2-Isoxazolidineethanols: An NMR study of the effect of intramolecular H-bonding on the population of nitrogen invertomers and inversion process

Article abstract of DOI:10.1016/j.saa.2007.07.032

A series of (βR,5R)- and (βR,5S)-2,5-disubstituted isoxazolidines: 5-(substituent)-β-phenyl-2-isoxazolidineethanols, have been prepared by asymmetric nitrone cycloaddition reactions and their NMR spectra recorded over a wide range of temperatures. The spe

Full text of DOI:10.1016/j.saa.2007.07.032

Available online at www.sciencedirect.com
Spectrochimica Acta Part A 70 (2008) 482–490
2-Isoxazolidineethanols: An NMR study of the effect of intramolecular
H-bonding on the population of nitrogen invertomers and inversion process
Shaikh A. Ali∗, Muhammad Z.N. Iman, Mohamed I.M. Wazeer, Mohammed B. Fettouhi
Chemistry Department, King Fahd University of Petroleum and Minerals, Dhahran 31261, Saudi Arabia
Received 8 May 2007; received in revised form 18 July 2007; accepted 24 July 2007
Abstract
A series of (βR,5R)- and (βR,5S)-2,5-disubstituted isoxazolidines: 5-(substituent)-␤-phenyl-2-isoxazolidineethanols, have been prepared by
asymmetric nitrone cycloaddition reactions and their NMR spectra recorded over a wide range of temperatures. The spectra at low temperatures
indicate the presence of the (βR,5S) diasteromer almost exclusively as a single invertomer having trans disposition of the substituents at N(2) and
C(5), while the (βR,5R) diasteromer remained as a mixture of two interconverting invertomers in deuterated chloroform. The effect of H-bonding
– intramolecular in CDCl3 and intermolecular in CD3OD – on the population ratio of the invertomers and nitrogen inversion process has been
investigated. The nitrogen inversion barriers are determined using complete line-shape analysis, and their dependence on solvent is discussed. Due
to steric factor the trans-invertomers are found to be more stable than their cis counterparts.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Isoxazoldines; Nitrogen inversion; Invertomers; Inversion barriers
1. Introduction
ence of heteroatom oxygen slows down the lone pair inversion in
nitrogen to such an extent that at ambient or lower temperatures
the presence of two interconverting diastereomeric isoxazo-
lidines trans and cis could be identified by NMR spectroscopy.
We have prepared a series of 2,5-disubstituted isoxazolidines
1 and 2 using nitrone-alkene cycloaddition reactions to study
the configurational aspects as well as nitrogen inversion process
by NMR spectroscopy (Scheme 2). The study would provide
us with the opportunity to ascertain the stereochemistry of the
cycloaddition products and examine the effect of H-bonding –
intramolecular in CDCl3 and intermolecular in CD3OD – on
the population ratio of the invertomers: trans-1 versus cis-1 and
trans-2 versus cis-2 (Scheme 2).
Among a plethora of functional groups the nitrone function-
ality has etched a place of distinction in the synthesis of a great
many natural products of biological interest [1]. In recent years,
the focus has been shifted towards asymmetric nitrone cycload-
dition reactions [2–7]. Intramolecularly H-bonded (R)-chiral
nitrone as shown in Scheme 1 is receiving increasing attention
and becoming very popular owing to its efficacy in transfer-
ring chirality to the newly created stereocenters at C(5) of the
cycloadducts isoxazolidines 1 and 2 in a highly stereoselective
manner [8,9]. The facile reductive cleavage of the isoxazolidines
provides a simple and efficient access to a variety of syntheti-
cally important chiral 1,3-aminoalcohols. However, it is often
difficult to assign the configuration at C(5) by spectroscopic
analysis owing to complications arising out of slow nitrogen
inversion process involving cis- and trans-invertomers as well
as the pseudorotation in a five-membered ring.
2. Experimental
2.1. Compounds studied
The previous studies [10–12] of nitrogen inversion in bicyclic
isoxazolidines have greatly contributed to the understanding of
several reaction processes involving isoxazolidines. The pres-
A total of 12 compounds have been studied in the current
work. The structures of these compounds are given in Scheme 2.
2.2. Physical methods
The variable temperature 1H NMR spectra were recorded on
a JEOL Lambda NMR spectrometer operating at 500.0 MHz.
∗
Corresponding author. Fax: +966 3 860 4277.
E-mail address: shaikh@kfupm.edu.sa (S.A. Ali).
1386-1425/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.saa.2007.07.032

S.A. Ali et al. / Spectrochimica Acta Part A 70 (2008) 482–490
483
Scheme 1.
Most of the compounds were studied as 25 mg/cm3 solutions in
CDCl3 and CD3OD with TMS as internal standard. 1H and 13
having (R)(S) and (R)(R) configuration, respectively (Scheme 2).
Since the configuration of 2b is known by X-ray analysis as
(R)(R), the 1b must then have the (R)(S) configuration. The
obtainment of the acetyl derivatives 1c and 2c from 1b and
2b thus assures the stereochemistry of the former pair (vide
infra). The C(5) configuration of most of the other isoxazo-
lidines has been established by chemical conversion to known
structures [13]. The low temperature 1H and 13C NMR data of
these compounds are given below.
C
NMR spectra were measured in CDCl3 using TMS as internal
standard on a JEOL LA 500 MHz spectrometer. Multiplicities of
the carbons were determined using DEPT experiments. Elemen-
tal analysis was carried out on a EuroVector Elemental Analyzer
ModelEA3000. IRspectrawererecordedonaPerkin-Elmer16F
PC FTIR spectrometer. X-ray crystallographic analysis was car-
ried out on a Bruker-AXS Smart Apex system equipped with
˚
graphite-monochromatized Mo K␣ radiation (λ = 0.71073 A).
