Reactivity of activated versus nonactivated 2-(bromomethyl)aziridines with respect to sodium methoxide: A combined computational and experimental study

Article abstract of DOI:10.1021/jo201255z

The difference in reactivity between the activated 2-bromomethyl-1- tosylaziridine and the nonactivated 1-benzyl-2-(bromomethyl)aziridine with respect to sodium methoxide was analyzed by means of DFT calculations within the supermolecule approach, taking into account explicit solvent molecules. In addition, the reactivity of epibromohydrin with regard to sodium methoxide was assessed as well. The barriers for direct displacement of bromide by methoxide in methanol are comparable for all three heterocyclic species under study. However, ring opening was found to be only feasible for the epoxide and the activated aziridine, and not for the nonactivated aziridine. According to these computational analyses, the synthesis of chiral 2-substituted 1-tosylaziridines can take place with inversion (through ring opening/ring closure) or retention (through direct bromide displacement) of configuration upon treatment of the corresponding 2-(bromomethyl)aziridines with 1 equiv of a nucleophile, whereas chiral 2-substituted 1-benzylaziridines are selectively obtained with retention of configuration (via direct bromide displacement). Furthermore, the computational results showed that explicit accounting for solvent molecules is required to describe the free energy profile correctly. To verify the computational findings experimentally, chiral 1-benzyl-2-(bromomethyl)aziridines and 2-bromomethyl-1-tosylaziridines were treated with sodium methoxide in methanol. The presented work concerning the reactivity of 2-bromomethyl-1- tosylaziridine stands in contrast to the behavior of the corresponding 1-tosyl-2-(tosyloxymethyl)aziridine with respect to nucleophiles, which undergoes a clean ring-opening/ring-closure process with inversion of configuration at the asymmetric aziridine carbon atom.

Full text of DOI:10.1021/jo201255z

Article
pubs.acs.org/joc
Reactivity of Activated versus Nonactivated 2-
(Bromomethyl)aziridines with respect to Sodium Methoxide: A
Combined Computational and Experimental Study
Hannelore Goossens,† Karel Vervisch,‡ Saron Catak,*,† Sonja Stankovic,‡ Matthias D’hooghe,‡
́
Frank De Proft,§ Paul Geerlings,§ Norbert De Kimpe,*,‡ Michel Waroquier,†
and Veronique Van Speybroeck*,†
†Center for Molecular Modeling, Ghent University, Member of QCMM-Alliance Ghent-Brussels, Technologiepark 903, B-9052
Zwijnaarde, Belgium
‡Department of Sustainable Organic Chemistry and Technology, Faculty of Bioscience Engineering, Ghent University, Coupure Links
653, B-9000 Ghent, Belgium
§Eenheid Algemene Chemie (ALGC), Vrije Universiteit Brussel (VUB), Member of QCMM-Alliance Ghent-Brussels, Pleinlaan 2,
1050 Brussels, Belgium
S
* Supporting Information
ABSTRACT: The difference in reactivity between the
activated 2-bromomethyl-1-tosylaziridine and the nonactivated
1-benzyl-2-(bromomethyl)aziridine with respect to sodium
methoxide was analyzed by means of DFT calculations within
the supermolecule approach, taking into account explicit
solvent molecules. In addition, the reactivity of epibromohydrin with regard to sodium methoxide was assessed as well. The
barriers for direct displacement of bromide by methoxide in methanol are comparable for all three heterocyclic species under
study. However, ring opening was found to be only feasible for the epoxide and the activated aziridine, and not for the
nonactivated aziridine. According to these computational analyses, the synthesis of chiral 2-substituted 1-tosylaziridines can take
place with inversion (through ring opening/ring closure) or retention (through direct bromide displacement) of configuration
upon treatment of the corresponding 2-(bromomethyl)aziridines with 1 equiv of a nucleophile, whereas chiral 2-substituted 1-
benzylaziridines are selectively obtained with retention of configuration (via direct bromide displacement). Furthermore, the
computational results showed that explicit accounting for solvent molecules is required to describe the free energy profile
correctly. To verify the computational findings experimentally, chiral 1-benzyl-2-(bromomethyl)aziridines and 2-bromomethyl-1-
tosylaziridines were treated with sodium methoxide in methanol. The presented work concerning the reactivity of 2-
bromomethyl-1-tosylaziridine stands in contrast to the behavior of the corresponding 1-tosyl-2-(tosyloxymethyl)aziridine with
respect to nucleophiles, which undergoes a clean ring-opening/ring-closure process with inversion of configuration at the
asymmetric aziridine carbon atom.
philic carbon atoms C2, C3, and C4 (Scheme 1). The regio-
and chemoselectivity of nucleophilic attack is of particular
INTRODUCTION
■
The aziridine moiety represents one of the most valuable three-
membered ring systems in organic chemistry due to the
uncommon combination of reactivity, synthetic flexibility and
atom economy.1−9 Indeed, ring strain renders aziridines
susceptible to ring-opening reactions that dominate their
chemistry and makes them useful synthetic intermediates in
the arsenal of the organic chemist.10
Scheme 1. 2-(Halomethyl)aziridines 1 and Epihalohydrins 2
(X = Halogen)
Among 2-(halomethyl)aziridines 1, 2-(bromomethyl)-
aziridines (X = Br) in particular comprise a peculiar class of
aziridines and constrained β-halo amines with high synthetic
potential.11−13 Although structurally related to their oxygen
analogues epihalohydrins 2,14,15 aziridines 1 have been studied
to a far lesser extent in the chemical literature. In both systems,
the electrophilic reactivity of the constrained heterocycle can be
assessed relative to the haloalkyl moiety, meaning that
nucleophiles can discriminate between three different electro-
importance in the design of synthetic protocols toward the
preparation of valuable target compounds starting from
aziridines 1 and epoxides 2, which underlines the need for a
Received: June 15, 2011
Published: September 30, 2011
© 2011 American Chemical Society
8698
dx.doi.org/10.1021/jo201255z|J. Org. Chem. 2011, 76, 8698−8709

