Rational design of chiral lithium amides for asymmetric alkylation reactions - NMR spectroscopic studies of mixed lithium amide/alkyllithium complexes

Article abstract of DOI:10.1002/(SICI)1521-3765(19990802)5:8<2348::AID-CHEM2348>3.0.CO;2-A

Treatment of solutions of chiral lithium amides, containing internally coordinating groups, in diethyl ether (DEE) with alkyllithiums results in the formation of chiral lithium amide/alkyllithium mixed dimers. We report the use of eight different chiral l

Full text of DOI:10.1002/(SICI)1521-3765(19990802)5:8<2348::AID-CHEM2348>3.0.CO;2-A

FULL PAPER
Rational Design of Chiral Lithium Amides for Asymmetric Alkylation
ReactionsÐNMR Spectroscopic Studies of Mixed Lithium
Amide/Alkyllithium Complexes
Per I. Arvidsson, Göran Hilmersson, and Öjvind Davidsson*[a]
Dedicated to Prof. Per Ahlberg on the occasion of his 60th birthday
Abstract: Treatment of solutions of chi-
ral lithium amides, containing internally
coordinating groups, in diethyl ether
(DEE) with alkyllithiums results in the
formation of chiral lithium amide/alkyl-
lithium mixed dimers. We report the use
of eight different chiral lithium amides/
n-butyllithium mixed dimers in the
asymmetric alkylation of benzaldehyde
in DEE solution at 1168C. The addi-
tion product, (S)-1-phenyl-1-pentanol,
was formed with enantiomeric excesses
(eeꢀs) ranging from 0 to 82%. In DEE/
dimethoxy methane solvent mixtures
the stereoselectivity was improved and
gave the product in 91% ee. Compar-
ison of the ligand structures revealed
that only one chiral center was necessary
to obtain high enantioselectivity. Re-
placement of the internally coordinating
methoxy group with a pyrrolidine group
lowered the ee from 82% to 24%. Low-
temperature 6Li NMR studies of mix-
tures of these reagents in DEE revealed
large differences in the concentration of
the reactive mixed dimers formed. Full
complexation of nBuLi to the chiral
lithium amides is not necessary in order
to obtain high enantioselectivity, as the
result of an increased reactivity of the
complexed nBuLi.
Keywords: alkylations ´ asymmetric
synthesis
amides
´
chiral
lithium
´
mixed complexes
´
NMR spectroscopy
chiral lithium amides (i.e., chiral inducers). Thus, chiral
lithium amides induce asymmetry into the otherwise sym-
metric organolithium compounds. Provided that such chiral
complexes are formed and react with benzaldehyde, it is
possible to achieve asymmetric induction (Scheme 1).
Introduction
Alkyllithium reagents are extensively used in organic syn-
thesis and several different chemical transformations are
achieved by the use of these reagents.[1] Possibly the most
important is the formation of C C bonds, which can be
achieved either through lithiation ± substitution, direct alkyl-
ation by an alkyllithium compound, or by aldol reactions.[2]
However, only a few examples have been reported regarding
the use of alkyllithium reagents in stereoselective C C bond
formation. The asymmetric inducers often contain lithium
amide or lithium alkoxide functionalities and one or several
stereogenic centers. Unfortunately, many of these compounds
are of limited synthetic application, since their preparation
requires substantial synthetic effort.[3, 4] Furthermore, only a
few of these chiral ligands or auxiliaries gave products with
enantiomeric excesses exceeding 60%.[5]
O
H
OH
*
2 nBuLi
RR*NH
RR*NLi/nBuLi
H
mixed complex
1-phenyl-1-pentanol
chiral amine
Scheme 1. The use of chiral lithium amides for asymmetric alkylation.
Hogeveen and Eleveld reported the asymmetric addition of
n-butyllithium (nBuLi) to benzaldehyde in the presence of the
chiral lithium amide base Li-1 in 1984. In diethyl ether (DEE)
at 1168C, they isolated (S)-1-phenyl-1-pentanol in 74% ee.
A 1:1 solvent mixture of diethyl ether and dimethoxy methane
(DMM) was reported to give an enantioselectivity of 90% in
the same reaction. When THF or toluene was used as the
solvent instead of DEE, the optical yields dropped to 68%
and 19%, respectively. This result shows the strong solvent
dependence of these reactions. The optimum ratio of benzal-
dehyde, nBuLi and 1 was found to be 1:6.7:4.
Because organolithium compounds are known to form
aggregates in solution, it is believed that chiral mixed
complexes are formed between nonchiral alkyllithiums and
[a] Prof. Ö. Davidsson, Dr. P. I. Arvidsson, Dr. G Hilmersson
Organic Chemistry, Department of Chemistry, Göteborg University
SE-412 96 Göteborg (Sweden)
Only a few investigations of chiral lithium amide/alkyllithi-
um mixed complexes have been reported.[6] In a previous study,
Fax: (‡46)317723840
Email: ojje@oc.chalmers.se
2348
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0508-2348 $ 17.50+.50/0
Chem. Eur. J. 1999, 5, No. 8

2348±2355
OMe
OMe
OMe
Me
Li
N
Me
N
Li
N
Li
N
Li
Me
N
Li
Li-4
Li-1
Li-2
Li-3
OMe
OMe
N
N
Me
N
Li
N
Li
Me
N
N
Li
Li
Li-5
Li-6
Li-7
Li-8
we showed that nBuLi is present in the above reaction mixture
at 908C, both as homoaggregated tetramers and as the mixed
aggregate Li-1/nBuLi. Based on these results, we suggested
that the mixed complex Li-1/nBuLi is the reactive species
responsible for the reported asymmetric alkylation reaction.
Homotetramers of nBuLi readily react with benzaldehyde
to yield a racemic mixture of the addition product.[7] This is
also clearly reflected in the choice of optimum experimental
conditions, as the molar ratios of nBuLi and chiral lithium
amide are critical for high stereoselectivity in the reaction.
The large solvent dependence might be explained by the
solvent-dependent equilibrium between homoaggregated Li-
1, homoaggregated nBuLi, and Li-1/nBuLi.
additional resonance, indicative of either a monomer or a
symmetrical internally coordinated dimer (structures a or b in
Figure 1). 15N 6Li NMR coupling studies performed with
[15N,6Li]-labeled Li-4 in our laboratory have shown that Li-4,
and probably also Li-8, forms symmetrical internally coordi-
nated dimers (structure b in Figure 1) in DEE solution. Thus,
we conclude that monomers are not observed for any of these
chelating lithium amides in DEE solution.
