Mechanistic insights into the palladium-induced domino reaction leading to ketones from benzyl β-ketoesters: First characterization of the enol as an intermediate

Article abstract of DOI:10.1021/jo049464w

The monitoring by UV spectroscopy of the Pd-catalyzed hydrogenolysis in acetonitrile of 2-methyl-2-benzyloxycarbonyl-1-indanone and 2-methyl-2- benzyloxycarbonyl-1-tetralone showed the successive formation of corresponding β-ketoacids and enols to deliver finally the ketones. Some factors which influence the stability of the intermediates are determined. In contrast to the above benzyl β-ketoesters, the enol was not detected from benzyl (2-methylinden-3-yl) carbonate.

Full text of DOI:10.1021/jo049464w

Mech a n istic In sigh ts in to th e P a lla d iu m -In d u ced Dom in o
Rea ction Lea d in g to Keton es fr om Ben zyl â-Ketoester s: F ir st
Ch a r a cter iza tion of th e En ol a s a n In ter m ed ia te
J ean-Franc¸ois Detalle, Abdelkhalek Riahi, Vincent Steinmetz, Franc¸oise He´nin,* and
J acques Muzart*
Unite´ Mixte de Recherche “Re´actions Se´lectives et Applications”, CNRS-Universite´ de Reims
Champagne-Ardenne, B.P. 1039, 51687 Reims Cedex 2, France
jacques.muzart@univ-reims.fr; francoise.henin@univ-reims.fr.
Received April 1, 2004
The monitoring by UV spectroscopy of the Pd-catalyzed hydrogenolysis in acetonitrile of 2-methyl-
2-benzyloxycarbonyl-1-indanone and 2-methyl-2-benzyloxycarbonyl-1-tetralone showed the succes-
sive formation of corresponding â-ketoacids and enols to deliver finally the ketones. Some factors
which influence the stability of the intermediates are determined. In contrast to the above benzyl
â-ketoesters, the enol was not detected from benzyl (2-methylinden-3-yl) carbonate.
1. In tr od u ction
amino alcohol to induce asymmetry in the last step.7,8
Interestingly, only catalytic amounts of the amino alco-
hol7,8 were required with benzyl â-ketoesters as sub-
strates.
Previously, we have disclosed that laser irradiation of
2-methyl-2-isobutyl-1-indanone (1) in deuterated aceto-
nitrile produced enol E-5, which was characterized by 1H
NMR and by its quantitative conversion into 2-methyl-
1-indanone (Scheme 2).9 A flash-photolysis study of 1 has
The transformation of allyl â-ketoesters to the corre-
sponding ketones is a one-pot reaction easily carried out
in the presence of an hydride source and catalytic
amounts of palladium. Since its discovery,1 this domino
reaction2 has been widely expanded.3 When ammonium
formates are the hydride source, Tsuji and Mandai3b have
proposed a palladium C-bound enolate as intermediate
(Scheme 1, path a) while Shimizu and Ishii4 postulated
a ketoacid (Scheme 1, path b).
The hydrogenolysis of benzyl â-ketoesters over pal-
ladium is also a well-known reaction to provide the
corresponding ketones (eq 1).5 The literature did not
contain any mechanistic scheme for this domino hydro-
genolysis/decarboxylation reaction; enolic species related
to those of Scheme 1 are nevertheless envisageable.6
allowed us to observe an UV absorption band (λmax
)
265-270 nm) attributable to the formation of E-5.10 UV
spectroscopy is indeed a very useful technique to char-
acterize enols,11 and subsequently, E-5 was also detected
with use of conventional photolysis and an UV spectro-
photometer.10
The above observations induced us to study by UV
spectroscopy the mechanism of the reactions of 2-methyl-
2-benzyloxycarbonyl-1-indanone (KE-5) and 2-methyl-2-
benzyloxycarbonyl-1-tetralone (KE-6) under hydrogen
atmosphere at room temperature in the presence of
catalytic amounts of palladium on carbon. Here, we
The enolic intermediates suspected for these reactions
led us to disclose enantioselective versions using a chiral
(7) J amal Aboulhoda, S.; He´nin, F.; Muzart, J .; Thorey, C.; Behnen,
W.; Martens, J .; Mehler, T. Tetrahedron: Asymmetry 1994, 5, 1321-
1326.
