Catalytic dehydrative allylation of alcohols

Article abstract of DOI:10.1002/anie.200462513

(Chemical Equation Presented) An environmentally benign synthesis of allyl ethers has been developed which can be applied to a highly chemoselective protection of hydroxyl groups. A highly efficient [CpRu]-2-quinolinecarboxylic acid (L) catalytic system was successfully employed for the dehydrative allylation of various alcohols without additional activators and solvent (see scheme; R = alkyl, aryl, multifunctional alkyl, Cp = C₅H ₅).

Full text of DOI:10.1002/anie.200462513

Communications
Synthetic Methods
sufficiently developed to be useful from a generic synthetic
perspective.[8] Herein, we report a highly efficient catalytic
system for the direct dehydrative allylation of alcohols, which
is effective even with only 1.0 equivalents of 2.
Catalytic Dehydrative Allylation of Alcohols**
We have developed a new catalytic system that consists of
a cationic [CpRu] (Cp = cyclopentadienyl) complex and 2-
pyridinecarboxylic acid derivatives for direct allyl ether
cleavage in alcoholic solvents.[2] Taking into account the
reversibility of the cleavage reaction, the catalytic cycle
shown in Scheme 1 is assumed to operate during the
Hajime Saburi, Shinji Tanaka, and Masato Kitamura*
Allyl ethers are not only recognized as important starting
materials for a wide range of organic reactions, including the
1,3 hydrogen shift, [3.3]-sigmatropic rearrangement, and
polymerization,[1] but also constitute one of the most useful
protecting groups for alcohols.[2] In contrast to the extensively
studied allyl ether cleavage, the formation of allyl ethers is
still under development and relies mainly on a Williamson-
type ether synthesis that uses metal alkoxides and allyl halides
or their equivalents.[3] Salt waste-free allylation of alcohols is
desirable, and several catalytic methods utilizing allyl esters
have been reported.[4] Ideally, the catalysis should directly
convert a 1:1 mixture of alcohols 1 and 2-propen-1-ol (2) into
allyl ethers 3 [Eq. (1)] in a solvent-free system without the
Scheme 1. Supposed catalytic cycle for allyl ether formation.
allylation. In the formation of 6 the catalyst precursor 5
captures 2 through an Ru–olefin interaction together with a
hydrogen bond interaction between the COOH moiety and
the OH group. In this catalyst–substrate complex, the hydro-
gen bond interaction and the strong coordination of the s-
donating sp2 nitrogen atom of the pyridine moiety and the
monoanionic h5-Cp ligand to the Ru atom synergistically
enhance the electrophilicity of 2 and the nucleophilicity of the
Ru center. This effect accelerates the oxidative addition of
the RuII center onto 2 to generate a cationic [CpRuIV(C3H5)]+
carboxylate species 7. Nucleophilic attack of 1 onto the p-allyl
carbon, assisted by the good leaving ability of the carboxylate
ligand, gives a catalyst–product complex 8. Liberation of the
product 3 revives the chain carrier 6. All the elementary steps
are essentially reversible, but the phase separation of water
from the reaction system together with waterꢀs poor nucle-
ophilicity may force the equilibrium to the allyl ether side.
Based on the above concept, the reaction conditions for
the allylation of 2-phenylethan-1-ol (1a) with 2 were exam-
ined. Representative results of screening reactions are listed
in Table 1. The allylation proceeds at 708C with 1.0 equiv-
alents of 2 and 0.0005 equivalents of [CpRu(CH3CN)3]PF6
(9)[9] and 2-quinolinecarboxylic acid (10) each to give 3a in
90% yield after 6 hours. Formation of the 1,3-hydrogen
shifted isomer, 1-propenyl ether, was not observed. The
substrate/catalyst (S/C) ratio of 10000 is acceptable, which
approaches a turnover number (TON) of 6500 and a turnover
frequency (TOF) of 5200 at 26% conversion. At 508C, it
requires 42 hours to attain 90% yield. When heated to reflux,
the shift of the equilibrium to the product side is retarded. The
ligand acceleration efficiencies of 1-isoquinolinecarboxylic
need for any additional stoichiometric activators. As water
should be the only co-product this dehydrative allylation is
efficient, environmentally friendly, and simple to operate.
