported. In general, aromatic decarboxylation
is difficult due to the unstable intermediates that are formed
during the course of the reaction. As a result, the methods
available for decarboxylation all require harsh, forcing
conditions.
The Barton protocol does not permit direct protonation in
aromatic systems although an indirect approach via radical
decarboxylative bromination is viable. This three-step
protocol (ester formation, decarboxylative bromination, and
bromine hydrogenolysis) still requires high temperatures and
is not suitable for highly hindered esters or substrates with
hydrogenolytically sensitive functionality.
5
Prototypical methods include heating in the presence of a
1
strong acid and heating with a copper catalyst and quino-
In this Letter, we report the development of a single-step
palladium-catalyzed decarboxylation utilizing trifluoroacetic
acid as the proton source that proceeds at <100 °C. Our
investigations were motivated by the need to remove
efficiently the carboxylate groups from 1 as part of our
2
line. The first method proceeds through ipso protonation
of the aromatic ring and thus requires electron-rich systems
and temperatures of at least 100 °C. In the copper/quinoline
method, the copper catalyst coordinates to the aromatic ring
and helps stabilize the anion that results upon loss of carbon
dioxide. Again these reactions only proceed at very high
6
7
studies of the perylenequinone natural products (Scheme
). Unfortunately, decarboxylation reactions with protic acid
1
3
temperatures (>160 °C). While mercury mediated decar-
were unsuccessful and those under the copper/quinoline
conditions resulted in low yields and racemization of the
biaryl stereochemistry.
boxylations can proceed under milder conditions, the re-
quirement for stoichiometric mercury salts and the interven-
4
tion of toxic organomercury(II) intermediates is restrictive.
(
1) (a) Olah, G. A.; Laali, K.; Mehrotra, A. K. J. Org. Chem. 1983, 48,
(4) (a) Gilman, H.; Wright, G. F. J. Am. Chem. Soc. 1933, 55, 3302-
3314. (b) Deacon, G. B.; O’Donoghue, M. F.; Stretton, G. N.; Miller, J. M.
J. Organomet. Chem. 1982, 233, C1-C3.
3
4
359-3360. (b) Horper, W.; Marner, F.-J. Phytochemistry 1996, 41, 451-
56.
(
2) (a) Cohen, T.; Schambach, R. A. J. Am. Chem. Soc. 1970, 92, 3189-
(5) (a) Barton, D. H. R.; Lacher, B.; Zard, S. Z. Tetrahedron Lett. 1985,
26, 5939-5942. (b) Barton, D. H. R.; Lacher, B.; Zard, S. Z. Tetrahedron
1987, 43, 4321-4328.
3
4
8
190. (b) Cohen, T.; Berninger, R. W.; Wood, J. T. J. Org. Chem. 1978,
3, 837-848. (c) Pulgarin, C.; Tabacchi, R. HelV. Chim. Acta 1988, 71,
76-880.
(6) Mulrooney, C. A.; Li, X.; DiVirgilio, E. S.; Kozlowski, M. C. J.
Am. Chem. Soc. 2003, 125, 6856-6857.
(7) Review: Weiss, U.; Merlini, L.; Nasini, G. Prog. Chem. Org. Nat.
Prod. 1987, 52, 1-71.
(3) Microwave heating can also be employed: Frederiksen, L. B.;
Grobosch, T. H.; Jones, J. R.; Lu, S. Y.; Zhao, C. C. J. Chem. Res. Synop.
000, 42-43.
2
1
0.1021/ol070749f CCC: $37.00
© 2007 American Chemical Society
Published on Web 06/02/2007

Having solved the atropoisomerization issues that were
our initial concern, our attention turned to reducing the
amount of palladium required for an effective reaction. To
do this required identification of a hydride or proton source
that (1) can be employed in the presence of palladium(II)
and (2) will allow catalytic turnover. The former condition
excludes most hydride sources such as hydrogen, borohy-
drides, etc. which cause rapid reduction of palladium(II) to
palladium(0) before decarboxylation can occur. The latter
condition either requires a reoxidation step if a reductive
elimination pathway occurs or Brønsted acid to allow
protonolysis of the aryl palladium intermediate. This last
proposal proved to be the most effective. Upon generation
of the aryl palladium intermediate from 2a with stoichio-
metric palladium(II) trifluoroacetate, addition of trifluoro-
acetic acid produced the decarboxylated material 3a (Scheme
3). Since a closed catalytic cycle was anticipated, we were
Scheme 1. Perylenequinone Family Common Intermediate
Recently, the Myers group reported a palladium-catalyzed
8,9
decarboxylative Heck coupling. Inspired by this precedent;
we proposed that a catalytic decarboxylation could be
achieved if an appropriate hydride or proton source could
be discovered. To begin with, we explored stoichiometric
palladium decarboxylation by treatment with palladium(II)
trifluoroacetate and silver carbonate followed by introduction
of a hydrogen atmosphere to reduce the resultant aryl
palladium species. Rewardingly, this protocol produced high
yields of decarboxylated product in model systems (Scheme
Scheme 3. Decarboxylation of the Model System
2) at much lower temperatures compared to classical
gratified that the amount of palladium catalyst could indeed
be lowered to 20 mol % without any effect on the yield of
Scheme 2. Initial Decarboxylation Results
3
a (Scheme 3).