(8) (a) Muzart, J .; He´nin, F.; J amal Aboulhoda, S. Tetrahedron:
Asymmetry 1997, 8, 381-389. (b) J amal Aboulhoda, S.; Reiners, I.;
Wilken, J .; He´nin, F.; Martens, J .; Muzart, J . Tetrahedron: Asymmetry
1998, 9, 1847-1850. (c) Roy, O.; Diekmann, M.; Riahi, A.; He´nin, F.;
Muzart, J . Chem. Commun. 2001, 533-534, errata p 1418. (d) Roy,
O.; Riahi, A.; He´nin, F.; Muzart, J . Eur. J . Org. Chem. 2002, 3986-
3994.
(1) Tsuji, J .; Nisar, M.; Shimizu, I. J . Org. Chem. 1985, 50, 3416-
3417.
(2) For a highlight definition of a domino reaction, see: Tietze, L.
F. Chem. Rev. 1996, 96, 115-136.
(3) For reviews, see: (a) Tsuji, J . Palladium Reagents and Catalysts;
Wiley: Chichester, UK, 1995. (b) Tsuji, J .; Mandai, T. Synthesis 1996,
1-24. (c) Guibe´, F. Tetrahedron 1998, 54, 2967-3042.
(4) Shimizu, I.; Ishii, H. Tetrahedron 1994, 50, 487-495.
(5) (a) Sato, M.; Sakaki, J .; Sugita, Y.; Yasuda, S.; Sakoda, H.;
Kaneko, C. Tetrahedron 1991, 47, 5689-5708. (b) Amat, M.; Llor, N.;
Bosch, J .; Solans, X. Tetrahedron 1997, 53, 719-730. (c) Damour, D.;
Barreau, M.; Fardin, V.; Dhaleine, F.; Vuilhorgne, M.; Mignani, S.
Synlett 1998, 153-156. (d) Shinkai, H.; Ozeki, H.; Motomura, T.; Ohta,
T.; Furukawa, N.; Uchida, I. J . Med. Chem. 1998, 41, 5420-5428. (e)
Donelly, D. M. X.; Finet, J .-P.; Guiry, P. J .; Nesbitt, K. Tetrahedron
2001, 57, 413-423.
(9) He´nin, F.; Muzart, J .; Pe`te, J . P.; M’Boungou-M’Passi, A.; Rau,
H. Angew. Chem., Int. Ed. Engl. 1991, 30, 416-418.
(10) Le´tinois, S. Ph.D. Thesis, Reims, 1999.
(11) (a) Haspra, P.; Sutter, A.; Wirz, J . Angew. Chem., Int. Ed. Engl.
1979, 18, 617-619. (b) Chiang, Y.; Kresge, A. J .; Capponi, M.; Wirz,
J . Helv. Chim. Acta 1986, 69, 1331-1332. (c) Toullec, J . In The
Chemistry of Enols; Rappoport, Z., Ed.; Wiley: Chichester, UK, 1990;
pp 323-398. (d) J efferson, E. A.; Keeffe, J . R.; Kresge, A. J . J . Chem.
Soc., Perkin Trans. 2 1995, 2041-2046. (e) Kresge, A. J . Chem. Soc.
Rev. 1996, 275-280. (f) Ogino, K.; Tamiya, H.; Kimura, Y.; Azuma,
H.; Tagaki, W. J . Chem. Soc., Perkin Trans. 2 1996, 979-984. (g)
Iglesias, E. J . Org. Chem. 2003, 68, 2680-2688.
(6) The formation of toluene from hydrogenolysis of the O-benzyl
group has been highlighted: Beaupe`re, D.; Boutbaiha, I.; Wadouchi,
A.; Frechou, C.; Demailly, G.; Uzan, R. New J . Chem. 1992, 16, 405-
411.
10.1021/jo049464w CCC: $27.50 © 2004 American Chemical Society
Published on Web 09/04/2004
6528
J . Org. Chem. 2004, 69, 6528-6532

Formation of Ketones from Benzyl â-Ketoesters
SCHEME 1
SCHEME 2
expect to detect E-5 by UV analysis if this enol was
formed in the course of the Pd-induced domino reaction.
