C1-symmetric versus C2-symmetric ligands in enantioselective copper-bis(oxazoline)-catalyzed cyclopropanation reactions

Article abstract of DOI:10.1002/chem.200700681

A thorough experimental and theoretical study of the enantioselective cyclopropanation of alkenes catalyzed by chiral bis(oxazoline)- and azabis(oxazoline)-copper complexes, which comprise a new family of ligands that lack C₂ symmetry, has been conducted. Surprisingly high enantioselectivities were observed with some of these ligands, which were rationalized on the basis of molecular modeling studies. The course of the asymmetric induction in connection with ligand symmetry and the implications for supported enantioselective catalysts are discussed.

Full text of DOI:10.1002/chem.200700681

DOI: 10.1002/chem.200700681
C1-Symmetric Versus C2-Symmetric Ligands in Enantioselective Copper–
Bis(oxazoline)-Catalyzed CycloACHTREpUNG ropanation Reactions
JosØ M Fraile,[a] JosØ I. García,*[a] Anja Gissibl,[b] JosØ A. Mayoral,[a] Elísabet Pires,[a]
Oliver Reiser,[b] Marta Roldµn,[a] and Isabel Villalba[a]
Abstract: A thorough experimental and theoretical study of the enantioselective
cyclopropanation of alkenes catalyzed by chiral bis(oxazoline)– and azabis(oxazo-
line)–copper complexes, which comprise a new family of ligands that lack C2 sym-
metry, has been conducted. Surprisingly high enantioselectivities were observed
with some of these ligands, which were rationalized on the basis of molecular mod-
eling studies. The course of the asymmetric induction in connection with ligand
symmetry and the implications for supported enantioselective catalysts are dis-
cussed.
Keywords: asymmetric catalysis
·
bis(oxazolines) computational
·
mechanistic models · copper · cyclo-
propanation
Introduction
Enantioselective homogeneous catalysis is one of the most
efficient ways of introducing asymmetry in organic synthe-
sis.[1] Although there are some significant exceptions,[2] C2-
symmetric ligands are usually preferred over C1-symmetric
(which are asymmetric in the sense that they lackany sym-
metry elements) ligands for catalytic enantioselective trans-
formations. Those C1-symmetric ligands that are successful
in catalytic applications are generally both electronically
and sterically asymmetric, for instance, salicylaldimines[2a]
and phosphinooxazolines (Figure 1).[2b] There are evident
advantages in using C2-symmetric ligands: Arguably most
importantly, fewer reaction channels are possible for the re-
action, which simplifies the prediction of chiral induction.
Furthermore, the synthesis of the ligands is often simpler.
Figure 1. Some examples of chiral ligands based on the oxazoline motif
with different degrees of electronic and steric asymmetry.
Oxazoline-based ligands, such as bis(oxazoline) (Box),
azabis(oxazoline) (azaBox) and pyridinebis(oxazoline)
(Pybox) ligands, have attracted much attention because they
have been successfully used in many different enantioselec-
tive organic reactions.[3] In the vast majority of cases all of
these ligands display C2 symmetry. Given that all of these li-
gands have two electronically and sterically equivalent coor-
dinating centers, there exists the possibility of modifying the
steric surroundings in the proximity of one of these centers,
thus leading to electronically equivalent, but sterically differ-
ent coordinating points. These ligands would be “halfway”
between the above-mentioned C1-symmetric ligands and the
usual C2-symmetric ligands. Some illustrative examples
based on the oxazoline motif are shown in Figure 1.
In general, in those cases in which C2-symmetric ligands
lead to good enantioselectivities, by using sterically nonequi-
valent but electronically equivalent (e.g., in the sense of a
close similarity of the coordinating groups), analogues re-
sults in a dramatic worsening of the results. Analogously,
the use of an asymmetric pyridine–oxazoline in the copper-
[a] Dr. J. M. Fraile, Prof. Dr. J. I. García, Prof. Dr. J. A. Mayoral,
Dr. E. Pires, M. Roldµn, I. Villalba
Departamento de Química Orgµnica, ICMA
CSIC-Universidad de Zaragoza and IUCH
Universidad de Zaragoza
Pedro Cerbuna 12
50009 Zaragoza (Spain)
Fax : (+34)976-762-077
E-mail: jig@unizar.es
[b] Dr. A. Gissibl, Prof. Dr. O. Reiser
Institut für Organische Chemie der Universität Regensburg
Universitätsstrasse 31, 93040 Regensburg (Germany)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
8830
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 8830 – 8839

FULL PAPER
catalyzed cyclopropanation re-
action of styrene with ethyl di-
azoacetate leads to virtually
racemic products.[4] Similar ob-
servations have also been de-
scribed for chiral unsymmetri-
cal 2,2’-bipyridyl ligands in the
satisfactory performance of this technique in relation to the
chemistry of transition metals,[9] particularly in the systems
studied in this work.[10] Full geometrical optimizations using
the 6-31G(d) basis set were carried out with the Gaussian 03
package.[11] Analytical frequencies were calculated at the
same level as that used for the geometry optimizations and
the nature of the stationary points was determined in each
case according to the appropriate number of negative eigen-
values of the Hessian matrix. Scaled frequencies were not
considered in full QM calculations because significant errors
in the calculated thermodynamic properties are not found at
this theoretical level.[12] Unless otherwise stated, only E0 +
ZPE energies (ZPE: zero-point energy) were used to discuss
the relative stabilities of the chemical structures considered.
Electronic energies, enthalpies, and Gibbs free energies of
the different conformations of all the structures considered
are available in the Supporting Information.
Figure 2. Structures of the
Nishiyama asymmetric pybox
ligands.
same
reaction.[5]
However,
there is at least one case in
which the use of sterically non-
equivalent ligands results in
enantioselectivities that are comparable to those obtained
with the corresponding C2-symmetric analogues, namely, the
so-called single-chiral Pybox ligands described by Nishiyama
et al. (Figure 2).[6]
When these ligands are used in the ruthenium-catalyzed
cyclopropanation reaction of styrene with alkyl diazoace-
tates (Scheme 1), very good enantioselectivities are obtained
for the trans-cyclopropanes (up to 94% ee).[6]
Experimental catalysis results: Figure 3 shows the structures
of the asymmetric ligands used in the catalytic experiments.
