Systematic study of halide-induced ring opening of 2-substituted aziridinium salts and theoretical rationalization of the reaction pathways

Article abstract of DOI:10.1002/ejoc.201000486

The ring-opening reactions of 2-alkyl-substituted l,1-bis(arylmethyl)- and 1-methyl-1-(l-phenylethyl)aziridinium salts with fluoride, chloride, bromide and iodide in acetonitrile have been, evaluated for the first: time in a systematic way. The reactions

Full text of DOI:10.1002/ejoc.201000486

FULL PAPER
DOI: 10.1002/ejoc.201000486
Systematic Study of Halide-Induced Ring Opening of 2-Substituted
Aziridinium Salts and Theoretical Rationalization of the Reaction Pathways
[a]
[b,c]
Matthias D’hooghe,[a] Saron Catak,[b,c] Sonja Stankovic, Michel Waroquier,
Yongeun Kim,[d] Hyun-Joon Ha,*[d] Veronique Van Speybroeck,*[b,c] and
Norbert De Kimpe*[a]
´
Dedicated to Professor Carmen Nájera on the occasion of her 60th birthday
Keywords: Nitrogen heterocycles / Aziridines / Halides / Regioselectivity / Reaction mechanisms / Density functional
calculations
The ring-opening reactions of 2-alkyl-substituted 1,1-bis(aryl-
methyl)- and 1-methyl-1-(1-phenylethyl)aziridinium salts
with fluoride, chloride, bromide and iodide in acetonitrile
have been evaluated for the first time in a systematic way.
The reactions with fluoride afforded regioisomeric mixtures
of primary and secondary fluorides, whereas secondary β-
chloro, β-bromo and β-iodo amines were obtained as the sole
reaction products from the corresponding halides by regio-
specific ring opening at the substituted position. Both experi-
mental and computational results revealed that the reaction
outcomes in the cases of chloride, bromide and iodide were
dictated by product stability through thermodynamic control
involving rearrangement of the initially formed primary ha-
lides to the more stable secondary halides. The ring opening
of the same aziridinium salts with fluoride, however, was
shown to be mediated by steric interactions (kinetic control),
with the corresponding primary β-fluoro amines being ob-
tained as the main reaction products. Only for 2-acylaziridin-
ium ions was the reaction outcome shown to be under full
substrate control, affording secondary β-fluoro, β-chloro, β-
bromo and β-iodo amines through exclusive attack at the ac-
tivated α-carbonyl carbon atom.
reactivity and applications of non-activated aziridines are
different and often complementary to those of activated
aziridines and epoxides, providing interesting opportunities
for the selective synthesis of a variety of functionalized
amines. The most commonly applied methodology in this
respect involves the formation of highly electrophilic azirid-
inium intermediates by N-alkylation, N-acylation, N-pro-
tonation or N-complexation with Lewis acids, which can
then easily be opened by different types of nucleophiles.
The ring opening of aziridinium salts with halides is a
convenient approach towards β-halo amines, which are gen-
erally recognized as useful building blocks in organic chem-
istry[2] and valuable targets in medicinal chemistry (nitrogen
mustards, chemotherapy agents).[3] If 2-substituted azirid-
ines are used for the synthesis of the corresponding β-halo
amines, the issue of regioselectivity in the ring opening of
the intermediate aziridinium salts becomes important as
two regioisomeric β-halo amines can be obtained. As de-
picted in Scheme 1, the ring opening of aziridinium salts 1
can occur at either the unsubstituted (path a) or substituted
aziridine carbon atom (path b), leading either to primary
halides 2 (path a) or secondary halides 3 (path b).
Introduction
The aziridine moiety represents one of the most valuable
three-membered ring systems in organic chemistry[1] and
the regiocontrolled ring opening of C-substituted aziridines
is a powerful approach towards the preparation of a large
variety of functionalized nitrogen-containing target com-
pounds. The ring opening of activated aziridines, that is,
aziridines bearing an electron-withdrawing group at the ni-
trogen atom, has been widely reported in the literature.[1d]
However, non-activated aziridines, that is, aziridines with
an electron-donating substituent at the nitrogen atom, have
to be activated prior to ring opening, but so far this has
only been evaluated to a limited extent. Nevertheless, the
[a] Department of Organic Chemistry, Faculty of Bioscience
Engineering, Ghent University,
Coupure Links 653, 9000 Ghent, Belgium
[b] Center for Molecular Modeling, Ghent University,
Technologiepark 903, 9052 Zwijnaarde, Belgium
[c] QCMM-Alliance Ghent-Brussels,
Ghent, Belgium
[d] Department of Chemistry and Protein Research Center for
Bio-Industry, Hankuk University of Foreign Studies,
Yongin 449-791, Korea
A number of reports on the synthesis of β-halo amines
by the ring opening of aziridinium salts with halides are
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201000486.
View this journal online at
wileyonlinelibrary.com
4920
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 4920–4931

Halide-Induced Ring Opening of 2-Substituted Aziridinium Salts
of the nucleophile and substrate on the regioselectivity of
these ring-opening reactions has been investigated and the
observed reaction paths have been rationalized by detailed
computational analysis.
Results and Discussion
In a first approach, the systematic study of halide-
induced ring opening of intermediate 2-aryloxymethyl-
1,1-bis(arylmethyl)aziridinium salts was contemplated. As
reported before, 2-(aryloxymethyl)aziridines 4 can be pre-
Scheme 1. Regioselectivity in the ring opening of 2-subsitituted
aziridinium salts 1.
available in the literature.[4] In most cases, 2-vinyl- and 2- pared in high yields and purity upon treatment of the corre-
arylaziridinium salts have been evaluated in which the re- sponding 2-(bromomethyl)aziridines[6] with 2 equiv. of the
gioselectivity is substrate-dictated due to the presence of a appropriate potassium phenolate in a DMF/acetone (1:1)
pronounced electrophilic centre at the substituted aziridine solvent system at reflux for 10–20 h.[7] Subsequent treat-
carbon atom. The use of 2-alkyl-substituted aziridinium ment of the aziridines 4 with 1 equiv. of benzyl bromide in
ions has somewhat been neglected in this respect, probably acetonitrile is known to afford secondary bromides 5 as the
because of the potential influence of different parameters sole reaction products in high yields after heating at reflux
such as the type of nucleophile, substrate and solvent on for 5 hours.[7] To provide an entry to the corresponding
the reaction outcome.
fluorides, chlorides and iodides as well, β-bromo amines 5
Although the issue of regioselectivity has been addressed were treated with different halide sources. Thus, both novel
in a number of reports, no systematic study has been per- β-chloro amines 6 and β-iodo amines 7 were prepared as
formed up to now in which aziridinium substrates have the sole reaction products by heating either with 10 equiv.
been subjected to ring opening with fluoride, chloride, bro- of tetraethylammonium chloride or with 20 equiv. of so-
mide and iodide. In the work reported herein two different dium iodide, respectively, at reflux in acetonitrile for
types of 2-substituted aziridinium salts, namely in situ gen- 3 hours (Scheme 2, Table 1). Detailed spectroscopic analysis
erated and preformed aziridinium ions, were used as elec- ruled out the formation of other regioisomers.
trophiles in a systematic study of ring opening with fluo-
The conversion of β-bromo amines 5 into β-chloro
ride, chloride, bromide and iodide in acetonitrile. As a con- amines 6 using 20 equiv. of NaCl instead of tetraethylam-
tinuation of our interest in the study of the ring-opening monium chloride in acetonitrile proceeded very sluggishly;
reactions of 2-substituted aziridinium salts,[5] the influence no conversion occurred after heating at reflux for 4 h and
Scheme 2. Synthesis of β-bromo amines 5, β-chloro amines 6, β-iodo amines 7 and β-fluoro amines 8 and 9.
Eur. J. Org. Chem. 2010, 4920–4931
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4921

H.-J. Ha, V. Van Speybroeck, N. De Kimpe et al.
Table 1. Synthesis of β-bromo amines 5, β-chloro amines 6, β-iodo amines 7 and β-fluoro amines 8 and 9.
FULL PAPER
Entry
R1
R2
5 (% yield)
6 (% yield)
7 (% yield)
8 (% yield)
9 (% yield)
Ratio[a] 8/9
1
2
3
4
2-Cl
4-Cl
4-Cl
4-OMe
H
H
Cl
H
5a (71)
5b (86)
5c (85)
5d (84)
6a (82)
6b (79)
6c (83)
6d (84)
7a (89)
7b (88)
7c (82)
7d (79)
8a (54)
8b (42)
8c (60)
8d (61)
9a (10)
9b (8)
9c (10)
9d (14)
5:1
5:1
6:1
6:1
1
[a] Ratio determined by H NMR analysis.
only partial conversion was observed after 60 h. On the previous theoretical studies, it was demonstrated that prod-
other hand, the reaction of β-bromo amines 5 with uct stabilities in the case of bromide seem to dictate the
10 equiv. of tetrabutylammonium iodide in acetonitrile ap- outcome of the reaction through thermodynamic control,
peared to be less successful in comparison with the use of whereas in the case of fluoride differences in barriers were
sodium iodide as only 50% conversion took place after shown to be mainly due to the differences in interaction
heating at reflux for 7 h. If 15 equiv. of sodium iodide were energies, which indicates that steric interactions dictate the
used instead of 20 equiv., a longer reaction time (5 h) was reaction outcome.[5i]
required to drive the reaction to completion.
Apparently, the chloride-[8] and iodide-promoted ring
When β-bromo amines 5 were treated with 2 equiv. of opening of aziridinium ions 10 is controlled by the same
tetrabutylammonium fluoride in acetonitrile, however, a factors as those influencing the bromide-induced ring open-
mixture of regioisomeric fluorides 8 and 9 was obtained ing with attack at the substituted position (X = Cl and I,
after heating at reflux for 15 h (Scheme 2, Table 1).[5c] In path b, Scheme 3). Thus, it can be concluded that chloride-
this case, primary fluorides 8 were formed as the major re- , bromide- and iodide-promoted ring opening of aziridin-
action products in addition to minor amounts of secondary ium ions 10 is under thermodynamic control, eventually
fluorides 9 (ratio 8/9 5–6:1). To test the reaction outcome leading to the more stable secondary halides 12 as the final
as a function of reaction time and temperature, prolonged reaction products. On the other hand, ring opening with
reaction times and elevated temperatures were also evalu- fluoride is kinetically controlled, which can be attributed to
ated. In particular, heating at reflux for 3 d instead of 15 h the poor leaving-group capacity of fluoride in comparison
did not affect the isomeric distribution (ratio 8/9: 5–6:1) and with the other halides, which prevents thermodynamic
the same conclusion was drawn after heating at reflux for equilibration.
25 h in DMF. These observations indicate that the product
distribution between primary and secondary fluorides 8 and strate in the above ring-opening reactions, the synthesis of
9 is not under thermodynamic control. another type of aziridinium salt was devised. In addition to
To provide an insight into the potential role of the sub-
From a mechanistic point of view, the formation of β- non-isolable aziridinium intermediates 10, stable 1-methyl-
bromo amines 5, β-chloro amines 6, β-iodo amines 7 and aziridinium triflates 14 were prepared by N-methylation of
β-fluoro amines 8 and 9 proceeds through the ring opening chiral aziridines 13 by treatment with 1.1 equiv. of methyl
of the same intermediate aziridinium salts 10 with different trifluoromethanesulfonate in acetonitrile for 10 min
halides (Scheme 3).
(Scheme 4). These were then evaluated as electrophiles for
As observed and investigated before, quaternization and the halide-induced ring-opening reactions. Chiral substrates
subsequent ring opening of 2-(aryloxymethyl)aziridines 4 13a,b were prepared starting from the corresponding com-
using benzyl bromide produces β-bromo amines 12 through mercially available (R)-2-hydroxymethyl-1-[(R)-1-phenyl-
the regiospecific ring opening of aziridinium salts 10 at the ethyl]aziridine according to literature protocols.[5c,5g,5h]
substituted aziridine carbon atom (X = Br, path b,
In this work, different tetrabutylammonium halides were
Scheme 3).[5b,5i,7] Furthermore, in addition to preliminary used as halide sources for the ring opening of the aziridin-
findings using other types of substrates,[5c] the ring opening ium triflates 14. First, the reactions of 2-(methoxymethyl)-
of aziridinium intermediates 10 with fluoride afforded a aziridinium 14a (R = CH2OMe) with 1.5 equiv. of tetrabu-
mixture of regioisomers in which primary fluorides 11 are tylammonium fluoride, chloride, bromide or iodide in ace-
predominant (X = F, path a, Scheme 3), which indicates a tonitrile at room temperature for 1 hour afforded the corre-
change in regioselectivity in comparison with bromide. In sponding β-halo amines in good yields. Interestingly, the
Scheme 3. Regioselectivity in the ring opening of aziridinium salts 10 with different halides.
4922
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 4920–4931