The Journal of Organic Chemistry
Article
thorough investigation and rationalization of the reaction
outcome.
If both pathways are competitive, a mixture of both
enantiomers will be obtained. It is clear that a deeper
understanding of this mechanism is of high importance
whenever the synthesis of chiral targets starting from chiral
aziridines 5 is contemplated.
The reactivity of epoxides 2 with respect to hydroxide and
water has previously been evaluated theoretically by means of
ab initio calculations, both in the gas phase and in continuum
solvent, pointing to the conclusion that epoxides 2 are
preferentially attacked at the unhindered carbon atom
C3.16,17 When compared to epoxides, the chemistry of
aziridines is further complicated by the presence of an
additional valency on the heteroatom, and since the mid-
1960s, aziridines have been classified as “activated” or
“nonactivated’” according to whether or not quaternization
toward an aziridinium intermediate is required for nucleophilic
ring-opening reactions.18 This classification is intimately related
to the nature of the N-substituent, i.e., its electron-withdrawing
or electron-donating properties. From experimental data, it is
clear that the nature of the N-substituent has a profound
influence on the reactivity of 2-(bromomethyl)aziridines upon
treatment with different nucleophiles. However, up to now, no
detailed theoretical and experimental evaluation of the
reactivity of aziridines 1 with respect to nucleophiles in terms
of the underlying mechanistic pathways has been performed.
One of the most striking features of 2-(bromomethyl)-
aziridines 3 is their general reactivity toward nucleophiles.
Independent of the nature of the N-substituent, the
corresponding 2-substituted aziridines 4 are isolated upon
treatment of aziridines 3 with 1 equiv of a large variety of
different nucleophiles (Scheme 2). Heteroatom-centered as
Sound experimental evidence has been provided in the
literature with respect to the use of the similar 2-
(tosyloxymethyl)aziridines in reaction with nucleophiles. It
was demonstrated that the activated 1-tosyl-2-
(tosyloxymethyl)aziridine undergoes selective ring opening at
the less hindered carbon atom of the aziridine moiety upon
treatment with organocuprates, immediately followed by ring
closure with simultaneous displacement of the tosylate
(pathway b, Scheme 3).21 Furthermore, for N,O-bis-
(diphenylphosphinyl)-2-(hydroxymethyl)aziridine, an aziridine
with a different electron-withdrawing substituent at nitrogen,
similar results have been published.22 On the other hand, it has
been proven that the substitution of nonactivated 1-(α-
methylbenzyl)-2-(tosyloxymethyl)aziridine with sodium meth-
oxide can proceed via direct SN2 substitution with retention of
configuration, leaving the aziridine moiety untouched (pathway
a, Scheme 3) or that it can yield two different isomers, coming
from both pathways a and b in Scheme 3, depending on the
orientation of the tosyloxymethyl group with respect to the N-
substituent.23,24 Therefore, the stereochemical outcome of
substitution reactions of chiral 2-(bromomethyl)aziridines is
influenced by many factors, such as the identity and strength of
the nucleophile, the leaving group capacity, the electron-
withdrawing or electron-donating character of the N-
substituent, and even the stereochemistry of the aziridine.
In the present paper, the reactivity of the activated 2-
bromomethyl-1-tosylaziridine 9 and the nonactivated 1-benzyl-
2-(bromomethyl)aziridine 10 (Scheme 4) with respect to the
nucleophile sodium methoxide in methanol will be investigated
for the first time from a theoretical point of view. Furthermore,
their reactivity will be compared to the reactivity of their
oxygen analogue epibromohydrin 11. For all three species
under study, the propensity of methoxide for nucleophilic
attack at the three different electrophilic carbon atoms will be
assessed. Next, the reactivity of the activated aziridine 9 and the
nonactivated 1-(α-methylbenzyl)-2-(bromomethyl)aziridine 12
with regard to sodium methoxide in methanol is evaluated
experimentally, particularly focusing on the competition
between direct nucleophilic displacement of bromide (pathway
a, Scheme 3) and a ring-opening/ring-closure process (pathway
b, Scheme 3), to verify the theoretical findings. Furthermore,
the reactivity of the activated 1-tosyl-2-(tosyloxymethyl)-
aziridine 13 with regard to sodium methoxide in methanol is
investigated experimentally in order to compare its reactive
behavior with that of its brominated counterpart 9.
Scheme 2. Treatment of 2-(Bromomethyl)aziridines 3 with
Nucleophiles
well as carbon-centered nucleophiles can be applied successfully
in this reaction, enabling the synthesis of a wide range of
aziridines 4 as interesting synthons for further elaboration.19,20
Although the net reaction comprises displacement of
bromide by a nucleophile, the underlying mechanism of this
transformation requires a more detailed investigation. As
depicted in Scheme 3, chiral (2S)-2-(bromomethyl)aziridines
Scheme 3. Nucleophilic Substitution of (2S)-2-
(Bromomethyl)aziridines 5 via Direct Substitution
(Pathway a) and Ring Opening/Ring Closure at the
Unhindered Aziridine Carbon Atom (Pathway b)
RESULTS AND DISCUSSION
■
1. Theoretical Results. In this section, the reactivity of
activated aziridine 9, nonactivated aziridine 10, and epoxide 11
(Scheme 4) with respect to sodium methoxide will be
investigated theoretically. In order to discriminate between
the three electrophilic carbon atoms in these three-membered
rings, the nucleophilic attack of methoxide at all electrophilic
centers (pathways a, b and c, Scheme 5) is investigated by
means of density functional theory (DFT) calculations within
the supermolecule approach (vide infra).
5 can undergo a direct SN2 nucleophilic substitution at the
halogenated carbon atom toward (2S)-aziridines 6 (pathway a
retention of configuration) or, alternatively, the nucleophile can
attack the unsubstituted aziridine carbon atom resulting in a
ring-opened intermediate 7, which is prone to undergo ring
closure toward the substituted (2R)-aziridines 8 (pathway b
inversion of configuration).
Attack of methoxide at the brominated carbon atom of
aziridines 3 affords the corresponding 2-(methoxymethyl)-
8699
dx.doi.org/10.1021/jo201255z|J. Org. Chem. 2011, 76, 8698−8709