We have recently described a general method for the syn-
Scheme 1. A typical cyclopropanation reaction.
We have recently studied this system from a theoretical
point of view, and we have shown that the existence of two
highly disfavored reaction channels out of a possible eight is
sufficient to explain the unexpectedly high enantioselectivity
observed.[7]
Figure 3. Structures of the asymmetric Box and azaBox ligands used in
the catalytic experiments.
The aim of the workreported herein was to explore the
enantioselection mechanisms in connection with ligand sym-
metry, which bears in mind the fact that the possibility of
obtaining good enantioselectivities with electronically equiv-
alent, but sterically nonequivalent, ligands would open the
door to new strategies for supported chiral catalysts. For our
study we chose chiral bis(oxazolines) (Box), a class of ligand
that has been recognized as of the utmost importance in
asymmetric catalysis. To expand the scope of these ligands,
which have been applied almost exclusively in their C2-sym-
metric versions to date, a thorough experimental and theo-
retical study has been carried out by using a completely new
family of chiral C1-symmetric Box ligands.
thesis of C1-symmetric bis(oxazoline) (Box) ligands[13] in
connection with their possible use in supported asymmetric
catalysis. Moreover, a strategy to synthesize a C1-symmetric
azabis(oxazoline) (azaBox) ligand has previously been de-
scribed.[14] Following these strategies, the unsymmetrical li-
gands 5a–j were synthesized. For comparison with the usual
C2-symmetric Box ligands, we included Box ligands in which
one sterically demanding substituent was missing (5a–c,g),
two different substituents were present that gave pseudo-C2-
symmetric-type ligands 5d–f, and in which one side of the
Box ligand was sterically blocked by introducing an achiral
(5h) or chiral (5i,j) quaternary center. Some of the corre-
sponding C2-symmetric ligands have also been included in
the study and are labeled with an additional “s” letter, for
instance, 5as for PhBox, 5bs for IndanylBox, and 5cs for
tBuBox.
Results and Discussion
Computational methods: All QM calculations were carried
The benchmarkreaction used throughout this workwas
the cyclopropanation of styrene (2) with ethyl diazoacetate
out by using the B3LYP hybrid functional[8] because of the
Chem. Eur. J. 2007, 13, 8830 – 8839
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8831

J. I. García et al.
(3a, Scheme 1) catalyzed by 5–Cu
A
tion, the observed level of enantioselection would roughly
correspond to one quarter of the possible reaction channels
being disfavored and the rest being roughly equally favored
(Figure 4).
chloromethane. A 1:1 molar ratio of 2 and 3 was used in all
cases for the sake of comparison with previous mechanistic
studies, which represented more challenging reaction condi-
tions with respect to yields based on the diazo ester as the
alkene is usually employed in large excess (5 equiv or
more).
Table 1 summarizes the main results obtained in these cy-
clopropanation reactions. Some comparative results ob-
Table 1. Results of the cyclopropanation reaction of styrene with ethyl
diazoacetate catalyzed by chiral Box–CuOTf complexes.
Ligand
Yield [%][a]
trans/cis
ee (trans) [%][b]
ee (cis) [%][b]
5as
5a
5bs
5b
5cs
5c
5d
5e
5 f
5gs
5g
5h
5is
5i
33
42
69
52
72
58
74
73
79
82
65
42
72
64
69
68:32
71:29
60:40
69:31
71:29
68:32
67:33
64:36
72:28
73:27
73:27
71:29
60:40
69:31
70:30
60
20
85
33
94
29
84
83
82
92
23
85
À12
À8
48
51
8
81
25
91
8
79
75
69
84
9
68
À31
À24
31
Figure 4. A possible mechanism to explain the low, but significant, enan-
tioselectivities observed with monosubstituted bis(oxazoline) ligands.
Only one of four possible reaction channels is disfavored.
5j
In the case of C1-symmetric ligands with two substituted
oxazoline rings (5d, 5e, 5 f, and 5h) the enantioselectivities
observed were closer to those obtained with the C2-symmet-
ric ligands (>80% ee for the trans-cyclopropanes and 70–
80% ee in the case of the cis-cyclopropanes), even if the bis-
[a] In all cases, the consumption of the diazo compound is the total yield,
so the yield is also a measure of the chemoselectivity of the reaction.
[b] Positive values indicate that (1R)-cyclopropanes are the major com-
pounds.
AHCTRE(UNG oxazoline) ligand bears only one stereogenic center, as in
tained by using analogous C2-symmetric Box ligands have
also been included.
the case of 5h. We can conclude that when one of the oxa-
zoline rings bears a tert-butyl group, any substitution on the
second ring (even with a single methyl group) leads to good
enantioselectivity levels in the benchmarkcyclopropanation
reaction. This result, apart from the mechanistic implications
that we will discuss below, opens up the possibility of de-
signing and synthesizing new asymmetric bis(oxazoline)-
type ligands that are susceptible to being anchored to a sup-
port through a single substituent in the 4-position, an immo-
bilization strategy not yet explored, without a foreseeable
loss of enantioselectivity. Workin this direction is currently
ongoing in our group.