Halide-Induced Ring Opening of 2-Substituted Aziridinium Salts
which underwent slow conversion into the secondary β-
chloro amine 16d upon standing at room temperature. Al-
though purification by column chromatography on silica
gel failed, the initially formed reaction product could be
identified as 2-amino-3-chloro-1-methoxypropane 15d by
1H NMR analysis. Clearly this primary chloride comprises
the kinetically controlled reaction product obtained
through ring opening of the aziridinium ion 14a at the un-
substituted position (path a, Scheme 4), which then re-
arranges into the more stable secondary chloride through a
thermodynamic equilibrium. The same observation was
made by careful analysis of the reaction between aziridin-
ium triflate 14a and 1.5 equiv. of Me4N+Cl–. The product
distribution between aziridinium ion 14a, primary chloride
15d and secondary chloride 16d was evaluated by 1H NMR
spectroscopy (CDCl3; Figure 1). Note that these findings
comprise the first experimental proof of the occurrence of
a thermodynamic equilibrium in the halide-induced ring
opening of 2-alkyl-substituted aziridinium salts.
Scheme 4. Ring opening of 1-methylaziridinium triflates 14 with
tetrabutylammonium halides.
same conclusions were drawn as described above in relation
to the selective synthesis of secondary bromide 16a, iodide
16b and chloride 16d as the sole reaction products and the
regioisomeric mixture of primary and secondary fluorides
15c and 16c (3:1; Scheme 4, Table 2). These observations
further confirm the nucleophile-dependency of the ring-
opening reactions of 1,1,3-trialkylaziridinium ions with ha-
lides, showing chloride-, bromide- and iodide-mediated ring
opening under thermodynamic control and fluoride-in-
duced ring opening under kinetic control.
Table 2. Ring opening of 1-methylaziridinium triflates 14 with tet-
rabutylammonium halides.
Entry Substrate
R
X–
Product
% Yield
1
2
3
4
5
6
7
8
14a
14a
14a
14a
14b
14b
14b
14b
CH2OMe Br–
CH2OMe I–
CH2OMe F–
CH2OMe Cl–
16a
16b
47
52
Figure 1. Product distribution between 14a, 15d and 16d as a func-
15c + 16c
16d
77 (3:1)
73
tion of reaction time.
CO2Et
CO2Et
CO2Et
CO2Et
Br–
I–
16e
92[a]
90[a]
71
From these experiments it can be concluded that the
ring-opening reactions of 2-alkyl-substituted aziridinium
salts 17 with chloride, bromide and iodide proceed under
thermodynamic control, with product stabilities dictating
the outcome of the reactions. Thus, the initially formed ki-
netic primary halides 18 undergo rearrangement into the
thermodynamically more stable secondary halides 19
(Scheme 5). Fluoride-mediated ring opening, however, is
under kinetic control, with the reaction outcome only being
dictated by steric interactions.
16f
F–
16g
Cl–
16h
83
[a] Compounds decomposed during chromatographic purification.
Nonetheless, the nucleophile-dependency of the aziridin-
ium ring-opening reactions is over-ruled by substrate con-
trol if an activated aziridine carbon atom is present in the
substrate. For example, with the 2-(ethoxycarbonyl)aziridin-
ium salt 14b (R = CO2Et), only the corresponding second-
ary halides were obtained by ring opening at the activated
α-carbonyl atom after reaction with 1.5 equiv. of tetra-
butylammonium halide in acetonitrile at room temperature,
and no primary halides were retrieved (Scheme 4, Table 2).
Thus, α-halo esters 16e–h were formed through regiospecific
ring opening of aziridinium salts 14b at the substituted azir-
idine carbon atom (path b, Scheme 4).
Interestingly, when aziridinium triflate 14a was treated
with 1.5 equiv. of NaCl in acetonitrile (20 h, room temp.)
instead of Bu4N+Cl–, a different reaction product was ini-
Scheme 5. Thermodynamic control for chloride-, bromide- and
tially observed upon chromatographic analysis (TLC), iodide-induced ring-opening reactions of aziridinium salts.
Eur. J. Org. Chem. 2010, 4920–4931
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4923

H.-J. Ha, V. Van Speybroeck, N. De Kimpe et al.
FULL PAPER
Theoretical Rationalization
models should be used with caution as they may give unreli-
able results.[20] Nonetheless, the mixed implicit/explicit sol-
vent model has been employed in addition to the supermol-
ecule approach for comparative purposes.
Computational studies have previously been performed
on fluoride- and bromide-mediated ring-opening reactions
of various N,N-dibenzylaziridinium ions,[5b,5d,5i,5j] however,
a systematic evaluation is necessary for the rationalization
of the observed trend in the halide series. To elucidate the
factors causing the differences in regioselectivity, a thor-
ough computational analysis was performed on the halide-
mediated ring opening of 14a. Iodide-induced ring-opening
reactions have been excluded from this study because they
show similar experimental regioselectivity to bromide, and
the main aim of this computational study was to elucidate
the factors causing the difference in experimental regio-
selectivities.
Nucleophilic Ring-Opening Mechanisms
Nucleophilic ring opening can occur by attack at the un-
hindered (path a) or hindered (path b) aziridine carbon
atoms (Scheme 1). Both reaction pathways were modelled
for fluoride, chloride and bromide. Subsequent comparison
of the reaction barriers and relative stabilities of the prod-
ucts should help us to understand the factors controlling
regioselectivity. Halide-induced ring openings were mod-
elled with the use of explicit solvent molecules; three aceto-
nitrile molecules have been used to solvate the fluoride,
chloride and bromide ions that attack the aziridinium ring
as this was previously shown to be adequate for the sol-
Computational Methodology
Reaction pathways for the nucleophilic attack of halides vation of the halide ion.[5i]
on aziridinium ion 14a were obtained at the B3LYP/6-
The transition-state geometries for the SN2 attack of ha-
31++G(d,p) level of theory.[9,10] Stationary points were lides on both aziridine carbon atoms of 14a are illustrated
characterized as minima or first-order saddle points from in Figure 2. Acetonitrile molecules stabilize the halide ions
frequency calculations. Intrinsic reaction coordinate through charge–dipole interactions; typical X···H3C–CN
(IRC)[11] calculations followed by full geometry optimiza- distances for fluoride, chloride and bromide are 2.0, 2.6 and
tions were used to verify the reactant complexes (ion–dipole 2.75 Å, respectively. These critical distances provide a mea-
complex) and products yielded by each transition state. En- sure of the ring opening within the aziridine as well as nu-
ergies were further refined by MPW1B95[12] single-point cleophile-to-aziridine attack and are considerably different
calculations as this method was recently shown to success- for both pathways and for all halides.
fully reproduce SCS-MP2[13] results in the ring opening of
The relative energies show that the barrier for path a is
aziridines with benzyl bromide.[5i] The effect of a polar envi- systematically lower than that for path b for all three ha-
ronment has been taken into account by the use of self- lides studied, that is, the kinetically favoured pathway is al-
consistent reaction field (SCRF) theory.[14] Solvation free ways the unhindered one as the steric properties would sug-
energies in acetonitrile (ε = 35.688) were obtained by gest. Differences in the activation energies between paths a
using the conductor-like polarizable continuum (C-PCM) (unhindered) and b (hindered) are shown in Table 3. On the
model.[15] All DFT calculations were carried out with the other hand, the differences in reaction energies show that
Gaussian 03 program package.[16]
path b leads to the thermodynamically more stable product
Explicit acetonitrile molecules were used to solvate ha- in all three cases. As mentioned earlier, the supermolecule
lide ions as calculations of gas-phase systems, in which bare (Figure 3) has also been embedded in a dielectric contin-
halide ions attack the electrophilic aziridinium ion, were uum to take into account long-range interactions with the
previously shown to give unrealistic results and were in- bulk solvent. This mixed implicit/explicit solvent approach
capable of representing the real system at hand.[4i] Accord- is shown to have the same trend as the discrete continuum
ingly, previous modelling studies on the bromide-induced model (supermolecule), which is solely solvated by explicit
ring opening of aziridinium ions effectively made use of ex- solvent molecules.
plicit solvent molecules as this had previously been shown
The potential energy surfaces (PES) for the halide-in-
to be vital in reproducing realistic potential energy sur- duced nucleophilic ring opening of 14a through paths a (un-
faces.[5d,5i] The use of discrete solvent molecules to stabilize hindered) and b (hindered) for all three halides are illus-
chemically active species in reactions is a well-established trated in Figure 3. There is a substantial barrier difference
methodology[5b,5d,5i,5j,17] and was used by us in a recent between the two pathways for fluoride attack that favours
study, which successfully pointed to the correct regioselec- the unhindered route (path a), which also leads to the ex-
tive outcome in aziridinium ring-opening reactions.[4i] How- perimentally observed major product F-P-a (15c). Although
ever, this approach only accounts for short-range interac- F-P-b (16c) is shown to be thermodynamically more stable
tions such as in ion–dipole complexes and does not take (Figure 3), because fluoride is a rather poor leaving group,
into account potential long-range interactions with solvent thermodynamic equilibration requires a back-reaction bar-
molecules, which are likely to be significant when anionic rier of approximately 130 kJ/mol to be overcome. This ex-
nucleophiles are involved, as is the case herein. In view of plains the mixture of products (15c and 16c) observed for
this, the supermolecule was also placed in a dielectric con- the fluoride-mediated ring opening of 14a. For chloride, re-
tinuum,[18] which led to a mixed implicit/explicit solvent action barriers for both pathways are approximately 15 kJ/
model.[19] A recent comparative study has shown that mixed mol lower than in the fluoride case, which is consistent with
4924
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 4920–4931