Silica gel chromatographic separations were performed with
Silica gel 100 from Fluka Chemie AG (Buchs, Switzerland).
2.3.1. (βR,5S)-5-Phenyl-β-phenyl-2-isoxazolidineethanol
(1a) and
(βR,5R)-5-phenyl-β-phenyl-2-isoxazolidineethanol (2a)
The isoxazolidines were prepared and purified by chromatog-
raphy to obtain a pure sample of a non-separable mixture of
isomers 1a and 2a, in a ratio of 73:27 [13]. This mixture was
used for the study. Careful analysis of the spectrum, measured in
CDCl3, at low temperature revealed the presence of two inver-
tomers for each of the isoxazolidines 1a and 2a. The spectrum
2.3. Synthesis of isoxazolidines (1 and 2)
The isoxazolidines (1, 2) a, b, d–f were prepared using proce-
dure as described [13]. The optically pure (R)-nitrone underwent
cycloaddition reactions with monosubstituted alkenes gave in
each case a mixture of two diastereomeric isoxazolidines 1 and 2
Scheme 2. Compounds studied.

484
S.A. Ali et al. / Spectrochimica Acta Part A 70 (2008) 482–490
was analyzed to obtain the signals for the individual isoazoli-
dine/invertomer as given below.
48.90, 52.64, 63.59, 68.87, 72.93, 127.69 (2C), 128.20 (2C),
128.80, 135.20, 174.12. Several nonoverlapping minor signals
indicated the major/minor invertomers in a ratio of 55:45.
In CD3OD (−30 ◦C), several nonoverlapping minor signals
indicatingthemajor/minorinaratioof70:30. C(5)Hofthemajor
and minor appeared at δ 4.63 and 4.48, respectively. Methyl
singlets appeared at 3.76 (major) and 3.69 (minor).
1a—Major invertomer: ␦H(CDCl3, −30 ◦C) 2.20 (1H, m),
2.77 (1H, m), 2.98 (1H, m), 3.28 (1H, m), 3.70 (1H, OH), 3.76
(1H, m), 4.05 (1H, m), 4.22 (1H, m), 5.39 (1H, t, J = 7.4 Hz),
7.40 (10H, m.; δC(CDCl3, −30 ◦C) 36.74, 53.94, 68.73, 70.83,
78.93, 128.8 (2C) and 129.8 (2C), 137.68 and 140.59.
Minor invertomer: C5(H) signal for minor invertomer appear
at 4.97 ppm (t, J = 7.5 Hz) in a ratio of 94:6. The 13C spec-
trum also revealed the presence of weak signals for the minor
invertomer of 1a at 37.22, 51.17, 63.51, 70.28, 78.59 ppm.
In CD3OD (−30 ◦C), several nonoverlapping minor signals
indicating the major/minor in a ratio of 79:21. C(5)H of the
major and minor appeared at δ 5.29 and 4.86 ppm, respectively.
2a—Major invertomer: δH(CDCl3, −30 ◦C) 2.1 (1H, m), 2.6
(1H, m), 2.65 (1H, m), 3.15 (1H, m), 3.55 (1H, m), 3.70 (1H,
OH), 4.00 (1H, m), 4.17 (1H, m), 5.16 (1H, t, J = 7.5 Hz), 7.4
(10H, m); δC(CDCl3, −30 ◦C) 36.26, 55.07, 68.42, 74.17, 80.22,
128.8–129.8 overlapping signals of aromatic carbons, 137.57
and 140.13.
2.3.3. (1R,5S)-Methyl 2-(2-acetyloxy-1-phenylethyl)-5-
isoxazolidinecarboxylate (1c) and (1R,5R)-methyl 2-
(2-acetyloxy-1-phenylethyl)-5-isoxazolidinecarboxylate (2c)
1c: A solution of the cycloadduct 1b (100 mg) and acetic
anhydride (0.5 g) in CHCl3 (2 cm3) was heated in a closed ves-
sel under N2 at 70 ◦C for 5 h. After removal of the solvent and
excess acetic anhydride by a gentle stream of N2, the resid-
ual liquid was taken up in ether (30 cm3) and washed with
5% NaHCO3 solution (10 cm3). The organic layer was concen-
trated and chromatographed using a mixture of hexane/ether
as eluant to afford the acetyl derivative 1c in 75% yield as a
colourless liquid—found: C, 61.2; H, 6.4; N, 4.7.; C15H19NO5
requires C, 61.42; H, 6.53; N, 4.78%; νmax.(neat) 3030, 2954,
2850, 1731, 1495, 1454, 1383, 1231, 1046, 833, 760 and
Minor invertomer: C5(H) signal for the minor invertomer of
minor isomer 2a appeared at 4.72 ppm (t, J = 7.1 Hz) in a ratio
of 91:9. The 13C spectrum also revealed the presence of weak
signals for the minor invertomer at 37.71, 52.13, 63.68, 71.35,
78.28 ppm.
703 cm−1
.
Major invertomer: δH(CDCl3, −30 ◦C) 1.99 (3H, s), 2.42
(1H, m), 2.61 (1H, m), 2.74 (1H, m), 3.02 (1H, m), 3.80 (3H,
s), 3.90 (1H, m), 4.38 (1H, m), 4.72 (1H, dd, J = 3.95, 11.0 Hz),
4.79 (1H, dd, J = 4.25, 9.75), 7.33 (5H, m); δC(CDCl3, −30 ◦C)
21.10, 31.87, 52.10, 52.70, 66.50, 68.46, 75.07, 128.24 (2C),
128.35, 128.61 (2C), 137.77, 171.09, 172.45.