The trans/cis diastereoselectivity values fall within a close
range, between 60:40 and 73:27, irrespective of the symme-
try of the chiral ligand and of the nature of the substitution
of the oxazoline rings. This fact confirms that the trans/cis
selectivity is mainly directed by the intermolecular steric in-
teractions between the ester group of the carbene intermedi-
ate and the incoming alkene in the transition state and that
it is basically unaffected by the nature of the chiral ligand.[15]
In contrast, large variations in enantioselectivity were ob-
served when the results obtained with C2-symmetric ligands
were compared with those of ligands that bear a single sub-
stituent (5as with 5a, 5bs with 5b, 5cs with 5c, and 5gs with
5g), except in the case of the ligands that bear quaternary
stereogenic carbon atoms (5is and 5i). In the former cases, a
marked decrease in the enantioselectivity, in both the trans-
and cis-cyclopropanes, was observed. However, note that
the enantioselectivity in the trans-cyclopropanes was still sig-
nificant (ca. 20–30% ee) when C1-symmetric ligands that
bear an unsubstituted oxazoline ring were used. This result
is qualitatively analogous to that described by Nishiyama
et al. in the case of Pybox ligands[6] and can probably be ex-
plained by the same mechanism, that is, the existence of par-
ticularly unfavorable reaction channels that lead to the for-
mation of the minor enantiomer.[7] In a simple interpreta-
Again, one possible explanation for these results comes
from the existence of two Si reaction channels that are dif-
ferently disfavored because of the different size of the bis-
AHCTREUNG
experimental results observed, at least in a qualitative
manner. However, if strict additivity (in energy terms) of
steric effects (which has recently been reported for ligand
5is[16]) is assumed, in the case of the aza-tBuMe2Box (5h)
ligand, a decrease in enantioselectivity should be expected
with regard to the tBuMeBox (5d) ligand as the second
methyl group in 5h would disfavor one of the Re reaction
channels to the same extent as the other methyl group disfa-
vors one of the Si reaction channels. As shown in Table 1,
8832
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 8830 – 8839

Cyclopropanation Reactions
FULL PAPER
We started our theoretical
study with the asymmetric
ligand most dissimilar to C2-
symmetric
tBuBox
(5cs),
namely, the tBuHBox (5c,
Figure 3). With this ligand we
considered the full catalytic
cycle previously elucidated for
this reaction and it is shown in
Scheme 2.[10,16]
A
fundamental difference
exists between the reaction
pathways that involve the C1-
and C2-symmetric ligands. As
noted previously by other au-
thors,[6,19] there are two possible
dispositions for the Cu–carbene
intermediate, in our case 9c,
namely those with the ester
Figure 5. A schematic representation of the possible reaction channels in the case of chiral ligands substituted
in both oxazoline rings, and the possible steric interactions that disfavor some of the corresponding TSs.
almost identical enantioselectivities in the major trans-cyclo-
propanes are observed with both chiral ligands. It is there-
fore clear that these models do not offer a complete view of
the stereodifferentiation mechanism.
A special case is that of bis(oxazolines) that bear quater-
nary carbon stereogenic centers. Unlike the other ligands
tested, in this case the presence or absence of C2 symmetry
(5is vs. 5i) does not introduce significant changes in the
enantioselectivies obtained.
The question arises as to how the enantioselectivity de-
pends on the substitution pattern of the asymmetric bis(oxa-
zoline) ligands. Furthermore, the possible steric interaction
between the incoming alkene and the Box substituent in the
highly congested ligands cannot be discarded. To test the re-
liability of the aforementioned simple models and to better
understand these systems, a thorough theoretical mechanis-
tic study was therefore undertaken.
Theoretical results: Several computational mechanistic stud-
ies of the mechanism of copper-catalyzed cyclopropanation
reactions have recently been published.[10,16–20] In particular,
Scheme 2. Full catalytic cycle for the cyclopropanation reaction of ethyl-
ene (2b) with methyl diazoacetate (3b) catalyzed by the 5c–CuI com-
plex.
the enantiodifferentiation mechanism in the case of bis(oxa-
zoline)–Cu catalysts has been studied by our group[10] and
by Norrby and co-workers,[17] by using both QM and QM/
MM methods to model the chiral ligand. It has been
shown[16] that the enantioselection mechanisms can be ap-
propriately studied by using simplified models in which the
ethyl diazoacetate (3a) is substituted by methyl diazoacetate
(3b) and the styrene (2a) is substituted by ethylene (2b) as
the main steric interactions responsible for the enantioselec-
tion are retained in the simplified model, and consequently
we adopted this approach in this study.
group and the bis(oxazoline) substituent either on the same
side (syn) or on different sides (anti) of the chelate complex
plane (Figure 6). This duplicates the number of possible re-
action channels. In these conditions it is important to deter-
mine if Curtin–Hammett conditions apply to the systems. If
the interconversion between the syn and anti forms of the
Cu–carbene intermediate is faster than the addition of the
carbene to the alkene double bond, then Curtin–Hammett
conditions are met and the stereoselectivity of the reactions
will only depend on the relative energies of the different ad-
dition transition states (TSs) 10c. On the other hand, if the
We recently reported[16] the results of a theoretical and ex-
perimental study of the prototypical C2-symmetric tBuBox
ligand (5cs). We have used these results as a reference for
the results obtained with asymmetric ligands.
Chem. Eur. J. 2007, 13, 8830 – 8839
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8833

J. I. García et al.
Figure 6. Different possible structures for the Box–Cu–carbene inter-
mediates.
interconversion between the syn and anti forms is slower
than the cyclopropanation step, then the stereoselectivity
will depend on the detailed kinetics of the whole reaction,
which include the nitrogen extrusion step via TS 8c. Of
course, when the ligand is C2-symmetric, the syn–anti inter-
À
conversion by Cu Ccarbene rotation is equivalent to the half-
rotation of the whole complex, thus leading to equivalent
structures, and the above considerations do not apply.
To gain an insight into this reaction feature, we calculated
all of the possible TSs 8c and the subsequent Cu–carbene
intermediates 9c, the syn–anti rotation transition state 9cr,
and all the possible cyclopropanation TSs 10c. In all cases,
the two possible conformations of the ester group (labeled I
and II) were taken into account. A scheme illustrating the
possible reaction channels that lead to the cyclopropane
product is shown in Figure 7. Of course, in the case of ethyl-
ene, no enantioselectivity is possible, but as mentioned
before, we can take the Re and Si reaction channels as a
measure of the enantioselectivity in real systems. Typical
structures calculated for the different reaction intermediates
and transition-state structures are shown in Figure 8 and the
main energy results are gathered in Table 2.
Figure 7. Possible reaction channels that lead to (1R)- and (1S)-cyclopro-
panes in the enantioselective reactions.