Halide-Induced Ring Opening of 2-Substituted Aziridinium Salts
Figure 2. Transition-state geometries optimized at the B3LYP/6-31++G(d,p) level of theory for halide-induced nucleophilic ring opening
of aziridinium ion 14a through paths a and b. Some critical distances are given in units of Å.
the strength of the nucleophile; the softer the nucleophile imately 75 kJ/mol) and thermodynamic equilibration is
the lower the barrier for attack. The reaction energies are feasible, leading to the energetically more stable final prod-
also significantly different; this is in line with relative leav- uct Cl-P-b (16d). This explains the initial observation of 15d
ing-group abilities. For chloride, Cl-P-b (16d) is more stable during the ring-opening reaction with chloride. For bro-
than Cl-P-a (15d) by 20 kJ/mol, whereas Cl-TS-a is more mide, the difference in energy between Br-TS-a and Br-TS-
feasible than Cl-TS-b by the same amount. Once again the b is smaller than for the corresponding chlorides. Product
kinetic route suggests path a, leading to 15d (Cl-P-a), how- stabilities favour path b and again the back-reaction barri-
ever, the back-reaction barriers are reasonably low (approx- ers are quite low, easily allowing equilibration to yield the
Eur. J. Org. Chem. 2010, 4920–4931
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4925

H.-J. Ha, V. Van Speybroeck, N. De Kimpe et al.
FULL PAPER
Figure 3. Free-energy surfaces (FES) for the halide-induced nucleophilic ring opening of 14a through paths a (unhindered) and b (hin-
dered) determined at the MPW1B95/6-31++G(d,p)//B3LYP/6-31++G(d,p) level of theory. Relative energies are given in kJ/mol.
Table 3. Differences in electronic (∆∆E‡) and free (∆∆G‡) energies tion model. To further verify the relationship between the
of activation and electronic (∆∆Erxn) and free (∆∆Grxn) energies of
reaction for halide-induced nucleophilic ring opening of 14a by
ergies of the transition states the activation strain model of
paths a (unhindered) and b (hindered).[a,b]
structural distinctions and differences in the relative free en-
chemical reactivity by Bickelhaupt,[21] also known as the
MPW1B95/6-31++G(d,p)[c]
CPCM[e] (ε = 36.6)
distortion/interaction model by Houk,[22]was employed.
The distortion/interaction model breaks down the acti-
vation energy (∆E‡) into the distortion (∆E‡dist) and inter-
action (∆E‡int) energies between the distorted fragments
[Equation (1)]; the former is associated with the strain
caused by deforming the individual reactants and the latter
is the favourable interaction between the deformed reac-
tants. The fragment distortion and interaction energies de-
termined at the MPW1B95/6-31++G(d,p) level are given in
Table 4.
∆∆E‡ ∆∆G‡[d] ∆∆Erxn ∆∆Grxn[d] ∆∆E‡ ∆∆G‡[f] ∆∆Erxn ∆∆Grxn[f]
F– 23.9
Cl– 21.4
Br– 12.0
16.2
18.5
11.3
–17.2
–18.0
–12.5
–12.3
–19.5
–11.1
10.8
15.8
16.3
10.2
8.7
11.1
–16.7
–20.1
–18.6
–23.3
–24.3
–15.8
‡
‡
[a] Energies in kJ/mol. [b] ∆∆E(G)‡ = ∆E(G)b – ∆E(G)a ;
∆∆E(G)rxn = ∆E(G)b – ∆E(G)arxn. [c] B3LYP/6-31++G(d,p) geo-
rxn
metries. [d] Thermal free-energy corrections from B3LYP/
6-31++G(d,p) calculations at 1 atm and 298 K. [e] Bond radii with
explicit hydrogen spheres. [f] Non-electrostatic terms included.
more stable product Br-P-b (16a; Table 3). Because Br-TS-
b is lower in energy than Cl-TS-b, it is likely that equilibra-
tion takes place much faster with the bromide and the ki-
netic product Br-P-a (15a) is therefore not experimentally
observed.
∆E‡ = ∆E‡ + ∆E‡
(1)
dist
int
Table 4. Differences in the reaction barriers (∆∆E‡) and the distor-
tion (∆∆E‡dist) and interaction energies (∆∆E‡int) between paths a
and b for the hydride and halide-induced nucleophilic ring opening
of 14a as determined at the MPW1B95/6-31++G(d,p) level of theo-
ry.[a,b]
The overall picture for halide-induced ring opening
shows that the unhindered route (path a) is always kinet-
ically preferred, however, the hindered route leads to the
thermodynamic product. The eventual outcome depends on
the softness and leaving-group ability of the nucleophile
(halide). If the nucleophile is a good leaving group (soft
nucleophile, bromide), back-reaction barriers are suffi-
ciently low to allow equilibration and the thermodynamic
product will prevail. If the nucleophile is a poor leaving
group (hard nucleophile, fluoride), the back reaction is un-
likely and the kinetic route will dictate the reaction out-
come. In the case of bromide, equilibration is so rapid that
the initial formation of the kinetic product (Br-P-a) is not
observed. In the case of chloride, however, equilibration is
slower and therefore the kinetic product (15d) is initially
observed during the reaction. Theoretical results are in per-
fect agreement with experimental findings and also in ac-
cord with the well-known trend in nucleophile strength and
leaving-group ability of the halide series.
∆∆E‡ [kJ/mol]
∆∆E‡ [kJ/mol]
∆∆E‡ [kJ/mol]
dist
int
F–
23.9
21.4
12.0
7.5
9.5
13.0
16.4
11.9
–1.1
Cl–
Br–
‡
‡
[a] ∆∆E‡ = ∆Eb – ∆Ea . [b] B3LYP/6-31++G(d,p) geometries.
The distortion energy is shown to increase on going from
fluoride to bromide. The penalty for distorting the aziridin-
ium ring increases as the nucleophile gets larger, which is
also reflected in the difference in elongation in the aziridin-
ium ring (Figure 2) and is an indication of the difference in
progression along the reaction coordinate; the transition
state is increasingly product-like going down the halide
series. There is a substantial difference in ∆∆E‡
for the
dist
halides studied even though the substrate is the same in
all three cases. This indicates that transition states occur at
different positions along the reaction coordinate, which in
turn also explains the stronger TS interaction ∆∆E‡int, as
suggested earlier by Bickelhaupt.[23]
Distortion/Interaction Model
Distortion/interaction calculations have revealed that for
Efforts to rationalize the experimentally observed reac- bromide the main contribution to the difference in acti-
tion outcomes also led to a comparative analysis of the vation barriers comes from the strain caused by the defor-
transition-state structures by using the distortion/interac- mation of the reactant, whereas no difference in orbital in-
4926
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 4920–4931