On the basis of a number of mechanistic experiments
performed in our lab as well as work reported by the Myers
8
group, the following mechanism for catalytic decarboxyla-
tion is proposed (Scheme 4). Coordination of the palladium
Scheme 4. Proposed Mechanism
methods,1,2 and to another report with a palladium catalyst.10
As a result, no loss of enantioselectivity was seen with these
conditions in thermally atroposensitive substrates such as 4.
(
8) (a) Myers, A. G.; Tanaka, D.; Mannion, M. R. J. Am. Chem. Soc.
2
6
002, 124, 11250-11251. (b) Tanaka, D.; Myers, A. G. Org. Lett. 2004,
to two molecules of DMSO and the substrate (2a) results in
the loss of a molecule of trifluoracetic acid. A four-membered
transition state with loss of carbon dioxide provides an
avenue for palladium insertion onto the aryl ring to give the
aryl palladium species (7). At this point the aryl palladium
intermediate (7) can undergo protonolysis with trifluoroacetic
acid to yield the product (3a) and reform the palladium(II)
trifluoroacetate catalyst.
, 433-436. (c) Tanaka, D.; Romeril, S. P.; Myers, A. G. J. Am. Chem.
Soc. 2005, 127, 10323-10333.
(9) For recent related work on decarboxylative cross-coupling reactions,
see: (a) Heim, A.; Terpin, A.; Steglich, W. Angew. Chem., Int. Ed. 1997,
3
3
6, 155-156. (b) Goossen, L. J.; Deng, G.; Levy, L. M. Science 2006,
13, 662-664. (c) Forgione, P.; Brochu, M. C.; St-Onge, M.; Thesen, K.
H.; Bailey, M. D.; Bilodeau, F. J. Am. Chem. Soc. 2006, 128, 11350-
1
1351. (d) Goossen, L. J.; Rodriguez, N.; Melzer, B.; Linder, C.; Deng,
G.; Levy, L. M. J. Am. Chem. Soc. 2007, 129, 4824-4833. For decar-
boxylation as a side reaction with palladium catalysts, see: (e) Lee, J. M.;
Chang, S. Tetrahedron Lett. 2006, 47, 1375-1379.
To further optimize and extend this process, the rate-
determining step in the above mechanistic cycle needed to
(
10) Matsubara, S.; Yokota, Y.; Oshima, K. Org. Lett. 2004, 6, 2071-
2
073.
2442
Org. Lett., Vol. 9, No. 13, 2007

1
be determined. Thus, H NMR spectroscopy was employed
to monitor the formation of the aryl palladium intermediate
(7) and the product (3a) during the course of the reaction.
In the first experiment, substrate 2a was treated with
stoichiometric palladium(II) trifluoroacetate in the presence
of excess methanesulfonic acid. Under these conditions,
carboxylate exchange occurred rapidly and adduct 6 was the
only species observed at the beginning of the reaction. Upon
heating to 65 °C, formation of the aryl palladium species
(
7) was directly related to the disappearance of adduct 6
during the initial stages of the reaction (Figure 1). The
Figure 2. Protonolysis of the aryl palladium intermediate corre-
sponding to 7 from 2a (reaction conditions: initial concentration
[
2a] ) 88 mM and [Pd(O
2
CCF
3
)
2
] ) 105 mM, CH
3 3
SO H (10
equiv), 0.75 mL d -DMSO, 65 °C (aryl palladium formed at 70 °C
6
but protonolysis at 65 °C)).
of a strong acid such as trifluoroacetic acid, and no base
such as Ag CO .
2 3
Surprisingly, when the methodology was extended to other
substrates it was found that two ortho substituents were
required to obtain high yields of the decarboxylation
Figure 1. NMR monitoring of the decarboxylation of 2a (reaction
conditions: initial concentration [2a] ) 88 mM and [Pd(O
2 3 2
CCF ) ]
)
105 mM, CH SO H (10 equiv), 0.75 mL d -DMSO, 65 °C).