The Journal of Organic Chemistry
Article
Scheme 4. 2-Bromomethyl-1-tosylaziridine 9, 1-Benzyl-2-(bromomethyl)aziridine 10, Epibromohydrin 11, 1-(α-Methylbenzyl)-
2-(bromomethyl)aziridine 12, and 1-Tosyl-2-(tosyloxymethyl)aziridine 13
Scheme 5. Nucleophilic Attack of Methoxide at All Electrophilic Centers of Aziridines 3 and Epoxide 11
aziridines 14 through direct nucleophilic displacement (path-
way a), whereas attack at the less hindered aziridine carbon
atom results in intermediate β-halo amines 15 via ring opening
(pathway b), which are readily converted into 2-
(methoxymethyl)aziridines 17 via ring closure (pathway d).
Alternatively, ring opening at the substituted aziridine carbon
atom would afford intermediate γ-halo amines 16 (pathway c),
which might be converted to 3-methoxyazetidines 18 via ring
closure (pathway e), although this behavior has not been
observed in experimental studies so far. On the other hand, the
rearrangement of aziridines 3 (R = Ts) to 3-aminoazetidines is
known in the literature, albeit based on a different mechanistic
pathway.25 The same pathways can be considered for epoxide
11, resulting in 2-(methoxymethyl)epoxides 19 or 22 via
pathways a and b, respectively, and 3-methoxyoxetane 23 via
pathway c.
Computational Methodology. The B3LYP/6-31++G(d,p)
level of theory was used for geometry optimizations.26,27
Stationary points were characterized as minima (ground states)
or first-order saddle points (transition states) via frequency
calculations. IRC (intrinsic reaction coordinate) calcula-
tions28−30 followed by full geometry optimizations were used
to verify the corresponding reactant and product complexes.
The B3LYP functional has been proven to produce good
geometries but is less accurate for energy calculations.31
Therefore energies were refined with the following methods:
MPW1B95,32 which has been shown to be efficient for
aziridines,33−37 and BMK38 and MPW1K,39,40 which are
known for their good performance for describing kinetics of
reactions in general. Previous studies have demonstrated the
differences between theory and experiment for sulfur systems
and concluded that the addition of d and f polarization
functions on the sulfur basis set gives reliable energies.41−43
The MPW1B95, BMK, and MPW1K functionals were used
with a combined basis set consisting of the 6-311+G(3df,2p)
basis set for sulfur and the 6-31++G(d,p) basis set for all other
atoms, since this method was shown to be adequate for
calculations on sulfur-containing systems.44 All thermal free
energy corrections reported were taken from B3LYP/6-31+
+G(d,p) optimizations at 1 atm and 298 K. All computations
were performed with the Gaussian 03 and Gaussian 09 program
packages.45,46
Since the reactions under study take place in methanol,
which has the potential to form hydrogen bonds with the
reactive substrate, it is essential to investigate the influence of
the solvent molecular environment on the reactions. The
simplest method consists of using a continuum model,47−50
where the solvent is modeled as a continuous medium
characterized by a dielectric constant. However, in this case
where explicit hydrogen bonds are possible, this methology is
not preferred and a discrete solvent model (also called
supermolecule approach or microsolvation),34−37,44,51 in
which discrete solvent molecules are placed around the
chemically active species are expected to be more reliable.
Ideally, the reactive species could be simulated by means of
molecular dynamics calculations in a solvent box;52,53 however,
this approach is computationally very expensive and cannot be
routinely applied. Furthermore, to account for potential long-
range interactions, this “supermolecule” can be placed in a
dielectric continuum, leading to a mixed implicit/explicit
model.54−56 Previous studies have shown that explicit solvation
alone can give reliable results provided that it is used with
sufficient care.36 First of all, the number and orientation of the
solvent molecules needs to be correctly chosen. For the
8700
dx.doi.org/10.1021/jo201255z|J. Org. Chem. 2011, 76, 8698−8709

The Journal of Organic Chemistry
Article
Figure 1. Transition state geometries for nucleophilic attack of methoxide at the electrophilic centers of 9−11, without explicit solvent molecules.
B3LYP/6-31++G(d,p) optimized structures. Some critical distances are given in Å. TS denotes Transition State. a, b, c denote pathways shown in
Scheme 5.
reactions under study, where a protic solvent is used, the
mobility of the solvent molecules is already restricted by H-
bonds, but still a variety of orientations is possible. We will
explicitly address the influence of the number and orientation
of the methanol molecules on the reactive pathways in the
Theoretical Results section. Another point of controversy is
related to the choice of the reference state for the reactants.
One could take either the separate reactants or the reactant
complexes. Since the energy of the reactant complexes is
usually considerably lower than the total energy of separate
reactants, due to favorable complexation, activation barriers
were calculated from the reactant complexes and not from the
separate reactants. Although, we will show that for the present
study, the particular choice of the reference state does not alter
the relative difference in activation barriers for the competitive
pathways of nucleophilic attack. As we started from the reactant
complexes, it was not necessary to take into account BSSE
(basis set superposition error) corrections.57
Nucleophilic Attack of Methoxide at the Electrophilic
Centers of 3-Membered Heterocycles. Analysis without
Explicit Solvent Molecules. A thorough conformational
analysis was performed on the three species under study to
identify the most plausible conformers, which were then used
to model the methoxide attack at the electrophilic centers.
Furthermore, nitrogen inversion barriers were calculated for
aziridines 9 and 10. The barriers (ΔG⧧ = 45.8 and 66.3 kJ/mol
8701
dx.doi.org/10.1021/jo201255z|J. Org. Chem. 2011, 76, 8698−8709

The Journal of Organic Chemistry
Article
(MPW1B95) for 9 and 10, respectively) are too low to allow
isolation of the individual invertomers and rapid inversion is
expected, leading to thermodynamic equilibration in favor of
the more stable trans invertomers. The conformational results
and the investigation of the inversion barriers are given on
pages S11−S14 and S15−S17 of the Supporting Information,
respectively.
Transition state geometries and free energy profiles for attack
of methoxide at the three electrophilic centers (Scheme 5) of 9,
10, and 11 without explicit solvent molecules are shown in
Figures 1−4. In Figures 2 and 4, the subsequent ring-closing
Figure 4. Gibbs free energy profiles (MPW1K/6-31++G(d,p)//
B3LYP/6-31++G(d,p)) for nucleophilic attack of methoxide on the
electrophilic centers of epoxide 11, without explicit solvent molecules.
a−e denote pathways shown in Scheme 5. Some critical distances are
given in Å. Free energies in kJ/mol at 298 K and 1 atm.
reactions for the corresponding intermediates 15, 16 and 20, 21
derived from activated aziridine 9 and epoxide 11, respectively,
are also taken up for completeness (pathways d and e, Scheme
5). As will be shown below, these ring-closing reactions are less
important for nonactivated aziridine 10, since free energies of
activation for ring opening are too high.
Energy refinements with the MPW1B95, BMK, and MPW1K
functionals give the same trends in terms of the selectivity of
methoxide attack at the electrophilic centers of the three-
membered rings under study (see Table S6 of the Supporting
Information for energies at all levels of theory). Barriers
indicated in the following text are at the MPW1K level of
theory. The calculations reveal a clear preference for pathway a
(direct substitution at the brominated carbon atom by
methoxide) for 9−11. The difference in free energy of
activation between pathway a and pathway b (ring opening at
the less hindered carbon atom) is much smaller for activated
aziridine 9 and epoxide 11 than for nonactivated aziridine 10
(ΔΔG⧧ = 16.3 and 28.6 versus 66.2 kJ/mol). In addition, the
difference in free energy of activation between pathway b and
pathway c (ring opening at the substituted carbon atom) is
much smaller for activated aziridine 9 than for nonactivated
aziridine 10 and epoxide 11 (ΔΔG⧧ = 9.2 versus 20.9 and 19.9
kJ/mol). Ont he basis of these gas-phase calculations, pathway
c cannot be excluded for 9, although this behavior has never
been observed in experimental studies.
Figure 2. Gibbs free energy profiles (MPW1K/6-31++G(d,p)//
B3LYP/6-31++G(d,p)) for nucleophilic attack of methoxide on the
electrophilic centers of aziridine 9, without explicit solvent molecules.
6-311+G(3df,2p) basis set on the sulfur atom. T denotes N-substituent
= tosyl. a−e denote pathways shown in Scheme 5. Some critical
distances are given in Å. Free energies in kJ/mol at 298 K and 1 atm.
Solvent Approach: Explicit Solvent Molecules. Nucleo-
philic substitution reactions are known to be influenced by the
solvent environment.47,58 Therefore, the gas-phase results are
extended toward a discrete solvent approach, as explained in
the Computational Methodology section. The number of
solvent molecules needed to accurately describe the chemical
problem at hand was determined by studying the convergence
behavior of the energy of solvation in terms of a systematically
increasing number of solvent molecules. The number of
methanol molecules on the methoxide oxygen atom, the
bromine atom, and the ring heteroatoms were varied. A
supermolecule model with five explicit methanol molecules, as
shown in Figure 5 for the prereactive complex of aziridine 10,
was deduced (see pages S19−S22 of the Supporting
Figure 3. Gibbs free energy profiles (MPW1K/6-31++G(d,p)//
B3LYP/6-31++G(d,p)) for nucleophilic attack of methoxide on the
electrophilic centers of aziridine 10, without explicit solvent molecules.
B denotes N-substituent = benzyl. a−e denote pathways shown in
Scheme 5. Some critical distances are given in Å. Free energies in kJ/
mol at 298 K and 1 atm.
8702
dx.doi.org/10.1021/jo201255z|J. Org. Chem. 2011, 76, 8698−8709