Halide-Induced Ring Opening of 2-Substituted Aziridinium Salts
7.22–7.33 (2 m, 3 H and 6 H) ppm. 13C NMR (75 MHz, CDCl3): δ
= 31.8, 38.0, 63.5, 70.0, 114.6, 120.9, 128.5, 129.3, 129.4, 132.8,
teractions between the competing pathways was present.
However, for fluoride and chloride, the differences in the
energy barriers of the two competing pathways are signifi-
cantly affected by the differences in the interaction energies
between the two transition states, which indicates that fluo-
ride and chloride ions are much more influenced by the
steric constraints around the aziridine carbon under attack.
This is certainly not the case for bromide, which is consider-
ably further from the aziridinium moiety in the transition
state (bromide–aziridine carbon distances in the transition
state are approximately 2.6 Å, Figure 2), which suggests
that bromide is not affected by the difference in steric ef-
fects.
137.4, 158.6 ppm. IR (neat): ν
= 2921, 1599, 1491, 1240, 1086,
˜
max
1034, 1015, 805, 752, 691 cm–1. MS (70 eV): m/z (%) = 274/6 (100)
[M + 1]+. C16H16ClNO (273.76): calcd. C 70.20, H 5.89, N 5.12;
found C 70.31, H 6.04, N 5.21.
2-[(4-Chlorophenoxy)methyl]-1-[(4-chlorophenyl)methyl]aziridine
1
(4c): Yield 82%, yellow liquid. Rf = 0.10 (hexane/EtOAc, 4:1). H
NMR (300 MHz, CDCl3): δ = 1.55 (d, J = 6.6 Hz, 1 H), 1.84 (d,
J = 3.3 Hz, 1 H), 1.91–1.98 (m, 1 H), 3.42 and 3.48 (2 d, J =
13.2 Hz, 2 H), 3.83 and 3.99 (2 dd, J = 10.4, 6.6, 4.4 Hz, 2 H),
6.77–6.80, 7.18–7.21 and 7.27–7.32 (3ϫm, 2 H, 2H and 4 H) ppm.
13C NMR (75 MHz, CDCl3): δ = 31.7, 37.9, 63.5, 70.4, 115.9,
125.8, 128.5, 129.29, 129.34, 133.0, 137.3, 157.3 ppm. IR (neat):
ν
max = 2986, 2923, 2830, 1596, 1489, 1284, 1240, 1171, 1089, 1015,
˜
822, 806, 668 cm–1. MS (70 eV): m/z (%) = 308/10/12 (100)
[M + 1]+. C16H15Cl2NO (308.21): calcd. C 62.35, H 4.91, N 4.54;
found C 62.42, H 5.23, N 4.45.
Conclusions
The ring-opening reactions of both non-isolable and
stable 2-alkyl-substituted aziridinium salts with fluoride,
chloride, bromide and iodide have been studied for the first
time in a systematic way and the results indicate an inherent
difference in reactivity between the reactions of fluoride on
the one hand and chloride, bromide and iodide on the
other. Both experimental and computational evidence
shows that product stabilities dictate the reaction outcome
through thermodynamic control in the cases of chloride,
bromide and iodide by rearrangement of the initially
formed primary halides to the more stable secondary ha-
lides through a thermodynamic equilibrium. The ring open-
ing of the same aziridinium salts with fluoride, however,
was shown to be mediated by steric interactions (kinetic
control) as the difference in the activation barriers was
mainly due to differences in interaction energies. For 2-acyl-
aziridinium ions, the reactions were shown to occur under
full substrate control to afford secondary β-halo amines by
attack at the activated α-carbonyl carbon atom regardless
of the nature of the halide.
1-[(4-Methoxyphenyl)methyl]-2-(phenoxymethyl)aziridine (4d): Yield
87 %, light-yellow crystals. Rf = 0.11 (hexane/EtOAc, 4:1). 1H
NMR (300 MHz, CDCl3): δ = 1.52 (d, J = 6.6 Hz, 1 H), 1.79 (d,
J = 3.3 Hz, 1 H), 1.86–1.97 (m, 1 H), 3.37 and 3.44 (2 d, J =
13.2 Hz, 2 H), 3.76 (s, 3 H), 3.91 (d, J = 5.5 Hz, 2 H), 6.81–6.93 and
7.19–7.28 (2 m, 5 H and 4 H) ppm. 13C NMR (75 MHz, CDCl3): δ
= 31.8, 37.8, 55.2, 63.6, 70.1, 113.8, 114.6, 120.8, 129.3, 129.4,
130.0, 158.7, 158.8 ppm. IR (neat): ν
= 2933, 2836, 1609, 1511,
˜
max
1457, 1243, 1173, 1030, 1018, 809, 760 cm–1. MS (70 eV): m/z (%)
= 270 (100) [M + 1]+. C17H19NO2 (269.34): calcd. C 75.81, H 7.11,
N 5.20; found C 75.67, H 7.27, N 5.08.
Synthesis of N-(3-Aryloxy-2-bromopropyl)amines 5: As a represen-
tative example the synthesis of N-benzyl-N-(2-bromo-3-phen-
oxypropyl)-N-(4-chlorobenzyl)amine (5b) is described here. Benzyl
bromide (1.71 g, 1 equiv.) was added to a solution of 1-[(4-chlo-
rophenyl)methyl]-2-(phenoxymethyl)aziridine (4b; 2.73 g, 10 mmol)
in acetonitrile (50 mL) at room temperature whilst stirring and the
resulting mixture was heated at reflux for 5 h. Afterwards the reac-
tion mixture was poured into water (50 mL) and extracted with
Et2O (3ϫ50 mL). Drying (MgSO4), filtration of the drying agent
and evaporation of the solvent afforded N-benzyl-N-(2-bromo-3-
phenoxypropyl)-N-(4-chlorobenzyl)amine (5b), which was purified
by column chromatography on silica gel (hexane/EtOAc, 1:1) to
obtain an analytically pure sample (86% yield, 3.83 g).
Experimental Section
N-Benzyl-N-(2-bromo-3-phenoxypropyl)-N-(4-chlorobenzyl)amine
Synthesis of 1-Arylmethyl-2-(aryloxymethyl)aziridines 4:[7] As a
representative example, the synthesis of 1-[(4-methoxyphenyl)-
methyl]-2-(phenoxymethyl)aziridine (4d) is described here. 2-Bro-
momethyl-1-[(4-methoxyphenyl)methyl]aziridine[6] (1.28 g, 5 mmol)
was added to a mixture of phenol (1.03 g, 2.2 equiv.) and K2CO3
(3.45 g, 5 equiv.) in a solvent mixture containing acetone and DMF
(50 mL, 1:1 v/v) and the resulting mixture was heated at reflux for
15 h. Afterwards the reaction mixture was poured into brine
(50 mL) and extracted with Et2O (3ϫ50 mL). Drying (MgSO4),
filtration of the drying agent and evaporation of the solvent af-
forded 1-[(4-methoxyphenyl)methyl]-2-(phenoxymethyl)aziridine
(4d), which was purified by column chromatography on silica gel
(hexane/EtOAc, 4:1) to give an analytically pure sample (87% yield,
1.17 g).
1
(5b): Yield 86%, colourless oil. Rf = 0.76 (hexane/EtOAc, 1:1). H
NMR (300 MHz, CDCl3): δ = 2.91 and 3.10 (2 dd, J = 13.8, 7.2,
5.5 Hz, 2 H), 3.53–3.71 (m, 4 H), 4.07–4.23 (m, 3 H), 6.77–6.80,
6.94–6.99 and 7.22–7.36 (3 m, 2 H, 1 H and 11 H) ppm. 13C NMR
(75 MHz, CDCl3): δ = 48.9, 57.8, 58.5, 59.2, 70.0, 114.6, 121.2,
127.4, 128.4, 128.5, 129.0, 129.5, 130.3, 132.9, 137.3, 138.5,
158.1 ppm. IR (neat): ν
= 2925, 2826, 1736, 1598, 1587, 1491,
˜
max
1453, 1240, 1088, 801, 752, 692 cm–1. MS (70 eV): m/z (%) = 364/
6 (100), 444/6/8 (15) [M + 1]+. C23H23BrClNO (444.80): calcd. C
62.11, H 5.21, N 3.15; found C 62.29, H 5.41, N 3.04.
N-Benzyl-N-[2-bromo-3-(4-chlorophenoxy)propyl]-N-(4-chloro-
benzyl)amine (5c): Yield 85%, colourless oil. Rf = 0.74 (hexane/
EtOAc, 1:1). 1H NMR (300 MHz, CDCl3): δ = 2.90 and 3.08 (2
dd, J = 13.8, 7.4, 5.8 Hz, 2 H), 3.52–3.72 (m, 4 H), 4.02–4.14 (m,
3 H), 6.67–6.71 and 7.20–7.38 (2 m, 2 H and 11 H) ppm. 13C NMR
(75 MHz, CDCl3): δ = 48.5, 57.7, 58.6, 59.4, 70.2, 115.9, 126.2,
127.4, 128.45, 128.53, 129.0, 129.4, 130.3, 133.0, 137.3, 138.5,
1-[(4-Chlorophenyl)methyl]-2-(phenoxymethyl)aziridine (4b): Yield
85%, yellow liquid. Rf = 0.28 (hexane/EtOAc, 4:1). 1H NMR
(300 MHz, CDCl3): δ = 1.55 (d, J = 6.6 Hz, 1 H), 1.85 (d, J =
3.3 Hz, 1 H), 1.94–2.03 (m, 1 H), 3.43 and 3.49 (2 d, J = 13.8 Hz,
2 H), 3.89 and 3.99 (2 dd, J = 10.3, 6.3, 5.0 Hz, 2 H), 6.86–6.96 and
156.8 ppm. IR (neat): ν
= 2920, 2826, 1596, 1490, 1453, 1240,
˜
max
1090, 821, 801, 740, 697 cm–1. MS (70 eV): m/z (%) = 398/400/402
Eur. J. Org. Chem. 2010, 4920–4931
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4927