Given that entropic effects are presumably more impor-
tant in the bimolecular cyclopropanation reaction than in
passes through TS 9cr, is very similar to the cyclopropana-
tion activation barriers (from 11.4 to 17.9 kcalmolÀ1), simi-
larly to that described for salicylaldimine–CuI systems.[19] Al-
though these values point to a non-Curtin–Hammett behav-
ior of the reaction, no clear conclusion can be made on the
basis of these calculations given the limitations in accuracy
that result from the theoretical level and the simplified
model used. A possibility would lie in a comparison of the
calculated enantioselectivities, by using each approximation,
with the experimental values obtained for the reaction of
styrene with ethyl diazoacetate catalyzed by the tBuHBox–
À
the carbene rotation around the Cu C bond, we considered
first the Gibbs free-energy surface. As can be seen in
Table 2, the nitrogen extrusion step is the rate-determining
step of the catalytic cycle, as previously described for other
similar systems.[10,16,21] The calculated activation energies for
the cyclopropanation step indicate a very fast reaction, and
compare well with the experimental[22] and theoretical[23]
values determined for the cyclopropanation reactions with
electrophilic carbenes, as well as with the experimental
value determined for the reaction of a neutral carbene com-
plex derived from a less electrophilic diazo compound
(DG° =20 kcalmolÀ1).[20] Concerning the mechanistic issue
raised above, the results are not conclusive because the cal-
culated carbene rotation barrier (14.5 kcalmolÀ1), which
CuACHTRE(UNG OTf)2 complex. As shown in Table 1, these values are
29% ee for the trans-cyclopropanes and 8% ee for the cis-
cyclopropanes. By using the calculated Gibbs free energies,
the values estimated for the “enantioselectivity” of the reac-
tion of ethylene with methyl diazoacetate (taking into ac-
8834
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 8830 – 8839

Cyclopropanation Reactions
FULL PAPER
count the Re and Si reaction
channels) are 70% ee assuming
Curtin–Hammett
conditions
and 66% ee assuming non-
Curtin–Hammett behavior (i.e.,
that the carbene rotation is
slower than the ethylene cyclo-
propanation) (Table 2).[24] In
both approaches, the enantiose-
lectivities are grossly overesti-
mated and do not allow differ-
entiation between the two
mechanistic alternatives.
We have previously shown[16]
that E0 +ZPE values often lead
to calculated selectivities that
are closer to the experimental
values in similar systems, so we
repeated our estimation of the
enantioselectivity by using this
parameter. The new calculated
values are 8 and 18% ee, re-
spectively, with and without as-
suming Curtin–Hammett condi-
Figure 8. Some selected calculated (at the B3LYP/6-31G(d) theoretical level of theory) geometries of the main
intermediates and transition-state structures of the reaction of ethylene with methyl diazoacetates, catalyzed
by the tBuHBox–CuI (5c-Cu) complex.
Table 2. Calculated [B3LYP/6-31G(d)] relative energies[a] and activation barriers [kcalmolÀ1] of the catalytic cycle of the cyclopropanation reaction of
ethylene with methyl diazoacetate, catalyzed by the tBuHBox–CuI complex. Estimated %ee values with and without assuming Curtin–Hammett condi-
tions.
Structure[b]
DE[c]
DDE
DE°
ee [%][d]
Non C–H
DG[c]
DDG
DG°
ee [%][d]
Non C–H
C–H
C–H
2b
3b
N2
6c
0.0
–
–
–
–
–
–
–
–
0.0
0.0
0.0
0.0
–
–
–
–
–
–
–
–
0.0
0.0
0.0
8c-anti-I
8c-anti-II
8c-syn-I
8c-syn-II
20.4
19.4
19.6
26.5
1.0
0.0
0.2
7.1
20.4
19.6
19.4
26.5
20.7
20.0
19.7
25.8
1.0
0.3
0.0
6.1
21.1
20.4
20.1
26.2
9c-anti-I
9c-anti-II
9c-syn-I
9c-syn-II
9cr
6.8
6.2
6.7
7.0
20.9
0.6
0.0
0.5
0.8
–
–
–
–
–
14.7
À3.0
À3.2
À2.7
À2.4
11.4
0.1
0.0
0.4
0.7
–
–
–
–
14.5
10c-anti-I-Re
10c-anti-II-Re
10c-syn-I-Re
10c-syn-II-Re
10c-anti-I-Si
10c-anti-II-Si
10c-syn-I-Si
10c-syn-II-Si
6.6
8.2
6.4
8.3
8.6
6.2
12.6
12.3
0.4
2.1
0.2
2.1
2.4
0.0
6.4
6.1
0.4
2.1
0.2
2.1
2.4
0.0
6.4
6.1
8.8
9.8
8.2
9.7
10.4
9.0
0.6
1.6
0.0
1.4
2.1
0.8
4.8
6.5
12.0
13.0
11.4
12.8
13.5
12.2
16.2
17.9
8[e]
18[e]
70[e]
66[e]
13.0
14.8
4b
À55.3
–
–
À51.9
–
–
[a] Energy values include ZPE corrections at the same theoretical level. [b] syn and anti stand for the relative positions of the tert-butyl and ester groups,
Re and Si stand for the carbene carbon face approached by ethylene and I and II stand for the conformation of the ester group (see Figure 4). [c] The in-
itial complex tBuHBox–Cu–ethylene (6c), ethylene (2b), methyl diazoacetate (3b), and dinitrogen have been arbitrarily chosen as the zero level in the
relative energy calculations. [d] The experimental values for the reaction of styrene with ethyl diazoacetate using the same ligand are 29% ee (trans-cy-
clopropanes) and 8% ee (cis-cyclopropanes). [e] Re attackis favored (leading to (1 R)-cyclopropanes in the case of a substituted alkene).
Chem. Eur. J. 2007, 13, 8830 – 8839
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8835

J. I. García et al.
Table 3. Calculated [B3LYP/6-31G(d)] relative energies and activation
barriers for the cyclopropanation transition-state structures of the reac-
tion of ethylene with methyl diazoacetate, catalyzed by the 5d–CuI and
5h–CuI complexes.
tions (Table 2). These values are, effectively, much closer to
the experimental observations, but unfortunately, also do
not allow discrimination between the two mechanistic alter-
natives. We realize that both mechanistic approaches lead to
similar calculated enantioselectivities. This fact can be ra-
tionalized through the structural similarities of TSs 8 (gov-
erning the non-Curtin–Hammett behavior) and 10 (govern-
ing the Curtin–Hammett behavior) such that the steric inter-
actions responsible for their relative stabilities should be
similar in both cases (compare, for instance, the anti-II and
syn-II structures for 8c and 10c in Figure 8 and their relative
energies in Table 2). Therefore, in a first approach, we will
henceforth assume Curtin–Hammett behavior for the cata-
lytic systems to facilitate the calculation of the enantioselec-
tivity.
