Catalytic, nucleophilic allylation of aldehydes with allyl acetate

Article abstract of DOI:10.1021/ol8028725

(Chemical Equation Presented) A new catalytic allylation of aldehydes has been developed that employs allyl acetate as the allylating reagent. Under catalysis by ruthenium trichloride (3 mol %) in the presence of carbon monoxide (30 psi), water (1.5 equiv), and Methylamine (0.1 equiv), a wide range of aromatic, olefinic, and aliphatic aldehydes are efficiently allylated under mild conditions (70 °C, 24-48 h). The stoichiometric byproducts of this reaction are carbon dioxide and acetic acid.

Full text of DOI:10.1021/ol8028725

ORGANIC
LETTERS
2009
Vol. 11, No. 3
781-784
Catalytic, Nucleophilic Allylation of
Aldehydes with Allyl Acetate
Scott E. Denmark* and Son T. Nguyen
Roger Adams Laboratory, Department of Chemistry, UniVersity of Illinois, 600 South
Mathews AVenue, Urbana, Illinois 61801
sdenmark@illinois.edu
Received December 13, 2008
ABSTRACT
A new catalytic allylation of aldehydes has been developed that employs allyl acetate as the allylating reagent. Under catalysis by ruthenium
trichloride (3 mol %) in the presence of carbon monoxide (30 psi), water (1.5 equiv), and triethylamine (0.1 equiv), a wide range of aromatic,
olefinic, and aliphatic aldehydes are efficiently allylated under mild conditions (70 °C, 24-48 h). The stoichiometric byproducts of this reaction
are carbon dioxide and acetic acid.
The allylation of carbonyl compounds is now universally
recognized as one of the premier methods for carbon-carbon
bond formation.1 The most prominent reasons for the
popularity of the method include: (1) the extreme diversity
of reagent reactivity based on allylmetal reagent, (2) the high
degree of both diastereo- and enantioselectivity observed in
many cases,2 and (3) the latent functionality in the homoal-
lylic alcohol product which makes the reaction ideal for
synthetic planning.1d Although many reagent variations are
known, the reactions generally involve the use of discrete
allylmetallic (or nonmetallic) reagents, either directly or in
combination with Lewis acidic or basic catalysts. In some
cases, the allylmetallic reagent is prepared in situ by the
combination of an allyl source (halide, acetate, alcohol) with
a stoichiometric amount of a metal salt.3 In addition, a
number of reports have described using one metal to catalyze
the formation of the allylmetallic reagent and a stoichiometric
amount of a second metal or metallic reagent to turn over
the catalyst.4 Although this approach avoids the preparation
of the discrete allylmetal species, a large amount of inorganic
waste is nevertheless generated which must be removed in
the purification step and reduces the overall atom efficiency.
To address these issues, we sought to develop a new
process for carbonyl allylation that is catalytic in metal,
(3) (a) Fu¨rstner, A. Chem. ReV. 1999, 99, 991. (b) Karbo, R. B.; Cook,
G. R. Curr. Org. Chem. 2007, 11, 1287–1309. (c) Pan, C.; Wang, Z. Coord.
Chem. ReV. 2008, 252, 736.
(4) For representative examples, see Pd(pincer)/diboron: (a) Selander,
N.; Kipke, A.; Sebelius, S.; Szabo, K. J. Am. Chem. Soc. 2007, 129, 13723–
13731. Co(acac)2/SnCl2: (b) Chaudhuri, M. K.; Dehury, S. K.; Hussain, S
Tetrahedron Lett. 2005, 46, 6247–6251. PdCl2/SnCl2: (c) Takahara, J. P.;
Masuyama, Y.; Kurusu, Y. J. Am. Chem. Soc. 1992, 114, 2577–2586.