3
3
6
Table 1. Catalytic Palladium Decarboxylation
-
1
disappearance of 6 was a first-order process (k ) 0.25 h ,
1/2 ) 2.8 h). As the reaction progressed, aryl palladium
t
species 7 began to react forming the product 3a via a zero-
order process (k ) 2.73 mM/h, t1/2 ) 17.7 h). Overall,
formation of aryl palladium 7 is much faster than the final
protonolysis step (7 to 3a). This result clarifies the need for
excess acid to achieve a reasonable reaction time. While the
mechanism in Scheme 4 does not require any additional acid
for turnover, the slow step (7 to 3a) does depend on the acid
concentration.
To further probe the protonation reaction, the aryl pal-
ladium intermediate (7) was generated in the absence of acid
and then treated with 10 equiv of methanesulfonic acid.
1
Monitoring by H NMR spectroscopy revealed a linear
relationship between formation of product (3a) and con-
sumption of intermediate 7 (Figure 2). The much slower rate
of this reaction (reaction of 7 was apparently zero order, k
)
2.69 mM/h, t1/2 ) 13 h) even with excess acid explains
8
why the decarboxylative Heck reaction can be so successful.
Namely, the aryl palladium species is sufficiently stable
that migratory insertion into the alkene can occur before
2 3
protonation. Additionally, the basic Ag CO employed in the
8
Heck process likely serves to scavenge any free acid further
1
preventing protonation. Overall, the results of the H NMR
experiments indicated that protonation is the slow step. On
this basis, the following general parameters were devel-
oped: e20 mol % of palladium(II) trifluoroacetate, an excess
a
Reaction conditions: Pd(O2CCF3)2 (0.2 equiv), 70 °C, CF3CO2H (10
equiv), 5% DMSO/DMF. b Isolated yield. c 80 °C.
Org. Lett., Vol. 9, No. 13, 2007
2443

found to be good with use of the catalytic palladium
conditions as illustrated in Table 1.
Table 2. Stoichiometric Palladium Decarboxylation
Further substrates were also explored with stoichiometric
palladium as shown in Table 2. For entries 6 and 7, the
potassium carboxylate salts were employed which precluded
turnover. Even so, these salts were found to undergo more
rapid reaction than the free acid under comparable conditions.
The substrates from entries 8 and 9 were not explored with
catalytic palladium as they are very late stage intermediates.
For these compounds, the aryl palladium species was
deliberately formed and then reduced with sodium borohy-
dride. In fact, many other hydrogen sources were found to
be effective when this approach was employed. For example,
methanesulfonic acid, acetic acid, nitric acid, and p-nitroben-
zoic acid were all effective with catalytic palladium. With
stoichiometric palladium, diethylsilane, triethylsilane, triiso-
propylsilane, triphenylsilane, diethylmethylsilane, borane-
THF complex, PHMS/KF, Mg/NH OAc, sodium formate,
4
ammonium formate, sodium methoxide, and potassium
carbonate/methanol were all found to be viable.
In conclusion, a catalytic decarboxylation method has been
realized for electron-rich bis-ortho-substituted aromatic
compounds. This mild and efficient decarboxylation can be
employed with sensitive substrates. This method works
especially well with substrates that contain sterically hindered
carboxylic acids. Further details about this and other aspects
will be reported in due course.
a
Reaction conditions: Pd(O2CCF3)2 (1.2 equiv), 70 °C, CF3CO2H (10
Acknowledgment. This research was supported by the
National Institutes of Health (CA-109164). We thank 3D
Pharmaceuticals (C.A.M.), the Division of Organic Chem-
istry of the ACS (C.A.M.), Eli Lilly (E.M.B), and the
Department of Chemistry at the University of Pennsylvania
(E.M.B) for graduate fellowships.
b
c
equiv), 5% DMSO/DMF. Isolated yields. Pd(O2CCF3)2 (2.5 equiv),
Ag2CO3 (6 equiv), 75 °C, followed by NaBH4. No loss of enantioselec-
tivity. No loss of diastereoselectivity.
d
e
products. Further NMR experiments elucidated a secondary
pathway with mono-ortho-substituted substrates comprised
of C-H insertion into positions adjacent to the carboxyl
group. Unfortunately, no conditions were found to completely
prevent this undesired reaction from occurring. Nevertheless,
the substrate scope of bis-ortho-substituted compounds was
Supporting Information Available: Experimental details
and characterization of all new compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
OL070749F
2444
Org. Lett., Vol. 9, No. 13, 2007