Lewis base catalyzed addition of trimethylsilyl cyanide to aldehydes

Article abstract of DOI:10.1021/jo060153q

A variety of achiral Lewis bases were found to catalyze the addition of TMSCN to the aldehydes. Among them, phosphines and amines were the most efficient catalysts. In addition, several chiral amines and phosphines were examined in a catalytic, asymmetric

Full text of DOI:10.1021/jo060153q

of carbonyl compounds fall into four main classes: (1) enzymes
(oxynitrilase), (2) peptides, (3) Lewis acids (primarily transition
metal complexes), and (4) combinations of Lewis acids and
Lewis bases. The majority of these synthetic agents have been
developed for the addition of trimethylsilyl cyanide (TMSCN)
as a safer alternative to HCN or KCN.7 The enantioselective
silylcyanation of aldehydes 1 can be promoted by many chiral
Lewis acidic complexes to provide products with good yields
and enantioselectivities (Scheme 1).6 Typically, these catalysts
are generated by the combination of a strong Lewis acid with
a chiral ligand either in situ or in a separate preparation. The
complexation of the Lewis acid and chiral ligand usually leads
to deactivation of the catalyst due to the basicity of the donor
atoms of the ligands. Therefore, these catalytic systems often
suffer from low reaction rates.
Lewis Base Catalyzed Addition of Trimethylsilyl
Cyanide to Aldehydes
Scott E. Denmark* and Won-jin Chung
Department of Chemistry, Roger Adams Laboratory,
UniVersity of Illinois, Urbana, Illinois 61801
denmark@scs.uiuc.edu
ReceiVed January 24, 2006
SCHEME 1
In recent years, Shibasaki and co-workers developed Lewis
acid-Lewis base bifunctional catalysts for the addition of
TMSCN to aldehydes.8 This new bifunctional catalyst consists
of a Lewis acidic Al(III) center and Lewis basic phosphine oxide
functionality. The catalyst is designed to activate both the
electrophile and nucleophile at defined positions simultaneously.
The authors suggest that the aluminum moiety electrophilically
activates the aldehyde by complexation, and the phosphine oxide
moiety activates the TMSCN by hypercoordination. These
combined effects lead to excellent chemical yields and high
enantioselectivities of the cyanohydrin product. However, the
reaction requires very long reaction times, and TMSCN must
be added slowly via syringe pump for several hours to achieve
high enantioselectivity.
A variety of achiral Lewis bases were found to catalyze the
addition of TMSCN to the aldehydes. Among them, phos-
phines and amines were the most efficient catalysts. In
addition, several chiral amines and phosphines were exam-
ined in a catalytic, asymmetric addition of TMSCN to
benzaldehyde albeit with low enantioselectivity. A mecha-
nistic study revealed that the reaction was first order in
aldehyde, first order in Lewis base, and zeroth order in
TMSCN, suggesting the complex formation of TMSCN and
Lewis base formation of complex i. However, there are at
least two possible scenarios for this catalytic process, and
in view of the low selectivities observed, it is not clear which
mechanism is operative.
Nucleophilic activation of TMSCN with triphenylphosphine
was first reported by Evans in 1977.9 A systematic investigation
of Lewis base catalysis of this reaction was carried out by
Mukaiyama and co-workers.10 These studies demonstrated that
achiral Lewis bases such as amines, phosphines, arsines, and
stibines can catalyze the addition of TMSCN to aldehydes.
Subsequently, Kagan showed that chiral lithium alkoxides
derived from binol and salens can catalyze the asymmetric
The addition of cyanide to aldehydes and ketones is one of
the oldest known carbon-carbon bond-forming reactions. First
reported in 1832 by Winkler,1 this reaction is the foundation of
the Kiliani-Fisher synthesis of carbohydrates and as such also
represents one of the first stereoselective reactions.2 Over the
intervening decades, cyanohydrins have demonstrated consider-
able synthetic potential as useful building blocks in organic
synthesis.3 Both the hydroxyl and nitrile groups of the cyano-
hydrins can be further transformed into a variety of useful
functional units. After activation, the hydroxyl group can be
displaced by a wide range of nucleophiles to form functionalized
nitriles.4 Furthermore, the nitrile function can be hydrolyzed to
create many different types of R-hydroxy carbonyl compounds
or reduced to form 1,2-amino alcohol derivatives.5
(5) (a) Tanaka, K.; Mori, A.; Inoue, S. J. Org. Chem. 1990, 55, 181-
185. (b) Brown, R. F. C.; Donohue, A. C.; Jackson, W. R.; McCarthy, T.