X
R1
R1
Li
N
R2
R2
N
N
R2
Li
We believe that in order to control the stereoselectivity, the
underlying mechanisms for the reactions and the induction of
asymmetry must be well understood. Knowledge about the
structure of the initial complexes and their dynamics in
solution are therefore crucial. We report here our systematic
search for new chiral inducers based upon NMR spectroscopy
and enantiomeric product distributions in the asymmetric
alkylation of benzaldehyde. The structure, dynamics, reac-
tivity, and selectivity of mixed dimers formed from the chiral
lithium amides, obtained from the amines 1 to 8, and
alkyllithium reagents were investigated.
Li
R1
X
X
a
b
R1
R1
Li
R2
R2
N
N
X
Li
X
c
Results and Discussion
Figure 1. a) Monomer. b) Symmetrical internally coordinated dimer.
c) Unsymmetrical internally coordinated dimer.
The structure of lithium amide/alkyllithium complexes in
solution: The chiral lithium amide Li-1 has already been
extensively studied by NMR spectroscopy.[8] The addition of
one equivalent of nBu6Li to the amines 1, 2, 4, 5, 6, 7, and 8,
respectively, results in the formation of a single species
according to 1H and 13C NMR spectra at 808C in DEE. The
dilithiated species Li-3 may be prepared from Li-1 and excess
nBu6Li, as previously described.[8] The 6Li NMR spectra of the
amides displays two signals in 1:1 ratio, one significantly
broader than the other; this indicates a lower coordination
number at this lithium.[9] Furthermore, the amides Li-2 to Li-8
are all structural analogues of Li-1 and exhibit similar 1H, 13C,
The formation of new species upon the addition of nBuLi to
the respective lithium amides is apparent according to NMR.
In addition to the signals for the chiral lithium amide dimers,
6
the Li NMR spectra show new signals, one for tetrameric
nBuLi at d ˆ 1.9 and another two (in a 1:1 ratio) that originate
from a mixed complex between a chiral lithium amide and
nBuLi. The 13C NMR spectra display two sets of resonances
for the chiral lithium amides and two sets of resonances for
nBuLi. This indicates the presence of tetrameric nBuLi,
lithium amide dimers, and a mixed complex between nBuLi
and the chiral lithium amide. The mixed complexes are all
dimers, since the 13C NMR spectrum displays quintets
1J(13C,6Li) ꢀ 8 Hz for the a carbons of the complexed nBuLi.
6
and Li NMR spectra. Therefore, it is likely that these chiral
lithium amides all form C2-symmetric dimers with unsym-
metrical internal coordination in DEE solution (structure c in
Figure 1). However, the 6Li NMR of Li-4 and Li-8 displays an
6
In the Li NMR spectrum of the homodimer Li-2 with an
additional 1.1 equivalents of nBuLi, the mixed dimer of Li-2/
Chem. Eur. J. 1999, 5, No. 8
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0508-2349 $ 17.50+.50/0
2349

Ö. Davidsson et al.
FULL PAPER
nBuLi is the predominant species; only traces of the
tetrameric nBuLi and homodimers of Li-2 are observed. This
indicates a larger preference for Li-2 to form mixed dimers
with nBuLi compared with Li-1. The lithium amide Li-6
showed intermediate behavior towards mixed complex for-
mation; the predominant species is the mixed dimer Li-6/
nBuLi, coexisting with small amounts of homodimers of Li-6
and homotetramers of nBuLi (Figure 2).
group in the alkyllithium important for the formation of
mixed complexes with lithium amides? The tendency for Li-1
to form mixed complexes with alkyllithiums of differing
bulkiness was investigated. A series of NMR experiments was
performed starting with the alkyllithium reagent with the least
steric hindrance, MeLi, followed by the more congested
alkyllithium compounds nBuLi, sBuLi, and tBuLi. In DEE
solution at 808C, homoaggregates of MeLi, nBuLi, and
sBuLi are predominantly tetrameric, whereas for tBuLi they
are predominantly dimeric. The equilibrium complexation
constants were determined from the relative 6Li NMR
Table 2. The equilibrium constants between alkyllithium (R Li)x aggre-
gates, (Li-1)2 dimers, and mixed dimers R-Li/Li-1.[a]
1
1
Equilibrium
K [moll
]
dG [kJmol ]
(MeLi)4 ‡ 2(Li-1)2„K 4(Li-1)/MeLi
(nBuLi)4 ‡ 2(Li-1)2„K 4(Li-1)/nBuLi
(sBuLi)4 ‡ 2(Li-1)2„K 4(Li-1)/sBuLi
(tBuLi)2 ‡ (Li-1)2„K 2(Li-1)/tBuLi
0.001 Æ 0.0009
1.22 Æ 0.41
150 Æ 70
11
0.32
8.0
1.5
0.4 Æ 0.3[b]
[a] Determined from the relative 6Li NMR intensities. [b] This equilibrium
constant has no units.
intensities at 808C (Table 2). From the equilibrium con-
stants in Table 2, it is clear that sBuLi is superior to the other
alkyllithium reagents in the formation of mixed dimers with
the chiral lithium amide Li-1.
The equilibrium constant between homo- and heterocom-
plexes is determined by the steric requirements of the
substituents on the a carbon. The introduction of branching
on the carbanion carbon, as found in sBuLi, adds substantial
steric requirements for the carbanion close to the lithium
atoms, which disfavors sBuLi homotetramers compared with
the Li-1/sBuLi mixed dimer. However, tBuLi is only sparingly
found in these mixed complexes. We suggest that this
unpredictable behavior is due to the fact that tBuLi is
homodimeric in DEE. This will result in an equilibrium
between homodimers and mixed dimers. This is in contrast to
the previously used alkyllithiums where there is an equili-
brium between homotetramers and mixed dimers. Thus, the
interactions are different and comparisons of the equilibrium
constants are, therefore, complex.
Figure 2. The 6Li NMR spectra ( 908C) of mixtures of nBuLi with (Li-
1)2, (Li-2)2, and (Li-6)2 showing the formation of mixed complexes.
