Determination of the absolute stereochemistry and the activation barriers of thermally interconvertible heterocyclic compounds bearing a naphthyl substituent

Article abstract of DOI:10.1016/j.tetasy.2005.09.019

The enantiotopic methyl signals of the compounds studied were resolved in the presence of the optically active chiral auxiliary (S)-(+)-2,2,2- trifluoroanthryl ethanol, [(S)-TFAE] via complex formation between (S)-TFAE and the respective compounds. Two different solvation models were proposed for both M and P conformations leading to the assignments of the ¹H NMR signals and thus absolute conformations. The solvation models proposed also explained the strong temperature dependence of the ¹H NMR signals upon cooling. The activation barriers for interconversion between the enantiomers of the compounds studied have been determined by either temperature dependent NMR or enantioresolution on a chiral sorbent via HPLC.

Full text of DOI:10.1016/j.tetasy.2005.09.019

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 16 (2005) 3752–3761
Determination of the absolute stereochemistry and the
activation barriers of thermally interconvertible
heterocyclic compounds bearing a naphthyl substituent
¨
_
*
Oznur Demir-Ordu, Esra M u¨ jde Yılmaz and Ilknur Do g˘ an
Bo g˘ azi c¸ i University, Chemistry Department, Bebek, 34342 Istanbul, Turkey
Received 14September 2005; accepted 28 September 2005
Available online 9 November 2005
Abstract—The enantiotopic methyl signals of the compounds studied were resolved in the presence of the optically active chiral aux-
iliary (S)-(+)-2,2,2-trifluoroanthryl ethanol, [(S)-TFAE] via complex formation between (S)-TFAE and the respective compounds.
1
Two different solvation models were proposed for both M and P conformations leading to the assignments of the H NMR signals
1
and thus absolute conformations. The solvation models proposed also explained the strong temperature dependence of the H NMR
signals upon cooling. The activation barriers for interconversion between the enantiomers of the compounds studied have been
determined by either temperature dependent NMR or enantioresolution on a chiral sorbent via HPLC.
ꢀ
2005 Elsevier Ltd. All rights reserved.
1. Introduction
oxygen present and (3) the existing p–p interaction
between the aromatic rings.
Supramolecular interactions such as hydrogen bonding
and p stacking between aromatic rings play an impor-
tant role in diverse areas such as the stereochemistry
The importance of a p–p interaction between the o-
substituted phenyl ring of the compounds studied,
and the anthryl ring of the chiral auxiliary led us to syn-
thesize a series of axially chiral heterocyclic compounds
bearing a naphthyl ring (Scheme 1), which can be
involved in enhanced p stacking interactions with the
6
1
2,3
of organic reactions, and host–guest chemistry. The
dependence of spectroscopic properties on these type
of interactions has been used for the determination of
4
,5
absolute stereochemistries for many years. Diastereo-
metric association complexes formed through these kind
of interactions of chiral molecules show non-identical
spectral behaviour on the NMR time scale. With knowl-
edge of the structures of the solvates and on the basis of
non-equivalance, absolute stereochemistries can be
determined.
R
R
R
R
X1
2
X1
2
5
5
4
4
Y
3
O
Y
3
O
N
N
Recently, we proposed a solvation model for the deter-
mination of the absolute conformations of the 5,5-
6
dimethyl-3-(o-aryl)-2,4-oxazolidinedione enantiomers.
In this model, the solute enantiomers interact with (S)-
(
+)-2,2,2-trifluoroanthryl ethanol, (S)-TFAE, in three
P
M
ways: (1) a two-point interaction complex formed with
the lactone part of the ring, (2) an association of the
(
(
(
(
(
(
±)-1 R=CH
3
, X=Y=O
, X=O, Y=S
S)-TFAE hydroxyl proton with the amide carbonyl
±)-2 R=CH
3
±)-3 R=H, X=Y=O
±)-4 R=H, X=Y=S
±)-5 R=H, X=S, Y=O
*
Corresponding author. Tel.: +90 212 359 6620; fax: +90 212 287
467; e-mail: dogan@boun.edu.tr
2
Scheme 1. The compounds synthesized.
0957-4166/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2005.09.019