D. Tetrahedron 1994, 50, 13739-13752. (c) Ziegler, T.; Ho¨rsch, B.;
Effenberger, F. Synthesis 1990, 575-578. (d) Kanerva, L. T. Acta Chem.
Scand. 1996, 50, 234-242.
(6) (a) North, M. Synlett 1993, 807-820. (b) North, M. Tetrahedron:
Asymmetry 2003, 14, 147-176. (c) Synthesis and application of nonracemic
cyanohydrins and R-amino nitriles. Tetrahedron Symposium-in-Print; North,
M., Ed. 2004, 60, 10383-10568. (d) Brunel, J.-M.; Holmes, I. P. Angew.
Chem., Int. Ed. 2004, 43, 2752-2778.
(7) (a) Weber, W. P. Silicon Reagents for Organic Synthesis; Springer-
Verlag: Berlin, 1983. (b) Colvin, E. W. Silicon Reagents in Organic
Synthesis; Academic Press: London, 1981; and references cites therein.
(8) (a) Hamashima, Y.; Sawada, D.; Kanai, M.; Shibasaki, M. J. Am.
Chem. Soc. 1999, 121, 2641-2642. (b) Kanai, M.; Hamashima, Y.;
Shibasaki, M. Tetrahedron Lett. 2000, 41, 2405-2409. (c) Hamashima,
Y.; Sawada, D.; Nogami, H.; Kanai, M.; Shibasaki, M. Tetrahedron 2001,
57, 805-814. (d) Gro¨ger, H. Chem. Eur. J. 2001, 7, 5246-5251.
(9) Evans, D. A.; Wong, R. Y. J. Org. Chem. 1977, 42, 350-352.
(10) Kobayashi, S.; Tsuchiya, Y.; Mukaiyama, T. Chem. Lett. 1991, 4,
537-540.
In light of the synthetic utility of this class of compounds,
the preparation of enantiomerically enriched cyanohydrins has
been extensively investigated and comprehensively reviewed.6
The reagents (catalysts) that effect enantioselective cyanation
(1) Winkler, F. W. Liebigs Ann. Chem. 1832, 4, 246-249.
(2) Kiliani, H. Chem. Ber. 1888, 21, 915-919.
(3) Gregory, R. J. H. Chem. ReV. 1999, 99, 3649-3682.
(4) (a) Effenberger, F.; Kremser, A.; Stelzer, U. Tetrahedron: Asymmetry
1996, 7, 607-618. (b) Effenberger, F.; Stelzer, U. Tetrahedron: Asymmetry
1995, 6, 283-286.
10.1021/jo060153q CCC: $33.50 © 2006 American Chemical Society
Published on Web 04/07/2006
4002
J. Org. Chem. 2006, 71, 4002-4005

again measured in situ using a ReactIR instrument (Figure 2).
Although n-Bu3P was the most effective catalyst found, amines
such as DMAP, Et3N, and NMI were comparable. With the
exception of HMPA, oxygen centered bases showed poor
catalytic activities. Clearly, the rate of reaction correlates well
with the donor ability of the Lewis bases.17
The generality of this trend (increasing reaction rate with
increasing nucleophilicity of the Lewis base) was examined for
three additional aldehydes (E)-cinnamaldehyde (1b), hydrocin-
namaldehyde (1c), and phenylpropagyl aldehyde (1d) with a
limited selection of Lewis bases (Figure 2). Hydrocinnamalde-
hyde and phenylpropagyl aldehyde were very reactive substrates,
and these reactions had to be carried out at -30 °C using 0.1
mol % of catalyst to allow for differences in the efficiency of
the catalysts to be discerned. Interestingly, the trend in catalytic
activity with respect to the donor character of the Lewis base
remained consistent with all of these aldehydes, namely n-Bu3P
> DMAP > Et3N.
