Enantioselective alkynylation of aldehydes catalyzed by [2.2]paracyclophane-based ligands

Article abstract of DOI:10.1021/ol049596b

[2.2]Paracyclophane-based ketimine ligands were evaluated as catalysts for the enantioselective addition of in situ-prepared alkynylzinc reagents to aldehydes. The initial high activity and enantioselectivity of these ligands could be improved by an addit

Full text of DOI:10.1021/ol049596b

ORGANIC
LETTERS
2
004
Vol. 6, No. 13
113-2116
Enantioselective Alkynylation of
Aldehydes Catalyzed by
2
[2.2]Paracyclophane-Based Ligands
Stefan Dahmen*
cynora GmbH, Kaiserstr. 100, D-52134 Herzogenrath, Germany
dahmen@cynora.de
Received March 4, 2004 (Revised Manuscript Received May 11, 2004)
ABSTRACT
[2.2]Paracyclophane-based ketimine ligands were evaluated as catalysts for the enantioselective addition of in situ-prepared alkynylzinc reagents
to aldehydes. The initial high activity and enantioselectivity of these ligands could be improved by an additive screening. The final protocol
gives chiral propargyl alcohols in up to >98% ee.
Alkynylzinc additions to aldehydes represent an interesting
class of reactions of zinc reagent mainly for the following
reasons. First, the products, chiral propargylic alcohols, are
of N-methylephedrine.4,5 Excellent results were obtained for
aliphatic aldehydes, and the method could even be run
solvent-free with catalytic amounts of ligand and metal.4d
Recently, Xu and Pu took advantage of the fact that, in
the presence of a suitable catalyst or ligand, the deprotonation
of a terminal alkyne by diethylzinc occurs even at room
temperature in THF. In the presence of bulky BINOL-based
ligands, good to excellent enantioselectivities were achieved
for aromatic aldehydes.
Despite the achievements made in this field of zinc
chemistry, the alkynylation of aldehydes has not yet reached
the level of practicability that is required for a synthetically
useful catalytic reaction. In all examples cited above, high
loadings of ligands (10 to >100%) had to be employed to
achieve good selectivity. In many cases, the gram amount
of the employed ligand exceeded the amount of substrate.
1
important precursors to many chiral organic compounds.
Second, the process as such can be conducted using several
2
different protocols that apparently all rely on different
6
mechanisms. In their initial report on the addition of
alkynylzinc reagents to aldehydes, Ishizaki and Hoshino used
a protocol to prepare the active zinc reagent by refluxing a
terminal alkyne with diethyl or dimethylzinc in THF.
Depending on the stoichiometry, alkyl-alkinylzinc or di-
3
alkynylzinc species are formed. The latter are scarcely
soluble in unpolar solvents and usually give a lower
selectivity in the addition to aldehydes in the presence of
chiral pyridyl alcohol ligands.
A mechanistically different approach by Carreira and co-
workers initially used stoichiometric amounts of N-methyl-
(
4) (a) Frantz, D. E.; F a¨ ssler, R.; Carreira, E. M. J. Am. Chem. Soc. 2000,
ephedrine and Zn(OTf)
2 3
in the presence of Et N to prepare
1
22, 1806. (b) Boyall, D.; L o´ pez, F.; Sasaki, H.; Frantz, D.; Carreira, E.
the zinc reagent, which most likely is a alkynylzinc complex
M. Org. Lett. 2000, 26, 4233-4236. (c) Frantz, D. E.; F a¨ ssler, R.; Tomooka,
C. S.; Carreira, E. N. Acc. Chem. Res. 2000, 33, 373. (d) Anand, N. K.;
Carreira, E. N. J. Am. Chem. Soc. 2001, 123, 9687.
(5) Zn(OTf)2 is 10 times more expensive than diethylzinc: Zn(OTf)2,
10 g ) 46.80 $ (1 euro/mmol); Et2Zn in hexane, 100 g (810 mmol) )
80.90 euro (0.1 euro/mmol). Prices were taken from Acros Organics.
(6) Xu, M.-H.; Pu, L. Org. Lett. 2002, 4, 4555.
(1) Modern Acetylene Chemistry; Stang, P. J., Diederich, F., Eds.;
VCH: Weinheim, 1995.
(
(
2) For a recent review, see: Pu, L. Tetrahedron 2003, 59, 9873.
3) Ishizaki, M.; Hoshino, O. Tetrahedron: Asymmetry 1994, 5, 1901.
1
0.1021/ol049596b CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/28/2004