¨
O. Demir-Ordu et al. / Tetrahedron: Asymmetry 16 (2005) 3752–3761
3753
anthryl ring of the (S)-TFAE. In these compounds, axial
compounds (± )-1 and (± )-2 also gave two distinct sing-
chirality arises from hindered rotation around the
lets in toluene-d , with same chemical shift difference of
8
Csp
2
–Nsp
2
single bond and makes the compounds
0.6 ppm. These results proved that there was a restricted
7
enantiomers.
rotation around the Csp
chiral axis.
2
–Nsp , bond making this bond a
2
Herein, we are concerned about the details of interac-
tions between enantiomeric naphthyl substituted com-
pounds (Scheme 1) and (S)-TFAE revealed by a
1
2.2. HNMR in the presence of a chiral auxiliary
1
strong temperature dependence of the H NMR spectra
The existence of the two enantiomeric forms for
1
of the diastereomeric association complexes, which was
not observed for the ortho-aryl substituted compounds
of the same series. This led us to investigate the possibil-
ity of determining the absolute conformations of the
enantiomers by an NMR method in the presence of an
optically active auxiliary. Herein, we also report the acti-
vation barriers to the hindered rotation of the com-
pounds studied, determined either by temperature
dependent NMR or by thermal racemization on a chiral
sorbent via HPLC.
compounds (± )-1–5 was also proven by H NMR spec-
troscopy in the presence of 6 equiv of (S)-(+)-2,2,2-tri-
fluoroanthryl ethanol, (S)-TFAE. Enantiomeric groups
become diastereotopic by non-equivalent interactions
6
with (S)-TFAE as proposed previously, and thus dis-
1
played unequal H NMR chemical shifts.
1
In the H NMR spectra of (± )-1 the expected four sing-
lets of the C-5 methyl groups were observed in the pres-
ence of (S)-TFAE with a chemical shift difference of
0.01 ppm for the upfield and 0.02 ppm for the downfield
pair of the signals in toluene-d (Fig. 1). For compound
8
2
. Results and discussion
(± )-2 however, only one enantiomeric pair for the C-5
methyl groups was distinguished under the same condi-
tions. The two separated singlets appeared downfield
with a chemical shift difference of 0.01 ppm in toluene-
d8 (Fig. 1).
1 13
.1. Hand C NMR
2
The naphthyl bearing heterocyclic compounds exist as
two enantiomeric M and P atropisomers (Scheme 1),
which exhibit identical NMR spectra in an achiral sol-
vent. However, due to hindered rotation around the
–Nsp single bond, the C-5 methyl protons of the
compounds 5,5-dimethyl-3-(a-naphthyl)-2,4-oxazolidin-
edione (± )-1 and 5,5-dimethyl-3-(a-naphthyl)-2-thioxo-
-oxazolidinone (± )-2 and the C-5 protons of the
compounds 3-(a-naphthyl)-2,4-oxazolidinedione (± )-3,
-(a-naphthyl)-rhodanine (± )-4, 3-(a-naphthyl)-2,4-
In the presence of (S)-TFAE, the two expected AB
splittings were differentiated on the NMR time scale
for the C-5 methylene protons of compounds (± )-3
and (± )-5, with the chemical shift difference being
equal to 0.02 ppm for the downfield and 0.01 ppm
for upfield pair of the signals (Fig. 1). The two AB
splittings for the C-5 methylene protons of compound
(± )-4 were also observed with a chemical shift differ-
ence of 0.02 ppm for the downfield signals; however
the upfield part of the spectrum could not be resolved
(Fig. 1D).
Csp
2
2
4
3
thiazolidinedione (± )-5 are diastereotopically related
and should have unequal chemical shifts if the barrier
to rotation is slow on the NMR time scale. Analysis
1
13
of the H and C NMR spectra of these compounds
did show these magnetically non-equivalent protons
With these results in hand, it was concluded that the
existence of the –C@S group on C-2 [compounds (± )-2
and (± )-4] decreases the resolution of the enantiomeric
pairs of these compounds compared to the –C@O ana-
logues. The observed decrease in the resolution for the
enantiomeric upfield signals may arise from a structural
change in the solvation model that will be discussed
later.
(
(
Table 1). The two methyl groups on C-5 in compounds
± )-1 and (± )-2 gave two separate singlets in toluene-d8
with a chemical shift difference of 0.04ppm. The pro-
tons at C-5 of (± )-3, (± )-4 and (± )-5 on the other hand,
showed AB type splittings, the chemical shift differences
being equal to 0.06, 0.07 and 0.06 ppm, respectively. The
1
3
anisochronous C nuclei of the C-5 methyl groups in
1
13
Table 1. H and C NMR spectroscopic data for the synthesized compounds in the presence and absence of the optically active auxiliary, (S)-TFAE
at 30 ꢁC
1
1
13
Compound no.
Medium
H NMR, ppm, C-5 methyl
H NMR, ppm, C-5 protons
C NMR, ppm, C-5 methyl
(± )-1
(± )-2
(± )-3
(± )-4
(± )-5
Toluene-d
Toluene-d
Toluene-d
Toluene-d
Toluene-d
Toluene-d
Toluene-d
Toluene-d
Toluene-d
Toluene-d
8
8
8
8
8
8
8
8
8
8
1.28 and 1.24—
1.35, 1.33, 1.30, 1.29
1.27 and 1.23
1.37, 1.36 and 1.31
—
—
—
—
—
—
28.9 and 28.3
a
a
a
a
a
+(S)-TFAE
+(S)-TFAE
+(S)-TFAE
+(S)-TFAE
+(S)-TFAE
—
—
—
—
28.6 and 28.0
—
—
—
3.07 and 3.01
3.06, 3.04, 2.98, 2.97
3.04and 2.97
2.88, 2.86, 2.79, 2.78
3.08 and 3.02
3.08, 3.06, 2.99, 2.98
—
—
—
—
a
1:6 equiv of (S)-TFAE were used.