H.-J. Ha, V. Van Speybroeck, N. De Kimpe et al.
FULL PAPER
(100) [M + 1]+. C23H22BrCl2NO (479.24): calcd. C 57.64, H 4.63,
N 2.92; found C 57.79, H 4.95, N 2.81.
(434.79): calcd. C 63.54, H 5.10, N 3.22; found C 63.78, H 5.41, N
3.38.
N-Benzyl-N-(2-bromo-3-phenoxypropyl)-N-(4-methoxybenzyl)amine N-Benzyl-N-(2-chloro-3-phenoxypropyl)-N-(4-methoxybenzyl)amine
1
1
(5d): Yield 84%, colourless oil. Rf = 0.76 (hexane/EtOAc, 1:1). H
(6d): Yield 84%, colourless oil. Rf = 0.75 (hexane/EtOAc, 1:1). H
NMR (300 MHz, CDCl3): δ = 2.91 and 3.08 (2 dd, J = 13.8, 8.3, NMR (300 MHz, CDCl3): δ = 2.80 and 3.00 (2 dd, J = 13.6, 8.0,
5.4 Hz, 2 H), 3.49–3.74 (m, 4 H), 3.77 (s, 3 H), 4.01–4.24 (m, 3 H),
6.78–6.83, 6.92–6.97 and 7.15–7.39 (3 m, 4 H, 1 H and 9 H) ppm.
13C NMR (75 MHz, CDCl3): δ = 49.3, 55.2, 57.8, 58.6, 59.3, 70.2,
113.7, 114.7, 121.1, 127.2, 128.3, 129.0, 129.4, 130.2, 130.8, 138.9,
5.5 Hz, 2 H), 3.48–3.75 (m, 4 H), 3.77 (s, 3 H), 3.86–3.95 (m, 1 H),
4.09–4.17 (m, 2 H), 6.78–6.84, 6.92–6.97 and 7.21–7.32 (3 m, 4 H,
1 H and 9 H) ppm. 13C NMR (75 MHz, CDCl3): δ = 55.2, 57.2,
58.7, 59.3, 70.0, 113.7, 114.7, 121.1, 127.2, 128.3, 129.0, 129.4,
158.3, 158.9 ppm. IR (neat): ν
= 2932, 2833, 1599, 1509, 1495,
130.1, 130.8, 138.9, 158.3, 158.8 ppm. IR (neat): νmax = 2932, 2833,
˜
˜
max
1453, 1241, 1172, 1034, 813, 752, 691 cm–1. MS (70 eV): m/z (%) = 2342, 1599, 1510, 1495, 1454, 1241, 1172, 1035, 813, 752, 741,
360 (100), 440/2 (30) [M + 1]+. C24H26BrNO2 (440.38): calcd. C
65.46, H 5.95, N 3.18; found C 65.62, H 6.13, N 3.14.
691 cm–1. MS (70 eV): m/z (%) = 360 (100), 396/8 (49) [M + 1]+.
C24H26ClNO2 (395.93): calcd. C 72.81, H 6.62, N 3.54; found C
72.94, H 6.82, N 3.40.
Synthesis of N-(2-Chloro-3-aryloxypropyl)amines 6: As a represen-
tative example the synthesis of N-benzyl-N-(2-chloro-3-phen-
oxypropyl)-N-(4-methoxybenzyl)amine (6d) is described here. Tet-
raethylammonium chloride (1.66 g, 10 equiv.) was added to a solu-
tion of N-benzyl-N-(2-bromo-3-phenoxypropyl)-N-(4-methoxy-
benzyl)amine (5d; 0.44 g, 1 mmol) in acetonitrile (20 mL) at room
temperature whilst stirring and the resulting mixture was heated at
reflux for 3 h. Afterwards the reaction mixture was poured into
Synthesis of N-(2-Iodo-3-aryloxypropyl)amines (7): As a representa-
tive example the synthesis of N-benzyl-N-(2-chlorobenzyl)-N-(2-
iodo-3-phenoxypropyl)amine (7a) is described here. Sodium iodide
(3.00 g, 20 equiv.) was added to a solution of N-benzyl-N-(2-chlo-
robenzyl)-N-(2-bromo-3-phenoxypropyl)amine (5a; 0.44 g,
1 mmol) in acetonitrile (20 mL) at room temperature whilst stirring
and the resulting mixture was heated at reflux for 3 h. Afterwards
water (50 mL) and extracted with Et2O (3 ϫ 50 mL). Drying the reaction mixture was poured into water (50 mL) and extracted
(MgSO4), filtration of the drying agent and evaporation of the sol-
vent afforded N-benzyl-N-(2-chloro-3-phenoxypropyl)-N-(4-meth-
oxybenzyl)amine (6d), which was purified by column chromatog-
raphy on silica gel (hexane/EtOAc, 1:1) to obtain an analytically
pure sample (84% yield, 0.33 g).
with Et2O (3ϫ50 mL). Drying (MgSO4), filtration of the drying
agent and evaporation of the solvent afforded N-benzyl-N-(2-chlo-
robenzyl)-N-(2-iodo-3-phenoxypropyl)amine (7a), which was puri-
fied by column chromatography on silica gel (hexane/EtOAc, 1:1)
to obtain an analytically pure sample (89% yield, 0.40 g).
N-Benzyl-N-(2-chlorobenzyl)-N-(2-chloro-3-phenoxypropyl)amine
(6a): Yield 82%, colourless oil. Rf = 0.78 (hexane/EtOAc, 1:1). H
N-Benzyl-N-(2-chlorobenzyl)-N-(2-iodo-3-phenoxypropyl)amine
(7a): Yield 89%, colourless oil. Rf = 0.76 (hexane/EtOAc, 1:1). H
1
1
NMR (300 MHz, CDCl3): δ = 2.87 and 3.06 (2 dd, J = 13.8, 7.2, NMR (300 MHz, CDCl3): δ = 3.01 and 3.07 (2 dd, J = 14.0, 8.5,
5.5 Hz, 2 H), 3.61–3.84 (m, 4 H), 3.90–3.96 (m, 1 H), 4.10–4.18 (m, 6.9 Hz, 2 H), 3.63 and 3.70 (2 d, J = 13.2 Hz, 2 H), 3.74 and 3.81
2 H), 6.76–6.78, 6.92–6.97, 7.13–7.39 and 7.51–7.54 (4 m, 2 H, 1
H, 10 H and 1 H) ppm. 13C NMR (75 MHz, CDCl3): δ = 57.0,
56.5, 57.5, 59.6, 69.9, 114.6, 121.1, 126.7, 127.3, 128.3, 128.4, 129.1,
(2 d, J = 14.1 Hz, 2 H), 4.05–4.18 (m, 2 H), 4.22–4.30 (m, 1 H),
6.76–6.79, 6.93–6.97, 7.12–7.41 and 7.54–7.57 (4 m, 2 H, 1 H, 10
H and 1 H) ppm. 13C NMR (75 MHz, CDCl3): δ = 28.7, 56.1, 59.4,
59.6, 71.0, 114.8, 121.1, 126.7, 127.3, 128.3, 128.4, 129.2, 129.4,
129.4, 129.6, 131.0, 134.3, 136.4, 138.5, 158.3 ppm. IR (neat): ν
˜
max
= 2923, 2849, 1599, 1495, 1453, 1241, 1037, 749, 690 cm–1. MS
129.5, 131.2, 134.2, 136.3, 138.4, 158.1 ppm. IR (neat): νmax = 2921,
˜
(70 eV): m/z (%) = 400/2/4 (100) [M + 1]+, 364/6 (75). C23H23Cl2NO 2849, 1598, 1494, 1453, 1239, 1029, 749, 690 cm–1. MS (70 eV): m/z
(400.35): calcd. C 69.00, H 5.79, N 3.50; found C 69.17, H 5.97, N
3.72.
(%) = 364/6 (100), 492/4 (5) [M + 1]+. C23H23ClINO (491.80):
calcd. C 56.17, H 4.71, N 2.85; found C 55.96, H 4.83, N 3.01.
N-Benzyl-N-(4-chlorobenzyl)-N-(2-chloro-3-phenoxypropyl)amine
(6b): Yield 79%, colourless oil. Rf = 0.79 (hexane/EtOAc, 1:1). H
N-Benzyl-N-(4-chlorobenzyl)-N-(2-iodo-3-phenoxypropyl)amine
(7b): Yield 88%, colourless oil. Rf = 0.77 (hexane/EtOAc, 1:1). H
1
1
NMR (300 MHz, CDCl3): δ = 2.81 and 3.02 (2 dd, J = 13.8, 7.2, NMR (300 MHz, CDCl3): δ = 2.93 and 3.00 (2 dd, J = 13.8, 7.7,
6.0 Hz, 2 H), 3.55 and 3.68 (2 d, J = 13.5 Hz, 2 H), 3.59 and 3.64 7.2 Hz, 2 H), 3.54 and 3.63 (2 d, J = 13.8 Hz, 2 H), 3.59 (d, J =
(2 d, J = 13.8 Hz, 2 H), 3.96–4.01 (m, 1 H), 4.05–4.18 (m, 2 H), 7.7 Hz, 2 H), 4.08–4.20 (m, 2 H), 4.22–4.32 (m, 1 H), 6.77–6.81,
6.76–6.79, 6.93–6.98 and 7.21–7.38 (3 m, 2 H, 1 H and 11 H) ppm.
13C NMR (75 MHz, CDCl3): δ = 57.1, 57.4, 58.7, 59.4, 70.0, 114.7,
121.3, 127.5, 128.5, 128.6, 129.1, 129.6, 130.4, 133.0, 137.5, 138.7,
6.94–6.99 and 7.21–7.42 (3 m, 2 H, 1 H and 11 H) ppm. 13C NMR
(75 MHz, CDCl3): δ = 28.7, 58.3, 59.0, 59.3, 71.1, 114.8, 121.3,
127.5, 128.5, 128.6, 129.1, 129.6, 130.4, 133.0, 137.4, 138.5,
158.3 ppm. IR (neat): ν
= 2924, 2828, 1599, 1588, 1491, 1453,
158.1 ppm. IR (neat): ν
= 2924, 2803, 1598, 1587, 1491, 1453,
˜
˜
max
max
1241, 1088, 801, 751, 691 cm–1. MS (70 eV): m/z (%) = 400/2/4 (100)
[M + 1]+, 364/6 (60). C23H23Cl2NO (400.35): calcd. C 69.00, H
5.79, N 3.50; found C 68.92, H 5.94, N 3.55.
1239, 1088, 800, 751, 690 cm–1. MS (70 eV): m/z (%) = 364/6 (100),
492/4 (8) [M + 1]+. C23H23ClINO (491.80): calcd. C 56.17, H 4.71,
N 2.85; found C 56.19, H 4.88, N 3.04.
N-Benzyl-N-(4-chlorobenzyl)-N-[2-chloro-3-(4-chlorophenoxy)-
propyl]amine (6c): Yield 83%, colourless oil: Rf = 0.80 (hexane/
EtOAc, 1:1). 1H NMR (300 MHz, CDCl3): δ = 2.80 and 3.00 (2
dd, J = 13.6, 7.4, 5.5 Hz, 2 H), 3.51–3.71 (m, 4 H), 3.91–3.96 (m,
1 H), 4.02–4.13 (m, 2 H), 6.65–6.68 and 7.18–7.31 (2 m, 2 H and
11 H) ppm. 13C NMR (75 MHz, CDCl3): δ = 56.7, 57.1, 58.7, 59.4,
70.0, 115.8, 126.1, 127.4, 128.4, 128.5, 128.9, 129.3, 130.2, 132.9,
N-Benzyl-N-(4-chlorobenzyl)-N-[3-(4-chlorophenoxy)-2-iodopropyl]-
amine (7c): Yield 82%, colourless oil. Rf = 0.77 (hexane/EtOAc,
1:1). 1H NMR (300 MHz, CDCl3): δ = 2.92 and 2.99 (2 dd, J =
14.0, 8.0, 6.9 Hz, 2 H), 3.52 and 3.64 (2 d, J = 13.5 Hz, 2 H), 3.58
(s, 2 H), 3.38–3.70 (m, 4 H), 4.03–4.14 (m, 2 H), 4.18–4.26 (m, 1
H), 6.67–6.70 and 7.18–7.38 (2 m, 2 H and 11 H) ppm. 13C NMR
(75 MHz, CDCl3): δ = 28.0, 58.3, 59.0, 59.1, 71.1, 115.9, 126.1,
127.4, 128.4, 128.5, 129.0, 129.3, 130.3, 132.9, 137.2, 138.4,
137.3, 138.5, 156.8 ppm. IR (neat): ν
= 2925, 2828, 2359, 1596,
˜
max
1490, 1453, 1285, 1241, 1090, 821, 801, 740, 698 cm–1. MS (70 eV):
m/z (%) = 434/36/38/40 (100) [M + 1]+, 398/400 (90). C23H22Cl3NO 1238, 1090, 821, 800, 737, 698 cm–1. MS (70 eV): m/z (%) = no
156.6 ppm. IR (neat): ν
= 2924, 2804, 2361, 1596, 1489, 1452,
˜
max
4928
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 4920–4931