The Journal of Organic Chemistry
Article
One concern regarding this approach is whether one selected
minimum on the free energy surface is capable of accurately
describing the behavior of the substrate in its true molecular
environment. We investigated this by scanning the free energy
surface for other minima of the supermolecule having a fixed
number of coordinating methanol molecules. For the protic
solvent under study, explicit hydrogen bonds are formed with
the substrate and thus their mobility with respect to the
substrate is a priori limited. The free energy results of the
alternative conformers are added in Table S7 of the Supporting
Information. The results show that energy variations among
various conformers are small (a maximum of 10 kJ/mol)
compared to the free energies of activation investigated in this
study and more importantly, compared to the difference in
energies of activation for nucleophilic attack of the methoxide
on the various electrophilic centers (a minimum of 20 kJ/mol).
Transition state geometries and free energy profiles for attack
of methoxide at the three electrophilic centers of 9−11 with
five explicit methanol molecules are shown in Figures 7−10.
Typical methoxide O···HOMe, Br···HOMe, epoxide
O···HOMe, and aziridine N···HOMe distances are around
1.6, 2.5, 1.8, and 1.8 Å, respectively. Transition state and
reactant critical distances (Å) and bond elongation percentages
for nucleophilic attack of methoxide at the electrophilic centers
of 9−11, with and without explicit solvent molecules, are
shown in Table 1. Critical distances are different for the
transition states with explicit solvent molecules compared to
the gas-phase results for all three species under study and for all
three different nucleophilic attack trajectories. Solvated
transition states are more “product-like”, which can be seen
by the extent of displacement of bromide (d(C4−Br) in TS-a)
and ring opening (d(C3−N) in TS-b and d(C2−N) in TS-c).
Bond elongation percentages were calculated with respect to
reactants 9−11 in the gas phase and are higher by
approximately 5% for transition states with explicit solvent
molecules versus transition states from gas-phase calculations.
Solvation has significantly changed the landscape of the
energy profiles (Figures 8−10 versus Figures 2−4). All
activation energies have increased in the solvated systems.
This is due to the stabilization of methoxide. The selectivity for
attack at the various electrophilic centers is qualitatively the
same at different electronic levels of theory; for completeness,
these results are given in Table S8 of the Supporting
Information. Barriers indicated in the following text are at the
MPW1K level of theory. Pathway a (direct substitution at the
brominated carbon atom by methoxide) is favored for all three
species under study (ΔG⧧ = 45.5, 38.1, and 53.9 kJ/mol for 9,
10 and 11, respectively). Furthermore, the difference in free
energy of activation between pathway a and pathway b (ring
opening at the less hindered aziridine carbon atom) is much
smaller for activated aziridine 9 and epoxide 11 than for
nonactivated aziridine 10 (ΔΔG⧧ = 17.5 and 26.1 versus 78.8
kJ/mol), making pathway b feasible for 9 and 11, and not for
10. Furthermore, activation energies for the subsequent ring-
closing reactions (pathways d and e) make these reactions
feasible (ΔG⧧ = 44.2 and 33.9 kJ/mol for 9 and 11,
respectively). Calculations with explicit solvent molecules
reveal a significant difference in free energy of activation
between pathway b and pathway c (ring opening at the
substituted aziridine carbon atom) for both activated aziridine 9
and epoxide 11 (ΔΔG⧧ = 15.6 and 18.9 kJ/mol, respectively).
This fact indisputably shows the necessity of taking into
account explicit solvation to obtain the correct reaction profiles.
Figure 5. Prereactive complex of aziridine 10 with methoxide, solvated
by five explicit methanol molecules. B3LYP/6-31++G(d,p) optimized
structure. Some critical distances are given in angstroms.
Information for the complete investigation with figures). The
bromine atoms and the methoxide anion are solvated with two
methanol molecules, whereas the aziridine nitrogen or epoxide
oxygen is solvated with one methanol molecule.
In parallel with this preliminary investigation on the number
of solvating molecules, the selected supermolecule was further
investigated by studying the Gibbs free energies of activation
for nucleophilic attack of methoxide at the various electrophilic
centers of aziridine 9 with a different number of explicit
methanol molecules. The results of these calculations are
shown in Figure 6 and give further evidence for the previously
Figure 6. Relative Gibbs free energies of activation (ΔG⧧) for
nucleophilic attack of methoxide at the electrophilic centers of
aziridine 9, with diffent amounts of explicit methanol molecules.
MPW1K/6-31++G(d,p)//B3LYP/6-31++G(d,p) energies with 6-
311+G(3df,2p) basis set on the sulfur atom. Free energies at 298 K
and 1 atm.
proposed supermolecule. The Gibbs free energies of activation
nicely converge by adding an extra methanol molecule on the
bromine atom and the nitrogen atom to the supermolecule with
three methanol molecules (three on the methoxide oxygen and
one on the bromine atom).
8703
dx.doi.org/10.1021/jo201255z|J. Org. Chem. 2011, 76, 8698−8709