In CD3OD (−30 ◦C), several nonoverlapping minor signals
indicating the major/minor of 2a in a ratio of 85:15. C(5)H of
the major and minor appeared at δ 5.03 and 4.83, respectively.
2.3.2. (1R,5S)-Methyl 2-(2-hydroxy-1-phenylethyl)-5-
isoxazolidinecarboxylate (1b) and (1R,5R,)-methyl 2-
(2-hydroxy-1-phenylethyl)-5-isoxazolidinecarboxylate (2b)
The isoxazolidines were prepared and separated by chro-
matography to obtain pure samples of isomers 1b and 2b [13].
Spectra, measured in CDCl3, at low temperature revealed the
presence of a single invertomer for 1b, and two invertomers for
2b.
Minor invertomer: δH(CDCl3, −30 ◦C) minor invertomer
revealed the presence of nonoverlapping acetyl protons at δ
2.09, CO2Me at 3.71 and C(5)H at 4.50 ppm. The ratio of the
invertomers was found to 87:13 by integration. The spectrum at
−30 ◦C revealed the presence of the following 13C signals for
the nonaromatic carbons of the minor invertomer: δC(CDCl3,
−30 ◦C) 21.08, 33.78, 52.10, 52.47, 66.15, 69.30, 74.67.
In CD3OD (−30 ◦C), several nonoverlapping minor signals
indicating the major/minor in a ratio of 91:9. Acetyl Me of
the major and minor appeared at δ 1.89 and 2.01, respectively.
Methyl singlets of CO2Me appeared at 3.74 (major) and 3.65
(minor).
1b—A single invertomer: δH(CDCl3, −30 ◦C) 2.42 (1H, m),
2.61 (2H, m), 3.01 (1H, m), 3.69 (1H, OH), 3.78 (1H, m),
3.83 (3H, s), 3.99 (2H, m), 4.76 (1H, dd, J = 4.0 and 9.5 Hz),
7.33 (5H, m); δC(CDCl3, −30 ◦C) 32.30, 52.50, 52.70, 67.10,
71.26, 74.37, 127.9 (2C), 128.1, 128.6 (2C), 138.19, 173.08. 1H
as well as 13C NMR spectra did not reveal the presence of a
minor invertomer. The proton signals were sharp even at room
temperature.
2c: The cycloadduct 2b was acetylated using procedure
as discussed above to give 2c in 73% yield as a colourless
liquid—found: C, 61.3; H, 6.5; N, 4.7; C15H19NO5 requires
C, 61.42; H, 6.53; N, 4.78%); νmax.(neat) 3030, 2992, 2955,
2853, 1738, 1444, 1380, 1228, 1045, 915, 842, 759, 733 and
In CD3OD (−30 ◦C), several nonoverlapping minor signals
indicatingthemajor/minorinaratioof89:11. C(5)Hofthemajor
and minor appeared at δ 4.78 and 4.35, respectively. Methyl
singlets appeared at 3.75 (major) and 3.66 (minor).
2b—Major invertomer: δH(CDCl3, −30 ◦C) 2.46 (2H, m),
3.12 (2H, m), 3.85 (3H, s), 3.65–4.40 (4H, several m), 4.71 (1H,
t, J = 7.6 Hz), 7.30 (5H, m); δC(CDCl3, −30 ◦C) 32.39, 52.91,
53.01, 68.07, 73.08, 76.68, 128.20 (2C), 128.80 (2C), 129.55,
138.26, 172.68.
705 cm−1
.
Major invertomer: δH(CDCl3, −30 ◦C) 2.0 (3H, s), 2.40 (1H,
m), 2.63 (1H, m), 2.84 (1H, td, J = 6.4, 10.7 Hz), 3.06 (1H, dt,
J = 7.0, 10.7 Hz), 3.82 (3H, s), 4.09 (1H, dd, J = 3.5, 6.3 Hz), 4.54
(1H, dd, J = 6.0, 11.0 Hz), 4.66(2H, m), 7.37(5H, m);δC(CDCl3,
−30 ◦C) 21.16, 32.65, 52.59, 52.74, 65.98, 69.15, 76.40, 128.24
(2C), 128.35, 128.61 (2C), 137.80, 171.12, 172.94.
Minor invertomer: δH(CDCl3, −30 ◦C) 2.09 (3H, s), 2.54
(2H, m), 2.72 (1H, q, J = 8.4 Hz), 3.49 (1H, m), 3.77 (3H,
s), 4.01 (1H, dd, J = 4.6, 7.0 Hz), 4.32 (1H, dd, J = 7.3,
11.9 Hz), 4.36 (1H, dd, J = 4.9, 11.9 Hz), 4.50 (1H, m), 7.37
Minor invertomer: δH(CDCl3, −30 ◦C) 2.46 (2H, m), 2.62
(1H, m), 2.92 (1H, m), 3.81 (3H, s), 3.65–4.40 (4H, complex m),
4.54 (1H,t, J = 6.6 Hz), 7.30 (5H, m); δC(CDCl3, −30 ◦C) 33.01,

S.A. Ali et al. / Spectrochimica Acta Part A 70 (2008) 482–490
485
(5H, m). The nonoverlapping signals for minor invertomer:
δC(CDCl3, −30 ◦C)33.09, 65.77, 73.63, 137.29, 170.99, 173.21.
Major/minor ratio was determined to be 70:30.
(3C), 29.37, 53.32, 64.44, 68.97, 70.92, 78.22, 127.81 (2C),
128.01, 129.13 (2C), 137.85.
In CD3OD (−30 ◦C), several nonoverlapping minor signals
including the t-Bu singlets at 0.96 (major) and 0.91 (minor)
and C(5)H at 4.31 (major) and 4.15 (minor) indicated the
major/minor of 1e in a ratio of 87:13.