Next, we considered two different substitution patterns in
the second oxazoline ring, namely, the presence of a single
methyl group (as in 5d), and that of two methyl groups (as
in 5h), which lead to a nonstereogenic, quaternary carbon
atom in the 4-position of the oxazoline ring. From an experi-
mental point of view, both substitutions led to (unexpect-
edly) good enantioselectivities when the corresponding li-
gands were used in the benchmarkcyclopropanation reac-
tion.
TS[a]
DDE°
ee
DDG°
ee
[kcalmolÀ1
]
[%][b]
[kcalmolÀ1
]
[%][b]
10d-anti-I-Re
10d-anti-II-Re
10d-syn-I-Re
10d-syn-II-Re
10d-anti-I-Si
10d-anti-II-Si
10d-syn-I-Si
10d-syn-II-Si
10h-anti-I-Re
10h-anti-II-Re
10h-syn-I-Re
10h-syn-II-Re
10h-anti-I-Si
10h-anti-II-Si
10h-syn-I-Si
10h-syn-II-Si
0.5
2.4
0.0
1.7
1.8
2.3
5.3
5.0
0.8
2.8
0.0
1.4
2.6
2.1
9.3
8.5
92
0.4
2.2
0.0
1.7
2.1
1.5
4.9
4.2
0.3
2.1
0.0
1.0
1.7
1.8
9.2
8.4
88
94
90
[a] syn and anti stand for the relative positions of the tert-butyl and ester
groups, Re and Si stand for the carbene carbon face approached by ethyl-
ene, and I and II stand for the conformation of the ester group (see
Figure 4). [b] Estimated ee values assuming Curtin–Hammett conditions.
Concerning the tBuMeBox ligand (5d), the relative ener-
gies of the corresponding cyclopropanation reactions are
gathered in Table 3, as well as the estimated enantioselectiv-
ities assuming Curtin–Hammett behavior. Some representa-
tive structures are shown in Figure 9.
As can be seen, Re approaches of the alkene are always
favored over the corresponding Si approach. In particular,
and similarly to what happened in the case of the 5c ligand,
the syn-Si reaction channels are especially disfavored owing
to the steric interaction between the ester group and the
tert-butyl substituent of the ligand that results in an impor-
tant deformation of the chelate complex (see the 10d-syn-
II-Si structure in Figure 9). A similar steric interaction exists
between the ester and methyl
reaction channels correspond to the 10d-syn-I-Re and 10d-
anti-I-Re TSs, both leading to (1R)-cyclopropanes, in agree-
ment with the experimental results. The estimated enantio-
selectivity for this ligand is around 90%, which is similar to
the approximate 80% ee observed experimentally in the re-
action of styrene and higher than the value calculated in the
case of the 5c ligand, that is again in complete agreement
with experimental observations.
Concerning the aza-tBuMe2Box ligand (5h), the relative
energies of the corresponding cyclopropanation TSs are
gathered in Table 3, as well as the estimated enantioselectiv-
ities assuming Curtin–Hammett behavior. Some representa-
tive structures are shown in Figure 9.
groups in TS 10d-anti-Si, which
also leads to a deformation of
the chelate complex. If we com-
pare the relative energies of the
TSs 10c and 10d, we realize
that the main difference be-
tween the two sets of data
indeed lies in the relative
energy of 10c-anti-II-Si, which
is very low (even corresponding
to the minimum value in the
case of E0 +ZPE energies) and
leads to an important reaction
channel for the formation of
(1S)-cyclopropanes, and hence,
to
a decrease in the global
enantioselectivity of the reac-
tion. In the case of the 5d
ligand, the two clearly favored (5d-Cu) and the aza-tBuMe2Box–CuI (5h-Cu) complexes.
Figure 9. Some selected calculated (at the B3LYP/6-31G(d) theoretical level of theory) geometries of the tran-
sition-state structures of the reaction of ethylene with methyl diazoacetates, catalyzed by the tBuMeBox–CuI
8836
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 8830 – 8839

Cyclopropanation Reactions
FULL PAPER
Both Re approaches of the alkene are clearly favored
over the Si approaches. Similarly to what happened in previ-
ous cases, the syn-Si reaction channel is highly disfavored
(even more than for the rest of ligands investigated) owing
to the steric interaction between the ester group and the
tert-butyl substituent of the ligand. The presence of a second
methyl group in the ligand does not appear to introduce
new steric interactions that modify the enantioselection of
the ligand (as assumed for the syn-Re TS in Figure 5), as is
shown by comparing the relative energies of TSs syn-I-Re,
anti-I-Re, and anti-II-Si for the 5d and 5h ligands (0.0, 0.4,
and 2.1 vs. 0.0, 0.3, and 1.8 kcalmolÀ1 in the case of Gibbs
free energies, and 0.0, 0.5, and 1.8 vs. 0.0, 0.8, and
2.1 kcalmolÀ1 in the case of E0 +ZPE energies). In fact, in
10h-syn-I-Re it can be seen that the ester group adopts a po-
sition that is intermediate between the two methyl groups,
which avoids the presumed steric interaction with the upper
methyl group. This effect is probably favored by the lesser
distortion of the azaBox ligand in adopting a boat confor-
mation. Also, note that a steric interaction between the in-
coming alkene and the ligand substituents is not detected in
any example, which is in agreement with previous computa-
tional studies.[10,16] These theoretical results are in excellent
agreement with the similar enantioselectivities obtained ex-
perimentally in the reactions carried out by using the 5d
and 5h ligands and highlight the excessive simplicity of the
stereoselection models previously presented.
This agreement between experiment and theory will allow
us to theoretically investigate the behavior of new ligands
before their synthesis and testing, which facilitates the
design of tailored catalytic systems for this reaction.