The equilibrium constants between the lithium amide
homodimers, (nBuLi)4, and lithium amide/nBuLi mixed
dimers were estimated from 6Li NMR intensities and the
known concentrations of added nBuLi and chiral amine at
908C (Table 1).
Table 1. Equilibrium constants for formation of mixed nBuLi/lithium amide
dimers from tetrameric nBuLi and dimers of the various lithium amides.[a]
1
Equilibrium
K [moll
]
(nBuLi)4 ‡ 2(Li-1)2„K 4(Li-1)/nBuLi
(nBuLi)4 ‡ 2(Li-2)2„K 4(Li-2)/nBuLi
(nBuLi)4 ‡ 2(Li-6)2„K 4(Li-6)/nBuLi
(nBuLi)4 ‡ 2(Li-8)2„K 4(Li-8)/nBuLi
4
10000
800
Diastereotopic a protons on nBuLi and diasteromeric sBuLi
complexes: The a protons of nBuLi in the mixed complexes
Li-2/nBuLi (Figure 3) and Li-6/nBuLi display different chem-
ical shifts. Such large differences in chemical shifts are not
observed with Li-1/nBuLi, although this complex also gives
high enantioselectivity in the alkylation of benzaldehyde. We
interpret this as a rigidification of the mixed complexes Li-2/
nBuLi and Li-6/nBuLi relative to Li-1/nBuLi, which reveals
the diastereotopicity of the two methylene a protons on
nBuLi due to nearby inherent chirality and/or slow rotation
about the Ca Cb bond in nBuLi.
0.14
[a] Determined from the relative 6Li NMR intensities.
The ability of the various chiral lithium amides to complex
nBuLi is evidently very different. A reason for this could be
differences in steric interactions within the lithium amide/
nBuLi complexes and the corresponding lithium amide
homodimers. Compound (Li-8)2 with intramolecular pyrro-
lidine coordination, showed only a weak tendency to coor-
dinate nBuLi.
The 13C NMR spectrum of the Li-1/sBuLi complex showed
the presence of two Li-1/sBuLi complexes in a 2:1 ratio at low
temperatures. This indicates a difference in the stability of the
two diastereomeric Li-1/sBuLi complexes, and corresponds to
1
Difference in complexation ability of lithium amides towards
different alkyllithium compounds: Is the size of the alkyl
a free energy difference of ꢀ1.7 kJmol (Scheme 2 and
Figure 4).
2350
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0508-2350 $ 17.50+.50/0
Chem. Eur. J. 1999, 5, No. 8

Chiral Lithium Amides
2348±2355
Solvent dependence upon complex formation: Addition of
THF to a DEE solution of the mixed complex Li-1/nBuLi
showed that the equilibrium between homo and hetero
complexes is solvent dependent (Scheme 3). The addition of
small amounts of THF to (Li-1)2 in DEE results in dramatic
changes in the 6Li NMR chemical shifts.[9] The 6Li NMR
spectra in Figure 5 shows the effect of THF addition to the
mixture of (Li-1)2, Li-1/nBuLi, and (nBuLi)4 in DEE. In
agreement with our previous observations, addition of 5% THF
6
Figure 3. 1H, 13C, and Li NMR spectra of Li-2/nBuLi.
O
O
Slow (T<-35°C)
Li
Li
N
N
Li
Li
Et2O
Scheme 2. The two diastereomeric complexes of Li-1/sBuLi.
Figure 5. The 6Li NMR spectra of Li-1/nBuLi in DEE with increasing
concentration of THF.
Figure 4. 13C NMR spectra of selected carbons from the two diastereo-
meric complexes, Li-1/sBuLi, showing the intensity ratio 2:1.
causes the DEE-solvated (Li-1)2 (d ˆ 2.72 and 2.90) to
disappear, while the THF-solvated (Li-1)2 (d ˆ 2.70 and
3.14) is formed together with the THF-solvated monomer
(d ˆ 2.28). The signals for Li-1/nBuLi at d ˆ 2.40 and 3.68
However, the activation energy for racemization, estimated
from the coalescence of the
NMR signals at
358C, is
(THF)n
O
O
N
O
N
Li
Li
H
H
1
K
Li
much larger (ꢀ50 kJmol ).
We ascribe this behavior of the
complexed sBuLi to the inher-
ent chirality of the environment.
N
+
n•THF
+
+
(nBuLi)4• 4THF
H
H
Li
(DEE)n
Scheme 3. The solvent dependence of the equilibrium between homo- and heterocomplexes.
Li
(THF)n
Chem. Eur. J. 1999, 5, No. 8
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0508-2351 $ 17.50+.50/0
2351

Ö. Davidsson et al.
FULL PAPER
6
were also affected by THF; the most downfield Li NMR
signal at d ˆ 3.68 was observed to shift slightly further
downfield, possibly caused by the replacement of one DEE
molecule by one THF molecule in the solvated Li-1/nBuLi
mixed complex.
At THF concentrations above 8%, the 6Li signals from the
THF-monosolvated Li-1/nBuLi shifted upfield towards that
of Li-1/nBuLi in pure THF (Figure 5). At 55% THF in DEE,
no homodimers of (Li-1)2 were observed. The most downfield
6Li resonance from Li-1/nBuLi (originally at d ˆ 3.68) was
observed at ꢀ0.5 ppm upfield (now at d ˆ 3.14). These
changes in the chemical shifts are likely to be caused by an
increase in the coordination number for one of the lithium
atoms in Li-1/nBuLi. This signal also showed a large temper-
ature dependence in the 6Li NMR spectra. Upon lowering the
6
temperature from 708C to 1108C, the Li NMR signal at
d ˆ 3.14 shifted upfield to d ˆ 2.42. The chemical shift for this
6
6Li NMR signal seems to merge with the Li NMR chemical
6
shift of the tetra-coordinated lithium in Li-1. The second Li
NMR signal at d ˆ 2.40 from Li-1/nBuLi in DEE showed only
small THF concentration-dependent variations, which implies
a preserved coordination sphere around this lithium atom.
This indicates that the lithium is solvated by the oxygen in
the methoxy group and that this coordination also persists in
THF.