The development of KE-5 (c0 ) 2.5 × 10-2 M) in the
presence of Pd/C and hydrogen was monitored by UV
spectroscopy. The absorption peak at λmax ) 245 nm of
the successive aliquots (c0,d ) 6.2 × 10-5 M after filtration
and dilution) decreased with time and this phenomena
was accompanied by the appearance of an absorption
peak centered at λ ) 268 nm (Figure 2). The latter
report that ketoacids KA-5 and KA-6 and enols E-5 and
E-6 (Scheme 3) are readily produced under these condi-
tions. These results led us to extend the analysis to the
hydrogenolysis of benzyl enol carbonate EC-5
2. Resu lts a n d Discu ssion .
2.1. Hyd r ogen olysis of 2-Meth yl-2-ben zyloxyca r -
bon yl-1-in d a n on e. As shown in Figure 1, the UV
F IGURE 2. UV absorption spectra of the hydrogenolysis
reaction progress of 2-methyl-2-benzyloxycarbonyl-1-indanone
(KE-5): formation of 2-methyl-2-inden-3-ol (E-5).
reached a maximum intensity at t ) 60 min; this
corresponded also to the extinction of the maxima at 288
and 295 nm.
F IGURE 1. UV absorption spectra of 2-methyl-2-benzyloxy-
carbonyl-1-indanone (KE-5) and 2-methyl-1-indanone (K-5).
From our previous observations,10 we assigned the
band at λmax ) 268 nm to enol E-5. When such a sample
of E-5 is left in the UV cuvette, the UV absorptions
evolved slowly to afford the UV spectrum of ketone K-5,
15 h being required when c0,d ) 1.1 × 10-4 M. In contrast,
the addition of (-)-ephedrine to E-5 (c0,d ) 1.1 × 10-4 M)
led the complete transformation of E-5 to K-5 in less than
6 min (Figure S1, Supporting Information). Without
ephedrine, the transformation to K-5 took 10 min for an
unfiltered sample of E-5 (c0 ) 2.23 × 10-2 M). Therefore,
it appears that (-)-ephedrine like the Pd catalyst ac-
spectra, in particular the λmax values of KE-5 (c ) 7.8 ×
10-5 M, λmax ) 244 (ꢀ ) 8110), 288 (ꢀ ) 1610), 295 (ꢀ )
1600) nm) and K-5 (c ) 7. 8 × 10-5 M, λmax ) 240 (ꢀ )
10590), 285 (ꢀ ) 2260), 292 (ꢀ ) 2150) nm) are very
similar. The very low stability of ketoacid KA-5 in
solution precludes its purification and the recording of a
reliable UV spectrum. Nevertheless, the UV spectrum of
a crude sample recorded quickly after its preparation
showed a λmax = 248 nm. Since the absorbances at 265-
270 nm of these three compounds were very low, we could
SCHEME 3
J . Org. Chem, Vol. 69, No. 20, 2004 6529

Detalle et al.
celerated the E-5 f K-5 reaction. KE-5, KA-5, K-5, and
PhMe having a low absorbance at 268 nm, the extinction
coefficient of E-5 is estimated to be = 8800 dm3 mol-1
cm-1
.
The curves of Figure 2 suggested the presence of three
isobestic points (λ ) 230, 255, and 291 nm) and conse-
quently a KE-5 f E-5 reaction without any intermediate!
A close inspection showed, however, that for each of them,
the crossing of the curves was not exactly at the same
point. Having noted above that ketoacid KA-5 has a very
low stability particularly in diluted solution, we suspected
its formation as a transient species. This led us to repeat
the hydrogenolysis of KE-5; after reaction for 15 min, the
heterogeneous mixture was filtered, cooled to 0 °C, and
laced with a solution of diazomethane. After the addition
of a small amount of silicagel, filtration, and evaporation
of the solvent, the 1H NMR spectrum of the crude mixture
showed the presence of KE-5, K-5, and 2-methyl-2-
methoxycarbonyl-1-indanone. Consequently, KA-5 is ef-
fectively a reaction intermediate. In agreement with this
assertion, the monitoring of the evolution of an acetoni-
trile solution of a partially degraded sample of KA-5 has
allowed the detection of E-5 by UV spectroscopy.