The Journal of Organic Chemistry
Article
Figure 7. Transition state geometries for nucleophilic attack of methoxide at the electrophilic centers of 9−11, with five explicit methanol molecules.
B3LYP/6-31++G(d,p)-optimized structures. Some critical distances are given in angstroms. TS denotes transition state. a−c denote pathways shown
in Scheme 5. 5MeOH denotes five explicit methanol molecules.
For transparency, relative Gibbs free energies of activation
(ΔG⧧) and differences in relative Gibbs free energies of
activation (ΔΔG⧧) for nucleophilic attack of methoxide at the
electrophilic centers of 9−11 with and without explicit solvent
molecules are summarized in Table 2. In addition, Table S9 of
the Supporting Information shows the individual contributions
of the enthalpy and entropy to the Gibbs free energies. These
results demonstrate that the major contribution to the
activation barriers originates from enthalpy, entropy contribu-
tions are merely the same. This is to be expected as the selected
supermolecule is very similar in all reactive pathways. In the
present study, the particular choice of the reference state does
not alter the difference in activation barriers for the competitive
pathways of nucleophilic attack. Furthermore, Table S10 of the
Supporting Information shows a comparison between the
separate reactants and the reactant complexes as the reference
state for the reactants. For the present study the particular
choice does not alter the results on the competitive pathways
for nucleophilic attack and since the energy of the reactant
complexes is usually considerably lower than the total energy of
separate reactants due to favorable complexation, activation
barriers were calculated from the reactant complexes and not
from the separate reactants.
Both gas-phase calculations and calculations with explicit
solvent molecules reveal a clear preference for direct
substitution at the brominated carbon atom (pathway a) for
all species under study. This is in contrast with the expectations
for activated aziridine 9, since experiments have shown that the
8704
dx.doi.org/10.1021/jo201255z|J. Org. Chem. 2011, 76, 8698−8709

The Journal of Organic Chemistry
Article
Figure 8. Gibbs free energy profiles (MPW1K/6-31++G(d,p)//B3LYP/6-31++G(d,p)) for nucleophilic attack of methoxide on the electrophilic
centers of aziridine 9, with five explicit methanol molecules. 6-311+G(3df,2p) basis set on the sulfur atom. T denotes N-substituent = tosyl. a−e
denote pathways shown in Scheme 5. 5MeOH denotes five explicit methanol molecules. Some critical distances are given in angstroms. Free
energies in kJ/mol at 298 K and 1 atm.
Figure 10. Gibbs free energy profiles (MPW1K/6-31++G(d,p)//
B3LYP/6-31++G(d,p)) for nucleophilic attack of methoxide on the
electrophilic centers of epoxide 11 with five explicit methanol
molecules. a−e denote pathways shown in Scheme 5. Some critical
distances are given in angstroms. 5MeOH denotes five explicit
methanol molecules. Free energies in kJ/mol at 298 K and 1 atm.
B3LYP/6-31++G(d,p)) for nucleophilic attack of methoxide on the
Figure 9. Gibbs free energy profiles (MPW1K/6-31++G(d,p)//
electrophilic centers of aziridine 10 with five explicit methanol
2. Experimental Results. In addition to computational
studies, the reactivity of activated and nonactivated 2-
(bromomethyl)aziridines with respect to sodium methoxide
in methanol was evaluated experimentally, particularly focusing
on the competition between ring opening/ring closure and
direct nucleophilic displacement of bromide, in order to
confirm theoretical predictions discussed earlier.
At first, racemic 2-bromomethyl-1-tosylaziridine rac-9 was
prepared from allylamine 24 according to a literature protocol
(involving consecutive treatment with HBr, Br2, and TsCl in
water)19 and transformed into the corresponding racemic 2-
(methoxymethyl)aziridine rac-25 upon treatment with sodium
methoxide in methanol (Scheme 6).19 Both rac-9 and rac-25
were then used to study the discrimination between both
enantiomeric pairs by means of 1H NMR analysis (CDCl3), for
which the use of 5 equiv of the chiral shift reagent Pirkle
alcohol resulted in the observation of distinct diastereotopic
proton signals pertaining to one of the enantiotopic aziridine
protons [(HtransCH)N].
molecules. B denotes N-substituent = benzyl. a−e denote pathways
shown in Scheme 5. 5MeOH denotes five explicit methanol molecules.
Some critical distances are given in angstroms. Free energies in kJ/mol
at 298 K and 1 atm.
closely related activated 1-tosyl-2-(tosyloxymethyl)aziridine
undergoes a selective ring-opening/ring-closure process upon
treatment with cuprates (pathway b).21 Although former ab
initio calculations of the reactivity of epihalohydrins with
respect to hydroxide and water pointed to the conclusion that
epichlorohydrins are preferentially opened at the unhindered
epoxide carbon atom (pathway b),16,17 epoxide 11 is known to
behave differently in that respect. All energies of activation have
increased in the solvated systems because of stabilization of the
reactive methoxide. On the basis of the calculations with
explicit solvent molecules, ring-opening reactions are feasible
for activated aziridine 9 and epoxide 11, although ring opening
at the substituted aziridine carbon (pathway c) of 9 and at both
epoxide carbon atoms of 11 are less probable since they are
quite high in energy. Ring-opening reactions are not feasible for
nonactivated aziridine 10, as initially expected.
Second, the enantiomerically pure 2(S)-(bromomethyl)-
aziridine 2(S)-9 was prepared through conversion of 2(R)-
(tosyloxymethyl)aziridine 2(R)-13, which was synthesized by
8705
dx.doi.org/10.1021/jo201255z|J. Org. Chem. 2011, 76, 8698−8709