Addition of THF to a mixture of Li-8 and nBuLi in DEE
increases the fraction of mixed dimers, because of the increase
in the complexation constant in THF. The unsymmetrical
internally coordinated dimer (Li-8)2 (structure c in Figure 1)
decreased upon formation of the symmetrical internally
coordinated dimer (structure b in Figure 1). Mixed dimers
Li-8/nBuLi are strongly favored in solutions of DEE contain-
ing traces of THF, as compared with a pure solution of DEE.
Fast ligand exchange (DEE and THF) was observed for Li-
1/nBuLi at all accessible temperatures (down to 1108C).
Addition of TMEDA to a DEE solution of Li-1/nBuLi, (Li-
1)2, and (nBuLi)4 resulted in a large fraction of TMEDA-
solvated monomers of Li-1, namely, Li-1/TMEDA and
TMEDA-solvated nBuLi dimers (Figure 6).
Figure 6. The 6Li NMR spectra of Li-1/nBuLi in DEE with increasing
concentrations of TMEDA.
effect of DMM on the equilibrium between Li-1, nBuLi, and
Li-1/nBuLi. However, we did not observe any structural
changes in the 6Li NMR spectra for Li-1/nBuLi upon addition
of 50% (v/v) DMM to the DEE solution.
No TMEDA-solvated Li-1/nBuLi mixed complexes were
obtained. Similar results were found upon the addition of
THF or TMEDA, respectively, to the other mixed-dimer
complexes in DEE. This shows that the coordination at the
most solvent-sensitive lithium cation in Li-1/nBuLi has a
lower energy barrier to coordination. Structurally, this implies
that the internal methoxy-oxygen coordination has to be very
strong, as it persists even in the presence of a strong polar
solvent, such as THF. Tentatively, this suggests that solvents,
as well as substrates, initially coordinate to the nonmethoxy-
coordinated lithium cation. A further consequence of the
strong internal methoxy coordination is that the substituents
at the methylene carbons between the amide nitrogen and the
methoxy group will be locked in a well-defined spatial
arrangement. Consequently, this will result in favorable and
unfavorable diastereomeric transition states upon substrate
complex formation.
However, the addition of DMM to a DEE solution of
nBuLi and Li-6 resulted in the disappearance of the signal
from the unsymmetrical internally coordinated dimer (Li-6)2
and a single 6Li NMR resonance was observed instead (either
from a symmetrical internally coordinated dimer or monomer
of Li-6). The signals from Li-6/nBuLi showed no indication of
specific solvation by DMM. This indicates that DMM only has
a marginal effect on the above equilibrium and structures.
Nucleophilic asymmetric addition of nBuLi to benzaldehyde:
Most of the asymmetric alkylation reactions were performed
in DEE solutions. This solvent does not give the highest
possible enantioselectivity in the asymmetric alkylation
reactions; however, it did enable us to use the information
obtained from the NMR studies of the reaction mixtures to
partially explain the selectivity in the reactions.
The reported improved stereoselectivity obtained in the
asymmetric alkylation of benzaldehyde in a 1:1 solvent
mixture of DEE and DMM encouraged us to investigate the
In order to test our method, we reproduced the experiment
of Hogeveen and Eleveld;[3a] however, we used GC on a chiral
stationary phase, instead of optical rotation, to accurately
2352
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0508-2352 $ 17.50+.50/0
Chem. Eur. J. 1999, 5, No. 8

Chiral Lithium Amides
2348±2355
determine the productsꢀ ee. The chiral lithium amides (Li-1
to Li-8) were generated by the addition of nBuLi (1.45 equiv)
to the respective amine precursors (1.0 equiv). After 30 mi-
evidently, not very large for the different conformations and a
net cancellation of the different effects occurs.
Hogeveen and Eleveld reported in their paper that a 1:1
solvent mixture of DMM/DEE in the alkylation reaction
results in a 90% ee of the chiral (S)-alcohol. However, after
repeating this experiment several times we conclude that the
(S)-alcohol is formed at best in 72% ee, indicating that DMM
has a limited effect, not only on the equilibrium and structure,
but also on the stereoselectivity in the addition reaction.
Hogeveen and Eleveld determined the ee from the optical
rotation, whereas we have determined the ee from GC.
Surprisingly, the addition of DMM to the reaction performed
with Li-6 as a chiral inducer resulted in a large increase in
stereoselectivity; (S)-1-phenyl-1-pentanol was formed in
91% ee. The reason for this improved stereoselectivity was
not determined unambiguously; however, it is probably an
effect of increased solvation of the transition state with
DMM.
nutes, the mixtures were cooled down to
1168C and
benzaldehyde (0.25 equiv) was added. The reaction mixture
was quenched at 1168C with MeOH and analyzed by GC.
We obtained the product (S)-1-phenyl-1-pentanol in ꢁ82%
ee (Table 3).
Table 3. Enantiomeric excess of 1-phenyl-1-pentanol, obtained
by asymmetric addition of nBuLi to benzaldehyde in the
presence of the chiral lithium amides Li-1 to Li-8.[a]
Amide
R1
R2
X
ee% (GC)
Me
Ph
OMe
72 (S)
Li-1
Ph
OMe
75 (S)
Li-2
Me
Li
It should also be mentioned that the alkoxide product may
affect the nBuLi aggregation. A study by Alberts and
Wynberg has shown that mixed chiral alkoxide/EtLi com-
plexes are capable of stereoselective autoinduction.[10] How-
ever, since the bidentate chiral lithium amides in this study are
better complexation agents for nBuLi than monodentate
alkoxides, stereoselective autoinduction is likely to be of
minor importance.
Ph
Ph
OMe
N
8 (S)
7 (R)
Li-3
Li-4
Me
N
ipropyl
Ph
Li-5
Li-6
Li-7
Me
ipropyl
Me
7 (R)
82 (S)
2 (S)
OMe
OMe
Ph
ipropyl
N
Li-8
Ph
26 (S)
Conclusions
[a] The following molar ratios were used: lithium amide
(1 equiv), nBuLi (0.45 equiv), and benzaldehyde (0.25 equiv).