As the initial survey clearly showed, n-Bu3P was a highly
effective catalyst for the trimethylsilylcyanation of all classes
of aldehydes surveyed. To determine the structural and chemical
requirements of an efficient catalyst, several phosphorus-
containing Lewis bases were investigated. The catalytic activities
of triphenylphosphine, triethyl phosphite, and hexamethylphos-
phorous triamide (HMPT) in the addition of TMSCN to benz-
aldehyde 1a were measured in situ using a ReactIR instrument.
Triphenylphosphine and triethyl phosphite showed moderate
reactivity, while HMPT showed reactivity similar to n-Bu3P.
3. Kinetic Study. To gain insight into the origin of the
significant catalysis by Lewis bases, a kinetic analysis of the
silylcyanation reaction was carried out. The integral method and
the method of initial rates were employed to determine the order
of the individual reagents in this reaction. First, to determine
the overall reaction order, the concentration of 1a was plotted
against time for three different scenarios; [1a], 1/[1a], and
ln[1a]. The best fit was obtained in plotting ln [1a] vs time.
Therefore, the overall reaction order is first order (see the
Supporting Information).
FIGURE 1. Solvent survey for the addition of TMSCN to benzalde-
hyde.
addition of TMSCN to aldehydes albeit with highly variable
enantioselectivities.11 More recently, Deng and co-workers
reported the use of cinchona alkaloids as chiral Lewis bases
for trimethylsilylcyanation of activated carbonyl compounds.12,13
All of these developments have significantly advanced the
frontier of enantioselective cyanohydrin synthesis. However,
there still is room for improvement, particularly with regard to
catalyst simplicity, reaction generality, and especially reaction
rate. In view of our continuing interest in the use of chiral Lewis
bases for asymmetric transformations,14 we have undertaken the
development of a general Lewis base catalyzed asymmetric tri-
methylsilylcyanation of aldehydes. Our initial efforts to develop
a novel catalytic system, focused on a thorough and quantitative
survey of various Lewis bases to establish their relative catalytic
efficiency. The rate equation was also of interest to clarify the
role of each component and determine if, mechanistically,
nucleophilic catalysis was amenable to asymmetric induction.
It was hoped that these fundamental investigations would guide
the design of an effective chiral Lewis base catalyst.
1. Survey of the Solvent. To develop an effective Lewis base
catalyzed process for the addition of TMSCN to aldehydes,
several experimental variables were investigated. Foremost
among the factors that can influence the rate of the reaction are
the solvents and catalyst structures. Therefore, a systematic study
of these two variables was performed. Initially, various solvents
were surveyed to determine what effect polarity and donicity15
have on the rate of addition of TMSCN to benzaldehyde (1a).
Triethylamine, which had been successfully employed previ-
ously, was chosen as the catalyst for these reactions.10 Both
the catalyzed and the uncatalyzed reaction rates were measured
in situ using a ReactIR instrument, and the catalyzed reaction
profiles are shown in Figure 1. Unfortunately, no clear trend
emerges from these data, as the relative rates do not correlate
with the polarity or the basicity of solvents.15,16 Nevertheless,
the rapid rate of the addition in acetonitrile compared to other
solvents clearly identified it as the solvent of choice for this
transformation. Furthermore, no reaction was observed in the
absence of triethylamine, thus demonstrating that CH3CN was
not catalyzing the reaction.
By varying the amount of 1a, the initial rates at different
concentrations of 1a could be measured. Examination of the
initial rates revealed that the reaction was first order in 1a.
Similar studies were done for the Et3N and TMSCN. The kinetic
plots revealed that the reaction was first order in Et3N and zeroth
order in TMSCN (see the Supporting Information).
4. Mechanism. On the basis of the results from these kinetic
studies, two catalytic cycles can be proposed (Figure 3).