CdSO4/Zn and SnCl2/Zn: (d) Zhou, C.; Zhou, Y.; Jiang, J.; Zhen, X.;
Zhiyong, W.; Zhang, J.; Wu, J.; Yin, H. Tetrahedron Lett. 2004, 45, 5537–
5540. Pd(Ph3P)4/SmI2: (e) Tabuchi, T.; Inanaga, J.; Yamaguchi, M.
Tetrahedron Lett. 1986, 27, 1195–1196. InCl3/Al or Zn: (f) Araki, S.; Jin,
S-J.; Idou, Y.; Butsugan, Y. Bull. Chem. Soc. Jpn. 1992, 65, 1736–1738.
[Ir(cod)Cl]2/SnCl2: (g) Roy, S.; Banerjee, M. J. Mol. Catal. A 2006, 246,
231–236. [RhCl(cod)]2/SnCl2: (h) Masuyama, Y.; Kaneko, Y.; Kurusu, Y.
Tetrahedron Lett. 2004, 45, 8969–8971. Pd(OAc)2/Et2Zn: (i) Kimura, M.;
Shimizu, M.; Shibata, K.; Tazoe, M.; Tamaru, Y. Angew. Chem., Int. Ed.
2003, 42, 3392–3395. Pd(OAc)2/Et3B: (j) Kimura, M.; Tomizawa, T.;
Horino, Y.; Tanaka, S.; Tamaru, Y. Tetrahedron Lett. 2000, 41, 3627–
3629.
(1) (a) Yamamoto, Y.; Asao, N. Chem. ReV. 1993, 93, 2207–2293. (b)
Helmchen, G.; Hoffmann, R.; Mulzer, J.; Schaumann, E. Eds. StereoselectiVe
Synthesis, Methods of Organic Chemistry (Houben-Weyl), E21; Thieme:
Stuttgart, 1996; Vol. 3; pp 1357-1602. (c) Denmark, S. E.; Almstead, N. G.
In Modern Carbonyl Chemistry; Otera, J., Ed.; Wiley-VCH: Weinheim,
2000; Chapter 10. (d) Chemler, S. R.; Roush, W. R. In Modern Carbonyl
Chemistry; Otera, J., Ed.; Wiley-VCH: Weinheim, 2000; Chapter 11, pp
403-490. (e) Gung, B. W. Org. React. 2004, 64, 1–113.
(2) (a) Denmark, S. E.; Fu, J. Chem. ReV. 2003, 103, 2763–2793. (b)
Denmark, S. E.; Fu, J. Chem. Commun. 2003, 167–170. (c) Denmark, S. E.;
Fu, J. J. Am. Chem. Soc. 2003, 125, 2208–2216. (d) Denmark, S. E.; Fu,
J.; Lawler, M. J. J. Org. Chem. 2006, 71, 1523–1536. (e) Denmark, S. E.;
Fu, J.; Coe, D. M.; Su, X.; Pratt, N. E; Griedel, B. D. J. Org. Chem. 2006,
71, 1513–1522.
10.1021/ol8028725 CCC: $40.75
Published on Web 12/31/2008
2009 American Chemical Society

environmentally benign, and amenable to large-scale ap-
plications. At the time we initiated this work, only three such
catalytic processes for carbonyl allylation were known: the
carbonyl-ene process,5 the hydrogenative coupling of dim-
ethylallene,6,7 and the ruthenium-catalyzed allylation of
aldehydes using allyl acetate.8 Whereas the first two pro-
cesses had appeared to be suitable only for highly reactive
aldehydes, the ruthenium-catalyzed process showed the
potential for a wider substrate scope. Curiously, however,
this method has not received much attention, most likely
because of the harsh conditions required.9 Herein, we disclose
our development of a practical, ruthenium-catalyzed carbonyl
allylation process using allyl acetate. Our studies revealed
critical insights into the roles of triethylamine, halide anion,
carbon monoxide, and water and enabled us to create a
simple, economical, and efficient procedure for allylation of
aliphatic, olefinic, and aromatic aldehydes.