Table 1. Reaction Optimization for the Addition of Phenylacetylene to Benzaldehydea
entry ligand (mol %)
solvent
1 mL of hexane
1 mL of hexane
1 mL of hexane
1 mL of hexane, 0.1 mL of THF
1 mL of hexane, 1 mL of THF
1 mL of hexane, 1 mL of toluene
1 mL of hexane
1 mL of hexane; 1 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 2 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 1 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 1 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 2 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 2 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 2 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 2 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 2 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 2 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 2 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 2 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 2 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 2 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 2 mL of toluene MeOPEG (1 mol %)
1 mL of hexane; 2 mL of toluene MeOPEG (1 mol %)
additive
temp (°C) protocolb yield of 3 (4) (%) ee of 3 (%)c
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
(Rp,S)-5 (2)
(Rp,S)-5 (2)
(Rp,S)-5 (2)
(Rp,S)-5 (2)
(Rp,S)-5 (2)
(Rp,S)-5 (2)
(Rp,S)-5 (2)
(Rp,S)-5 (2)
(Rp,S)-5 (2)
(Rp,S)-5 (2)
(Rp,S)-5 (2)
(Rp,S)-5 (2)
(Sp,S)-5 (2)
(Rp,S)-6 (2)
(Sp,S)-6 (2)
(Sp,S)-7 (2)
(-)-8 (5)
(-)-9 (5)
(S)-10 (5)
11 (10)
12 (10)
(+)-13 (5)
(-)-14 (2)
(Rp,S)-5 (5)
20
20
20
20
20
20
20
20
10
5
A
B
C
C
C
C
C
B
B
B
B
D
B
B
B
B
B
B
B
B
B
B
B
E
22 (73)
36 (60)
47 (50)
72 (22)
98
63 (33)
76 (20)
62 (33)
46 (50)
50
54 (R)
58 (R)
57 (R)
51 (R)
9 (R)
40 (R)
64 (R)
71 (R)
82 (R)
74 (R)
54 (R)
91 (R)
67 (S)
44 (R)
72 (S)
75 (S)
37 (S)
67 (R)
10 (S)
7 (S)
MeOPEG (1 mol %)
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
0
72
10
10
10
10
10
10
10
10
10
10
10
10
10
51
44
66
48
51
66
79
85
96
95
74
86
23 (S)
16 (S)
57 (R)
92 (R)
2 mL of toluene
DiMPEG (2.5 mol %)
90
a
All reactions were run on a 0.5 mmol scale. b Protocol A: ligand, solvent, additive (as indicated), diethylzinc (1 mL, 1 M in hexane), phenylacetylene
(
110 µL), 0.5 h, rt, then benzaldehyde (50 µL), 12 h. Protocol B: like protocol A, but all compounds were stirred for 1 h at room temperature before addition
of benzaldehyde. Protocol C: like protocol A, but all compounds were stirred for 3 h at room temperature before addition of benzaldehyde. Protocol D:
slow addition (5 h) of a mixture of phenylacetylene (110 µL) and diethylzinc (1 mL, 1 M in hexane) to a mixture of ligand, additive, benzaldehyde. Protocol
E: like protocol A, but dimethylzinc (0.5 mL, 2 M in toluene) was used instead of diethylzinc. c Absolute configurations were assigned by comparison of
HPLC traces and optical rotation values with known compounds and the assumption of a unanimous reaction pathway for all substrates.
9
This is most certainly due to the higher reactivity of the
alkynylzinc bond as compared to alkylzinc bonds, which also
manifests in the high selectivity for alkynyl vs alkyl transfer
when excess diethylzinc is used in the above-mentioned
protocols.
in organozinc reactions inspired us to take a closer look at
the alkynylzinc addition to aldehydes.
In an initial screening, conditions for the in situ formation
of alkynylzinc by deprotonation of phenylacetylene with
diethylzinc were elaborated (Table 1). In all runs listed in
7
Our experience with active zinc species and highly active
(
8) (a) Dahmen, S.; Br a¨ se, S. J. Am. Chem. Soc. 2002, 124, 5940. (b)
Hermanns, N.; Dahmen, S.; Bolm, C.; Br a¨ se, S. Angew. Chem., Int Ed.
8
substrates as well as recent observations of additive effects
2002, 41, 3692.
(7) Dahmen, S.; Br a¨ se, S. Org. Lett. 2001, 3, 4119.
(9) Bolm, C.; Rudolph, J. J. Am. Chem. Soc. 2002, 124, 14850.
2114
Org. Lett., Vol. 6, No. 13, 2004