The Journal of Organic Chemistry
Article
Table 1. Transition State and Reactant Critical Distances (Å) and Bond Elongation Percentages for Nucleophilic Attack of
a
Methoxide at the Electrophilic Centers of 9−11, with and without Explicit Solvent Molecules
a
Bond elongation percentages P (%) = (dTS − dreactant)/(dreactant) in bold.
Table 2. Summarizing Table: Relative Gibbs Free Energies of Activation (ΔG⧧) and Differences in Relative Gibbs Free Energies
of Activation (ΔΔG⧧) for Nucleophilic Attack of Methoxide at the Electrophilic Centers of 9−11, with and without Explicit
a c
−
Solvent Molecules
a
b
MPW1K/6-31++G(d,p)//B3LYP/6-31++G(d,p) energies with 6-311+G(3df,2p) basis set on the sulfur atom. Free energies in kJ/mol at 298 K
c
and 1 atm. ΔΔG⧧ = ΔG⧧ − ΔG⧧a; ΔΔG⧧ = ΔG⧧ − ΔG⧧
.
b
b−a
b
c−b
c
sodium borohydride in the presence of lithium chloride to
produce β-amino alcohol 28.59 Compound 28 was converted
into the corresponding aziridine 29 in 86% yield via a
Mitsunobu reaction using diisopropyl azodicarboxylate
(DIAD) and triphenylphosphine in THF. Further desilylation
of 29 upon treatment with tetra-n-butylammonium fluoride
(TBAF) in THF followed by tosylation of the alcohol with
tosylchloride in CH2Cl2 produced aziridine 2(R)-13 in 63%
yield over two steps (Scheme 7). Finally, 2(R)-13 was treated
with tetra-n-butylammonium bromide (TBAB) in acetonitrile,
affording the desired enantiomerically pure aziridine 2(S)-9 in
excellent yield (96%) (Scheme 7) through a ring-opening/ring-
closure protocol based on the reactivity profile of 13 as
Scheme 6. Synthesis of rac-2-Methoxymethyl-1-tosylaziridine
rac-25
means of a modified literature protocol (Scheme 7).21 To that
end, readily available (S)-serine methyl ester 26 was N-
tosylated, followed by protection of the free hydroxyl group
with tert-butyldimethylsilyl chloride toward α-N-tosylamino
ester 27 in 81% yield. Reduction of 27 was performed using
8706
dx.doi.org/10.1021/jo201255z|J. Org. Chem. 2011, 76, 8698−8709

The Journal of Organic Chemistry
Article
Scheme 7. Synthesis of 2(S)-Bromomethyl-1-tosylaziridine 2(S)-9
described in the literature19 and supported by 1H NMR analysis
using Pirkle alcohol.
In the next part, the reactivity of 2-(bromomethyl)aziridine 9
toward methoxide was assessed and compared with that of the
corresponding 2-(tosyloxymethyl)aziridine 13. First, 2(R)-13
was converted into 2(S)-25 upon treatment with NaOMe in
methanol (1 M) (Scheme 8). No racemization of the
The above-described experimental results verified the rather
unexpected conclusion from theoretical calculations that
competition between ring opening/ring closure and direct
nucleophilic displacement of bromide takes place upon
treatment of aziridine 9 with nucleophiles, whereas a
straightforward ring-opening/ring-closure process with inver-
sion of configuration at the asymmetric center occurs when
aziridine 13 is used instead.
In light of these surprising results, efforts were devoted to
evaluate the reactivity of nonactivated 2-(bromomethyl)-
aziridines with respect to sodium methoxide as well. To
discriminate between the two possible routes toward the
corresponding 2-(methoxymethyl)aziridine, i.e., direct bromide
displacement versus a ring-opening/ring-closure protocol, the
reference (1R,2R)-2-methoxymethyl-1-(α-methylbenzyl)-
aziridine 31 was synthesized from (1R,2R)-1-(α-methylben-
zyl)-2-(hydroxymethyl)aziridine 30 by a Williamson ether
synthesis using sodium hydride and iodomethane in THF in
94% yield (Scheme 9, method a).23 The alternative approach
Scheme 8. Preparation of 2-Methoxymethyl-1-tosylaziridine
25
Scheme 9. Synthesis of (1R,2R)-2-Methoxymethyl-1-(α-
methylbenzyl)aziridine 31
asymmetric aziridine carbon atom was observed upon 1H
NMR analysis using Pirkle alcohol, thus supporting the
reactivity profile of 13 with regard to nucleophiles (i.e.,
exclusively via a ring-opening/ring-closure protocol) as
described in the literature for organocuprates.21 Apparently,
also the oxygen nucleophile methoxide exhibits the same
selectivity. It should be mentioned that the formation 1,3-
dimethoxy-N-tosyl-2-propylamine was observed as well due to a
second addition of methoxide to the initially formed 2(S)-25.
A similar reaction was performed on 2(S)-9, providing a
scalemic mixture of 2(S)-25 and 2(R)-25 (er = 42:58 or vice
versa) after treatment with sodium methoxide in methanol (1
1
M) (Scheme 8), which was evidenced by H NMR analysis
using Pirkle alcohol. Additionally, 2(S)-9 was treated with
sodium methoxide in methanol at room temperature using
different concentrations, but no significant changes regarding
the enantiomeric ratio were observed (2 M NaOMe in MeOH:
er = 43:57; 0.5 M NaOMe in MeOH: er = 44:56). Again, the
presence of 1,3-dimethoxy-N-tosyl-2-propylamine was observed
in the reaction mixtures, and higher amounts of this side
product were formed when a higher concentration of sodium
methoxide was used.
for the synthesis of 31 commenced with tosylation of 30 using
tosyl chloride in CH2Cl2 in the presence of triethylamine and a
catalytic amount of DMAP, furnishing (1R,2R)-1-(α-methyl-
benzyl)-2-(tosyloxymethyl)aziridine 3212 in an excellent yield
8707
dx.doi.org/10.1021/jo201255z|J. Org. Chem. 2011, 76, 8698−8709