Catalytic cycle A involves an ionized cyanide intermediate i
held in a tight ion pair. After the silyl cation coordinates to the
(11) (a) Holmes, I. P.; Kagan, H. B. Tetrahedron Lett. 2000, 41, 7453-
7456. (b) Holmes, I. P.; Kagan, H. B. Tetrahedron Lett. 2000, 41, 7457-
7460.
(12) Tian, S.-K.; Hong, R.; Deng, L. J. Am. Chem. Soc. 2003, 125, 9900-
9901.
(13) For other examples, see: (a) Kitani, Y.; Kumamoto, T.; Isobe, T.;
Fukuda, K.; Ishikawa, T. AdV. Synth. Catal. 2005, 347, 1653-1658. (b)
Wen, Y.; Huang, X.; Huang, J.; Xiong, Y.; Qin, B.; Feng, X. Synlett 2005,
2445-2448.
(14) (a) Denmark, S. E.; Fujimori, S. In Modern Aldol Reactions;
Mahrwald, R., Ed.; Wiley-VCH: Weinheim, 2004; Vol. 2; Chapter 7. (b)
Denmark, S. E.; Stavenger, R. A. Acc. Chem. Res. 2000, 33, 432-440. (c)
Denmark, S. E.; Beutner, G. L.; Wynn, T.; Eastgate, M. D. J. Am. Chem.
Soc. 2005, 127, 3774-3789.
(15) Reichardt, C. SolVents and SolVent Effects in Organic Chemistry,
2nd ed.; VCH: Weinheim, 1988.
(16) Maria, P.-C.; Gal, J.-F. J. Phys. Chem. 1985, 89, 1296-1304.
(17) Bassindale, A. R.; Stout, T. Tetrahedron Lett. 1985, 26, 3403-3406.
2. Survey of the Lewis Base Catalyst. A wide range of
functionally and structurally diverse Lewis bases were examined
in acetonitrile. Included in this survey were tri-n-butylphosphine
(n-Bu3P), N,N-(dimethylamino)pyridine (DMAP), triethylamine
(Et3N), N-methylimidazole (NMI), hexamethylphosphoric tri-
amide (HMPA), N,N-dimethylformamide (DMF), dimethyl sul-
foxide (DMSO), pyridine N-oxide, triphenylphosphine oxide,
thiourea, and tetramethylurea. The rate of trimethylsilylcyanation
of benzaldehyde (1a) in the presence of these Lewis bases was
J. Org. Chem, Vol. 71, No. 10, 2006 4003

FIGURE 3. Proposed catalytic cycles.
Catalytic cycle B involves formation of a free cyanide anion
that can add to the aldehyde without influence of the silyl cation
bound Lewis base. In this cycle, the chiral information on the
Lewis base is not transferred to the newly forming stereocenter.
If the Lewis base-cyanide exchange is irreversible or the
equilibrium lies on the free cyanide anion, the reaction will be
first order in catalyst and zeroth order in TMSCN. Furthermore,
if the addition of the free cyanide to aldehyde is the rate
determining step, the reaction will be first order in aldehyde.
Therefore, to obtain asymmetric induction, the ion pair should
be stabilized to prevent dissociation, thereby allow for the
reaction to proceed through catalytic cycle A. The stability of
the ion pair i can be greatly influenced by changing reaction
conditions such as solvent, Lewis base, and temperature. Studies
of trimethylsilylcyanation of aldehydes in the presence of various
chiral Lewis bases is described below.
5. Survey of Chiral Lewis Bases. 5.1. Amines. On the basis
of the results from the survey of catalysts, chiral amines and
phosphines were investigated as potential candidates for asym-
metric trimethylsilylcyanation of benzaldehyde 1a. Because
amines are easier to handle than oxidation sensitive phosphines,
several readily available chiral amines were examined. Unfor-
tunately, trimethylsilylcyanation performed in the presence of
chiral amines, (S)-R-methylbenzylamine (4), (1S,2R)-N-meth-
ylephedrine (5), (-)-strychnine (6), (DHQ)2PHAL (7), and
(DHQD)2PHAL (8) afforded only racemic product (Chart 1).