Experimental Section
General methods: All reactions were carried out under nitrogen or argon
in oven-dried glassware. Dichloromethane was distilled from calcium hy-
dride. Ethanol was distilled from magnesium. Tetrahydrofuran was dis-
tilled from potassium and toluene was distilled from sodium. Amino
acids were used as commercially available. 1H and 13C NMR spectra
(CDCl3, d [ppm], J [Hz]) were obtained using a Bruker ARX-300 instru-
ment with TMS as the standard. Quantitative elemental analyses were
performed on a Perkin–Elmer 2400 instrument. Polarimetry was carried
out using a Jasco P-1020 instrument. Mass spectra were carried out using
a VG Autospec instrument.
Ligand synthesis: The asymmetric bis(oxazoline) ligands were prepared
as previously described.[13a] The aminooxazolines[25] and ethoxyoxazo-
lines[26,27] necessary for the azabis(oxazoline) syntheses were prepared ac-
cording to literature procedures. The asymmetric azabis(oxazoline) li-
gands were prepared by the following general procedure: Ethoxyoxazo-
line (1.2 mmol), aminooxazoline (1.0 mmol), and a catalytic amount of p-
toluenesulfonic acid (20 mg) were dissolved in toluene (20 mL) and
heated at reflux for 24 h. After this period, the solution was concentrated
in vacuo and purified by chromatography on silica gel by using ethyl ace-
tate as the eluent to give the desired azabis(oxazolines) in a yield of 57
to 69%.
[(S)-4-tert-Butyl-4,5-dihydrooxazol-2-yl](4,5-dihydrooxazol-2-yl)amine:
According to the general procedure, reaction of 2-ethoxy-4,5-dihydrooxa-
zole and (S)-4-tert-butyl-4,5-dihydrooxazol-2-ylamine gave the title com-
pound as a colorless solid in a yield of 62%. M.p. 103–1068C; [a]2D0 (c=
1.0 in CH3OH)=75.5; 1H NMR (300 MHz, CDCl3): d=4.39–4.29 (m,
3H), 4.19 (dd, J=8.8, 6.3 Hz, 1H; CH2CH), 3.59–3.75 (m, 3H), 0.91 ppm
Conclusion
(s, 9H, C
ACHTREUNG
Experimental catalytic studies of the cyclopropanation reac-
tion of styrene with ethyl diazoacetate catalyzed by a series
of C1-symmetric bis(oxazoline) ligands have shown that C2
symmetry is not mandatory to obtain high levels of enantio-
selectivity. Even ligands that bear only one stereogenic
center are able to induce stereoselectivity levels that are
close to the best ones obtained with the classical C2-symmet-
ric ligands. This finding, apart from being of mechanistic in-
terest, opens the door to new immobilization strategies
through covalent bonding of the bis(oxazoline) ligand to the
support through only one of the substituents in the 4-posi-
tion, with no foreseeable loss of enantioselectivity. Workin
this direction is currently ongoing in our group.
The computational mechanistic studies carried out do not
allow a decision to be made as to whether or not Curtin–
Hammett conditions are met in the catalytic reactions of
asymmetric ligand complexes, but both approximations lead
to sufficiently similar predictions that mechanistic issues can
still be discussed. Thus, the theoretical calculations carried
out on model systems show a very good agreement with ex-
perimental observations and have allowed us to gain an in-
sight into the different systems investigated. In particular,
enantioselectivity arises from differently favored reaction
channels, which lead to the formation of one or other cyclo-
propane enantiomer depending on the steric interactions be-
tween the ester group and the bis(oxazoline) substituents.
(NOCN), 67.5 (CH2CH2), 67.3 (CHCH2), 66.1 (CHCH2), 49.1ACHTREUNG
33.6 (CCATHER(UNG CH3)3), 25.3 ppm (CACHTREUNG
1479, 1439, 1385, 1250, 1058, 760 cmÀ1; MS (CI-MS): m/z: 212.1 [MH+];
elemental analysis calcd (%) for C10H17N3O2: C 56.85, H 8.11, N 19.89;
found: C 56.55, H 7.85, N 19.89.
[(S)-4-tert-Butyl-4,5-dihydrooxazol-2-yl](4,4-dimethyl-4,5-dihydrooxazol-
2-yl)amine: According to the general procedure, reaction of 2-ethoxy-4,4-
dimethyl-4,5-dihydrooxazole and (S)-4-tert-butyl-4,5-dihydrooxazol-2-yla-
mine gave the title compound as a colorless solid in a yield of 59%. M.p.
81–848C; [a]2D0 (c=1.0 in CH3OH)=73.8; 1H NMR (300 MHz, CDCl3):
d=4.34–4.32 (m, H), 4.22–4.12 (m, H), 4.01 (s, H), 4.84–3.72 (m, H), 1.32
(s, 3H), 1.30 (s, 3H), 0.89 ppm (s, 9H); 13C NMR (75.5 MHz, CDCl3): d=
165.7 (NOCN), 164.6 (NOCN), 77.8 (CCH2), 67.5 (CHCH2), 67.4
(CHCH2), 62.0, 33.7 (CACHTRE(UNG CH3)3), 29.0 (CACHETR(UGN CH3)2), 28.4 (CACHTRE(UNG CH3)2), 25.3 ppm
(CACHTRE(UNG CH3)3); IR: n˜ =2965, 2880, 1633, 1594, 1442, 1388, 1267, 1200, 1053,
999, 755 cmÀ1; MS (CI-MS): m/z: 239.2 [M +]; HRMS (EI-MS): m/z:
C
calcd for [M +]: 239.1634; found: 239.1636.
C
General procedure for the methylation of azabis(oxazolines): Azabis(ox-
azoline) (1 mmol) was dissolved in THF (10 mL), cooled to À788C, and
n-butyllithium (1.10 mmol, 690 mL of a 1.6n solution in hexane) was
added slowly. After stirring for 10 min, MeI (5.0 mmol, 710 mg) was
added dropwise, the solution was slowly warmed to room temperature
overnight and stirred for another 10 h. Then, aqueous Na2CO3 (5 mL)
was added and the mixture concentrated. The residue was diluted with
DCM (10 mL) and aqueous Na2CO3 (10 mL) and the phases were sepa-
rated. The aqueous phase was extracted twice with DCM, the combined
organic phases were dried with MgSO4, and the solvent was evaporated
to give the product in a yield of 98%.