The Journal of Organic Chemistry
Article
2-bromomethyl-1-(α-methylbenzyl)aziridine 34 (ratio ∼2/1, overall
yield 97%), which were separated by silica gel column chromatography
(petroleum ether/ethyl acetate 4:1).
(Scheme 9, method b). Compound 32 was then subjected to a
substitution reaction using tetraethylammonium bromide
(TEAB) in CH3CN, resulting in a mixture of (1R,2R)-1-(α-
methylbenzyl)-2-(bromomethyl)aziridine 33 and (1R,2S)-1-(α-
methylbenzyl)-2-(bromomethyl)aziridine 34 (ratio ∼2/1, over-
all yield 97%). This result showed that aziridine 32 underwent
both nucleophilic attack at the exocyclic methylene carbon with
displacement of the leaving group and attack at the less
hindered carbon atom of the aziridine moiety followed by ring
closure, which is in accordance with previous findings described
in the literature.24
After separation of both diastereomers by silica gel column
chromatography (petroleum ether/ethyl acetate 4:1), the major
isomer 33 was treated with an excess (11 equiv) of sodium
methoxide in methanol (1M) under reflux, resulting in a single
reaction product 31. The comparison of aziridines 31 obtained
via two different reaction pathways (a and b in Scheme 9) by
means of various techniques (1H NMR, 13C NMR, IR, and MS)
proved these compounds to be identical, showing that the last
step of method b occurs exclusively via direct nucleophilic
displacement of bromide by methoxide furnishing 2(R)-31 with
retention of configuration, whereas the epimeric 2(S)-31 would
have been formed if ring opening/ring closure had taken place.
In contrast to the rather unexpected results in the case of
activated aziridine 9, nonactivated aziridine 33 was shown to
exhibit a straightforward reactivity with respect to sodium
methoxide as initially expected.
(1R,2R)-2-(Bromomethyl)-1-(α-methylbenzyl)aziridine 33:
yield 65% (0.47 g); white solid; mp = 45.9−47.8 °C; Rf = 0.31
1
(petroleum ether/EtOAc 4:1); [α]28 = +15.5 (c = 0.40, CHCl3); H
D
NMR (300 MHz, CDCl3) δ 1.44 (3H, d, J = 6.6 Hz), 1.44 (1H, d, J =
6.6 Hz), 1.60 (1H, d, J = 2.8 Hz), 1.85−1.89 (1H, m), 2.42 (1H, q, J =
6.6 Hz), 3.17 and 3.44 (2H, 2 x dd, J = 10.5, 8.3, 5 Hz), 7.15−7.30
(5H, m); 13C NMR (75 MHz, ref = CDCl3) δ 23.7 (CH3), 35.3
(CH2), 35.9 (CH), 41.2 (CH2), 69.6 (CH), 126.9 (CH), 127.2 (CH),
128.5 (CH), 144.2 (C); IR (neat, cm−1) νmax = 3034, 2923, 2964,
2838, 1493, 1450, 1422, 1223, 1165, 958, 758, 702, 629; MS m/z 240/
2 (M+ + 1, 100); HRMS m/z (ESI) calcd for C11H15BrN [MH]+
240.0388, found 240.0380.
(1R,2S)-2-(Bromomethyl)-1-(α-methylbenzyl)aziridine 34:
yield 33% (0.24 g); light yellow oil; Rf = 0.51 (petroleum ether/
1
EtOAc 4:1); [α]28 = +80.2 (c = 0.55, CHCl3); H NMR (300 MHz,
D
CDCl3) δ 1.42 (3H, d, J = 6.6 Hz), 1.64 (1H, d, J = 6.1 Hz), 1.88 (1H,
d, J = 3.3 Hz), 1.80−1.90 (1H, m), 2.53 (1H, q, J = 6.6 Hz), 3.13 and
3.31 (2H, 2 x dd, J = 10.2, 6.9, 5.8 Hz), 7.24−7.37 (5H, m); 13C NMR
(75 MHz, ref = CDCl3) δ 23.1 (CH3), 34.8 (CH2), 35.5 (CH), 39.5
(CH2), 69.8 (CH), 127.0 (CH), 127.3 (CH), 128.4 (CH), 144.1 (C);
IR (neat, cm−1) νmax = 3026, 2968, 2926, 2837, 1493, 1448, 1352,
1240, 1222, 1069, 971, 755, 698, 640; MS m/z 240/2 (M+ + 1, 100).
ASSOCIATED CONTENT
* Supporting Information
■
S
Cartesian coordinates and energies of the optimized geometries
(B3LYP/6-31++G(d,p)) of ground states; Cartesian coordi-
nates, energies, imaginary and low frequencies of the optimized
geometries (B3LYP/6-31++G**) of transition states. This
material is available free of charge via the Internet at http://
pubs.acs.org.
CONCLUSIONS
■
The reactivity of activated and nonactivated 2-(bromomethyl)-
aziridines with regard to sodium methoxide has been evaluated
both theoretically and experimentally, pointing to the
conclusion that 1-benzyl-2-(bromomethyl)aziridines exclusively
undergo direct displacement of bromide, whereas for 2-
bromomethyl-1-tosylaziridines, competition between ring
opening/ring closure at the less hindered aziridine carbon
atom and direct displacement of the bromide was observed.
The formation of a mixture of (R)- and (S)-2-
(methoxymethyl)aziridine starting from an enantiopure 2-
bromomethyl-1-tosylaziridine stands in contrast to the behavior
of the corresponding 1-tosyl-2-(tosyloxymethyl)aziridine, which
undergoes a clean ring-opening/ring-closure process with
inversion of configuration at the asymmetric aziridine carbon
atom. In addition, we have critically evaluated the effect of
solvent environment on the computational results. We have
found that for the chemical problem at hand where the solvent
is able to make explicit hydrogen bonds with the reacting
substrate, the supermolecule approach gives a fair representa-
tion of the molecular environment. On the other hand, explicit
accounting for the methanol environment was found to be
essential to acquire an adequate representation of the free
energy surface and the competition between the various attack
modes of methoxide at the different electrophilic centers.
AUTHOR INFORMATION
Corresponding Author
*E-mail: saron.catak@UGent.be; norbert.dekimpe@UGent.be;
veronique.vanspeybroeck@UGent.be.
■
ACKNOWLEDGMENTS
■
This work was supported by the Research Foundation-Flanders
(FWO-Vlaanderen), the Agency for Innovation through
Science and Technology (IWT), and the Research Board of
Ghent University (BOF-GOA). The computational resources
(Stevin Supercomputer Infrastructure) and services used in this
work were provided by Ghent University. This work is
supported by the IAP-BELSPO program in the frame of IAP
6/27.
REFERENCES
■
(1) Lindstrom, U. M.; Somfai, P. Synthesis 1998, 109.
̈
(2) Zwanenburg, B.; ten Holte, P. Top. Curr. Chem. 2001, 94.
(3) Sweeney, J. B. Chem. Soc. Rev. 2002, 31, 247.
(4) Hu, X. E. Tetrahedron 2004, 60, 2701.
(5) Tanner, D. Angew. Chem., Int. Ed. Engl. 1994, 33, 599.
(6) Osborn, H. M. I.; Sweeney, J. Tetrahedron: Asymmetry 1997, 8,
EXPERIMENTAL SECTION
1693.
■
(7) McCoull, W.; Davis, F. A. Synthesis 2000, 1347.
(8) Watson, I. D. G.; Yu, L.; Yudin, A. K. Acc. Chem. Res. 2006, 39,
Synthesis of Chiral 2-Bromomethyl-1-(α-methylbenzyl)-
aziridines 33 and 34. To a solution of (1R,2R)-1-(α-methylben-
zyl)-2-(tosyloxymethyl)aziridine 3212 (0.99 g, 3 mmol) in MeCN (10
mL) was added tetraethylammonium bromide (2.30 g, 15 mmol, 5
equiv), and the reaction mixture was heated under reflux for 21 h.
Extraction with CH2Cl2 (3 × 15 mL), drying (MgSO4), filtration of
the drying agent, and evaporation of the solvent afforded a mixture of
(1R,2R)-2-bromomethyl-1-(α-methylbenzyl)aziridine 33 and (1R,2S)-
194.
(9) Singh, G. S.; D’hooghe, M.; De Kimpe, N. Chem. Rev. 2007, 107,
2080.
(10) Stankovic,
́
S.; D’hooghe, M.; Catak, S.; Waroquier, M.; Van
Speybroeck, V.; De Kimpe, N.; Ha, H.-J. Chem. Soc. Rev. 2011,
DOI:10.1039/c1cs15140a.
8708
dx.doi.org/10.1021/jo201255z|J. Org. Chem. 2011, 76, 8698−8709