F IGURE 3. Change of the UV absorption spectra of 2-carboxy-
2-methyl-1-tetralone (KA-6): formation of 2-methyl-1-tetralen-
1-ol (E-6).
2.2. Hyd r ogen olysis of 2-Meth yl-2-ben zyloxyca r -
bon yltetr a l-1-on e. The UV characteristics of KE-6 are
λmax ) 248 (ꢀ ) 9430), 291 (ꢀ ) 1260) nm and those of
K-6 are λmax ) 245 (ꢀ ) 8330), 288 (ꢀ ) 1610) nm. KA-6
being much more stable than KA-5, its UV characteristics
have been determined: λmax ) 248 (ꢀ ) 8830), 291 (ꢀ )
1260) nm. Since the UV spectra of KE-6 and KA-6 are
very similar, it was not possible to monitor the transfor-
mation KE-6 f KA-6 by UV spectroscopy. Nevertheless,
we observed by TLC that the Pd-catalyzed hydrogenolysis
of a solution of KE-6 in MeCN (c0 ) 2.18 × 10-2 M) was
complete at room temperature in 70 min; an aliquot was
thus withdrawn, filtered, and diluted: its UV spectrum
was almost identical with that of KA-6 (Figure S2,
Supporting Information). The above mother liquor and
a more concentrated solution of KA-6 (c ) 3.04 × 10-2
M) alike were almost not transformed in 70 min. In
contrast, a more diluted solution of KA-6 (c ) 6.5 × 10-5
M) led to the fast formation of K-6, which was deferred
by the addition of small amounts of AcOH. This stability
of KA-6 in relatively concentrated solutions or in the
presence of AcOH could be due to the formation of
stabilizing intermolecular hydrogen bonds.
Hence, further experiments have been performed in the
absence of the palladium catalyst with use of an UV
cuvette containing a diluted solution of KA-6 in MeCN
(c ) 6.5 × 10-5 M). Under these conditions, we observed
the decrease with time of the band centered at 248 nm
and the appearance of two absorptions which, after 120
min, were centered at 224 and 280 nm (Figure 3). Then,
an apparent reverse development was observed during
180 min, the bands at λmax ) 224 and 280 nm decreased
while an absorption band, finally centered at 245 nm,
appeared (Figure 4). This UV spectrum, obtained about
300 min after the dissolution of KA-6, was indeed the
UV spectrum of K-6. From the previous results, the λmax
) 280 nm was assigned to E-6.
F IGURE 4. Change of the UV absorption spectra of 2-methyl-
3,4-dihydro-1-naphthol (E-6): formation of 2-methyl-1-tetral-
one (K-6).
be relatively stable in polar solvents9,12 and (ii) the
Kamlet-Taft hydrogen bond accepting parameter, â,13
is convenient to correlate the effect of solvent with the
stability of enols.14 The strong difference of the â values
between acetonitrile (â ) 0.31) and hexane (â ) 0.00)
may explain why the enol was not observed in hexane.
Acetonitrile may stabilize the enol in serving as hydrogen
bond acceptor for intermolecular hydrogen bonding with
the OH of the enol.12c,14b,15
It is interesting to underline the presence of three clear
isobestic points in both Figures 3 (λ ) 214, 235, 259 nm)
and 4 (λ ) 213, 234, 256 nm). Those of Figure 3 indicate
that KA-6 would be completely transformed into E-6
before the E-6 f K-6 reaction takes place; this denotes
some stabilization of the enol by the ketoacid and could
be related to the observation of Miller who has disclosed
(12) (a) Toullec, J . In Advances in Physical Organic Chemistry; Gold,
V., Bethell, D., Eds.; Academic Press: London, UK, 1982; Vol. 18, pp
1-77. (b) Chin, C. S.; Lee, S. Y.; Park, J .; Kim, S. J . Am. Chem. Soc.
1988, 110, 8244-8245. (c) Correa, R. A.; Lindner, P. E.; Lemal, D. M.
J . Am. Chem. Soc. 1994, 116, 10795-10796.
(13) (a) Kamlet, M. J .; Abboud, J .-L. M.; Taft, R. W. Prog. Phys. Org.