Mixed 1:1 complexes between nBuLi and chiral lithium
amides are formed in solution, as shown by low-temperature
NMR spectroscopy. The equilibrium constants for these
complexes were determined by 6Li NMR spectroscopy. These
findings allowed a systematic investigation of the factors
controlling the stereoselectivity in the addition of nBuLi to
benzaldehyde. Addition of n-butyllithium to benzaldehyde in
the presence of Li-1 gave the product (S)-1-phenyl-1-pentanol
in 72% ee. From rational design, partially based on multi-
nuclear NMR studies of mixtures of alkyllithium and lithium
amides, the optimum chiral lithium amide Li-6 was synthe-
sized. The use of this new ligand for the asymmetric alkylation
of benzaldehyde gave the same product in 91% ee (GC).
Thus, the optimum chiral inducer in this system is Li-6/nBuLi.
We are currently investigating other asymmetric alkylation
reactions with Li-6/nBuLi.[11]
The reactions were carried out at
1168C in DEE. See
Figure 1 for definition of the substituents R1, R2, and X.
From the results in Table 3, some important structural
factors in the lithium amide, which controls the enantio-
selectivity of the alkylation reactions, can be identified. The
steric requirements of the R1 substituent are crucial for
the reaction, that is, the stereoselectivity is dependent on a
large bulky group at R1, a methyl group is too small (see the
entries with compounds Li-4, Li-5, and Li-7 in Table 3).
Sterically demanding R1-groups, such as those in Li-1, Li-2,
and Li-6, gave high enantioselectivities. Furthermore, com-
pound Li-6, in which one of the chiral centers of Li-1 has
been removed, did not result in a decrease in stereoselec-
tivity; instead, the product ee increased to 82%. This
observation shows that the asymmetric induction is con-
trolled by the chiral center between the R2 and the chelating
group.
Experimental Section
General: All glassware used in the syntheses was dried overnight in an oven
(1208C) when necessary. Glassware and syringes used for the NMR studies
and alkylation reactions were dried at 508C in a vacuum oven before
transfer into a glove box (Mecaplex GB80 equipped with a gas-purification
system to remove oxygen and moisture) and stored under a nitrogen
atmosphere. Typical moisture content was <0.5 ppm. Most manipulations
for the alkylation reactions were carried out in the glove box by means
of gas-tight syringes. Ethereal solvents, distilled under nitrogen from
sodium and benzophenone, were kept over 4 Š molecular sieves in
septum-sealed flasks inside the glove box. The concentrations of commer-
cially available nBuLi solutions (nBuLi, 1.6 or 2.5m solution in hexanes,
The enantioselectivity decreased and the configuration of
the 1-phenyl-1-pentanol was reversed when amides Li-4 and
Li-5 were used. Obviously, the pyrrolidine group forces the
benzaldehyde into a different geometry in the transition state.
Two factors are important, as seen for Li-8. The large iso-
propyl group in R1 would imply a good selectivity for the (S)-
alcohol; however, this effect is biased by the large pyrrolidine
group, which favors a transition state that results in the (R)-
alcohol. The effects controlling the enantioselectivities are,
Chem. Eur. J. 1999, 5, No. 8
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0508-2353 $ 17.50+.50/0
2353

Ö. Davidsson et al.
FULL PAPER
Aldrich) were determined by titration with biphenylmethanol.[12] The
enantiomeric purity of the synthesized chiral amines were verified to be
>95% ee, based on 1H NMR in the presence of (R)-1-phenyl-2,2,2-
trifluoroethanol.[13]
nitrogen inlet. The dispersing oil was removed by successive washing with
hexane. Dry THF (60 mL) and methyl iodide (5.1 mL, 81.8 mmol) was
added.
A
suspension of N-formyl-2-amino-2-phenylethanol (4.5 g,
27.3 mmol) in THF was added, and the resulting mixture was refluxed for
6 h. The mixture now had a white appearance and contained a lot of solid
material. Saturated Na2SO4 was added to the suspension, which dissolved
after the addition of water. The layers were separated, and the aqueous
layer was extracted with EtOAc (4 Â 150 mL). The combined organic
layers were dried (Na2SO4), and the solvent evaporated under a vacuum.
The resulting oil, which contained a mixture of mono- and dimethylated N-
formyl-2-amino-2-phenylethanol, was purified by flash chromatography
[silica gel, EtOAc/hexane 8:2; Rf ˆ 0.54 (monomethylated) and 0.25
(dimethylated)]. Yield of N-formyl-N-methyl-1-phenyl-2-methoxyethan-
amine: 3.6 g (69%).
Routine 1H and 13C 1D-NMR spectra were recorded on a Varian Unity
400 MHz instrument. Optical rotations were measured on
a Perkin
Elmer341LC polarimeter. Melting points were measured on a BüchiB-
545 melting point apparatus. Chromatographic analyses were carried out
on a Varian Star 3400CX gas chromatograph. All GC analyses were run on
a chiral stationary phase column (CP-Chirasil-DEXCB, 25 m, 0.32 mm)
from Chrompack. All analyses were performed at 1358C (injector: 2258C;
detector: 2508C) with He (2 mLmin 1) as the carrier gas. Mass spectra
(MS) were recorded on a Varian Saturn2000 GC-MS/MS operating in the
electronic ionization (EI) or chemical ionization (CI) mode. Methane was
used as the reagent gas for CI operation. The GC column connected to the
mass analyzer was a DB-5MS (J and W Scientific).
Attempts to hydrolyze the amide with KOH (101 %) failed, even after long
reaction times (>12 h). Instead, the N-formyl-N-methyl-1-phenyl-2-
methoxyethanamine (3.6 g, 18.7 mmol) was refluxed in a mixture of EtOH
(10 mL) and KOH (20 mL, 50% aqueous solution) for 12 h. The reaction
mixture was cooled and then extracted with DEE (3 Â 100 mL). The
combined ether extracts were dried (Na2SO4), and evaporation of the
solvent under vacuum gave an orange oil, which was purified by distillation
(b.p. 458C, oil bath 738C, 2.9 Â 10 2 mbar), to give (R)-N-methyl-1-phenyl-
Preparations: The following compounds were synthesized according to
literature methods: (R)-N-2-methoxy-1-phenyl-(S)-a-methylbenzylamine
(1),[10] (R)-N-2-methoxy-1-phenyl-(R)-a-isopropylbenzylamine (2),[10] (R)-
N-methyl-1-phenyl-2-(1-pyrrolidinyl)ethanamine (4),[14] and (R)-N-meth-
yl-1-isopropyl-2-(1-pyrrolidinyl)ethanamine (5).[16]
2-methoxyethanamine as a clear oil. Yield: 1.76 g (62%); [a]D23
ˆ
97.5
(R)-N-Isopropyl-O-methylphenylglycinol (6):[15] A solution of (R)-phenyl-
glycinol (10.1 g, 73.8 mmol), acetone (5.9 mL, 80 mmol), and a catalytic
amount of p-toluenesulfonic acid in benzene (150 mL) was refluxed for
3 days, while water was removed by means of a Deans ± Stark trap. Upon
cooling, the reaction mixture was washed with Na2CO3 (100 mL, 5%
aqueous solution) and brine (100 mL), dried (MgSO4), and concentrated
under vacuum to yield the imine as an oil, which slowly crystallized into
pale yellow crystals. Yield: 13.1 g (100%).