was performed in which controlled amounts of oxygen gas
and/or water were blended into the reaction mixtures. From
these experiments we found that water accelerated the
reaction but also caused significant consumption of allyl
acetate via an unproductive pathway (oxygen had no
beneficial effect). Next, the stoichiometries of Et3N and water
were systematically varied. Interestingly, when the amount
of Et3N was reduced to 0.1 equiv and water was adjusted to
1.5 equiv, the reaction proceeded to 95% conversion at 70
°C in 24 h. This result marked a critical departure from the
original conditions in which Et3N was believed to be the
stoichiometric reducing reagent.10
The conclusion that Et3N is not a hydride donor in this
system is supported by two observations: (i) 1H NMR
analysis of the reaction mixtures indicated that triethylamine
was not consumed;11 (ii) when Et3N was replaced by
quinuclidine, an amine that cannot function as a hydride
donor, the reaction still proceeded to comparable conversion
in the same reaction period. In addition to Et3N, other
secondary and tertiary amines such as i-Pr2EtN and i-Pr2NH
were also effective. However, in the absence of an amine,
no reaction occurred (Table 1, entries 1 and 2).
In the original report from Wantanabe and co-workers,
allyl acetate was the limiting reagent (Scheme 1).8 For most
synthetic applications, the aldehyde is the precious reagent;
thus, it was necessary to reevaluate the roles of all reaction
components and experimental variables.
Scheme 1. Originally Reported Allylation Conditions
Table 1. Catalytic Carbonyl Allylation: Effects of Triethylamine
and Chloride
Et3N
(equiv)
TBACl
(equiv)
conversiona
(%)
entry
Ru sources
To orient our efforts, the preliminary reactions were run
under the original conditions, employing benzaldehyde as a
model substrate with 1.1 equiv of allyl acetate at 70 °C in a
variety of solvents with RuCl3 as the catalyst. Because of
the high pressure of CO required, a six-well autoclave was
employed to perform the optimization. Interestingly, we
discovered that the reaction proceeded faster after the
autoclave was cooled and opened in air for monitoring. To
identify the origin of this effect, a battery of experiments
1
2
3
4
5
6
7
8
9
RuCl3·xH2O
RuCl3·xH2O
0
0.1
0
0.1
0
0.1
0
0
0
0
0
0
0
0.03
0
0.03
0
95
12
93
43
70
84
15
78
allylRu(CO)3Br
allylRu(CO)3Br
allylRu(CO)3OAc
allylRu(CO)3OAc
allylRu(CO)3OAc
b
Ru3(CO)12
0
0
b
Ru3(CO)12
a Conversions were calculated from 1H NMR integration values of the
reaction mixtures using hexamethylbenzene as internal standard. b 0.01 equiv
was used.
(5) (a) Mikami, K.; Shimizu, M. Chem. ReV. 1992, 92, 1021–1050. (b)
Mikami, K.; Terada, M. In ComprehensiVe Asymmetric Catalysis; Jacobsen,
E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer-Verlag: Heidelberg, 1999;
Chapter 32. (c) Johnson, J. S.; Evans, D. A. Acc. Chem. Res. 2000, 33,
325–335.
(6) (a) Skucas, E.; Bower, J. F.; Krische, M. J. J. Am. Chem. Soc. 2007,
129, 12678–12679. (b) Ngai, M.-Y.; Skucas, E.; Krische, M. J. Org. Lett.
To elucidate whether the amine is needed only in the initial
reduction of RuCl3 to Ru(0) at the beginning of the process12
2008, 10, 2705–2708
.
(7) During the completion and submission of this manuscript, Krische
and co-workers reported a hydrogenative allylation by using an alcohol as
the hydrogen donor and reducing reagent for the allylation of aldehydes
and alcohols: (a) Kim, I. S.; Ngai, M-Y.; Krische, M. J. J. Am. Chem. Soc.
2008, 130, 14891–14899. (b) Kim, I. S.; Ngai, M-Y.; Krische, M. J. J. Am.