Halide-Induced Ring Opening of 2-Substituted Aziridinium Salts
[M]+, 398/400/402 (100) [M – I]+. C23H22Cl2INO (526.24): calcd. C 127.2, 128.4, 128.5, 128.6, 129.5, 130.0, 132.7, 138.4, 139.6,
52.49, H 4.21, N 2.66; found C 52.36, H 4.18, N 2.53.
158.4 ppm. 19F NMR (CCl3F): δ = –227.30 (td, J = 47.3,
22.3 Hz) ppm. IR (neat): νmax = 2928, 2833, 1599, 1588, 1491, 1470,
˜
N-Benzyl-N-(2-iodo-3-phenoxypropyl)-N-(4-methoxybenzyl)amine
(7d): Yield 79%, colourless oil. Rf = 0.77 (hexane/EtOAc, 1:1). H
1241, 1088, 1014, 907, 753, 730, 691 cm–1. MS (70 eV): m/z (%) =
364/6 (100), 384/6 (77) [M + 1]+. C23H23ClFNO (383.89): calcd. C
71.96, H 6.04, N 3.65; found C 71.89, H 6.18, N 3.67.
1
NMR (300 MHz, CDCl3): δ = 2.95 and 3.01 (2 dd, J = 13.9, 8.5,
6.6 Hz, 2 H), 3.48–3.72 (m, 4 H), 3.76 (s, 3 H), 4.01–4.29 (m, 3 H),
6.78–6.83, 6.92–6.97 and 7.22–7.35 (3 m, 4 H, 1 H and 9 H) ppm.
13C NMR (75 MHz, CDCl3): δ = 29.1, 55.2, 58.3, 58.9, 59.3, 71.1,
113.7, 114.7, 121.1, 127.2, 128.3, 129.0, 129.4, 130.2, 130.7, 138.8,
N-Benzyl-N-(4-chlorobenzyl)-N-(2-fluoro-3-phenoxypropyl)amine
1
(9b): Yield 8%, colourless oil. Rf = 0.28 (hexane/EtOAc, 96:4). H
NMR (300 MHz, CDCl3): δ = 2.75–2.95 (m, 2 H), 3.65 and 3.67
(2 s, 4 H), 3.91–4.06 (m, 2 H), 4.80–4.86 and 4.93–5.02 (2 m, 1 H),
158.2, 158.8 ppm. IR (neat): ν
= 2928, 2833, 1598, 1509, 1495,
˜
max
1240, 1171, 1033, 812, 752, 734, 691 cm–1. MS (70 eV): m/z (%) = 6.79–6.82, 6.94–6.99 and 7.22–7.38 (3 m, 2 H, 1 H and 11 H) ppm.
360 (100), 488 (10) [M + 1]+. C24H26INO2 (487.38): calcd. C 59.14, 13C NMR (75 MHz, ref. = CDCl3): δ = 53.7 (d, J = 21.9 Hz), 58.7,
H 5.38, N 2.87; found C 59.30, H 5.62, N 3.00.
59.4, 68.2 (d, J = 23.1 Hz), 90.9 (d, J = 174.3 Hz), 114.6, 121.3,
127.4, 128.5, 128.6, 129.0, 129.6, 130.3, 132.9, 137.7, 138.9,
158.4 ppm. 19F NMR (CCl3F): –188.58 to –188.18 (m) ppm. IR
Synthesis of 2-Amino-3-aryloxy-1-fluoropropanes 8 and N-(2-
Fluoro-3-aryloxypropyl)amines 9: As a representative example the
synthesis of 2-[N-benzyl-N-(2-chlorobenzyl)amino]-1-fluoro-3-
phenoxypropane (8a) and N-benzyl-N-(2-chlorobenzyl)-N-(2-
fluoro-3-phenoxypropyl)amine (9a) is described here. TBAF
(2.61 g, 2 equiv.) was added to a solution of N-benzyl-N-(2-chlo-
robenzyl)-N-(2-bromo-3-phenoxypropyl)amine (5a; 2.22 g,
5 mmol) in acetonitrile (30 mL) at room temperature whilst stirring
and the resulting mixture was heated at reflux for 15 h. Extraction
with Et2O (3ϫ50 mL), drying (MgSO4), filtration of the drying
agent and evaporation of the solvent afforded a mixture of 2-[N-
benzyl-N-(2-chlorobenzyl)amino]-1-fluoro-3-phenoxypropane (8a)
and N-benzyl-N-(2-chlorobenzyl)-N-(2-fluoro-3-phenoxypropyl)-
amine (9a) in a ratio of 5:1. The two isomers were separated by
column chromatography (hexane/EtOAc, 97:3) to furnish com-
pounds 8a (54% yield, 1.04 g) and 9a (10% yield, 0.19 g) as analyti-
cally pure samples.
(neat): νmax = 2924, 2829, 1598, 1588, 1490, 1453, 1242, 1088, 1014,
˜
801, 752, 691 cm–1. MS (70 eV): m/z (%) = 364/6 (100), 384/6 (55)
[M + 1]+.
2-[N-Benzyl-N-(4-chlorobenzyl)amino]-3-(4-chlorophenoxy)-1-
fluoropropane (8c): Yield 60%, colourless oil. Rf = 0.33 (hexane/
EtOAc, 96:4). 1H NMR (300 MHz, CDCl3): δ = 3.25–3.40 (m, 1
H), 3.81 and 3.82 (2 s, 4 H), 4.11 (d, J = 6.0 Hz, 2 H), 4.71 (dd, J
= 47.3, 4.9 Hz, 2 H), 6.75–6.80 and 7.19–7.37 (2 m, 2 H and 11
H) ppm. 13C NMR (75 MHz, CDCl3): δ = 54.7, 55.3, 56.4 (d, J =
18.5 Hz), 66.0 (d, J = 7.0 Hz), 82.3 (d, J = 171.9 Hz), 115.7, 125.9,
127.2, 128.4, 128.5, 128.6, 129.4, 129.9, 132.7, 138.3, 139.4,
157.0 ppm. 19F NMR (CCl3F): δ = –227.31 (td, J = 47.3,
22.3 Hz) ppm. IR (neat): νmax = 2930, 2831, 1596, 1588, 1490, 1470,
˜
1241, 1090, 1014, 1006, 821, 737, 698 cm–1. MS (70 eV): m/z (%) =
418/20/22 (100) [M + 1]+. C23H22Cl2FNO (418.34): calcd. C 66.04,
H 5.30, N 3.35; found C 66.11, H 5.54, N 3.57.
2-[N-Benzyl-N-(2-chlorobenzyl)amino]-1-fluoro-3-phenoxypropane
(8a): Yield 54%, colourless oil. Rf = 0.17 (hexane/EtOAc, 97:3). 1H
N-Benzyl-N-(4-chlorobenzyl)-N-[3-(4-chlorophenoxy)-2-fluoro-
NMR (300 MHz, CDCl3): δ = 3.31–3.44 (m, 1 H), 3.88 and 4.01 propyl]amine (9c): Yield 10%, colourless oil. Rf = 0.27 (hexane/
(2 s, 4 H), 4.20 (d, J = 6.1 Hz, 2 H), 4.77 (dd, J = 47.4, 5.5 Hz, 2
EtOAc, 96:4). 1H NMR (300 MHz, CDCl3): δ = 2.74–2.96 (m, 2
H), 6.86–6.89, 6.91–6.97, 7.12–7.38 and 7.61–7.63 (4 m, 2 H, 1 H, H), 3.64 and 3.66 (2 s, 4 H), 3.90–4.05 (m, 2 H), 4.75–4.82 and
10 H and 1 H) ppm. 13C NMR (75 MHz, CDCl3): δ = 52.2, 55.5,
57.2 (d, J = 18.4 Hz), 65.3 (d, J = 5.8 Hz), 82.4 (d, J = 170.7 Hz),
114.4, 121.0, 126.8, 127.1, 128.1, 128.3, 128.7, 129.4, 129.5, 130.5,
4.91–5.03 (2 m, 1 H), 6.69–6.72 and 7.20–7.37 (2 m, 2 H and 11
H) ppm. 13C NMR (75 MHz, CDCl3): δ = 53.4 (d, J = 21.9 Hz),
58.7, 59.4, 68.5 (d, J = 24.2 Hz), 90.7 (d, J = 175.4 Hz), 115.7,
134.0, 137.1, 139.6, 158.5 ppm. 19F NMR (CCl3F): δ = –227.32 (td, 127.3, 128.4, 128.5, 128.9, 129.3, 129.6, 130.2, 132.9, 137.5, 138.7,
J = 48.0, 23.7 Hz) ppm. IR (neat): ν
= 3062, 3029, 2954, 1599,
156.9 ppm. 19F NMR (CCl3F): δ = –188.96 to –188.73 (m) ppm.
˜
max
1495, 1470, 1241, 1037, 751, 734, 691 cm–1. MS (70 eV): m/z (%) = IR (neat): ν
= 2925, 2828, 1594, 1488, 1452, 1240, 1090, 1014,
˜
max
384/6 (100) [M + 1]+. C23H23ClFNO (383.89): calcd. C 71.96, H
6.04, N 3.65; found C 71.82, H 6.19, N 3.56.
907, 822, 732, 698 cm–1. MS (70 eV): m/z (%) = 418/20/22 (100) [M
+ 1]+.
N-Benzyl-N-(2-chlorobenzyl)-N-(2-fluoro-3-phenoxypropyl)amine 2-[N-Benzyl-N-(4-methoxybenzyl)amino]-1-fluoro-3-phenoxy-
(9a): Yield 10%, colourless oil. Rf = 0.10 (hexane/EtOAc, 97:3). 1H
propane (8d): Yield 61%, colourless oil. Rf = 0.13 (hexane/EtOAc,
1
NMR (300 MHz, CDCl3): δ = 2.84–2.93 (m, 2 H), 3.71 and 3.82 97:3). H NMR (300 MHz, CDCl3): δ = 3.29–3.44 (m, 1 H), 3.78
(2 s, 4 H), 3.92–4.06 (m, 2 H), 4.80–4.87 and 4.96–5.03 (2 m, 1 H), (s, 5 H), 3.84 (s, 2 H), 4.15 (d, J = 6.6 Hz, 2 H), 4.71 (dd, J = 47.6,
6.78–6.81, 6.92–6.97, 7.15–7.40 and 7.52–7.55 (4 m, 2 H, 1 H, 10 5.2 Hz, 2 H), 6.83–6.87, 6.92–6.97 and 7.20–7.39 (3 m, 4 H, 1 H
H and 1 H) ppm. 13C NMR (75 MHz, CDCl3): δ = 54.0 (d, J = and 9 H) ppm. 13C NMR (75 MHz, CDCl3): δ = 54.7, 55.1, 55.2,
21.9 Hz), 56.5, 59.6, 68.3 (d, J = 23.1 Hz), 90.9 (d, J = 174.2 Hz), 56.2 (d, J = 18.5 Hz), 65.5 (d, J = 7.0 Hz), 82.6 (d, J = 171.9 Hz),
114.6, 121.2, 126.8, 127.4, 128.5, 129.1, 129.5, 129.6, 129.7, 131.0,
113.7, 114.4, 120.9, 127.0, 128.3, 128.6, 129.5, 129.8, 131.8, 140.0,
134.3, 136.7, 138.9, 158.5 ppm. 19F NMR (CCl3F): δ = –188.68 to
158.5, 158.7 ppm. 19F NMR (CCl3F): δ = –227.27 (td, J = 47.4,
–188.2 (m) ppm. IR (neat): ν
= 2923, 2850, 1598, 1588, 1494, 23.7 Hz) ppm. IR (neat): νmax = 2917, 2849, 1599, 1510, 1495, 1454,
˜
˜
max
1443, 1242, 1049, 1037, 750, 690 cm–1. MS (70 eV): m/z (%) = 384/
1241, 1171, 1035, 831, 819, 753, 737, 691 cm–1. MS (70 eV): m/z
(%) = 380 (100) [M + 1]+. C24H26FNO2 (379.47): calcd. C 75.96,
H 6.91, N 3.69; found C 75.77, H 7.03, N 3.51.
6 (82) [M + 1]+.
2-[N-Benzyl-N-(4-chlorobenzyl)amino]-1-fluoro-3-phenoxypropane
(8b): Yield 42%, colourless oil. Rf = 0.35 (hexane/EtOAc, 96:4). 1H
N-Benzyl-N-(2-fluoro-3-phenoxypropyl)-N-(4-methoxybenzyl)-
NMR (300 MHz, CDCl3): δ = 3.28–3.43 (m, 1 H), 3.82 and 3.83 amine (9d): Yield 14%, colourless oil. Rf = 0.09 (hexane/EtOAc,
1
(2 s, 4 H), 4.16 (d, J = 6.1 Hz, 2 H), 4.72 (dd, J = 47.4, 5.0 Hz, 2
H), 6.85–6.88, 6.94–6.99 and 7.22–7.38 (3 m, 2 H, 1 H and 11
H) ppm. 13C NMR (75 MHz, CDCl3): δ = 54.7, 55.3, 56.4 (d, J =
97:3). H NMR (300 MHz, CDCl3): δ = 2.73–2.94 (m, 2 H), 3.61
and 3.67 (2 s, 4 H), 3.78 (s, 3 H), 3.94–4.04 (m, 2 H), 4.77–4.83
and 4.93–5.00 (2 m, 1 H), 6.78–6.85, 6.93–6.98 and 7.22–7.36 (3 m,
18.5 Hz), 65.4 (d, J = 6.9 Hz), 82.5 (d, J = 172.0 Hz), 114.4, 121.0, 4 H, 1 H and 9 H) ppm. 13C NMR (75 MHz, CDCl3): δ = 53.6 (d,
Eur. J. Org. Chem. 2010, 4920–4931
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4929