¨
O. Demir-Ordu et al. / Tetrahedron: Asymmetry 16 (2005) 3752–3761
3754
1
Figure 1. The H NMR spectra of the compounds (± )-1–5 taken in the presence of 6 equiv of (S)-TFAE in toluene-d
8
, A = (± )-1, B = (± )-2,
C = (± )-3, D = (± )-4 and E = (± )-5.
6
2
.3. Determination of absolute stereochemistry
with the chiral auxiliary (S)-TFAE. For 5,5-dimethyl-
-(o-iodophenyl)-2,4-oxazolidinedione studied in previ-
3
6
The resolution of the enantiomeric resonances of 5,5-
dimethyl-3-(o-aryl)-2,4-oxazolidinediones was accom-
plished through the interaction of the enantiomers
ous work, four singlets corresponding to the C-5
methyl groups of M and P rotational isomers could be
distinguished by H NMR in the presence of (S)-TFAE.
1

¨
O. Demir-Ordu et al. / Tetrahedron: Asymmetry 16 (2005) 3752–3761
3755
For this compound the chemical shift difference between
the diastereotopic C-5 methyl protons (0.19 ppm in
C D ) did not change with the addition of (S)-TFAE
3 led us to consider that the origin of this resolution
was much the same as it was in our previous work on
5,5-dimethyl-3-(o-aryl)-2,4-oxazolidinediones. However
on lowering the temperature, the C-5 methyl and
6
6
while cooling a toluene-d solution in the NMR probe
8
1
from 30 to ꢀ70 ꢁC did not result in a significant differ-
ence in chemical shifts. Contrary to this, it was observed
for compound (± )-1 that the chemical shift differ-
ence between the diastereotopic C-5 methyl protons
were found to be slightly different in the absence
C-5 methylene H NMR signals of the complexed enan-
tiomers of the a-naphthyl derivatives showed a different
shielding behaviour, as will be discussed below, from
what had been observed for the o-iodophenyl derivative.
1
It is also interesting to note that the H NMR spectra of
(
0.040 ppm, in toluene-d ) and presence (0.046 ppm, in
all comparable signals through the series showed close
similarities. The C-5 methyl groups of compounds (± )-
1 and (± )-2 were shielded in the same way upon cooling
while the signals of the C-5 methylene for compounds
(± )-3, 4 and 5 were shielded in a similar way on cooling.
The fact that the temperature dependence of the C-5
methylene signals for compounds (± )-4 and (± )-5 was
similar with that of compound (± )-3 led us think that
all of these compounds could have the same solvation
model, although the enantiomers of (± )-2 and (± )-4
could not be completely resolved.
8
1
toluene-d ) of (S)-TFAE. Moreover the H NMR of this
compound in toluene-d , unlike the 5,5-dimethyl-3-
(
8
8
o-iodophenyl)-2,4-oxazolidinedione, showed a strong
8
,9
temperature dependence, in the 30 to ꢀ70 ꢁC range
(
Fig. 2). We presume that this may arise from the exis-
tence of a naphthyl group and hence p stacking. There-
fore to relate these results to a solvation model we
studied sterically congested, axially chiral heterocyclic
compounds (Scheme 1) bearing a naphthyl ring that
would facilitate studies of intermolecular interactions
between stacked aromatic groups. We surmise that the
stacking propensity of a naphthyl ring with an anthryl
group of (S)-TFAE would be higher than that of the
o-iodo substituted phenyl ring due to the high surface
area of naphthyl.10,11
By considering these results it can be said that the solva-
tion model for all of these compounds (± )-1–5 should
be the same within the series and is very similar to that
proposed for 5,5-dimethyl-3-(o-aryl)-2,4-oxazolidinedi-
6
ones. However, there must be a difference from the pre-
Among these compounds (± )-1–5 the enantiomers of
viously proposed model to account for the different