Chem. Soc. 2008, 130, 6340–6341. (c) Shibahara, F.; Bower, J. F.; Krische,
M. J. J. Am. Chem. Soc. 2008, 130, 14120–14122. (d) Shibahara, F.; Bower,
(10) For an isotope labeling study and mechanistic discussion, see:(a)
Kondo, T.; Ono, H.; Satake, N.; Mitsudo, T.; Watanabe, Y. Organometallics
1995, 14, 1945–1953. (b) Kondo, T.; Misudo, T. Curr. Org. Chem. 2002,
6, 1163–1179. (c) Sakaki, S.; Ohki, T.; Takayama, T.; Sugimoto, M.; Kondo,
T.; Mitsudo, T. Organometallics 2001, 20, 3145. (d) Trost, B. M.; Toste,
F. D.; Pinkerton, A. B. Chem. ReV. 2001, 101, 2067–2096.
(11) 1H NMR spectra of these reaction mixtures (without workup) only
showed signals arising from benzaldehyde, allyl acetate, 1a, triethylam-
momium acetate, and propene (in trace amounts).
J. F.; Krische, M. J. J. Am. Chem. Soc. 2008, 130, 6338–6339
.
(8) (a) Tsuji, Y.; Mukai, T.; Kondo, T.; Watanabe, Y. J. Organomet.
Chem. 1989, 369, C51-C53. For dehydrogenative allylation to prepare
enones, see: (b) Kondo, T.; Mukai, T.; Watanabe, Y. J. Org. Chem. 1991,
56, 487–489.
(12) For the role of base in the reduction of RuCl3 to Ru3(CO)12, see:
Roveda, C.; Cariati, E.; Lucenti, E.; Roberto, D. J. Organomet. Chem. 1999,
580, 117–127.
(9) For an intramolecular example, see: Yu, C-M.; Lee, S.; Hong, Y-T.;
Yoon, S-K. Tetrahedron Lett. 2004, 45, 6557–6561.
782
Org. Lett., Vol. 11, No. 3, 2009

or also needed throughout the catalytic cycle,
allylRu(CO)3Br13 and allylRu(CO)3OAc8a were prepared and
used in the catalytic reactions (Table 1). Addition of Et3N
in both cases significantly improved the conversions, espe-
cially in the reaction that employed allylRu(CO)3Br (Table
1, entries 3-6).
These results suggest that the amine is acting as a base
perhaps to neutralize HCl generated during the reduction of
RuCl3 (or HBr when allylRu(CO)3Br is used) and to partially
buffer the medium from becoming too acidic from the acetic
acid generated as a byproduct. An excess of Et3N, on the
other hand, causes the unproductive consumption of allyl
acetate possibly because Et3N competitively binds to the
allyl-Ru complex and thus inhibits the carbonyl allylation
process while favoring the undesired protonolysis of the
π-allyl complex to produce propene.14
π-allyl-Ru complex (II) which ultimately delivers the allyl
group to the aldehyde. It is possible that this process occurs
via the coordination of the aldehyde carbonyl to the
ruthenium center causing a shift of the π-allylruthenium from
the η3- to the η1- binding mode and the (η1-allyl)ruthenium
undergoes the nucleophilic attack at the carbonyl.21
A
plausible mechanism for the catalyst turnover involves
hydrolysis of the ruthenium alkoxide complex (III) to release
the homoallylic alcohol and generate a ruthenium hydroxide
species (IV). Insertion of CO into the Ru-OH bond of
complex IV, followed by extrusion of CO2 and subsequent
reductive elimination of acetic acid, regenerates the Ru(0)
catalyst. This mechanistic formulation allows a clearer insight
into the roles of CO and water; i.e., the combination of CO
and water must provide the stoichiometric reducing equiva-
lents (water gas shift reaction).22
The observation that reactions with RuCl3 and
allylRu(CO)3Br showed higher conversions than reactions
with allylRu(CO)3OAc (Table 1, entries 2, 4, and 6) indicated
