Dynamically Controlled Lithiation of Chiral Aziridines
FULL PAPER
Table 4. Relative energies [kcalmolꢁ1] for lithiated oxazolinylaziridines
(R,R)-9-Li and (R,S)-9-Li.
Molecule
(R,S)-9-N-Li-inv
(R,S)-9-h3-N-Li
(R,S)-9-O-Li
(R,R)-9-N-Li-inv
(R,R)-9-h3-N-Li
(R,R)-9-O-Li
G MP2[a,b]
G M062X[a]
G
11.81
0.00
4.52
10.41
0.00
7.15
10.33
0.00
5.89
11.88
0.00
8.50
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
[a] Relative energies in kcalmolꢁ1. [b] G MP2//M062X
Scheme 4. Possible pathways for the regioselective a-lithiation of oxazoli-
nylaziridines.
(d.r.=83:17 in [D8]toluene) by dynamic NMR spectroscopic
N double bond of the oxazoline ring and could likely benefit
from extra stabilization by the phenyl ring of the N-substitu-
ent (Figure 5). This type of h3-coordination has been already
observed by FTIR and NMR spectroscopic analysis on a-
lithiated oxazolinyloxiranes in THF solution.[29] The extra
stability from the interaction with the phenyl group is not
possible in the invertomer (R,S)-9-N-Li-inv.
The O-chelated structure (R,S)-9-O-Li was also found as
a local minimum, but was 5.9 kcalmolꢁ1 less stable than
(R,S)-9-h3-N-Li.[30] Optimization on the nitrogen invertomer
revealed that the equilibrium geometry (R,S)-9-N-Li-inv
was almost 10.3 kcalmolꢁ1 less stable than (R,S)-9-h3-N-Li
and that the Li ion could be chelated by the lone pair of
experiments and line-shape analysis[24] (Figure 3) by finding
¼
DG
values for the conversion of major into minor of
(298)
¼
DG (298) =16.8 and 16.3 kcalmolꢁ1, respectively. These
values were in the range expected for this type of N-alkyl-
substitued aziridine[25] and were used as a reference for the
estimation of the nitrogen inversion barrier of aziridines
(R,R)-9 and (R,S)-9, in which the very low intensity (<5%)
1
of the H NMR signals of the minor invertomers encumbers
an accurate analysis of the inversion process by dynamic
NMR analysis.[26]
Calculations with the modern M06-2X functional and post
HF MP2 method[27] were subsequently performed on neutral
and lithiated oxazolinylaziridines (R,R)-9, (R,S)-9, (R,R)-9-
Li, and (R,S)-9-Li. The search for the equilibrium geometry
was run for both the invertomers of the neutral aziridines
(R,R)-9 and (R,S)-9. According to 1D selective NOESY ex-
periments in CDCl3, [D8]THF, and [D8]toluene, even the
calculations demonstrated that for both (R,R)-9 and (R,S)-9
the most stable invertomer sets the N-substituent and the
oxazolinyl ring trans each other. In fact, it was found that
aziridines (R,R)-9 and (R,S)-9 were more stable than the
corresponding nitrogen invertomers (R,R)-9-inv and (R,S)-
9-inv of about 7.3 kcalmolꢁ1, respectively (Figure 4). The
difference in energy between (R,S)-9 and (R,R)-9 was only
ꢁ
electrons on the aziridine nitrogen atom (Li30 N4=
1.89 ꢄ). Similar conclusions were obtained from the analysis
with the MP2 method (Table 4).
The computed structures obtained in the analysis of
(R,R)-9-Li showed again that an h3-coordinated intermedi-
ate (R,R)-9-h3-N-Li (0 kcalmolꢁ1 rel. energy) was more
stable than (R,R)-9-O-Li and (R,R)-9-N-Li-inv by about 8.5
and 11.9 kcalmolꢁ1, respectively (Figure 6). A comparison
between neutral and lithiated aziridines (R,S)-9 and (R,S)-9-
ꢁ
ꢁ
h3-N-Li reveals that the C2 C5 and C5 N6 bond lengths
are 1.48 and 1.28 ꢄ, respectively, in the neutral aziridine
versus 1.43 and 1.31 ꢄ, respectively, in the lithiated inter-
mediate. Similar bond lengths have also been computed for
(R,R)-9 and (R,R)-9-h3-N-Li (1.48 and 1.28 ꢄ in the neutral
aziridine versus 1.43 and 1.30 ꢄ in the lithiated intermedi-
ate). These values suggest that the aziridine carbon atoms
likely do not change their hybridization upon lithiation and
that the negative charge could not be delocalized into the
oxazolinyl ring, which is in contrast with lithiated aziridine-
carboxylate compounds in which an enolate-like structure
has been taken into consideration.[8,9]
¼
DG (298) =0.26 kcalmolꢁ1.
On the basis of the computed structures obtained for the
neutral aziridines (R,R)-9 and (R,S)-9 and the estimated
value of the inversion barrier, it seems reasonable to assume
that the deprotonation reaction could involve invertomer A
(Scheme 4).
Computational analysis on lithiated oxazolinylaziridines
(R,R)-9 and (R,S)-9: With the aim of explaining the config-
urational stability observed with this type of lithiated oxazo-
linylaziridines, which is in striking contrast with the bias to
epimerization of lithiated oxazolinyloxiranes, a computa-
tional analysis with the hybrid GGA M06-2X functional
with the 6-31+G(d) basis set[28] was undertaken.
Gas-phase calculations on diastereoisomeric lithiated azir-
idines (R,S)-9-Li and (R,R)-9-Li were performed taking into
consideration also the corresponding nitrogen invertomers
(Table 4). The computational analysis of (R,S)-9-Li gave
(R,S)-9-h3-N-Li as equilibrium geometry (relative energy=
0 kcalmolꢁ1), in which the Li atom is coordinated to the C=
By considering the optimized structures (R,S)-9-h3-N-Li
and (R,R)-9-h3-N-Li, their energy difference was
5.4 kcalmolꢁ1, the former being more stable than the latter.
This result contrasts with the experimental evidence in THF
in which lithiated aziridine (R,R)-9-Li proved to be more
thermally stable than (R,S)-9-Li. Such a different stability
could be justified because these computed structures do not
take into account aggregation phenomena.[31]
Nevertheless, to validate the computational results in the
gas-phase experimentally, the lithiation of aziridines (R,S)-9
and (R,R)-9 was performed in a less polar and non-coordi-
Chem. Eur. J. 2011, 17, 4992 – 5003
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4997