(c ˆ 1.2 in CHCl3); 1H NMR (400 MHz, CDCl3): d ˆ 1.95 (br, 1H), 2.28 (s,
3H), 3.36 (s, 3H), 3.38 ± 3.48 (m, 2H), 3.75 (dd, 1H), 7.22 ± 7.37 (m, 5H);
13C NMR (100 MHz, CDCl3): d ˆ 34.6, 59.9, 65.0, 77.9, 127.6, 127.7, 128.5,
‡
140.5; MS (CI): m/z: 166 [M ‡1].
(R)-N-Isopropyl-1-phenyl-2-(1-pyrrolidinyl)ethanamine (8): This com-
pound was prepared from (R)-styrene oxide following the general method
of OꢀBrien et al:[17] (R)-styrene oxide (1.5 mL, 13.1 mmol) and pyrrolidine
(1.8 mL, 21.5 mmol) dissolved in EtOH (45 mL, 95%) was refluxed for 3 h.
The solvent was removed under vacuum and the flask, which contained a
slightly beige oil, was put under high vacuum (5 Â 10 2 mbar). Light beige
crystals had formed after 10 min. These crystals were dried for 1 h
(vacuum), before introduction of a dry nitrogen atmosphere. Dry DEE
(60 mL) and triethylamine (5.5 mL, 39.6 mmol) were added and the flask
was cooled to 08C. Methane sulfonyl chloride (1.23 mL, 15.9 mmol) was
added dropwise over a period of 40 min, during which time the suspension
turned yellow. Triethylamine (3.7 mL, 26.4 mmol) was added, and the
mixture was allowed to warm to room temperature before addition of
2-aminopropane (12.3 mL, 144 mmol) and water (7.5 mL). The two-phase
reaction mixture was vigorously stirred for 16 h. The mixture was trans-
ferred to a separating funnel, and the layers separated. The yellow aqueous
layer was extracted with DEE (3 Â 100 mL). The combined organic extracts
were washed with aqueous sodium hydrogen carbonate (5%, 100 mL) and
water (100 mL). The mixture was dried (Na2SO4), and evaporation of the
solvent under reduced pressure followed by Kügelrohr distillation (1008C,
Without further purification, the imine crystals (12.5 g, 71.0 mmol) were
dissolved in dry THF (200 mL) and catalytically hydrogenated for 12 h in a
Parr apparatus with H2 (4 atm) and with Pd/C (10%) as the catalyst. The
mixture was filtered through Celite, dried (MgSO4), and the solvent
evaporated under vacuum to give the amino alcohol as white acicular
crystals. Yield: 11.7 g (93%).
Sodium hydride (3.1 g, 55% dispersion in oil, 71 mmol) was added to a
round-bottomed flask equipped with a magnet, a dropping funnel, a
thermometer, and a nitrogen inlet. The dispersing oil was removed by
successive washing with hexane, and dry THF (20 mL) was added. The
amino alcohol (11.5 g, 64.2 mmol) dissolved in dry THF (100 mL) was
added over a period of 15 min to the hydride suspension. The mixture was
then heated to 408C for 30 min. Upon cooling to 58C, methyl iodide
(4.0 mL, 64.2 mmol) dissolved in dry THF (40 mL) was added to the
gelatinous, pale greenish-yellow solution. The reaction mixture was kept at
room temperature overnight. Slow addition of water (20 mL) gave a yellow
solution, which was extracted with DEE (3 Â 150 mL). The combined
organic extracts were washed with brine (150 mL), dried over MgSO4, and
evaporated under vacuum to yield an yellow oil. The oil was dissolved in
aqueous HCl (100 mL, 1m) and washed with DEE (150 mL). The aqueous
layer was made alkaline by addition of NaOH and was extracted with DEE
(3 Â 150 mL). The combined extracts were again dried and evaporated. The
resulting pale yellow oil was distilled at reduced pressure (658C, 1 Â
10 3 mbar) to give (R)-N-isopropyl-O-methylphenyl glycinol as a clear
7 Â 10 3 mbar) gave a very pale yellow oil. Yield: 2.18 g (72%); [a]2D3
ˆ
60.8 (c ˆ 1.2 in EtOH); 1H NMR (400 MHz, CDCl3): d ˆ 0.99 (d, 3H),
1.04 (b, 3H), 1.64 (br, 1H), 1.78 (m, 4H), 2.25 (dd, 1H), 2.42 (m, 2H), 2.60
(m, 3H), 2.80 (t, 1H), 3.85 (dd, 1H), 7.2 ± 7.4 (m, 5H); 13C NMR (100 MHz,
CDCl3): d ˆ 22.2, 23.8, 24.9, 46.2, 54.1, 59.4, 64.3, 127.1, 127.5, 128.4, 144.2;
‡
MS (CI): m/z: 233 [M ‡1].
Preparation of NMR samples: For the preparation of 6Li-labeled lithium
amides and details of the NMR sample preparations, see previously
published papers.[18]
oil. Yield: 9.83 g (76%); [a]D23
ˆ
79.1 (c ˆ 1.2 in CHCl3); 1H NMR
(400 MHz, CDCl3): d ˆ 0.99 (d, 3H), 1.04 (d, 3H), 1.62 (br, 1H), 2.64 (sept,
1H), 3.36 (s, 3H), 3.39 ± 3.47 (m, 2H), 4.02 (dd, 1H), 7.30 ± 7.38 (m, 5H);
13C NMR (100 MHz, CDCl3): d ˆ 22.2, 24.6, 45.9, 59.0, 60.1, 78.2, 127.4,
6
6
Quantitative Li NMR measurements: All quantitative Li NMR data for
the calculations of equilibrium constants were obtained from single
transient experiments on 6Li-labeled compounds. The signal-to-noise ratio
was typically ꢀ200 or better. The spectral widths were 2000 Hz.