In CD3OD (−30 ◦C), several nonoverlapping minor signals
indicating the major/minor in a ratio of 79:21. Acetyl Me of
the major and minor appeared at δ 1.90 and 2.00, respectively.
Methyl singlets of CO2Me appeared at 3.68 (major) and 3.60
(minor)
2e—Major invertomer: δH(CDCl3, −30 ◦C) 0.11 (3H, s),
0.12 (3H, s), 0.93 (9H, s), 2.04 (1H, m), 2.21 (1H, m), 2.58
(1H, q, J = 8.0 Hz), 2.98 (1H, m), 3.65–3.85 (5H, m), 4.13 (1H,
m), 4.29 (1H, m), 7.32 (5H, m); δC(CDCl3, −30 ◦C) (–)5.43,
(–)5.28, 18.24, 25.76 (3C), 29.31, 54.25, 64.30, 68.55, 73.65,
79.27, 127.70 (2C), 128.08, 129.23 (2C), 137.96. Major/minor
ratio was determined to be 97:3 as indicated by the presence of
minor tBu singlet.
2.3.4. (βR,5S)-5-(Hydroxymethyl)-β-phenyl-2-
isoxazolidineethanol (1d) and (βR,5R)-5-(hydroxymethyl)-
β-phenyl-2-isoxazolidineethanol (2d)
The isoxazolidines 1d and 2d were prepared by LiAlH4
reduction of the isoxazolidines 1b and 2b, respectively [13].
Spectra, measured in CDCl3, at low temperature revealed the
presence of a single invertomer for 1d, and two invertomers for
2d.
In CD3OD (−30 ◦C), several nonoverlapping minor signals
including the t-Bu singlets at δ 0.94 (major) and 0.84 (minor) in
a ratio 81:19.
1d—A single invertomer: δH(CDCl3, −30 ◦C) 2.11 (1H, m),
2.35 (1H, m), 2.73 (1H, m), 3.03 (1H, td, J = 7.8, 10.7 Hz), 3.67
(1H, dd, J = 4.5, 12.0 Hz), 3.75–4.05 (5H, m), 4.11 (1H, dd, 6.9,
11.1 Hz), 4.52 (1H, dtd, J = 2.15, 4.7, 8.7 Hz), 7.32 (5H, m).
δC(CDCl3, −30 ◦C) 28.81, 53.80, 64.50, 67.89, 71.28, 78.20,
127.82 (2C), 128.25, 128.79 (2C), 137.59.
2.3.6. (βR,5S)-5-[2-[[(1,1-Dimethylethyl)dimethylsilyl]
oxy]ethyl]-β-phenyl-2-isoxazolidineethanol (1f) and
(βR,5R)-5-[2-[[(1,1-dimethylethyl)dimethylsilyl]oxy]ethyl]-
β-phenyl-2-isoxazolidineethanol (2f)
The isoxazolidines were prepared and separated by chro-
matography to obtain pure samples of isomers 1f and 2f [13].
Spectra, measured in CDCl3, at low temperature revealed the
presence of a single invertomer for 1f, and two invertomers for
2f.
In CD3OD (−30 ◦C), several nonoverlapping minor signals
including C(5)H at δ 4.39 (major) and 4.25 (minor) indicated
the major/minor in a ratio of 88:12.
2d—Major invertomer: δH(CDCl3, −30 ◦C) 2.19 (2H, app.
q, J = 7.2 Hz), 2.67 (1H, td, J = 7.1, 10.4 Hz), 3.02 (1H, td,
J = 7.1, 10.4 Hz), 3.64 (1H, dd, J = 3.7, 12.5 Hz), 3.73 (1H, d,
J = 11.6 Hz), 3.92 (1H, dd, J = 2.2, 8.9 Hz), 4.20 (1H, dd, 8.2,
11.9 Hz), 4.0–4.4 (2H, br, OHs), 4.34 (1H, m), 7.32 (5H, m);
δC(CDCl3, −30 ◦C) 28.49, 54.39, 62.58, 68.40, 72.80, 80.36,
127.75 (2C), 128.22, 128.80 (2C), 137.75.
1f—A single invertomer: δH(CDCl3, −30 ◦C) 0.082 (3H, s),
0.087 (3H, s), 0.91 (9H, s), 1.84 (2H, m), 1.91 (1H, m), 2.45
(1H, m), 2.80 (1H, m), 3.05 (1H, m), 3.6–3.8 (4H, m), 3.89 (1H,
d, J = 6.1 Hz), 4.14 (1H, dd, J = 8.0, 11.1 Hz), 4.56 (1H, m), 7.32
(5H, m); δC(CDCl3, −30 ◦C) (–)5.50 (2C), 18.29, 25.81 (3C),
33.23, 37.61, 53.25, 59.92, 69.05, 70.88, 74.59, 127.78 (2C),
128.07, 128.69 (2C), 137.89.
Following are the nonoverlapping signals for the Minor inver-
tomer:δH(CDCl3, −30 ◦C)2.03(1H, m), 2.09(1H, m), 2.85(1H,
m), 3.33 (1H, m), 3.55 (1H, m), 3.86 (1H, m), 4.07 (1H, m), 4.12
(1H, m); δC(CDCl3, −30 ◦C) 29.94, 52.06, 63.68, 71.32, 80.26,
128.45, 128.99, 136.72. Major/minor ratio was determined to be
84:16.
In CD3OD (−30 ◦C), several nonoverlapping minor signals
including the t-Bu singlets at 0.83 (major) and 0.76 (minor)
indicated the presence of the invertomers of 1f in a ratio 82:18.