H.-J. Ha, V. Van Speybroeck, N. De Kimpe et al.
FULL PAPER
J = 23.1 Hz), 55.3, 58.8, 59.4, 68.5 (d, J = 23.1 Hz), 91.0 (d, J = N-[(S)-2-Chloro-3-methoxypropyl]-N-methyl-N-[(R)-1-phenylethyl]-
174.2 Hz), 113.8, 114.6, 121.1, 127.2, 128.4, 129.0, 129.5, 130.2,
amine (16d): Yield 73%, colourless oil. [α]2D5 = +22 (c = 1.20,
CHCl3). H NMR (400 MHz, CDCl3): δ = 1.36 (d, J = 6.8 Hz, 3
1
131.1, 139.3, 158.5, 158.8 ppm. 19F NMR (CCl3F): δ = –188.52 to
–188.05 (m) ppm. IR (neat): ν
= 2951, 2834, 1599, 1510, 1495, H), 2.44 (s, 3 H), 2.57 (dd, J = 13.6, 6.0 Hz, 1 H), 2.86 (dd, J =
˜
max
1453, 1243, 1172, 1035, 812, 752, 742, 691 cm–1. MS (70 eV): m/z
13.2, 8.4 Hz, 1 H), 3.38 (s, 3 H), 3.53 (dd, J = 10.8, 6.4 Hz, 1 H),
3.66 (q, J = 6.8 Hz, 1 H), 3.73 (dd, J = 10.4, 4.0 Hz, 1 H), 3.99–4.05
(m, 1 H), 7.21–7.34 (m, 5 H) ppm. 13C NMR (100 MHz, CDCl3): δ
= 17.7, 39.4, 58.1, 58.6, 59.3, 63.9, 74.9, 127.2, 127.9, 128.4,
143.4 ppm. HRMS (ESI): calcd. for C13H20ClNONa [M +Na]+
264.1131; found 264.1129.
(%) = 380 (100) [M + 1]+.
General Procedure for the Synthesis of Chiral β-Halo Amines 15 and
16: Methyl trifluoromethanesulfonate (1.1 equiv.) was added to a
solution of either (R)-2-methoxymethyl-1-[(R)-1-phenylethyl]azirid-
ine (13a) or ethyl (R)-1-[(R)-1-phenylethyl]aziridine-2-carboxylate
(13b) in CH3CN (3 ) at room temperature under nitrogen in oven-
dried glassware. This solution was stirred for 10 min at room tem-
perature before the addition of tetra-n-butylammonium halide
(1.5 equiv.) at 0 °C. The resulting reaction mixture was stirred at
room temperature for 1 h. Subsequently the reaction mixture was
extracted with CH2Cl2 and water and the combined organic layers
were dried with anhydrous MgSO4. Filtration of the drying agent
and removal of the solvent in vacuo afforded the crude product,
which was purified by column chromatography on silica gel to pro-
vide an analytically pure sample.
Ethyl (S)-2-Bromo-3-{N-methyl-N-[(R)-1-phenylethyl]amino}-
propionate (16e): This compound was isolated with a small amount
of the dehydrobrominated product. Yield 92%, colourless oil. [α]D
= +32 (c = 0.524, CHCl3). 1H NMR (200 MHz, CDCl3): δ = 1.26–
1.37 (m, 6 H), 2.16 (s, 3 H), 2.70–2.79 (m, 1 H), 3.19–3.31 (m, 1 H),
3.58–3.88 (m, 1 H), 4.11–4.30 (m, 3 H), 7.18–7.38 (m, 5 H) ppm.
Ethyl (S)-2-Iodo-3-{N-methyl-N-[(R)-1-phenylethyl]amino}-
propionate (16f): This compound was isolated with a small amount
of the dehydroiodinated product. Yield 90%, colourless oil. [α]D
=
+19 (c = 2.53, CHCl3). 1H NMR (400 MHz, CDCl3): δ = 1.26–
1.31 (m, 3 H), 1.37 (d, J = 6.8 Hz, 3 H), 2.14 (s, 3 H), 2.75–2.86
(m, 1 H), 3.07–3.23 (m, 1 H), 3.72–3.86 (m, 1 H), 4.14–4.30 (m, 3
H), 7.22–7.38 (m, 5 H) ppm.
N-[(S)-2-Bromo-3-methoxypropyl]-N-methyl-N-[(R)-1-phenylethyl]-
amine (16a): Yield 47%, colourless oil. [α]2D5 = +13 (c = 0.722,
1
CHCl3). H NMR (400 MHz, CDCl3): δ = 1.36 (d, J = 6.8 Hz, 3
H), 2.21 (s, 3 H), 2.69 (dd, J = 13.6, 6.0 Hz, 1 H), 2.93 (dd, J =
13.2, 8.8 Hz, 1 H), 3.38 (s, 3 H), 3.6–3.67 (m, 2 H), 3.67 (dd, J =
10.8, 4.1 Hz, 1 H), 4.06–4.12 (m, 1 H), 7.22–7.32 (m, 5 H) ppm.
13C NMR (100 MHz, CDCl3): δ = 17.4, 38.9, 51.0, 58.3, 58.9, 63.5,
74.7, 127.0, 127.7, 128.1, 143.2 ppm. MS: m/z (%) = 287 (0.3), 285
(0.3), 270 (1), 272 (1), 148 (75), 105 (100). HRMS (ESI): calcd. for
C13H20BrNONa [M +Na]+ 308.0626; found 308.0624.
Ethyl (S)-2-Fluoro-3-{N-methyl-N-[(R)-1-phenylethyl]amino}-
propionate (16g): Yield 71%, colourless oil. [α]2D5 = +6 (c = 1.24,
CHCl3). H NMR (400 MHz, CDCl3): δ = 1.29 (t, J = 7.6 Hz, 3
1
H), 1.38 (d, J = 6.8 Hz, 3 H), 2.30 (s, 3 H), 2.98 (dd, J = 25.6,
4.4 Hz, 2 H), 3.75 (q, J = 6.8 Hz, 1 H), 4.23 (qd, J = 7.6, 5.6 Hz,
1 H), 5.03 (dt, J = 49.6, 4.8 Hz, 1 H), 7.23–7.32 (m, 5 H) ppm. 13
C
NMR (100 MHz, CDCl3): δ = 14.1, 17.4, 39.5, 39.5, 39.5, 55.2,
55.4, 61.4, 63.3, 88.4, 88.5, 90.27, 90.31, 126.9, 127.6, 128.1, 143.0,
168.8, 169.1 ppm. HRMS (ESI): calcd. for C14H20FNO2Na [M
+Na]+ 276.1376; found 276.1377.
N-[(S)-2-Iodo-3-methoxypropyl]-N-methyl-N-[(R)-1-phenylethyl]-
amine (16b): Yield 52%, colourless oil. [α]2D5 = –4 (c = 3.26, CHCl3).
1H NMR (400 MHz, CDCl3): δ = 1.36 (d, J = 6.8 Hz, 3 H), 2.17
(s, 3 H), 2.72–2.87 (m, 2 H), 3.38 (s, 3 H), 3.61–3.73 (m, 3 H),
4.16–4.22 (m, 1 H), 7.21–7.37 (m, 5 H) ppm. 13C NMR (100 MHz,
CDCl3): δ = 17.6, 31.8, 38.9, 59.0, 60.0, 63.6, 75.9, 127.2, 128.0,
128.4, 143.4 ppm. HRMS (ESI): calcd. for C13H20INONa
[M +Na]+ 356.0487; found 356.0488.
Ethyl (S)-2-Chloro-3-{N-methyl-N-[(R)-1-phenylethyl]amino}-
propionate (16h): Yield 83%, colourless oil. [α]2D5 = +16 (c = 1.472,
CHCl3). 1H NMR (200 MHz, CDCl3): δ = 1.26–1.37 (m, 6 H), 2.19
(s, 3 H), 2.76 (dd, J = 13.0, 5.4 Hz, 1 H), 3.22 (dd, J = 13.0, 9.8 Hz,
1 H), 3.79 (q, J = 6.8 Hz, 1 H), 4.19–4.30 (m, 3 H), 7.22–7.35 (m,
5 H) ppm. 13C NMR (50 MHz, CDCl3): δ = 14.0, 16.3, 38.4, 54.1,
58.1, 61.8, 63.1, 127.0, 127.6, 128.1, 142.4, 169.4 ppm. HRMS
(ESI): calcd. for C14H20ClNO2Na [M +Na]+ 292.1080; found
292.1084.
N-[(R)-1-Fluoromethyl-2-methoxyethyl]-N-methyl-N-[(R)-1-phenyl-
ethyl]amine (15c): Yield 58%, colourless oil. [α]2D5 = +51 (c = 0.758,
1
CHCl3). H NMR (200 MHz, CDCl3): δ = 1.37 (d, J = 6.8 Hz, 3
H), 2.32 (s, 3 H), 3.09–3.26 (m, 1 H), 3.29 (s, 3 H), 3.39–3.52 (m,
2 H), 3.87 (q, J = 6.8 Hz, 1 H), 4.36 (ddd, J = 46.8, 9.6, 4.6 Hz, 1
H), 4.40 (ddd, J = 46.8, 9.6, 4.6 Hz, 1 H), 7.17–7.37 (m, 5 H) ppm.
13C NMR (50 MHz, CDCl3): δ = 20.8, 34.2, 57.4, 57.8, 58.9, 62.4,
69.9, 70.1, 81.4, 84.8, 126.8, 127.3, 128.3, 145.4 ppm. HRMS (ESI):
calcd. for C13H20FNONa [M +Na]+ 248.1427; found 248.1424.
Supporting Information (see also the footnote on the first page of
this article): Cartesian coordinates, imaginary and low frequency
modes (B3LYP/6-31++G** optimized) for solvent-assisted ring-
opening transition states and single-point energies from
MPW1B95/6-31++G** calculations (Tables S1–S6).
N-[(S)-2-Fluoro-3-methoxypropyl]-N-methyl-N-[(R)-1-phenylethyl]-
amine (16c): Yield 19%, colourless oil. [α]2D5 = +25 (c = 0.86,
CHCl3). H NMR (200 MHz, CDCl3): δ = 1.37 (d, J = 6.8 Hz, 3
Acknowledgments
1
H), 2.27 (s, 3 H), 2.45–2.77 (m, 2 H), 3.36 (s, 3 H), 3.42–3.68 (m,
3 H), 4.53–4.87 (m, 1 H), 7.20–7.36 (m, 5 H) ppm. 13C NMR
(50 MHz, CDCl3): δ = 17.9, 39.6, 39.6, 54.5, 54.9, 59.3, 63.8, 73.0,
73.4, 90.2, 93.6, 126.9, 127.7, 128.1, 143.3 ppm. HRMS (ESI):
calcd. for C13H20FNONa [M +Na]+ 248.1427; found 248.1422.
The authors are indebted to the Fund for Scientific Research/Flan-
ders (Belgium) (FWO-Vlaanderen), Ghent University (GOA), the
Korea Science and Engineering Foundation (R01-2007-000-20037-
0) and the Center for Bioactive Molecular Hybrides for financial
support. The computational resources and services used in this
work were provided by Ghent University.
N-[(R)-1-Chloromethyl-2-methoxyethyl]-N-methyl-N-[(R)-1-phenyl-
ethyl]amine (15d): Spectral data derived from the crude reaction
mixture. Colourless oil. 1H NMR (400 MHz, CD3CN): δ = 1.36
(d, J = 6.8 Hz, 3 H), 2.39 (s, 3 H), 3.30 (s, 3 H), 3.23 (q, J = 6.8 Hz,
1 H), 3.48–3.68 (m, 3 H), 4.08–4.14 (m, 1 H), 7.21–7.35 (m, 5
H) ppm.
[1] a) U. M. Lindström, P. Somfai, Synthesis 1998, 109–117; b) B.
Zwanenburg, P. ten Holte, Top. Curr. Chem. 2001, 216, 93–124;
c) J. B. Sweeney, Chem. Soc. Rev. 2002, 31, 247–258; d) X. E.
Hu, Tetrahedron 2004, 60, 2701–2743; e) D. Tanner, Angew.
4930
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 4920–4931