temperature dependence of the signals.
2
,4-oxazolidinedione (± )-1 and 3 and 2,4-thiazolidinedi-
1
one (± )-5 derivatives could be differentiated by H NMR
spectroscopy in the presence of (S)-TFAE (Fig. 1). The
better resolution of the 2,4-oxazolidinediones (± )-1 and
In order to obtain a solvation model for our com-
1
pounds, all H NMR signals of pure (S)-TFAE and each
1
Figure 2. The temperature dependence of the H NMR signals of the methyl groups on C-5 of the compound (± )-1 upon cooling from 30 to ꢀ70 ꢁC,
S: solvent.

¨
O. Demir-Ordu et al. / Tetrahedron: Asymmetry 16 (2005) 3752–3761
3756
Figure 3. The NOESY spectrum of compound (± )-1 taken in the presence of (S)-TFAE. The cross-peaks between the naphthyl and the anthryl rings
are shown in circles.
of the racemates were compared with those of the diaste-
reomeric association complex. It was observed that
for all of the signals of complexed (S)-TFAE and the
enantiomers, with the exception of the hydroxyl group
of (S)-TFAE, were shielded with respect to the corre-
sponding pure compounds and these observed upfield
shifts and the downfield shift increased with decreasing
the solvent used or by enforced p–p interaction, the tolu-
ene-d solution of compound (± )-1 was cooled to 0 ꢁC
8
in the absence of (S)-TFAE. The fact that none of the
protons of compound (± )-1 indicated any shielding
effect on cooling to 0 ꢁC revealed the enhanced p–p
interaction between the anthryl group of the (S)-TFAE
and the naphthyl substituent of the enantiomers.
9
temperature.
In our previous work carried out for diastereomeric
The downfield shift observed for the hydroxyl proton
was thought to result from hydrogen bonding with race-
(5S)-methyl-3(o-aryl)-2,4-oxazolidinediones
it
was
found that the naphthyl substituent exerted a similar
1
1
13
mates. The upfield changes in chemical shifts observed
for the aromatic naphthyl and anthryl groups may arise
from ring current effects and suggests a face-to-face p
shielding effect as the iodine on the C-5 methyl group,
which is on the same side with the naphthyl or iodine.
Similarly, the enantiomers of compounds (± )-1–5 stud-
ied in the absence of (S)-TFAE could be differentiated
on the NMR time scale via the diastereotopic groups
1
0,12
stacked geometry (Fig. 3).
Also consistent with the
face-to-face p stacking, cross-peaks between some of
the protons of naphthyl and anthryl ring of (S)-TFAE
were observed in the NOESY spectrum (Fig. 3). The p
stacking tendency of the naphthyl substituent in a p–p
0
0
on C-5 (a–b or a –b ) (Fig. 1) depending on the position
of naphthyl substituent. In the presence of (S)-TFAE
however, the discrimination of the enantiotopic groups
2
10,11
0
0
interaction as a p-base, was thought to be higher,
(a–a and b–b ) was also possible.
and hence these compounds (± )-1–5 would form tighter
complexes with (S)-TFAE than that of the o-iodo substi-
tuted phenyl ring. The change in the chemical shift dif-
ference of the diastereotopic C-5 protons of compound
For the M and P conformations the hydroxyl group
of (S)-TFAE formed a hydrogen bond with the carbonyl
or thiocarbonyl groups; the carbinyl hydrogen of
(
± )-1 from 0.040 to 0.046 ppm upon complex formation,
might be further indication for an increase in anisotropic
shielding effect as a result of enhanced p stacking
between the aromatics. In order to find out whether
the observed shielding was influenced by the effect of
0
0
Notations a, b, a , b refer to the C-5 methyl protons for the
compounds (± ) 1–2 and the C-5 methylene protons for the
compounds (± ) 3–5.