Chem. 1981, 13, 485-630. (b) Kamlet, M. J .; Abboud, J .-L. M.;
Abraham, M. H.; Taft, R. W. J . Org. Chem. 1983, 48, 2877-2887. (c)
Kamlet, M. J .; Doherty, R. M.; Abraham, M. H.; Carr, P. W.; Doherty,
R. F.; Taft, R. W. J . Phys. Chem. 1987, 91, 1996-2004.
In contrast to the above experiments, the UV study of
the KA-6 f K-6 reaction in hexane (c ) 5.3 × 10-6 M)
did not allow us to detect the intermediate formation of
E-6. It has been reported that (i) enols of ketones may
(14) (a) Mills, S. G.; Beak, P. J . Org. Chem. 1985, 50, 1216-1224.
(b) Rochlin, E.; Rappoport, Z. J . Am. Chem. Soc. 1992, 114, 230-241.
(15) (a) Noyori, R.; Inoue, H.; Kato, M. Bull. Chem. Soc. J pn. 1976,
49, 3673-3678. (b) Rappoport, Z.; Biali, S. E. Acc. Chem. Res. 1988,
21, 442-449.
6530 J . Org. Chem., Vol. 69, No. 20, 2004

Formation of Ketones from Benzyl â-Ketoesters
SCHEME 4
carboxylate KA-P d and/or the palladium enolate E-P d
cannot be excluded. Since only 0.025 equiv of palladium
is used in the preparative experiments, KA-P d and E-P d
would be at best produced in minute amounts. Hence,
even if we consider the formation of KA-P d and E-P d
and then their hydrolysis in the course of the preparation
of the UV samples, the concentrations of KA and E
detected by UV analysis of samples withdrawn from the
palladium-induced domino reaction (Figures S2 and 2,
respectively) imply that notable amounts of KA and E
are already present in the mother solutions. These
remarks led us to assume that Scheme 4 depicts the main
mechanism pathway of the KE f K domino reaction; this
scheme does not exclude some assistance of the KA f E
and E f K steps by the Pd-catalyst when the whole
transformation KE f K is carried out in the presence of
palladium. One referee has suggested that the KA f E
transformation involves the ionization of KA and decar-
boxylation of KA- to afford E- which would be protonated
in the medium on the more electronegative oxygen; this
possibility, which is compatible with our results, is
depicted in Scheme 5.
that enols are formed more effectively in acetic acid than
in hexane.16 Whatever the explanation, it appears that
KA-6 may have a complementary role to MeCN in the
stabilization of E-6.
2.3. NMR An a lysis. As pointed out in the Introduc-
1
tion, E-5 has been previously characterized by H NMR
spectroscopy from the photolysis of 1.9 Using the same
procedure for the characterization of E-6 led to complex
spectra. In contrast, we have been able to detect E-6 by
NMR from the decarboxylation of KA-6 (Figure S3,
1
Supporting Information). The H NMR monitoring of a
From the literature, the ketonization of an enol occurs
through O-protonation, C-protonation, or a cyclic mech-
anism.17 The above E f K reaction being observed in
anhydrous acetonitrile in the absence of any additive
implies a self-ketonization under these experimental
conditions and Scheme 6 outlines the different possibili-
ties. To make a choice between these plausible mecha-
nistic pathways was out of the scope of the present work.
CD3CN solution of KA-6 (c ) 0.01 M) showed the
appearance of signals at 1.21 (d, J ) 6.5 Hz) and 1.39 (s)
ppm. When the signal of the methyl group of KA-6 (s at
1.44 ppm) completely disappeared, (-)-ephedrine was
added and we observed the disappearance of the singlet
at 1.39 ppm while the doublet at 1.21 ppm, which
corresponds to the methyl group of K-6, was maintained.
Consequently, the singlet at 1.39 ppm was due to the
methyl group of E-6.
2.4. Mech a n ism . Under the above conditions, we have
demonstrated the production of KA and E from KE in
relatively large amounts when acetonitrile was the
solvent. For the preparative experiments which lead to
K from KE, palladium is present in the mixture the
entire reaction time, and the formation of the palladium
3. Exten d ed An a lysis
These results urged us to monitor also the hydro-
genolysis of a 2.22 × 10-2 M acetonitrile solution of the
benzyl enol carbonate EC-5 by UV spectroscopy since we
have previously carried out this reaction in a chiral
SCHEME 5
SCHEME 6
J . Org. Chem, Vol. 69, No. 20, 2004 6531

Detalle et al.