a possible halide affect.15 Indeed, when tetrabutylammonium
chloride (TBACl) was added to a reaction catalyzed by
allylRu(CO)3OAc, the conversion improved from 43 to 84%
(Table 1, entries 5 and 7). More significantly, the conversion
improved from 15 to 78% when Ru3(CO)12 was used as the
precatalyst (Table 1, entries 8 and 9).16
The effect of halides on the structure and reactivity of
ruthenium carbonyl complexes has been the subject of
countless investigations since the early 1980s.17 Geoffroy18
has shown that halide anions (X) can displace one or more
CO ligands in the Ru3(CO)12 complex to produce anionic
species such as [Ru3(CO)11X]-,19 [Ru3(CO)10X]-, and
[Ru4(CO)13X]-. Furthermore, neutral Ru(0) complexes such
as Ru3(CO)12 and Ru(COD)(COT) do not react with allyl
acetate even in refluxing toluene.8a Amatore and Jutand have
established the role of chloride ion in palladium-catalyzed,
cross-coupling reactions in which anionic chloropalladium(0)
ate species participate in the oxidative addition step.20
By analogy, we postulate that the rate enhancement caused
by chloride observed in this process is due to the chloride-
ligated anionic complexes (I) formed by displacement of
carbon monoxide ligand(s) from the neutral Ru(0) species
by chloride (Figure 1). These anionic complexes should be
more nucleophilic than the neutral Ru(0) complexes and thus
can readily react with allyl acetate to form the requisite
Figure 1
allylation.
. Proposed mechanism for the RuCl3-catalyzed carbonyl
The final stages of the optimization concerned pressure,
temperature, and solvent. We examined reactions at pressures
from 15 to 200 psi of CO and found that the conversions
did not change at pressures above 30 psi. The reactions
proceeded even at 15 psi of CO but required longer time.
Thus, for the survey of aldehyde scope, the CO pressure was
kept at 30-35 psi. This was a significant improvement in
(13) Sbrana, G.; Braca, G.; Piacenti, F.; Pino, P. J. Organomet. Chem.
1968, 13, 240–242.
(14) Approximately 0.10 equiv of propene was observed by NMR
analysis in the crude reaction mixture when 3 equiv of triethylamine was
used.
(15) Fagnou, K.; Lautens, M. Angew. Chem., Int. Ed. 2002, 41, 26–47.
(16) Addition of TBAI also improved the conversions but to a lesser
extent.
(17) For reviews, see: (a) Kaesz, H. D. J. Organomet. Chem. 1990, 383,
413–420. (b) Lavigne, G. Eur. J. Inorg. Chem. 1999, 917–930. (c) Hill,
A. F. Angew. Chem., Int. Ed. 2000, 39, 130–133.
(21) For analogy in allylpaladium catalysts, see: (a) Nakamura, H.;
Iwama, H.; Yamamoto, Y. J. Am. Chem. Soc. 1996, 118, 6641–6647. (b)
Nakamura, H.; Aoyagi, K.; Shim, J.-G.; Yamamoto, Y. J. Am. Chem. Soc.
2001, 123, 372–377. (c) Szabo, K. J. Chem.sEur. J. 2000, 6, 4413–4421.
(d) Solin, N.; Kjellgren, J.; Szabo, K. J. J. Am. Chem. Soc. 2004, 126, 7026–
7033.
(18) Han, S.-H.; Geoffroy, G. L.; Dombek, B. D.; Rheingold, A. L.
Inorg. Chem. 1988, 27, 4355–4361.
(19) For an X-ray crystallographic determination of this structure (X )
Cl), see: Chin-Choy, T.; Harrison, W. T. A.; Stucky, G. D.; Keder, N.;
Ford, P. C. Inorg. Chem. 1989, 28, 2028–2029.
(20) (a) Amatore, C.; Jutand, A.; Suarez, A. J. Am. Chem. Soc. 1993,
115, 9531–9541. (b) Amatore, C.; Azzabi, M.; Jutand, A. J. Am. Chem.