2f—Major invertomer: δH(CDCl3, −30 ◦C) 0.086 (6H, s),
0.91 (9H, s), 1.87 (3H, m), 2.31 (1H, m), 2.51 (1H, m), 2.98
(1H, m), 3.6–3.8 (5H, m), 4.14 (1H, m), 4.29 (1H, m), 7.32 (5H,
m); δC(CDCl3, −30 ◦C) (–)5.50 (2C), 18.40, 25.96 (3C), 33.35,
38.11, 54.42, 60.41, 69.02, 71.08, 74.12, 127.81 (2C), 127.89,
128.76 (2C), 137.84.
In CD3OD (−30 ◦C), several nonoverlapping minor signals
including C(5)H at 4.17 (major) and 4.05 (minor) indicated the
major/minor in a ratio of 90:10.
2.3.5. (βR,5S)-5-[[[(1,1-Dimethylethyl)dimethylsilyl]oxy]
methyl]-β-phenyl-2-isoxazolidineethanol (1e) and (βR,5R)-
5-[[[(1,1-dimethylethyl)dimethylsilyl]oxy]methyl]-β-
phenyl-2-isoxazolidineethanol (2e)
The isoxazolidines were prepared and separated by chro-
matography to obtain pure samples of isomers 1e and 2e [13].
Spectra, measured in CDCl3, at low temperature revealed the
presence of a single invertomer for 1e, and two invertomers for
2e.
Minor invertomer: the presence of the minor invertomer was
displayed by minor signals at δ 0.87 (tBu) and 4.29 (C-5, H) in
a 92:8 ratio.
In CD3OD (−30 ◦C), several nonoverlapping minor signals
including the t-Bu singlets at 0.83 (major) and 0.76 (minor)
indicated the presence of the invertomers in a ratio of 93:7.
2.4. Inversion barrier calculations
1e—A single invertomer: δH(CDCl3, −30 ◦C) 0.093 (6H, s),
0.91 (9H, s), 1.92 (1H, m), 2.31 (1H, m), 2.73 (1H, m), 3.00 (1H,
td, J = 7.6, 10.6 Hz), 3.65(2H, m), 3.74(1H, dd, J = 6.3, 10.6 Hz),
3.87 (1H, d, J = 5.8 Hz), 4.15 (1H, dd, J = 7.65, 11.6 Hz), 4.46
(1H, m), 7.32 (5H, m); δC(CDCl3, −30 ◦C) (–)5.42, 18.25, 25.76
The variable temperature 1H NMR spectra were recorded on
a JEOL Lambda NMR spectrometer operating at 500.0 MHz.
Most of the compounds were studied as 25 mg/cm3 solutions in
CDCl3 and CD3OD with TMS as internal standard. Simulations
of exchange-affected proton spectra for all compounds were car-

486
S.A. Ali et al. / Spectrochimica Acta Part A 70 (2008) 482–490
ried out using a computer program AXEX [14], corresponding to
a two non-coupled sites exchange with unequal populations. For
1b(CD3OD), 2b(CDCl3 andCD3OD), 1c(CDCl3), 2c(CDCl3),
C(5)CO2Me resonances were utilized, whereas for 1c (CD3OD),
and 2c(CD3OD), the OAc singlet signals were utilized. For 1e
(CD3OD), 2e (CD3OD) and 1f (CD3OD) the tert-butyl singlets
were utilized. Simulations of exchange-affected triplets were
carried out by modifying the two-site exchange program [15].
The first order coupling to these protons is simply assumed as
giving overlapping two site exchanges with the same population
ratio and equal rates of exchange. Similar simulation was carried
out for the quartet signals of a C(3)H in 2d (CDCl3).
trans-2 (or cis-1 and cis-2) (Scheme 1) should have compara-
ble steric environments, and as such a large discrepancy in the
population ratio of the invertomers in 1b and 2b is quite unex-
pected. The current study helped us to provide a rational for this
observation (vide infra).
The nitrogen inversions barriers are determined using NMR
band shape analysis. Slow nitrogen inversion in most of the isox-
azolidines, especially in the series 2, has been observed to give
broadened peaks in 1H and 13C spectra recorded at ambient tem-
perature. On lowering the temperature, the spectral lines become
sharper and show two distinct forms of the compound. The 13
C
chemical shifts were assigned on the basis of DEPT experiment
results, general chemical shifts arguments and consideration of
substituent effects, and are given in Table 1.