However, the poor leaving ability of the hydroxy group as
well as the low nucleophilicity of the alcoholic oxygen
presents major difficulties.[5] Pioneering research in this area
has been carried out by the Showa Denko company. As it is
assumed that the reaction involves a p-allyl mechanism,
various combinations of palladium or platinum compounds
with mono- and bidentate phosphanes or phosphites were
systematically examined.[6] Several similar approaches have
been reported,[7] but to date none of them have been
[*] H. Saburi, S. Tanaka, Prof. Dr. M. Kitamura
Research Center for Materials Science
Department of Chemistry, Nagoya University
Chikusa, Nagoya 464-8602 (Japan)
Fax: (+81)52-789-2261
E-mail: kitamura@os.rcms.nagoya-u.ac.jp
[**] This work was aided by a Grant-in-Aid for Scientific Research
(No. 14078212) from the Ministry of Education, Science, Sports,
and Culture, Japan. We are grateful to T. Noda, K. Oyama, Y. Maeda,
and T. Okuno for their technical support in the production of
reaction vessels, NMR spectroscopic analysis, and X-ray
diffraction crystallographic analysis.
Supporting information for this article (preparation and
characterization of all substrates and products, general procedures
for allylations, 1H NMR spectroscopic analysis, and X-ray
crystallographic analysis) is available on the WWW under
http://www.angewandte.org or from the author.
1730
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200462513
Angew. Chem. Int. Ed. 2005, 44, 1730 –1732

Angewandte
Chemie
Table 1: Direct allylation of various monoalcohols with 2-propen-1-ol (2)
catalyzed by [CpRuPF6]–2-pyridinecarboxylic acid derivative combined
systems [cf. Eq. (1)].[a]
(11a) in 98% ee was allylated with 1.2 equivalents of 2 to give
(S)-allyl glycidyl ether (11b) in 98% ee and in 87% yield (S/C
= 100/1, CH2Cl2, reflux). No racemization was detectable by
Entry
Alcohol
Ligand
t [h][b]
Yield [%][b]
1
2
1a
1a
6 (3)
90 (93)
10
0.5 (0.5)
66 (58[c])
3
4
1a
1a
0.5 (0.5)
0.5 (0.5)
16 (11[c])
20 (1.3[c])
5
6
1a
1a
0.5 (0.5)
0.5 (0.5)
21 (5.7[c])
2 (0[c])
7
8
9
10
11
12
13
14
15
16
1b
1c
1d
1e
1 f
1g
1h
1i
10
10
10
10
10
10
10
10
10
10
24 (24)
12 (5)
3 (24)
12 (5)
12 (5)
12 (3)
10 (6)
3 (6)
76 (90)
84 (92)
29 (30)
90 [d] (97)
24 (62)
92 (91)
90 (94)
92 (94)
93 (92)
91 (97)
high performance liquid chromatographic (HPLC) analysis.
Optically active 11b is widely used as a chiral unit not only for
the synthesis of functionalized epoxy resins but also for a
variety of natural products. However, there are problems
associated with the preparation of (S)- or (R)-11b. The
Williamson-type allylation often causes a Payne rearrange-
ment, which decreases the optical purity.[10] Transformation
from chiral glycerol derivatives involves several steps.[11] The
hydroxy group at the anomeric C1 of 12a was also smoothly
allylated to give the allyloxy compound in 91% yield of
isolated product, and 1,5-free furanose 13a was diallylated in
90% yield of isolated product. tert-Butyl Fmoc-protected
phenylalanyl serine 14a was converted in a 98% yield to the
corresponding allyl ether, leaving the Fmoc and tert-butyl
ester intact (12a, 13a, or 14a = 100 mm, 2 equivalents of 2 for
each OH group, CH2Cl2, reflux).
Consistent with the proposed catalytic cycle in Scheme 1,
9 was mixed with 10 in a 1:1 ratio in [D6]acetone (each 10 mm)
at 278C. A new set of signals was observed [d = 8.01 (t, 1H, J
= 7.57 Hz), 8.15 (t, 1H, J = 7.57 Hz), 8.32 (m, 2H), 8.84 (d,
1H, J = 8.26 Hz), 9.14 ppm (d, 1H, J = 8.26 Hz)] that could be
assigned to 5 (S = CH3CN, R = 5,6-(CH)4).[12] The signals
immediately disappeared when 1 equivalent of 2 was intro-
duced at 278C to give another new set of signals characteristic
of 7 [R = 5,6-(CH)4] [p-allyl moiety: d = 4.40 (dd, 1H (syn), J
= 2.75, 5.85 Hz), 4.44 (dd, 1H (syn), J = 2.75, 6.20 Hz), 4.75 (d,
1H (anti), J = 9.64 Hz), 4.96 (d, 1H (anti)), 4.96–5.20 ppm (m,
1H (center)]. In solution, the complex is assumed to have an
endo p-allyl structure with a narrowed dihedral angle f
defined by the coordination plane and the p-allyl plane.[13]
This is supported by three observations: 1) the clear A2B2X
allyl signal pattern, 2) the two Hsyn protons resonating at a
higher magnetic field than Hanti protons, and 3) the observa-
tion of 2–5% and approximately 8% enhancements of the
two Hanti signals and the Hsyn signal on irradiating CpH
protons and C(8)H of the quinoline ring, respectively. Under
the reaction conditions described herein no catalyst–substrate
complex 6 was detected, which demonstrates the efficiency of
1j
1k
12 (6)
12 (6)
[a] Reactions were performed at 708C without solvent in a 2000:2000:1:1
ratio of 1/2/[CpRu(CH3CN)3]PF6(9)/ligand. The yields were determined
by GC analysis, see the Supporting Information for details. [b] The values
in parentheses are those obtained at reflux temperature in CH2Cl2
([1]=[2]=500 mm; [9]=[ligand]=1 mm). [c] S/C=100. [d] S/C=1000.