SCHEME 7
medium.8a,b,18 The experiments from KE-5 and KE-6 led
us to suspect the formation of hydrogen enol carbonate
HC-5 and E-5 as intermediates (Scheme 7, path a).
However, we have not been able to detect an intermediate
and we only observed the complete EC-5 f K-5 reaction
in less than 5 min (Figure S4, Supporting Information).
This leads to various hypotheses: (i) E-5 would be
consumed as soon as it is produced because it will not be
stabilized by intermolecular interactions with a ketoacid,
or (ii) E-5 is not produced, HC-5 leading directly to K-5
(Scheme 7, path b), or (iii) palladium is involved in all
steps and K-5 would be produced through hydridopalla-
dium enol carbonate HC-5-P d and hydridopalladium
enolate E-5-P d intermediates (Scheme 7, path c), or (iv)
HC-5-P d evolves to HC-5 (Scheme 7, path d). At the
present time, we have no argument to favor one of these
proposals.
similarly monitor the domino reaction of allyl â-ke-
toesters in the aim to settle between the proposals
depicted in Scheme 1.
5. Exp er im en ta l Section
Typ ica l UV An a lysis P r oced u r e: Hyd r ogen olysis of
KE-5. In a 25-mL round-bottomed flask containing a stir bar,
KE-5 (50 mg, 0.178 mmol), and acetonitrile (8 mL) was added
Pd/C (20 mg, 0.025 equiv). The flask was purged for 0.5-1
min by a slight flow of hydrogen, then connected to a rubber
balloon filled with hydrogen and the stirring was started. At
the times indicated in Figure 2, an aliquot (0.5 mL) was drawn
with a syringe and immediately injected into a flask through
an HPLC filter; 25 µL of the filtrate was placed in a 5-mL
volumetric flask that was filled with MeCN. After a brief
shaking to ensure mixing, a quartz cuvette (path length: 10
mm) was charged with this solution and immediately placed
in the cell holder of the UV spectrophotometer to perform the
analysis over the range 200-400 nm.
4. Con clu sion
We now have clear evidence that the palladium-
catalyzed hydrogenolysis of a benzyl â-ketoester such as
KE-5 or KE-6 involves, at least in MeCN as solvent, the
intermediate formation of the corresponding ketoacid and
the enol of the terminal ketone. Under similar experi-
mental conditions, the benzyl enol carbonate EC-5 led
also to the corresponding ketone but with a different
reactivity. Whatever the mechanism from the benzyl enol
carbonate, this difference of reactivity may explain the
differences of enantioselectivities obtained for the reac-
tions KE-5 f K-5 and EC-5 f K-5, when they are
carried out under identical palladium-catalyzed hydro-
genolysis conditions in the presence of the same enan-
tiopure amino alcohol.7,8a,b It now would be of interest to
Su p p or tin g In for m a tion Ava ila ble: Figures S1, S2, S3,
and S4, standard experimental procedures, and characteriza-
tion of products. This material is available free of charge via
the Internet at http://pubs.acs.org.
J O049464W
(16) Miller, A. R. J . Org. Chem. 1976, 41, 3599-3602.
(17) (a) Zimmerman, H. E. Acc. Chem. Res. 1987, 20, 263-268. (b)
Chiang, Y.; Kresge, A. J .; Santaballa, J . A.; Wirz, J . J . Am. Chem. Soc.
1988, 110, 5506-5510. (c) Keeffe, J . R.; Kresge, A. J . In The Chemistry
of Enols; Rappoport, Z., Ed.; Wiley: Chichester, UK, 1990; pp 399-
480. (d) Zimmerman, H. E.; Wang, P. Org. Lett. 2002, 4, 2593-2595
and cited references.
(18) He´nin, F.; Muzart, J . Tetrahedron: Asymmetry 1992, 3, 1161-
1164.
6532 J . Org. Chem., Vol. 69, No. 20, 2004