Although there are many reasons why a catalyzed reaction dis-
plays no enantioselectivity, two explanations immediately come
to the fore: (1) the high reactivity of TMSCN in CH3CN, which
CHART 1
FIGURE 2. Lewis base survey for the addition of TMSCN to various
aldehydes (only efficient Lewis bases are shown for 1b-d).
aldehyde, cyanide addition can occur under influence of the
chiral Lewis base. In this cycle, the chiral information from the
Lewis base can be transferred to the newly forming chiral center.
If the binding of the Lewis base to TMSCN is irreversible or
the equilibrium lies on the ion pair i, the reaction will be first
order in catalyst and zeroth order in TMSCN. If the coordination
of silyl group to aldehyde is the rate-determining step, the
reaction will be first order in aldehyde.
4004 J. Org. Chem., Vol. 71, No. 10, 2006

TABLE 1. Asymmetric Trimethylsilylcyanation of 1a with
(DHQ)2PHAL in Various Solventsa
phines was attenuated through introducing oxygens in place of
the nitrogens.
Although both compounds showed appreciable catalytic
activity, there was no asymmetric induction as racemic products
were isolated. Phosphoramidite 1119 and phosphite 1220b were
also surveyed, but no addition product could be isolated.
Despite unsatisfactory results from chiral Lewis base cata-
lyzed reactions, the nonracemic product obtained suggests that
improving enantioselectivity is mechanistically possible. Whereas
cinchona alkaloid monomers provided only racemic products,
modest enantioselectivity could be obtained with cinchona
alkaloid dimers. Because the reaction was first order in catalyst,
it is reasonable to assume that only one of the cinchona units
in the dimer is needed to catalyze the reaction. Therefore, it
can be postulated that only one cinchona alkaloid of the dimer
activates TMSCN and the other cinchona alkaloid transfers
chiral information to the reaction center, probably by shielding
one side of aldehyde. This design will be pursued.
entry
solvent
ꢀrb
time, h
yield, %
erc
1
2
hexane
toluene
1,4-dioxane
diethyl ether
CHCl3
1.88
2.38
3.31
4.20
4.81
6.02
7.58
8.93
8.93
35.94
17.8
16.7
14.8
18.2
15.1
17.3
13.7
5.0
63
59
37
53
73
21
54
57
17
74
49/51
34/66
41/59
51/49
37/63
46/54
48/52
40/60
42/58
51/49
3
4
5
6
ethyl acetate
THF
7
8
CH2Cl2
9d
10
CH2Cl2
5.3
4.0
CH3CN
a All reactions employed 1.05 equiv of TMSCN at 0.16 M concentration.
b Dielectric constant. c Determined by CSP-GC with G-TA column; the
absolute configuration was not determined. d Reaction at -78 °C.
Experimental Section
General Experimental Procedures. See the Supporting Infor-
mation.
would be characterized by an early transition state with little
steric interaction between the aldehyde and chiral catalyst or
(2) the ionization of the TMSCN by the Lewis base catalyst,
thus generating free cyanide which can add to the aldehyde in
an achiral addition reaction. Both of these pathways may be
rendered selective the use of other solvents that would either
slow the reaction or stabilize a hypervalent silicon intermediate
and disfavor ionization.
To test this hypothesis, several solvents were examined in
the addition of TMSCN to benzaldehyde 1a in the presence of
(DHQ)2PHAL (Table 1). Although moderate enantioselectivities
were achieved in several cases (entries 2, 5, and 8), the increase
in selectivity could not be related to the reaction rate or polarity
of the solvents.
Although the addition to 1a proceeded with low enantiose-
lectivity, other aldehyde structures were also tested to ensure
that this phenomenon was not substrate specific. Therefore, the
same solvents were surveyed in the addition to 1b and 1c in
the presence of (DHQ)2PHAL. For 1b, the products were always
racemic. However, for 1c, moderate enantioselectivity was
observed in 1,4-dioxane (65/35).18
5.2. Phosphines. Although trialkylphosphines are air-sensi-
tive, aminophosphines are less prone to oxidation and easier to
prepare. Therefore, chiral aminophosphines were the first choice
as chiral, phosphorus containing Lewis bases (Chart 2). Due to
the strong influence of solvent on the enantioselectivity that
was observed when catalyzed by amines, these reactions were
performed in a range of solvents. Chiral aminophosphines 919
and 1020 were synthesized and used as catalysts in the tri-
methylsilylcyanation of 1a in various solvents. However, under
these reaction conditions no enantioselectivity was observed.