[(S)-4-tert-Butyl-4,5-dihydrooxazol-2-yl](4,5-dihydrooxazol-2-yl)methyla-
mine (5g): According to the general procedure, reaction of [(S)-4-tert-
Chem. Eur. J. 2007, 13, 8830 – 8839
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8837

J. I. García et al.
butyl-4,5-dihydro-oxazol-2-yl](4,5-dihydrooxazol-2-yl)amine with MeI
[3] See for instance: a) A. K. Ghosh, M. Packiarajan, J. Cappiello, Tet-
rahedron: Asymmetry 1998, 9, 1–45; b) J. S. Johnson, D. A. Evans,
Acc. Chem. Res. 2000, 33, 325–335; c) G. Desimoni, G. Faita, P.
Quadrelli, Chem. Rev. 2003, 103, 3199–3154; d) H. Lebel, J.-F. Mar-
coux, C. Molinaro, A. B. Charette, Chem. Rev. 2003, 103, 977–1050.
[4] A. Cornejo, J. M. Fraile, J. I. García, M. J. Gil, C. I. Herrerías, G. Le-
garreta, V. Martínez-Merino, J. A. Mayoral, J. Mol. Catal. A 2003,
196, 101–108.
[5] M. P. A. Lyle, N. D. Draper, P. D. Wilson, Org. Biomol. Chem. 2006,
4, 877–885.
[6] H. Nishiyama, N. Soeda, T. Naito, Y. Motoyama, Tetrahedron:
Asymmetry 1998, 9, 2865–2869.
gave the title compound as a clear oil in a yield of 98%. [a]2D0 (c=1.0 in
1
CH3OH)=15.7; H NMR (300 MHz, CDCl3): d=4.27 (dd, J=9.6, 8.5 Hz,
1H; CH2CH), 4.13 (dd, J=8.5, 7.4 Hz, 1H; CH2CH), 3.92 (dd, J=9.6,
7.4 Hz, 1H), 3.83–3.70 (m, 1H; OCH2CH2), 3.68–3.55 (m, 1H;
OCH2CH2), 3.34–3.24 (m, 2H; OCH2CH2), 3.20 (s, 3H; CH3), 0.89 ppm
(s, 9H; CCH3); 13C NMR (75.5 MHz, CDCl3): d=158.6 (NOCN), 157.7
(NOCN), 73.2 (CHCH2), 70.5 (CHCH2), 68.9 (CH2CH2), 52.8 (CH2CH2),
36.9 (NCH3), 33.9 (CACHTREUNG(CH3)3), 25.6 ppm (CACHTREU(NG CH3)3); IR: n˜ =3432, 2965,
2890, 1635, 1599 1479, 1395, 1255, 1058, 765 cmÀ1; MS (CI-MS): m/z:
+
226.2 [MH ]; HRMS (EI-MS): m/z: calcd for [M +]: 225.1477; found:
C
225.1479.
[(S)-4-tert-Butyl-4,5-dihydrooxazol-2-yl](4,4-dimethyl-4,5-dihydrooxazol-
2-yl)methylamine (5h): According to the general procedure, reaction of
[(S)-4-tert-Butyl-4,5-dihydrooxazol-2-yl](4,4-dimethyl-4,5-dihydrooxazol-
2-yl)amine with MeI gave the title compound as a light yellow oil in a
yield of 97%. [a]2D0 (c=1.0 in CH3OH)=7.6; 1H NMR (300 MHz,
CDCl3): d=4.38–4.2 (m, 2H; CH2CH), 4.08 (dd, J=9.4, 7.9 Hz, 2H;
CH2C), 3.79 (dd, J=9.5, 6.8 Hz, 1H; CHCH2), 3.38 (s, 3H; NCH3), 1.3 (s,
[7] A. Cornejo, J. M. Fraile, J. I. García, M. J. Gil, V. Martínez-Merino,
J. A. Mayoral, Angew. Chem. 2005, 117, 462–465; Angew. Chem.
Int. Ed. 2005, 44, 458–461.
[8] a) C. Lee, W. Yang, R. Parr, Phys. Rev. B 1988, 37, 785–789;
b) A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[9] S. Niu, M. B. Hall, Chem. Rev. 2000, 100, 353–406.
[10] a) J. M. Fraile, J. I. García, V. Martínez-Merino, J. A. Mayoral, L.
Salvatella, J. Am. Chem. Soc. 2001, 123, 7616–7625; b) J. M. Fraile,
J. I. García, M. J. Gil, V. Martínez-Merino, J. A. Mayoral, L. Salva-
tella, Chem. Eur. J. 2004, 10, 758–765.
6H; CH3), 0.88 ppm (s, 9H; C
157.6 (NOCN), 156.6 (NOCN), 80.4 (CCH2), 73.2 (CHCH2), 70.5
(CCH2), 64.9 (CHCH2), 37.5 (NCH3), 33.9 (C(CH3)3), 28.6 (C(CH3)2),
28.5 (C(CH3)2), 25.5 ppm (C(CH3)3); IR: n˜ =3426, 2961, 1757, 1639, 1479,
(CH3)3); 13C NMR (75.5 MHz, CDCl3): d=
ACHTREUNG
A
ACHTREUNG
A
ACHTREUNG
[11] Gaussian 03, Revision C.02, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomer-
y, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyen-
gar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N.
Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J.
Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C.
Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cio-
slowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaro-
mi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wall-
ingford, CT, 2004.
[12] C. W. Bauschlicher, Jr., Chem. Phys. Lett. 1995, 246, 40–44.
[13] a) J. I. García, J. A. Mayoral, E. Pires, I. Villalba, Tetrahedron:
Asymmetry 2006, 17, 2270–2275; b) A. Cornejo, J. M. Fraile, J. I.
García, M. J. Gil, V. Martínez-Merino, J. A. Mayoral, E. Pires, I. Vil-
lalba, Synlett 2005, 2321–2324.
[14] H. Werner, R. Vicha, A. Gissibl, O. Reiser, J. Org. Chem. 2003, 68,
10166–10168.
[15] a) H. Fritschi, U. Leutenegger, A. Pfaltz, Helv. Chim. Acta 1988, 71,
1553–1565; b) A. Pfaltz, Comprehensive Asymmetric Catalysis, Vol.
II, Springer, Berlin, 1999, pp. 513–538.