Soc. 1991, 113, 8375–8384.
(22) (a) Laine, R. M.; Wilson, R. B. In Aspects of Homogeneous
Catalysis; Ugo, R., Ed.; Reidel (Kluwer): Dodrecht, 1984; Vol. 5. (b) Laine,
R. M.; Rinker, R. G.; Ford, P. C. J. Am. Chem. Soc. 1977, 99, 252–253.
Org. Lett., Vol. 11, No. 3, 2009
783

Table 2. Catalytic Carbonyl Allylation: Substrate Surveya,b
a Reaction conditions: (A): 1.2 equiv of allyl acetate, 70 °C, 24 h; (B): 1.2 equiv of allyl acetate, 80 °C, 24 h; (C): 1.5 equiv of allyl acetate, 80 °C, 48 h.
b Yields of isolated products. c dr 1.6:1
the procedure because it can now be easily adapted to
conventional reactors. We found that 70-80 °C was adequate
for reaction to be complete within 24-48 h for most
substrates. The reaction did proceed at temperatures as low
as 40 °C but again required longer time. Comparable yields
were also obtained when reactions were run in other solvents
such as tetrahydrofuran, tert-butyl alcohol, ethyl acetate, and
cyclohexanone; however, the reactions were slower in
strongly coordinating solvents such as N,N-dimethylforma-
mide and dimethyl sulfoxide.
Under the optimized conditions (A, Table 2), a wide range
of aldehyde substrates were tested. The only changes that
were needed to observe full conversion with slower acting
aldehydes was a minor increase in temperature (B: 80 °C)
and an increase in reaction time (C: 80 °C, 48 h). For the
aromatic aldehydes examined, the reaction was remarkably
insensitive to electronic effects and steric effects. Both
electron-rich (1b, 1c, 1d, 1h) and electron-poor substrates
(1o, 1p, 1q, 1r, 1s) reacted in good yield with little influence
of substituent location. Only the most electron-rich substrate
with a 4-dimethylamino group (1e) did not react completely
under conditions C. Functional group compatibility is quite
good considering the reducing conditions. Thus, nitro, ester,
and hydroxyl groups (1f) and a bromide (1g) are compatible.
Sterically hindered aldehydes (1i, 1j, 1k, 1z) reacted
smoothly as well. Heterocyclic aldehydes (1l, 1m, 1n) reacted
under the standard conditions to give good yields of the
allylation products. We were very pleased to find that olefinic
aldehydes (1t, 1u, 1v) reacted without problem. Perhaps most
satisfying was the ability to allylate aliphatic aldehydes
including linear (1w, 1x), branched (1y), and even the
hindered pivalaldehyde (1z) in good yields. Finally, glyc-
eraldehyde acetonide (2) reacted under the standard condi-
tions (dr, 1.6:1) illustrating the compatibility of heteroatom-
substituted aldehydes. Interestingly, alkynyl aldehydes were
unreactive under these conditions.
In conclusion, we have developed a new, operationally
simple, and highly efficient ruthenium-catalyzed carbonyl
allylation process using inexpensive and noncorrosive allyl
acetate, water, and a low pressure of carbon monoxide. The
role of each reaction component has been defined and
optimized. The reaction is catalytic in ruthenium trichloride,
and because the stoichiometric byproducts are carbon dioxide
and acetic acid, it is environmentally benign and readily
adaptable to large-scale operation. The method shows
excellent substrate generality, functional group compatibility,
and high tolerance to electronic and steric factors. Further
studies to determine structures of the catalytically active
species to introduce stereocontrolling ligands for the carbonyl
addition are underway.
Acknowledgment. We are grateful to Milliken and
Company for generous financial support.
Supporting Information Available: Experiment proce-
dures and compound characterization data. This material is
available free of charge via the Internet at http://pubs.acs.org.
OL8028725
784
Org. Lett., Vol. 11, No. 3, 2009