3. Results and discussion
1
Around −10 ◦C, the H NMR spectra of these compounds
show well-separated signals for the two invertomers. Integra-
tion of the relevant peaks gives the population trends in these
systems. The proton spectra were used in the calculation of barri-
ers in all compounds. The complete band shape analysis yielded
the rate constants and the free energy of activation using Eyring
equation. The activation parameters ꢀH= and ꢀS= were cal-
culated from plots of ln(k/T) versus 1/T. It is well known [16]
that NMR band shape fitting frequently gives rather large but
mutually compensating errors in ꢀH= and ꢀS= and as such
their values are not reported here. However, band shape fitting
is viewed as a method of getting rather accurate values of ꢀG=
(probably within 0.3 kJ/mol) in the vicinity of the coalescence
Each nitrone–alkene cycloaddition afforded a separable mix-
ture of diastereomeric isoxazolidines 1 and 2 having (R,S) and
(R,R) configurations [13], respectively (priority based on assum-
ing R1 as oxygenated substituents). Since the nitrone is optically
pure having ’R’ configuration, the isoxazolidines differ only in
the configuration of the C(5) substituents. During the course of
a structural investigation of the isomeric isoxazolidines 1b and
2b (R = CO2Me) it was interesting to observe the presence of a
single invertomer for one of the isoxazolidines, while the other
remained as an equilibrating mixture of two invertomers (cis and
trans) in a 55:45 ratio at −30 ◦C in CDCl3. This is surprising
since a look at their structures seems to covey that trans-1 and
Table 1
13C NMR chemical shifts of compounds studied in CDCl3 at −30 ◦C

S.A. Ali et al. / Spectrochimica Acta Part A 70 (2008) 482–490
487
Table 2
Free energy of activation (ꢀG=) for nitrogen inversion, ratio of the invertomers, and standard free energy change (ꢀG◦) for major ↔ minor isomerization in CDCl3
and CD3OD
Compound
CDCl3
CD3OD
ꢀG= (kJ/mol)a
Invertomer ratio
ꢀG◦ (kJ/mol)b
+5.3
+4.5
–
+0.38
+3.7
+1.7
–
+3.2
–
ꢀG= (kJ/mol)a
Invertomer ratio
ꢀG◦ (kJ/mol)b
1a
2a
1b
2b
1c
2c
1d
2d
1e
2e
1f
62.2
61.2
–
55.4
60.1
58.7
–
59.6
–
94:6
91:9
–
–
60.3
58.8
59.9
59.6
79:21
85:15
89:11
70:30
91:9
79:21
88:12
90:10
87:13
81:19
82:18
93:7
–
–
+4.0
+1.6
+4.5
+2.6
100:∼0
55:45
87:13
70:30
100:∼0
84:16
100:∼0
97:3
63.0
61.6
60.6
60.4
–
+4.3
+3.7
+2.8
+2.9
–
–
–
–
+6.9
–
–
100:∼0
2f
92:8
a
At 0 ◦C.
At −30 ◦C.
b
temperature. The ꢀG= values calculated at 0 ◦C are reported in
Table 2, along with the invertomer and ꢀG◦ values.
Even though III is depicted as pseudo diequatorially substi-
tuted, there will be considerable steric repulsion between the
substituents since staggering in the heterocyclic five-membered
ring is not as pronounced as in the six-membered systems. In
fact actual conformations of the trans isomer would lie some-
where between I and II and that of the cis-invertomer between
III and IV.
Both the cis-1,3-dimethylcyclopentane and cis-1,3-
dimethylcyclohexane are more stable than their trans
counterparts by an enthalpy difference of 2.3 kJ/mol and
7.1 kJ/mol, respectively. The slight preference for the cis
isomer in cyclopentane disappears in heterocyclic systems
like isoxazolidines, where the presence of two heteroatoms
in the ring skeleton would shorten the bond lengths thus
augmenting the steric congestion between the cis substituents.
The 2,5-disubstituted isoxaozolidines have thus been found to
have a slight preference for the trans-invertomers [17]. The
conformation of five-membered ring system is indeed very
complex to elucidate with some certainty. The complexity
arises from the fact that changing the size of the substituent
may lead to change in conformation (half chair/envelope/near
planar) and the flap of the envelope.
Earlier works [17,18] on 2,5-disubstituted isoxa-
zolidines revealed the trans-invertomer as the major
isomer. The 2-methyl-, 2-isopropyl- and 2-tbutyl-5-
tbutyldimethylsiloxymethylisoxazolidines were found to
have the trans- and cis-invertomers in a ratio of 53:47, 55:45
and 63:37, respectively. The compounds studied in this work
are sterically similar to the 2-isopropylisoxazolidines since
they also contain a secondary alkyl (i.e. phenylhydroxyethyl)
substituent at the 2-position (Scheme 2). It is interesting to note
that the isoxazolidines 1b, 1d–1f exist as a single invertomer
in each case, and display sharp NMR signals at ambient as
well as lower temperatures. However, all the isoxazolidines in
the series 2, revealed the presence of two invertomers at the
lower temperatures. For instance, while the isoxazolidine 1b
exists as a single invertomer, the corresponding diastereomer
2b shows the presence of two invertomers in a ratio 55:45,
respectively. This is surprising since the trans-1 and trans-2 (or
cis-1 and cis-2) (Scheme 2) seem to have comparable steric
environments, and as such a large difference in the population
ratio of the invertomers in 1b and 2b demands an explanation
as to the origin of such a descrepency. Similar trend is observed
for the other isoxazolidines; while most of the isoxazolidines
in series 1 remained exclusively as a single invertomer in each
case, the compounds 2 show the presence of two invertomers.
The analysis below will help us to explain the subtle differences
in the stereochemistry between the isoxazolidines 1 and 2
and provide rationale for identifying the major invertomers as
having trans configuration.
The dynamic cycle depicting the pseudorotation (Pr) and
nitrogen inversion (Ni) in isoxazolidines is shown in Scheme 3.
The likely configuration of the invertomers of the isoxazo-
lidines is shown in Scheme 4. The Newman projections around
the exocyclic C–N bond are shown in Scheme 5. Benzylic ’H’
Scheme 3. Pseudororation (Pr) and nitrogen inversion (Ni) in isoxazolidines (2).

488
S.A. Ali et al. / Spectrochimica Acta Part A 70 (2008) 482–490
Scheme 4.
is placed anti to the nitrogen lone pair in all the projections
since it will have the lower number of gauche interactions (two
in these cases). Molecular models reveal that these conforma-
tions do not help the OH group make effective H-bonding with
the nitrogen lone pair. The H-bonding in that case will be in a
less favourable five-membered ring structure and will develop
considerable eclipsing between the phenyl group and N–C(3)
bond of the isoxazolidine ring. Moreover, the nitrogen in the
isoxazolidine moiety is much less basic than in the ordinary
amines (vide infra) due to the presence of the adjacent oxygen,
and thus more reluctant to allow its lone pair to participate in
an H-bond formation. This would leave the possibility of H-
bonding with the ring oxygen in trans-1 and cis-2 invertomers
as shown in Schemes 4 and 5. The cis-1 and trans-2 cannot
form H-bonded structure. The trans-1 has the two advantages
– sterically favored trans disposition of the substituents as well
as H-bonding – rendering it overwhelmingly favored over the
cis-1 and as such it exists as the sole invertomer. For the isox-
azolidines 2, each invertomer enjoys one advantage; while the
cis-2 is H-bonded, the trans-2 enjoys the trans disposition of
the substituents. As a result, both the invertomers exist for the
isoxazolidines 2.