acid, 3-isoquinolinecarboxylic acid, and 2-pyridinecarboxylic
acid are 3–10 times lower than that of 10 in comparison to the
initial rates (entries 2–5). Saturation of the pyridine ring
retards catalytic activity (entry 6). Replacement of the
COOH group of 2-pyridinecarboxylic acid with COOCH3 or
CH2OH groups also abolishes the activity. Diphenylphos-
phanyl acetic acid gave an undesired 1,3-hydrogen shift-
derived compound, 1,1-di(2-phenylethoxy)propane, in 27%
yield after 24 hours. Secondary alcohols, such as 1b and 1c,
are allylated in high yields (entries 7 and 8). Alcohol 1e,
=
which has a C C bond at the 5-position, was converted into
the corresponding allyl ether without any isomerization
(entry 10). With geraniol (1g), only geranyl allyl ether (3g)
was produced among many other possible diallyl ethers
(entry 12). The low yields in the allylation of tertiary alcohol
1d and aryl alcohol 1 f (entries 9 and 11) may be ascribed to
the low nucleophilicity and the reversibility of the present
catalysis. The chemoselectivity of the reaction was high;
allylation was attained in > 90% yield without modifying the
benzyl, benzoyl, methoxymethyl, and tert-butyldiphenylsilyl
protecting groups (entries 13–16).
Furthermore, the present catalytic system is even more
effective with a solvent and can be applied to the synthesis of
multifunctional compounds, such as carbohydrates and pep-
tides, where solvents are essential. For example, (S)-glycidol
Angew. Chem. Int. Ed. 2005, 44, 1730 –1732
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1731

Communications
[3] A. W. Williamson, J. Chem. Soc. 1852, 4, 229.
the bifunctional property of 6 to form 7. The signals for 7 were
[4] Pd: a) R. Akiyama, S. Kobayashi, J. Am. Chem. Soc. 2003, 125,
3412 – 3413; b) A. Yamamoto, Adv. Organomet. Chem. 1992, 34,
111 – 147; c) R. C. Larock, N. H. Lee, Tetrahedron Lett. 1991, 32,
6315 – 6318; d) R. Lakhmiri, P. Lhoste, D. Sinou, Tetrahedron
Lett. 1989, 30, 4669 – 4672; e) J. Muzart, J.-P. GenÞt, A. Denis, J.
Organomet. Chem. 1987, 326, C23 – C28; f) I. Minami, I.
Shimizu, J. Tsuji, J. Organomet. Chem. 1985, 296, 269 – 280;
g) F. Gꢁibe, Y. S. Mꢀleux, Tetrahedron Lett. 1981, 22, 3591 – 3594;
h) K. Takahashi, A. Miyake, G. Hata, Bull. Chem. Soc. Jpn. 1972,
45, 230 – 236; Ir: i) H. Nakagawa, T. Hirabayashi, S. Sakaguchi,
Y. Ishii, J. Org. Chem. 2004, 69, 3474 – 3477.
unaffected by the introduction of 10 mole amounts each of 1a
and 2 at reflux while the allylation proceeded.[14] This
indicates that the p-allyl species 7 is at the resting state in
the catalysis. Complex 7 [R = 5,6-(CH)4] was isolated as a
single pale yellow crystal. The characteristic feature of the
endo p-allyl conformation at a small value of f is also seen in
the X-ray crystallographic measurements (Figure 1). The
[5] For a recent example of Pd-catalyzed allylation of metal
alkoxides with allyl esters, see: H. Kim, C. Lee, Org. Lett.
2002, 4, 4369 – 4371.
[6] a) J. Qꢁ, Y. Ishimura, N. Nagato, Nippon Kagaku Kaishi 1996, 9,
787 – 791; b) Y. Ishimura, J. Qꢁ, Jpn. Kokai Tokkyo Koho Jpn.
Pat. 05306246, 1993.
[7] Pd: Y. Kayaki, T. Koda, T. Ikariya, J. Org. Chem. 2004, 69, 2595 –
2597; Ru: R. C. van der Drift, M. Vailati, E. Bouwman, E. Drent,
J. Mol. Catal. A 2000, 159, 163 – 177; Ni: H. Bricout, J.-F.
Carpentier, A. Mortreux, J. Mol. Catal. A 1998, 136, 243 – 251.