Table 1, the conversion of starting benzaldehyde was >95%
and the amounts of alkynylation and ethylation product were
measured by a calibrated HPLC method.
Scheme 1. Equilibrium of Zinc Species
Stirring both starting materials in the presence of 2
mol % ligand (R ,S)-5 for 1 h at room temperature and then
p
adding the aldehyde resulted in the formation of 3 in 22%
yield and 54% ee (entry 1). If the stirring time before the
addition of aldehyde is prolonged to 2 or 3 h, 36 and 47%
yield of 3 are obtained (entries 2 and 3). Hence, the
deprotonation of phenylacetylene by diethylzinc in hexane
is a relatively slow step. This deprotonation, however, is
supported by basic additives. In the presence of 10 vol %
THF, 72% yield of 3 is obtained (entry 4). However, the ee
begins to drop. In a solvent mixture of hexane and THF (1/
redissolved by heating, addition of toluene, or ultrasound.
Presumably, the primarily formed mixed ethyl-alkynyl-zinc
species 15 disproportionates to give the (insoluble) di-
alkynylzinc 16 and diethylzinc. This behavior was only
observed in unpolar solvents such as hexane and toluene. In
1), 98% yield of 3 is formed in a disappointing 9% ee (entry
5).
Addition of 1 mol % MeOPEG (poly(ethylene glycol)
monomethyl ether, MW 2000) in a solvent mixture of hexane
(
(
1 mL of a 1 M diethylzinc solution in hexane) and toluene
1 to 2 mL) improved the enantioselectivity from 58 to 71%
1
1
THF, the mixed species was apparently stable.
This assumption is supported by the finding that a
suspension of the precipitated dialkynylzinc gives the desired
product 3 in high yield (90%) but lower ee (72% ee).12 The
same effect is observed when the dialkynylzinc is prepared
selectively by the reaction of diethylzinc and excess phenyl-
acetylene (2.2 equiv) at elevated temperature (94% yield,
ee at room temperature. Lowering the temperature to 10 °C
boosted the ee up to 82%. However, at 5 and 0 °C, the
enantioselectivity again dropped to 74 and 54% ee, respec-
tively (entries 10 and 11).
Under these conditions, a ligand screening was performed
employing a variety of commercially available N,O-ligands,
6
3% ee).
1
0
as well as a set of [2.2]paracyclophane-based ketimines.
Several attempts were made to improve the enantioselec-
As can be seen from Table 1, most simple amino alcohol
ligands do not represent promising catalysts for this reaction
tivity on the basis of the protocol proceeding via dialkynyl-
zinc species 16 (temperature, additives, ligands; results not
listed). As these attempts were unsuccessful, we chose to
go back to the initial protocol C and substitute diethylzinc
for the less reactive dimethylzinc. Gratifyingly, this change
in the zinc precursor not only led to drastically better alkynyl
vs alkyl addition selectivity (presumably due to the much
slower methyl transfer) but also increased the enantioselec-
tivity of the reaction. Finally, DiMPEG 2000 (poly(ethylene
glycol) dimethyl ether, MW 2000) proved to be slightly
superior to MeOPEG (entry 1 in Table 2).
(under the chosen conditions). An interesting observation is
that secondary amine 9 gives a reasonable 67% ee at 5
mol % ligand loading. Secondary amines usually do not give
good results in the diethylzinc addition to aldehydes.
Among the [2.2]paracyclophane-based ligands, initially
p
chosen (R ,S)-5 gave the best results. Using either a slow
addition technique (entry 12) or increasing the catalyst
loading to 5 mol % (entry 24, see below) gave rise to the
reaction product in >90% ee.
Having found the most effective ligand for the test
reaction, we again had to address the problem of selectivity
of alkynyl vs alkyl transfer. As the paracyclophane-based
ligands depicted in Table 1 are also highly potent ligands
for the dialkylzinc addition to aldehydes, the ethylation
product 4 is also obtained when larger amounts of diethylzinc
are present in the reaction mixture. As already mentioned
above, the deprotonation of phenylacetylene in unpolar
solvents is rather slow at room temperature in the absence
of basic additives. Basic additives and especially THF, on
the other hand, drastically diminished the ee. One way out
this problem would be the preformation of the desired active
zinc species 15 (Scheme 1) by (a) a longer deprotonation
time or (b) a deprotonation at higher temperature before the
addition of the aldehyde.
With this final protocol in hand, we examined the scope
of the reaction using a variety of aromatic aldehyde
substrates. In nearly all cases, enantioselectivities >90%
could be obtained using only 5 mol % ligand (Table 2, entries
1
6
-7). Only m-anisaldehyde gave a lower ee of 80% (entry
). Astonishingly, in the reaction with diethylzinc instead
of dimethylzinc, 90% ee is obtained for this substrate (entry
).
Cyclohexylaldehyde gives lower ee’s than the aromatic
substrates (entry 9). This supports the literature evidence that
for aliphatic substrates, the Zn(OTf) /Et N protocol seems
7
2
3
2
to be superior. Less acidic acetylenes such as 1-hexyne
require a higher deprotonation temperature of 40 °C (entry
10). The addition product is obtained in an unoptimized 84%
ee. TMS-acetylene could be employed giving rise to the
corresponding addition product in 78% ee (entry 11), while
However, if an equimolar mixture of diethylzinc and
phenylacetylene was stored over a prolonged period of time
(
11) This behavior is atypical for mixed zinc species. The equilibrium
could be influenced by the very low solubility of the dialkynylzinc reagents.
12) Clear supernant solution gives rise to the ethylation product in >80%
yield.
(2-3 days), a white precipitate formed that could not be
(
(10) Dahmen, S.; Br a¨ se, S. Chem. Commun. 2002, 26.
Org. Lett., Vol. 6, No. 13, 2004
2115