To confirm the stereochemistry, adduct 2b (the only crys-
talline compound among the isoxazolidines studied) was
subjected to X-ray crystallographic analysis; the ORTEP rep-
resentation is shown in Scheme 6. The isoxazolidine 2b exists
in CDCl3 as two invertomers in a ratio of 55:45. Needless to say,
the favourable invertomer in the solid state, as dictated by the
crystal packing forces, may not be the conformation of choice in
the solution. However, it was found that the isoxazolidine prefers
to be in the trans-2 invertomer in the solid as well as in solution.
It can be seen that the trans-2 invertomer is not intramolecu-
larely H-bonded and benzylic ’H’ is anti to the nitrogen lone
pair as anticipated in Schemes 4 and 5. The relative stabilization
imparted by the substituents in trans disposition in trans-2, and
H-bonding in cis-2, dictate the population ratio of the isoxazo-
lidines 2 and will vary depending on the size of the substuituent
at C(5). As evident from Table 2, the CO2Me substituent at
C(5) in 2b being smaller than the C(5) substituents in all the
other isoxazolidines 2, gave the highest percentage of the cis-2
invertomer in CDCl3.
The above discussion regarding H-bonding gets credence
when the hydroxyl group in 1b and 2b is acetylated to give 1c
and 2c, respectively. In the absence of H-bonding, the trans-1b
Scheme 6. ORTEP drawing of 2b.
Scheme 5.

S.A. Ali et al. / Spectrochimica Acta Part A 70 (2008) 482–490
489
Table 3
and cis-1b ratio of 100:∼0 is changed to 87:13 for the trans-1c
and cis-1c, respectively, while the trans-2b and cis-2b ratio of
55:45 is changed to 70:30 for the trans-2c and cis-2c. As can
be seen from Scheme 4, after acetylation the amount of trans-
1c and cis-2c decreases in comparison to trans-1b and cis-2b
owing to the absence of H-bonding.
1H NMR chemical shifts of C(5)H signals of the compounds studied in CDCl3
and CD3OD at −30 ◦C
Isoxazolidine
δ (ppm)
CDCl3
CD3OD
Major-trans
Minor-cis
Major-trans
Minor-cis
Assignment of major invertomers to the trans conformers
gets further credence by studying the population ratio in CD3OD
(Table 2). In methanol solvent, the nitrogen as well as oxygen
lone pairs will be involved in H-bonding with the solvent. As
such, in the absence of intramolecular H-bonding, the trans-1
and cis-2 will loose their advantages in favor of the cis-1 and
trans-2, respectively. This is exactly what is observed: while
most of the cis-1 invertomers are absent in CDCl3, a consider-
able proportion exists in CD3OD. However, for 2 in CD3OD the
results are mixed; while the population of cis-2 is expected to
decrease owing to the disruption of intramolecular H-bonding
(Schemes 4 and 5), the solvation of the nitrogen lone pair will
act in opposite way to increase its population. The overall effect
on the population ratio would then depend on the relative impor-
tance of the two effects.
1a
2a
1b
2b
1c
2c
1d
2d
1e
2e
1f
5.39
5.16
4.76
4.71
4.79
4.66
4.52
4.34
4.46
4.29
4.56
4.29
4.97
5.29
5.03
4.78
4.63
4.80
4.67
4.39
4.17
4.31
4.17
4.38
4.42
4.86
4.83
4.35
4.48
4.42
4.48
4.25
4.05
4.15
4.00
4.12
4.27
4.72
a
–
4.54
4.50
4.50
a
–
–
–
–
–
b
a
b
a
b
2f
–
a
No minor invertomer.
Could not be detected as a result of very weak signal or overlapping with
b
other signals.
The chemical shift difference between the isomers for a par-
ticular ring carbon is generally less than 1 ppm for most carbons
and as such the C-13 shifts are not very sensitive to the dif-
ference in the isomeric conformations (Table 1). This is not
surprising in view of the fact that the five-membered ring does
not have the well-defined conformation of six-membered sys-
for all the isoxazoldines. This is expected since the steric require-
ments to attain the sp2 hybridized transition state (through which
the nitrogen inversion occurs) remains more or less similar as
the substituents in the immediate vicinity of nitrogen remains the
same in all the isoxazolidines. The inversion barrier increases to
some extent in hydrogen bonding solvent CD3OD. Any increase
in the barrier in cyclic system is attributed to the extra energy
required for breaking of H-bonding prior to inversion [19]. The
H-bonding involving the nitrogen lone pair may be minimal
since the nitrogen in the isoxazolidine moiety is much less basic
than in the ordinary amines (cf. pKb of Me3N (4.19) versus pKb
of Me3N–OMe 10.35) [20]. The electron-withdrawing power
of the oxygen reduces the electron-donating power of the nitro-
gen by a large amount. The presence of electronegative oxygen
attached to the nitrogen results in more ’s’ character in the lone
pair orbital and as such the nitrogen would form very weak H-
bonds. A more effective intramolecular H-bonding to the ring
oxygen in CDCl3 or intermolecular H-bonding with CD3OD
could be achieved, however, in such a scenario there would be
only minimal effect, if any, on the inversion barriers [21].