[8] For acid-catalyzed direct allylation, see: a) E. Moffett, J. Am.
Chem. Soc. 1934, 56, 2009; b) M. J.-B. Senderens, Compt. Rend.
1925, 181, 698 – 701; for oxymetalation–dehydroxymetalation
using Cu, Pd, and Hg, see: c) W. Oguchi, H. Uchida, WO Patent
03/106024, 2003; d) C. M. Dumlao, J. W. Francis, P. M. Henry,
Organometallics 1991, 10, 1400 – 1405; e) W. H. Watanabe, L. E.
Conlon, J. C. H. Hwa, J. Org. Chem. 1958, 23, 1666 – 1668.
[9] T. P. Gill, K. R. Mann, Organometallics 1982, 1, 485 – 488; for a
recent efficient synthesis, see: E. P. Kꢁndig, F. R. Monnier, Adv.
Synth. Catal. 2004, 346, 901 – 904.
Figure 1. Molecular structure of [CpRu(p-C3H5)(2-quinolinecarboxyl-
ato)]PF6 [7; R=5,6-(CH)4] determined by X-ray crystallographic
analysis.[15]
[10] G. Uray, N. M. Maier, W. Lindner, J. Chromatogr. A 1994, 666,
41 – 53.
isolated RuIV complex acted as the allylation catalyst with
higher reactivity than that of the corresponding 2-pyridine-
carboxylic acid complex. The rate-determining reductive
elimination of RuIV to a RuII center could be accelerated by
a quinoline ring, which has a higher p-accepting ability than
pyridine.
In conclusion, we have developed an efficient catalytic
system for the dehydrative allylation of alcohols. The new
methodology is superior to conventional synthetic routes[3–8]
in many respects, and it increases the importance of allyl
ethers not only as basic compounds but also as protecting
groups in organic synthesis. Furthermore, a series of NMR
spectroscopic and X-ray crystallographic studies on a key p-
allyl intermediate has given insight into the probable reaction
mechanism.
[11] a) R. L. Pederson, K. K.-C. Liu, J. F. Rutan, L. Chen, C.-H.
Wong, J. Org. Chem. 1990, 55, 4897 – 4901; b) A. B. Mikkilineni,
P. Kumar, E. Abushanab, J. Org. Chem. 1988, 53, 6005 – 6009.
[12] About 10% of unassignable signals were also observed, see
Supporting Information for details.
[13] a) E. Rꢁba, W. Simanko, K. Mauthner, K. M. Soldouzi, C.
Slugovc, K. Mereiter, R. Schmid, K. Kirchner, Organometallics
1999, 18, 3843 – 3850; b) T. Kondo, H. Ono, N. Satake, T.
Mitsudo, Y. Watanabe, Organometallics 1995, 14, 1945 – 1953;
c) H. Nagashima, K. Mukai, Y. Shiota, K. Yamaguchi, K. Ara, T.
Fukahori, H. Suzuki, M. Akita, Y. Moro-oka, K. Itoh, Organo-
metallics 1990, 9, 799 – 807.
[14] The ratio of 3a, 2, and diallyl ether after 30 minutes was
approximately 40:20:20.
[15] Crystallographic analysis of 7 (R = 5,6-(CH)4) pale yellow prism;
¯
P1, a = 7.94(6), b = 8.3(1), c = 14.3(1) ꢂ, a = 106.2(5)8,
b
= 89.95(2)8, g = 90.0(3)8, V= 906.7(19) ꢂ3, Z = 2, R = 0.136,
Rw = 0.163. CCDC 251818 contains the supplementary crystal-
lographic data for this paper. These data can be obtained free of
charge from the Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
Received: November 4, 2004
Published online: February 3, 2005
Keywords: allyl ethers · allylation · homogeneous catalysis ·
.
quinolinecarboxylic acid · ruthenium
[1] a) J. Tsuji in Handbook of Organopalladium Chemistry for
Organic Synthesis, Vol. 5 (Ed.: E. Negishi), Wiley, New York,
2002, pp. 1669 – 1687; b) B. M. Trost, D. L. VanVranken, Chem.
Rev. 1996, 96, 395 – 422; c) S. A. Godleski in Comprehensive
Organic Synthesis, Vol. 4 (Eds.: B. M. Trost, I. Fleming),
Pergamon, Oxford, 1991, pp. 585 – 661.
[2] S. Tanaka, H. Saburi, Y. Ishibashi, M. Kitamura, Org. Lett. 2004,
6, 1873 – 1875, and references therein.
1732
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2005, 44, 1730 –1732