‡
127.8, 128.5, 141.7; MS (CI): m/z: 194 [M ‡1].
(R)-N-Methyl-1-phenyl-2-methoxyethanamine (7): The method described
by Rossiter et. al.[16] for the synthesis of the enantiomer of this compound
was followed with some modifications: (R)-Phenylglycinol (5.0 g,
36.5 mmol) and ethyl formate (3.8 mL, 45.6 mmol) was refluxed for 18 h.
Evaporation of the excess ethyl formate under vacuum resulted in a clear
oil. This oil was stripped with DEE (10 mL) to afford white crystals, which
were filtered and dried under a vacuum to give N-formyl-2-amino-2-
phenylethanol. Yield: 5.7 g, (95%); m.p. 100.68C.
Asymmetric alkylation of benzaldehyde: A septum-sealed flask containing
a magnet, the amine (0.30 mmol), and dry DEE (1.6 mL) was assembled
inside the glove box. The flask was taken out of the box and directly fitted
with a dry argon inlet. The flask was cooled to 08C and nBuLi (0.435 mmol)
was added with a syringe. After stirring for 15 minutes at 08C, the flask was
moved to a DEE/N2(l) bath ( 1168C) and the temperature was allowed to
equilibrate at 1168C for 45 min. Benzaldehyde (0.075 mmol) was added
with a syringe. The reaction was quenched after 1 h by the addition of
methanol (0.5 mL). The crude mixture, which contained the alkylation
product, was analyzed by chiral stationary-phase GC.
Sodium hydride (3.3 g, 55% dispersion in oil, 76 mmol) was added to a
round-bottomed flask equipped with a magnet, a thermometer, and a
2354
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0508-2354 $ 17.50+.50/0
Chem. Eur. J. 1999, 5, No. 8

Chiral Lithium Amides
2348±2355
Angew. Chem. 1995, 107, 2328; Angew. Chem. Int. Ed. Engl. 1995, 34,
2158; i) H. Nozaki, T. Aratani, T. Toraya, R. Noyori, Tetrahedron Lett.
1971, 27, 905; j) M. Paetow, M. Kotthaus, M. Grehl, R. Frohlich, D.
Hoppe, Synlett 1994, 1034; k) S. Thayumanavan, S. Lee, C. Liu, P.
Beak, J. Am. Chem. Soc. 1994, 116, 9755; l) M. Mizuno, M. Kanai, A.
Iida, K. Tomioka, Tetrahedron: Asymmetry 1996, 7, 2483; m) K.
Tomioka, Synthesis 1990, 541; n) P. Sommerfeld, D. Hoppe, Synlett
1992, 764; o) W. Guarineri, M. Grehl, D. Hoppe, Angew. Chem. 1994,
106, 1815; Angew. Chem. Int. Ed. Engl. 1994, 33, 1734; p) H. Ahrens,
M. Paetow, D. Hoppe, Tetrahedron Lett. 1992, 33, 5327; q) S. Wu, S.
Lee, P. Beak, J. Am. Chem.Soc. 1996, 118, 715; r) P. Beak, S. T.
Kerrick, S. Wu, J. Chu, J. Am. Chem. Soc. 1994, 116, 3231; s) S. T.
Kerrick, P. Beak, J. Am. Chem. Soc. 1991, 113, 9708; t) D. Hoppe, A.
Carstens, T. Krämer, Angew. Chem. 1990, 102, 1455; Angew. Chem.
Int. Ed. Engl. 1990, 29, 1424; u) D. Hoppe, T. Hense, Angew. Chem.
1997, 109, 2376; Angew. Chem. Int. Ed. Engl. 1997, 36, 2282.
[5] a) S. Kanoh, H. Muramoto, K. Maeda, N. Kawaguchi, M. Motoi, H.
Suda, Bull. Chem. Soc. Jpn. 1988, 61, 2244; b) M. Ye, S. Logaraj, L. M.
Jackman, K. Hillegrass, K. A. Hirsh, A. M. Bollinger, A. L. Grosz,
Tetrahedron 1994, 50, 6109; c) A. Basu, P. Beak, J. Am. Chem. Soc.
1996, 118, 1575, d) A. Corruble, J.-Y. Valnot, J. Maddaluno, Y. Prigent,
D. Davoust, P. Duhamel, J. Am. Chem. Soc. 1997, 119, 10042; e) K.
Tomioka, M. Nakajima, K. Koga, Chem. Lett. 1987, 65; f) T. D. Inch,
G. J. Lewis, G. L. Sainsbury, D. J. Sellers, Tetrahedron Lett. 1969, 3657;
g) D. Seebach, A. K. Beck, S. Roggo, A. Wonnacott, Chem. Ber. 1985,
118, 3673; h) M. T. Reetz, T. Kükenhöhner, P. Weinig, Tetrahedron
Lett. 1986, 27, 5711.
Acknowledgments
We are grateful to the Swedish Natural Science Research Council (k-AA/
ku11723303) and the Carl Tryggers Research Fundation for financial
support.
[1] M. B. Smith, Organic Synthesis, McGraw-Hill, New York, 1994,
pp. 719; Houben-Weyl, Methoden der Organischer Chemie, Vol E19d,
Carbanionen (Ed.: M. Hanack), Thieme, Stuttgart 1993.
[2] a) R. Noyori, M. Kitamura, Angew. Chem. 1991, 103, 34; Angew.
Chem. Int. Ed. Engl. 1991, 30, 49; b) E. Juaristi, B. Weber, D. Seebach,
Tetrahedron 1994, 50, 6117; c) A. K. Beck, J. Hansen, T. Matt, T.
Mukhopadhyay, M. Simson, D. Seebach, Synthesis 1993, 1271.
[3] a) M. B. Eleveld, H. Hogeveen, Tetrahedron Lett. 1984, 25, 5187; b) T.