The NMR study at lower temperatures has thus led success-
fully the foundation using which the stereochemistry of this
important class of chiral nitrone cycloaddition reactions might
be determined.
1
tems. While the use of H NMR coupling constant using the
Karplus equation is highly successful in assigning configuration
of six-membered rings, such applications often do not work well
with isoxazolidines or any five-membered ring systems. For the
isomers 1a and 2a, while the C-3, C-5 and N–C (external C
attached to N) of the trans-invertomers appear at lower fields,
the C-4 appears at slightly higher fields than the corresponding
cis invertomers (Table 1). The N–C of the cis-invertomer absorbs
upfield presumably due to the crowded environment of the car-
bon in a cis disposition with the C(5) substituents. Similar trends
are observed in the compounds 2b and 2d. Such trends are not
observed in the chemical shift values of 1c, 2c pair owing to the
absence of H-bonding. Further evidence that the major isomer
in the pairs 1e, 2e and 1f, 2f have the same configuration (i.e.
trans) comes the similarity in the chemical shifts values within
the pair.
It is well known that an axial proton in cyclohexane systems
appears highfield than the equatorial proton. Further evidence
that the major invertomers in all these isoxazolidines have
the same trans configuration comes from the signal of C(5)H
which, by virtue of being pseudo-equatorial in one of the trans-
invertomer (Scheme 4), invariably appears downfield in compar-
ison to the pseudoaxially disposed C(5)H in the cis-invertomers
(Table 3). The C(5)H of the trans-1, in turn, appears downfield
with respect to that of the corresponding trans-2 invertomer; a
diagnostic trend that would enable us to identify the stereochem-
istry of this important class of cycloaddition reactions.
Acknowledgement
The facilities provided by the King Fahd University of
Petroleum and Minerals, Dhahran, are gratefully acknowledged.
References
The nitrogen inversion barrier is expected to be high when an
oxygen atom is directly attached to the nitrogen as in isoxazo-
lidines [11,19]. The inversion barriers hover around 60 kJ/mol
[1] (a)J.J. Tufariello, in:A. Padwa(Ed.), 1,3-DipolarCycloadditionChemistry,
vol. 2, Wiley-Interscience, New York, 1984, pp. 83–168, Chapter 9;
(b) P.N. Confalone, E.M. Huie, Org. React. 36 (1988) 1–173.

490
S.A. Ali et al. / Spectrochimica Acta Part A 70 (2008) 482–490
[2] (a) X. Ding, K. Taniguchi, Y. Ukaji, K. Inomata, Chem. Lett. (2001)
[11] M.I.M. Wazeer, S.A. Ali, Magn. Reson. Chem. 31 (1993) 12–16.
[12] M.I.M. Wazeer, H.A. Al-Muallem, S.A. Ali, J. Phys. Org. Chem. 6 (1993)
326–332.
[13] S.A. Ali, M.Z.N. Iman, Tetrahedron 63 (2007) 9134–9145.
[14] The NMR Program Library, Science and Engineering Research Council,
Daresbury Laboratory, Chesire, UK.
[15] M.I.M. Wazeer, S.A. Ali, Canad. J. Appl. Spectrosc. 38 (1993) 22.
[16] J. Sanstrom, Dynamic NMR Spectroscopy, Academic Press, London, 1982.
[17] M.I.M. Wazeer, S.A. Ali, Canad. J. Appl. Spectrosc. 40 (1995) 53–60.
[18] M.I.M. Wazeer, S.M.A. Hashmi, S.A. Ali, Canad. J. Anal. Sci. Spectrosc.
42 (1997) 190–195.
[19] M. Raban, F.B. Jones Jr., E.H. Carlson, E. Banucci, N.A. LeBel, J. Org.
Chem. 35 (1970) 1496–1499, and references cited therein.
[20] T.C. Bissot, R.W. Parry, D.H. Campbell, J. Am. Chem. Soc. 79 (1956)
796–800.
468–469;
(b) W.S. Jen, J.J.M. Wiener, D.W.C. MacMillan, J. Am. Chem. Soc. 122
(2000) 9874–9875.
[3] K.V. Gothelf, K.A. Jorgensen, Chem. Commun. (2000) 1449–1458.
[4] (a) S. Karlsson, H.-E. Ho¨gberg, Org. Perp. Proceed. Int. 33 (2001) 103–172;
(b) G. Zecchi, G. Broggini, Synthesis (1999) 905–917;
(c) K.V. Gothelf, K.A. Jorgensen, Chem. Rev. 98 (1998) 863–909;
(d) M. Frederickson, Tetrahedron 53 (1997) 403–425.
[5] S. Kanemasa, Synlett (2002) 1371–1387.
[6] Y. Ukaji, K. Taniguchi, K. Sada, K. Inomata, Chem. Lett. (1997) 547–548.
[7] X. Ding, K. Taniguchi, Y. Hamamoto, K. Sada, S. Fujinami, Y. Ukaji, K.
Inomata, Bull. Chem. Soc. Jpn. 79 (2006) 1069–1083.
[8] R. Hanselmann, J. Zhou, P. Ma, P.N. Confalone, J. Org. Chem. 68 (2003)
8739–8741.
[9] Q. Zhao, F. Han, D.L. Romero, J. Org. Chem. 67 (2002) 3317–3322.
[10] S.A. Ali, M.I.M. Wazeer, Tetrahedron Lett. 34 (1993) 137–140.
[21] M.I.M. Wazeer, H.A. Al-Muallem, S.S. Fayyaz, S.A. Ali, Canad. J. Appl.
Spectrosc. 40 (1995) 27–30.