benzaldehyde (entry 13) and 1-hexyne to cyclohexylaldehyde
(entry 14) were both obtained with disappointing enantio-
selectivities (38 and 37% ee, respectively). The 1,2-addition
of 1-hexyne to an R,â-unsaturated aldehyde proceeded with
Table 2. Scope of Reactiona
8
0% ee (entry 15).
In retrospect, the described protocols are much more
sensitive to slight changes in the reaction conditions than
protocols for any other organozinc addition we have en-
countered so far. From an experimental point of view, the
reactivity of alkynylzinc species would be on the same order
of magnitude as diphenylzinc (alkylzinc < alkenylzinc <
phenylzinc e alkynylzinc).
In summary, different protocols for the enantioselective
addition of alkynylzinc reagents to aldehydes were compared.
On the basis of the protocol that allows for the lowest catalyst
loading, a screening of different commercially available N,O-
ligands as well as several [2.2]paracyclophane-based ketimine
ligands was conducted. Due to their high activity, the [2.2]-
paracyclophane-based ligands proved to be superior over
conventional amino alcohol ligands. The developed protocol
reflects the most economic catalyst system so far for the
addition of terminal alkynes to aromatic aldehydes in terms
of ligand loading and enantioselectivity.
entry
alkyne
aldehyde (R1) yield (%)
ee (%)
1
2
3
4
5
6
7b
8
9
PhCtCH
PhCtCH
PhCtCH
PhCtCH
PhCtCH
PhCtCH
PhCtCH
PhCtCH
PhCtCH
1-hexyne
C6H5
90
87
92
89
86
87
58
86
82
86
83
52
85
56
86
92 (R)
85 (R)
90 (R)
95 (R)
>98 (R)
80 (R)
90 (R)
94 (R)
77 (R)
84 (R)
78 (R)
85 (R)
3-Cl-C6H4
3-Me-C6H4
4-Cl-C6H4
2-Br-C6H4
3-MeO-C6H4
3-MeO-C6H4
1-Napht
c-C6H11
1
1
1
1
1
1
0c
1c
2c
C6H5
(CH3)3SiCtCH
(CH3)3CCtCH
PhCtCH
1-hexyne
C6H5
c-C6H11
n-C3H7
c-C6H11
C6H5CHdCH
3b,c
38 (R)
37d
80 (R)
4c
5c
1-hexyne
a
Ligand (Rp,S)-5 (5 mol %), toluene (1.0 mL), DiMPEG 2000 (2.5 mol
%), dimethylzinc (0.5 mL, 2 M in toluene), acetylene (1 mmol), 2 h, rt,
b
then aldehyde (0.5 mmol), 12 h, 10 °C. Diethylzinc was used instead of
dimethylzinc. Deprotonation was conducted at 40 °C instead of room
temperature. Analyzed as the (-)-camphanic acid ester.
c
Acknowledgment. J. Rudolph is acknowledged for
inspiring discussions on additive effects. T. Dahmen is
acknowledged for assistance, and M. Lormann is acknowl-
edged for carefully reading the manuscript
d
the addition of tert-butyl acetylene proceeded with 85% ee
(entry 12). The addition products of phenyl acetylene to
OL049596B
2116
Org. Lett., Vol. 6, No. 13, 2004