1432, 1385, 1194, 952, 712 cmÀ1; MS (EI-MS): m/z: 253.2 [M +]; HRMS
C
(EI-MS): m/z: calcd for [M +]: 253.1790; found: 253.1794.
C
Cyclopropanation reactions: The bis(oxazoline)–copper complexes were
prepared by dissolving the copper salt CuACHTRE(UNG OTf)2: (0.05 mmol) and the
corresponding ligand (0.05 mmol) in anhydrous dichloromethane (1 mL).
After stirring for 1 h, the insoluble materials were removed by microfil-
tration and the bluish-green solution was added to a 25 mL two-necked
round-bottomed flaskthat contained styrene (520.75 mg, 5 mmol) and n-
decane (100 mg) in CH2Cl2 (4 mL) under argon. Ethyl diazoacetate
(570.5 mg. 5 mmol) diluted in anhydrous CH2Cl2 (1 mL) was slowly
added (4 h) by using a syringe pump. The reaction was stirred at room
temperature for 24 h. After this time the solution was diluted (5 mL
CH2Cl2) and the results of the reaction were determined by gas chroma-
tography. FID from Hewlett-Packard 5890II; cross-linked methyl silicone
column: 25 m”0.2 mm”0.33 mm; helium as carrier gas: 20 psi; injector
temperature: 2308C; detector temperature: 2508C; oven temperature
program: 708C (3 min), 158CminÀ1 to 2008C (5 min); retention times:
ethyl diazoacetate 4.28 min, styrene 5.03 min, n-decane 6.93 min, cis-cy-
clopropanes 11.84 min, and trans-cyclopropanes 12.35 min. The asymmet-
ric inductions of the reactions were also determined by gas chromatogra-
phy. FID from Hewlett-Packard 5890II; Cyclodex B column: 30 m”
0.25 mm”0.25 mm; helium as carrier gas: 20 psi; injector temperature:
2308C; detector temperature: 2508C; oven temperature program: 1258C
isotherm; retention times: (1S,2R)-cyclopropane 28.9 min, (1R,2S)-cyclo-
propane 29.8 min, (1R,2R)-cyclopropane 34.3 min, and (1S,2S)-cyclopro-
pane 34.9 min.
[16] J. I. García, G. JimØnez-OsØs, V. Martínez-Merino, J. A. Mayoral, E.
Pires, I. Villalba, Chem. Eur. J. 2007, 13, 4064–4073.
[17] T. Rasmussen, J. F. Jensen, N. ðstergaard, D. Tanner, T. Ziegler, P.-
O. Norrby, Chem. Eur. J. 2002. 8. 177–184.
[18] a) B. F. Straub, P. Hofmann, Angew. Chem. 2001, 113, 1328–1330;
Angew. Chem. Int. Ed. 2001, 40, 1288–1290; b) B. F. Straub, I.
Gruber, F. Rominger, P. Hofmann, J. Organomet. Chem. 2003, 684,
124–143.
Acknowledgements
This workwas made possible by the financial support of the Ministerio
de Educación y Ciencia (projects CTQ2005–08016 and Consolider Ingen-
io 2010 CSD 2006–00). I.V. and M.R. thankthe Ministerio de Educación
y Ciencia and the Diputación General de Aragón for grants, respectively.
[19] K. Suenobu, M. Itagaki, E. Nakamura, J. Am. Chem. Soc. 2004, 126,
7271–7280.
[20] X. Dai, T. H. Warren, J. Am. Chem. Soc. 2004, 126, 10085–10094.
[21] a) R. G. Salomon, J. K. Kochi, J. Am. Chem. Soc. 1973, 95, 3300–
3310; b) M. Díaz-Requejo, T. R. Belderrain, M. C. Nicasio, F. Prieto,
P. J. PØrez, Organometallics 1999, 18, 2601–2609.
[22] N. J. Turro, G. F. Lehr, J. A. Butcher, Jr., J. Am. Chem. Soc. 1982,
104, 1754–1756.
[1] Comprehensive Asymmetric Catalysis (Eds.: E. N. Jacobsen, A.
Pfaltz, H. Yamamoto), Springer-Verlag, Berlin, 1999.
[2] See for instance: a) Arataniꢁs catalysts, used industrially in the pro-
duction of chrisantemic acid derivatives: T. Aratani, Pure Appl.
Chem. 1985, 57, 1839–1844; b) phosphinooxazolines: O. Reiser,
Angew. Chem. 1993, 105, 576–578; Angew. Chem. Int. Ed. Engl.
1993, 32, 547–549; .
[23] A. E. Keating, S. R. Merrigan, D. A. Singleton, K. N. Houk, J. Am.
Chem. Soc. 1999, 121, 3933–3938, and references therein.
8838
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 8830 – 8839

Cyclopropanation Reactions
FULL PAPER
[24] The %ee in non-Curtin–Hammett conditions was estimated as fol-
lows: First, the percentage of the reaction going through the reac-
tion channels that lead to syn and anti carbene intermediates was
calculated on the basis of the relative energies of TS 8c. Next, and
considering that the ester rotation is faster than the carbene addi-
tion to ethylene, the percentage of the reaction going through the
Re and Si reaction channels was calculated separately for the syn
and anti reaction channels on the basis of the relative energies of
TSs 10c-syn and 10c-anti. Finally, the percentage of R and S prod-
ucts, and hence the enantioselectivity, was estimated from the above
percentages. For instance, in the case of the E0 +ZPE energies,
37.6% of the reaction occurs through the syn reaction channel and
62.4% through the anti reaction channel. 100% of the syn carbene
intermediates give the R product, whereas only 34.6% of the anti
carbene intermediates give the R product. This allows an estimate
of 0.376”1.00+0.624”0.346=0.592 (ca. 59%) for the R product,
and hence 18% ee.
[25] R. R. Wittekind, J. D. Rosenau, G. I. Poos, J. Org. Chem. 1961, 26,
444–446.
[26] E. Le Gall, J.-P. Hurvois, S. Sinbandhit, Eur. J. Org. Chem. 1999,
2645–2653.
[27] M. Huche, G. Lhommet, J. Heterocycl. Chem. 1986, 23, 701–704.
Received: May 4, 2007
Published online: August 2, 2007
Chem. Eur. J. 2007, 13, 8830 – 8839
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8839