Halide-Induced Ring Opening of 2-Substituted Aziridinium Salts
Chem. Int. Ed. Engl. 1994, 33, 599–619; f) H. M. I. Osborn, J.
Sweeney, Tetrahedron: Asymmetry 1997, 8, 1693–1715; g)
W. M. McCoull, F. A. Davis, Synthesis 2000, 1347–1365; h)
I. D. G. Watson, L. Yu, A. K. Yudin, Acc. Chem. Res. 2006,
39, 194–206; i) A. Padwa, S. S. Murphree, ARKIVOC 2006, 3,
6–33; j) D. S. Tsang, S. Yang, F. A. Alphonse, A. K. Yudin,
Chem. Eur. J. 2008, 14, 886–894; k) G. S. Singh, M. D’hooghe,
N. De Kimpe, Chem. Rev. 2007, 107, 2080–2135.
[8]
a) M. Sivaprakasham, F. Couty, G. Evano, B. Srinivas, R. Srid-
har, K. Rama Rao, ARKIVOC 2007, 10, 71–93; b) P. Gmeiner,
D. Junge, A. Kartner, J. Org. Chem. 1994, 59, 6766–6776.
C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789.
A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
K. Fukui, Acc. Chem. Res. 2002, 35, 363–368.
Y. Zhao, D. G. Truhlar, J. Phys. Chem. A 2004, 108, 6908–6918.
M. Gerenkamp, S. Grimme, Chem. Phys. Lett. 2004, 392, 229–
235.
[9]
[10]
[11]
[12]
[13]
[2] a) Y.-Q. Fang, R. Karisch, M. Lautens, J. Org. Chem. 2007, 72,
1341–1346; b) F. Couty, O. David, B. Larmanjat, J. Marrot, J.
Org. Chem. 2007, 72, 1058–1061; c) B. C. H. May, J. A. Zorn,
J. Witkop, J. Sherrill, A. C. Wallace, G. Legname, S. B. Prus-
iner, F. E. Cohen, J. Med. Chem. 2007, 50, 65–73; d) M.
D’hooghe, N. De Kimpe, Tetrahedron 2008, 64, 3275–3285.
[3] a) S. J. Whittaker, F. M. Foss, Cancer Treat. Rev. 2007, 33, 146–
160; b) W. A. Denny, Curr. Med. Chem. 2001, 8, 533–544.
[4] a) J. L. Pierre, P. Baret, E. M. Rivoirard, J. Heterocycl. Chem.
1978, 15, 817–823; b) A. R. Bassindale, P. A. Kyle, M. C. Soob-
ramanien, P. G. Taylor, J. Chem. Soc. Perkin Trans. 1 2000,
439–448; c) K. Weber, S. Kuklinski, P. Gmeiner, Org. Lett.
2000, 2, 647–649; d) D. Gnecco, L. Orea, F. A. Galindo, R. G.
Enríquez, R. A. Toscano, W. F. Reynolds, Molecules 2000, 5,
998–1003; e) B. Crousse, S. Narizuka, D. Bonnet-Delpon, J.-P.
Begue, Synlett 2001, 679–681; f) L. Testa, M. Akssira, E. Za-
ballos-García, P. Arroyo, L. R. Domingo, J. Sepúlveda-Arques,
Tetrahedron 2003, 59, 677–683; g) T. N. Wade, J. Org. Chem.
1980, 45, 5328–5333; h) G. M. Alvernhe, C. M. Ennakoua,
S. M. Lacombe, A. J. Laurent, J. Org. Chem. 1981, 46, 4938–
4948; i) P. O’Brien, T. D. Towers, J. Org. Chem. 2002, 67, 304–
307; j) T. Katagiri, M. Takahashi, Y. Fujiwara, H. Ihara, K.
Uneyama, J. Org. Chem. 1999, 64, 7323–7329; k) K. Higashi-
yama, M. Matsumura, K. Hiroshi, T. Yamauchi, Heterocycles
2009, 78, 471–485; l) H. A. Song, M. Dadwal, Y. Lee, E. Mick,
H.-S. Chong, Angew. Chem. Int. Ed. 2009, 48, 1328–1330; m)
M. Kumar, S. K. Pandey, S. Gandhi, V. K. Singh, Tetrahedron
Lett. 2009, 50, 363–365; n) H.-S. Chong, H. A. Song, M. Dad-
wal, X. Sun, I. Sin, Y. Chen, J. Org. Chem. 2010, 75, 219–221;
o) T.-X. Métro, B. Duthion, D. Gomez Pardo, J. Cossy, Chem.
Soc. Rev. 2010, 39, 89–102.
[5] a) T. B. Sim, S. H. Kang, K. S. Lee, W. K. Lee, H. Yun, Y.
Dong, H.-J. Ha, J. Org. Chem. 2003, 68, 104–108; b) M.
D’hooghe, V. Van Speybroeck, M. Waroquier, N. De Kimpe,
Chem. Commun. 2006, 1554–1556; c) M. D’hooghe, N.
De Kimpe, Synlett 2006, 2089–2093; d) M. D’hooghe, V.
Van Speybroeck, A. Van Nieuwenhove, M. Waroquier, N.
De Kimpe, J. Org. Chem. 2007, 72, 4733–4740; e) M.
D’hooghe, K. Vervisch, A. Van Nieuwenhove, N. De Kimpe,
Tetrahedron Lett. 2007, 48, 1771–1774; f) M. D’hooghe, K.
Vervisch, N. De Kimpe, J. Org. Chem. 2007, 72, 7329–7332; g)
Y. Kim, H.-J. Ha, S. Y. Yun, W. K. Lee, Chem. Commun. 2008,
4363–4365; h) S. Y. Yun, S. Catak, W. K. Lee, M. D’hooghe,
N. De Kimpe, V. Van Speybroeck, M. Waroquier, Y. Kim, H.-
J. Ha, Chem. Commun. 2009, 2508–2510; i) S. Catak, M.
D’hooghe, N. De Kimpe, M. Waroquier, V. Van Speybroeck, J.
Org. Chem. 2010, 75, 885–896; j) S. Catak, M. D’hooghe, T.
Verstraelen, K. Hemelsoet, A. Van Nieuwenhove, H.-J. Ha, M.
Waroquier, N. De Kimpe, V. Van Speybroeck, J. Org. Chem.
2010, 75, 4530–4541.
[14]
[15]
J. Tomasi, B. Mennucci, R. Cammi, Chem. Rev. 2005, 105,
2999–3094.
a) V. Barone, M. Cossi, J. Phys. Chem. A 1998, 102, 1995–
2001; b) M. Cossi, N. Rega, G. Scalmani, V. Barone, J. Comput.
Chem. 2003, 24, 669–681.
[16]
M. J. Frisch, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R.
Cheeseman, J. A. Montgomery Jr., T. Vreven, K. N. Kudin,
J. C. Burant, J. M. Millam, S. S. Iyengar, 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. Sal-
vador, 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. Ba-
boul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Lia-
shenko, P. Piskorz, I. Komaromi, 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 03, Revision D.01,
Gaussian, Inc., Wallingford, CT, 2004.
a) S. Catak, G. Monard, V. Aviyente, M. F. Ruiz-López, J.
Phys. Chem. A 2006, 110, 8354–8365; b) S. Catak, G. Monard,
V. Aviyente, M. F. Ruiz-López, J. Phys. Chem. A 2008, 112,
8752–8761; c) S. Catak, G. Monard, V. Aviyente, M. F. Ruiz-
López, J. Phys. Chem. A 2009, 113, 1111–1120; d) V. Van Spe-
ybroeck, K. Moonen, K. Hemelsoet, C. V. Stevens, M. Waro-
quier, J. Am. Chem. Soc. 2006, 128, 8468–8478; e) B. De Sterck,
R. Vaneerdeweg, F. Du Prez, M. Waroquier, V. Van Speyb-
roeck, Macromolecules 2010, 43, 827–836.
a) C. J. Cramer, D. G. Truhlar, in Continuum Solvation Models
(Eds.: O. Tapia, J. Bertrán), Kluwer, Dordrecht, 1996, pp. 1–
80; b) J. Tomasi, B. Mennucci, R. Cammi, Chem. Rev. 2005,
105, 2999–3094.
a) E. F. da Silva, H. F. Svendsen, K. M. Merz, J. Phys. Chem.
A 2009, 113, 6404–6409; b) C. P. Kelly, C. J. Cramer, D. G.
Truhlar, J. Phys. Chem. A 2006, 110, 2493–2499; c) J. R. Pliego,
J. M. Riveros, J. Phys. Chem. A 2001, 105, 7241–7247.
S. C. L. Kamerlin, M. Haranczyk, A. Warshel, ChemPhysChem
2009, 10, 1125–1134.
a) A. P. Bento, F. M. Bickelhaupt, J. Org. Chem. 2008, 73,
7290–7299; b) F. M. Bickelhaupt, J. Comput. Chem. 1999, 20,
114–128.
[17]
[18]
[19]
[20]
[21]
[22]
[23]
a) D. H. Ess, K. N. Houk, J. Am. Chem. Soc. 2008, 130, 10187–
10198; b) D. H. Ess, K. N. Houk, J. Am. Chem. Soc. 2007, 129,
10646–10647.
G. T. d. Jong, F. M. Bickelhaupt, ChemPhysChem 2007, 8,
1170–1181.
[6] a) N. De Kimpe, R. Jolie, D. De Smaele, J. Chem. Soc., Chem.
Commun. 1994, 1221–1222; b) N. De Kimpe, D. De Smaele, Z.
Szakonyi, J. Org. Chem. 1997, 62, 2448–2452; c) M. D’hooghe,
A. Waterinckx, N. De Kimpe, J. Org. Chem. 2004, 70, 227–232.
[7] M. D’hooghe, A. Waterinckx, T. Vanlangendonck, N.
De Kimpe, Tetrahedron 2006, 62, 2295–2303.
Received: April 9, 2010
Published Online: July 15, 2010
Eur. J. Org. Chem. 2010, 4920–4931
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4931