The lack of asymmetric induction could result from the strong
nucleophilicity of the aminophosphine, generating free cyanide
ion from the ion pair. Therefore, the nucleophilicity of phos-
DMAP-Catalyzed Addition of TMSCN to Benzaldehyde.
Acetonitrile (20 mL) was added to a flame-dried, 50 mL flask
containing DMAP (12 mg, 0.098 mmol, 0.01 equiv) with magnetic
stir bar under N2. Benzaldehyde (1 mL, 9.841 mmol) and TMSCN
(1.325 mL, 9.939 mmol, 1.01 equiv) were added to the solution.
After 1 h, the solvent was removed under reduced pressure. The
residue was purified by column chromatography (silica gel, hexane/
ethyl acetate, 10/1) to give 1.920 g (95%) of 3a as a clear, colorless
oil.21 Data for 3a: 1H NMR (400 MHz, CDCl3) 7.49-7.39 (m, 5
H), 5.50 (s, 1 H), 0.24 (s, 9 H); 13C NMR (100 MHz, CDCl3) 136.2,
129.6, 129.5, 126.6, 119.4, 63.9, -0.0; GC (G-TA column, 110
kPa, 90 °C) tR 13.3 min (50%), 14.3 min (50%).
DMAP-Catalyzed Addition of TMSCN to Hydrocinnamal-
dehyde. Acetonitrile (3 mL) was added to a flame-dried, 10 mL
flask containing DMAP (1.9 mg, 0.015 mmol, 0.01 equiv) with a
magnetic stir bar under N2. Hydrocinnamaldehyde (200 µL, 1.519
mmol) and TMSCN (205 µL, 1.534 mmol, 1.01 equiv) were added
to the solution. After 20 min, the solvent was removed under
reduced pressure. The residue was purified by flash column
chromatography (silica gel, hexane/ethyl acetate, 10/1) to give 301
mg (85%) of 3c as a clear, colorless oil.21 Data for 3c: 1H NMR
(400 MHz, CDCl3) 7.33-7.19 (m, 5 H), 4.37 (dd, J ) 6.4, 6.4, 1
H), 2.80 (dd, J ) 3.8, 3.8, 2 H), 2.09-2.15 (m, 2 H), 0.21 (s, 9 H);
13C NMR (100 MHz, CDCl3) 139.9, 128.6, 128.4, 126.4, 119.9,
60.6, 37.6, 30.6, -0.4; GC (G-TA column, 110 kPa, 110 °C) tR
12.5 min (50%), 13.5 min (50%).
Acknowledgment. We are grateful to the National Science
Foundation (NSF CHE 041440) for generous financial support.
Supporting Information Available: Full experimental proce-
dures and characterization data (including representative 2D NMR
spectra) for all addition products. This material is available free of
charge via the Internet at http://pubs.acs.org.
JO060153Q
(18) Me3NO is as effective as DMAP catalyst for the trimethylsilylcya-
nation. A few chiral trialkylamine N-oxides were prepared, and they have
appreciable catalytic activity; no significant asymmetric induction was
observed.
CHART 2
(19) Hulst, R.; de Vries, N. K.; Feringa, B. L. Tetrahedron: Asymmetry
1994, 5, 699-708.
(20) (a) Merkulov, R. V.; Maslennikova, V. I.; Nifant’ev, E. E. Russ. J.
Gen. Chem. 1998, 68, 876-880. (b) Kolodiazhnyi, O. I.; Sheiko, S.;
Grishkun, E. V. Heteroatom Chem. 2000, 11, 138-143.
(21) Yang, W.-B.; Fang, J.-M. J. Org. Chem. 1998, 63, 1356-1359.
J. Org. Chem, Vol. 71, No. 10, 2006 4005