Mukaiyama, K. Soai, T. Sato, H. Shimizu, K. Suzuki, J. Am. Chem.
Soc. 1979, 101, 1455; c) L. Colombo, C. Gennari, G. Poli, C. Scolastico,
Tetrahedron 1982, 38, 2725; d) D. Seebach, H.-O. Kalinowski, B.
Bastani, G. Crass, H. Daum, H. Dörr, N. P. DuPreez, V. Ehrig, W.
Langer, C. Nüssler, H.-A. Oei, M. Schmidt, Helv. Chim. Acta 1977, 60,
301; e) J.-P. Mazaleyrat, D. J. Cram, J. Am. Chem. Soc. 1981, 103, 4585;
f) E. Kaufmann, P. von R. Schleyer, J. Am. Chem. Soc. 1985, 107, 5560;
g) K. Sorger, P. von R. Schleyer, R. Fleischer, D. Stalke, J. Am. Chem.
Soc. 1996, 118, 6924; h) K. Scharpwinkel, S. Matull, H. J. Schäfer,
Tetrahedron: Asymmetry 1996, 7, 2497; i) M. Ye, S. Logaraj, L. M.
Jackman, K. Hillegass, K. A. Hirsh, A. M. Bollinger, A. L. Grosz,
Tetrahedron 1994, 50, 6109; j) R. P. Davies, P. R. Raithby, R. Snaith,
Angew. Chem. 1997, 109, 1261; Angew. Chem. Int. Ed. Engl. 1997, 36,
1215; k) K. Ruthlandt-Senge, J. J. Ellison, R. J. Whemschulte, F.
Pauer, P. P. Power, J. Am. Chem. Soc. 1993, 115, 11353; l) P. Wijkens,
E. M. van Koten, M. D. Janssen, J. T. B. H. Jastrzebski, A. L. Spek, G.
van Koten, Angew. Chem. 1995, 107, 239; Angew. Chem. Int. Ed. Engl.
1995, 34, 219; m) L. M. Pratt, I. M. Khan, Tetrahedron: Asymmetry
1995, 6, 2165; n) M. Schlosser, S. Strunk, Tetrahedron Lett. 1984, 25,
741; o) L. Brandsma, H. D. Verkruijsse, C. Schade, P. von R. Schleyer,
J. Chem. Soc. Chem. Commun. 1986, 260; p) C. Schade, W. Bauer, P.
von R. Schleyer, J. Organomet. Chem. 1985, 295, C25; q) W. Bauer, L.
Lochmann, J. Am. Chem. Soc. 1992, 114, 7482; r) M. March, K. Harms,
L. Lochmann, G. Boche, Angew. Chem. 1990, 102, 334; Angew. Chem.
Int. Ed. Engl. 1990, 29, 308; s) P. G. Williard, G. J. Mac Ewen, J. Am.
Chem. Soc. 1989, 111, 7671; t) P. G. Williard, M. A. J. Nichols, J. Am.
Chem. Soc. 1991, 113, 9671.
[6] a) G. Hilmersson, Ö. Davidsson, J. Organomet. Chem. 1995, 489, 175;
b) P. G. Williard, C. Sun, J. Am. Chem. Soc. 1997, 119, 11693.
[7] J. F. McGarrity, C. A. Ogle, Z. Brich, H. R. Loosli, J. Am. Chem. Soc.
1985, 107, 1810.
Ê
[8] G. Hilmersson, P. I. Arvidsson, M. Hakansson, Ö. Davidsson, Organo-
metallics, 1997, 16, 3352, and unpublished results.
[9] G. Hilmersson, Ö. Davidsson, J. Org. Chem. 1995, 60, 7660.
[10] A. H. Alberts, H. Wynberg, J. Am. Chem. Soc. 1989, 111, 7265.
[11] P. I. Arvidsson, Ö. Davidsson, G. Hilmersson, Tetrahedron: Asymme-
try, 1999, 10, 527.
[12] E. Juaristi, A. Martinez-Richa, J. Org. Chem. 1983, 48, 2603.
[13] W. H. Pirkle, P. L. Rinaldi, J. Org. Chem. 1977, 42, 3217.
[14] D. Bhuniya, V. K. Singh, Synth. Comm. 1994, 24, 1475.
[15] An alternative route to this compound has been described in the
literature. a) R. Shirai, D. Sato, K. Aoki, M. Tanaka, H. Kawasaki, K.
Koga, Tetrahedron 1997, 53, 5963; b) A. Ando, T. Shioiri, Tetrahedron
1989, 45, 4969.
[16] B. E. Rossiter, M. Eguchi, G. Miao, N. M. Swingle, A. E. Hernandez,
D. Vickers, E. Fluckiger, R. G. Patterson, K. V. Reddy, Tetrahedron
1993, 49, 965.
[17] S. E. de Sousa, P. OꢀBrien, P. Poumellec, J. Chem. Soc. Perkin Trans. 1
1998, 1483.
[4] a) H. Nozaki, T. Aratani, T. Toraya, Tetrahedron Lett. 1968, 9, 4997;
b) D. Hoppe, F. Hintze, P. Tebben, Angew. Chem. 1990, 102, 1457;
Angew. Chem. Int. Ed. Engl. 1990, 29, 1422; c) P. Beak, H. Du, J. Am.
Chem. Soc. 1993, 115, 2516; d) D. J. Gallagher, S. T. Kerrick, P. Beak, J.
Am. Chem. Soc. 1992, 114, 5872; e) D. J. Gallagher, S. D. Wu, N. A.
Nikolic, P. Beak, J. Org. Chem. 1995, 60, 8148; f) J. Haller, T. Hense, D.
Hoppe, Synlett 1993, 726; g) D. Hoppe, F. Hintze, P. Tebben, M.
Paetow, H. Ahrens, J. Schwerdtfeger, P. Sommerfeld, J. Haller, W.
Guarnieri, S. Kolczewski, T. Hense, I. Hoppe, Pure Appl. Chem. 1994,
66, 1479; h) I. Hoppe, M. Marsch, K. Harms, G. Boche, D. Hoppe,
[18] See Refs. [6a], [8], and [9].
Received: December 7, 1998 [F1479]
Chem. Eur. J. 1999, 5, No. 8
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0508-2355 $ 17.50+.50/0
2355