The Journal of Organic Chemistry
Article
(11) D’hooghe, M.; Waterinckx, A.; De Kimpe, N. J. Org. Chem.
2005, 70, 227.
(53) Ensing, B.; De Vivo, M.; Liu, Z. W.; Moore, P.; Klein, M L. Acc.
Chem. Res. 2006, 39, 73.
(54) Pliego, J. R.; Riveros, J. M. J. Phys. Chem. A 2001, 105, 7241.
(55) Kelly, C. P.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. A 2006,
110, 2493.
(56) da Silva, E. F.; Svendsen, H. F.; Merz, K. M. J. Phys. Chem. A
2009, 113, 6404.
(57) Boys, S. F.; Bernardi, F. Mol. Phys. 1970, 19, 553.
(58) van Bochove, M. A.; Bickelhaupt, F. M. Eur. J. Org. Chem. 2008,
649.
(12) D’hooghe, M.; Van Speybroeck, V.; Waroquier, M.; De Kimpe,
N. Chem. Commun. 2006, 1554.
(13) Vervisch, K.; D’hooghe, M.; Tornroos, K. W.; De Kimpe, N.
̈
Org. Biomol. Chem. 2009, 7, 3271.
(14) Su, W.; Liu, C.; Shan, W. Synlett 2008, 725.
(15) Madhusudhan, G.; Om Reddy, G.; Rajesh, T.; Ramanatham, J.;
Dubey, P. K. Tetrahedron Lett. 2008, 49, 3060.
(16) Merrill, G. N. J. Phys. Org. Chem. 2007, 20, 19.
(17) Shields, E. S.; Merrill, G. N. J. Phys. Org. Chem. 2007, 20, 1058.
(18) Ham, G. E. J. Org. Chem. 1964, 29, 3052.
(59) Wipf, P.; Fritch, P. C. J. Org. Chem. 1994, 59, 4875.
(19) D’hooghe, M.; Rottiers, M.; Jolie, R.; De Kimpe, N. Synlett
2005, 931.
(20) D’hooghe, M.; De Kimpe, N. ARKIVOC 2008, No. ix, 6.
(21) Bergmeier, S. C.; Seth, P. P. J. Org. Chem. 1997, 62, 2671.
(22) Sweeney, J. B.; Cantrill, A. A. Tetrahedron 2003, 59, 3677.
(23) D’hooghe, M.; De Kimpe, N. Synlett 2004, 271.
(24) Han, S.-M.; Ma, S.-H.; Ha, H.-J.; Lee, W. K. Tetrahedron 2008,
64, 11110.
(25) Karikomi, M.; De Kimpe, N. Tetrahedron Lett. 2000, 41, 10295.
(26) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
(27) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(28) Hratchian, H. P.; Schlegel, H. B. J. Chem. Phys. 2004, 120, 9918.
(29) Hratchian, H. P.; Schlegel, H. B. J. Chem. Theory Comput. 2005,
1, 61.
(30) Fukui, K. Acc. Chem. Res. 1981, 14, 363.
(31) Izgorodina, E. I.; Coote, M. L. Chem. Phys. 2006, 324, 96−110.
(32) Zhao, Y.; Truhlar, D. G. J. Phys. Chem. A 2004, 108, 6908.
(33) Yun, S. Y.; Catak, S.; Lee, W. K.; D’hooghe, M.; De Kimpe, N.;
Van Speybroeck, V.; Waroquier, M.; Kim, Y.; Ha, H.-J. Chem.
Commun. 2009, 2508.
(34) Catak, S.; D’hooghe, M.; De Kimpe, N.; Waroquier, M.; Van
Speybroeck, V. J. Org. Chem. 2010, 75, 885.
(35) Catak, S.; D’hooghe, M.; Verstraelen, T.; Hemelsoet, K.; Van
Nieuwenhove, A.; Ha, H.-J.; Waroquier, M.; De Kimpe, N.; Van
Speybroeck, V. J. Org. Chem. 2010, 75, 4530.
(36) D’hooghe, M.; Catak, S.; Stankovic, S.; Waroquier, M.; Kim, Y.;
́
Ha, H.-J.; Van Speybroeck, V.; De Kimpe, N. Eur. J. Org. Chem. 2010,
4920.
́
(37) Stankovic, S.; Catak, S.; D’hooghe, M.; Goossens, H.; Tehrani,
K. A.; Bogaert, P.; Waroquier, M.; Van Speybroeck, V.; De Kimpe, N.
J. Org. Chem. 2011, 76, 2157.
(38) Boese, A. D.; Martin, J. M. L. J. Chem. Phys. 2004, 121, 3405.
(39) Lynch, B. J.; Fast, P. L.; Harris, M.; Truhlar, D. G. J. Phys. Chem.
A 2000, 104, 4811.
(40) Lynch, B. J.; Zhao, Y.; Truhlar, D. G. J. Phys. Chem. A 2003,
107, 1384.
(41) Bauschlicher, C. W.; Partridge, H. Chem. Phys. Lett. 1995, 240,
533.
(42) Bell, R. D.; Wilson, A. K. Chem. Phys. Lett. 2004, 394, 105.
(43) Yockel, S.; Wilson, A. K. Chem. Phys. Lett. 2006, 429, 645.
(44) Hermosilla, L.; Catak, S.; Van Speybroeck, V.; Waroquier, M.;
Vandenbergh, J.; Motmans, F.; Adriaensens, P.; Lutsen, L.; Cleij, T.;
Vanderzande, D. Macromolecules 2010, 43, 7424.
(45) Frisch, M. J.et al. Gaussian 03, Revision C.02; Gaussian, Inc.:
Wallingford, CT, 2004.
(46) Frisch, M. J.et al. Gaussian 09, Revision A.02; Gaussian, Inc.:
Wallingford, CT, 2009.
(47) Cramer, C. J.; Truhlar, D. G. Solvent Effects and Chemical
Reactivity; Kluwer: Dordrecht, 1996.
(48) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995.
(49) Cossi, M.; Rega, N.; Scalmani, G.; Barone, V. J. Comput. Chem.
2003, 24, 669.
(50) Takano, Y; Houk, K. N. J. Chem. Theory Comput. 2004, 70.
(51) Kelly, C. P.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. A 2006,
110, 2493.
(52) Leach, A. R. Molecular Modelling: Principles and Applications,
2nd ed.; Pearson Education: Harlow, 2001.
8709
dx.doi.org/10.1021/jo201255z|J. Org. Chem. 2011, 76, 8698−8709