¨
O. Demir-Ordu et al. / Tetrahedron: Asymmetry 16 (2005) 3752–3761
3757
(
S)-TFAE involved in hydrogen bonding with the oxy-
anthryl cannot approach the naphthyl due to the pres-
2
gen or sulfur atoms of the heterocycle formed a che-
late-like structure and the anthryl group of (S)-TFAE
formed a p–p interaction with the naphthyl substituent
of the enantiomers (Fig. 4, models A and C). As can
be seen in Figure 4, during the interaction of compound
ence of anthryl and the weak basicity of the naphthyl.
3
Therefore, either carbinyl hydrogen bonding (A and
C) or p–p interaction (B and D) should be preferred at
a given time, one being more populated than the other.
The models B and D were thought to be less populated
than the models A and C since they have less interac-
tions. Compared to model D, model B would be even
less populated due to the weaker p–p interaction. There-
fore, it was thought that model B may have little or no
(
± )-1 with (S)-TFAE, two more, different, solvation
models (B and D in Fig. 4) are possible for both con-
formers. In our basic models (A and C in Fig. 4) (S)-
TFAE forms a chelate via a hydrogen bond between
the C–H proton of the (S)-TFAE and the ring oxygen
and a hydrogen bond between the –OH group of the
1
effect on the observed chemical shifts. The observed H
NMR signals should be consistent with the time-aver-
aged structure of these solvates.
(
S)-TFAE and the –C@O on C-2. However, a p–p inter-
action would also be possible between the anthryl and
the naphthyl rings, thus destroying the chelate structure
without cleavage of the hydrogen bonding with the C-2
carbonyl group. In these models (B and D in Fig. 4),
0
0
In order to assign the pairs (a–b or a –b ) corresponding
to the signals of the diastereotopic methyl or methylene
protons, the second eluted enantiomers of (± )-2 and
anthryl1
(
a')
(b')
CH3
H C
(a')
(b')
CH3
3
F3C
H C
3
O
N
C
H
O
N
anthryl3
CF3
O
O
H
H
O
O
C H
O
O
H
H
O
C
F3C
M-(S) solvate
MODEL B
anthryl2
M-(S) solvate
MODEL A
anthryl1
(
a)
(
b)
H C
CH3
(b)
(a)
3
H C
CH3
3
F3C
O
C
O
N
O
N
H
anthryl3
CF3
H
O
H
O
O
C H
O
O
H
H
O
C
F3C
P-(S) solvate
MODEL D
anthryl2
P-(S) solvate
MODEL C
Figure 4. The proposed solvation models for the M and the P enantiomers of (± )-1.

¨
O. Demir-Ordu et al. / Tetrahedron: Asymmetry 16 (2005) 3752–3761
3758
0 0
increased in the following order: (b ) > (b) > (a ) > (a)
à
(
± )-4 resolved by HPLC on a Chiralpak AD-H was
added into a toluene-d8 solution containing racemic
compound and (S)-TFAE. With the NMR spectrum,
it was observed that the intensity of the signals belong-
ing to this enantiomer, which were the more shielded
of the C-5 methyl or methylene proton signals (Fig. 1B
and D) increased. For compounds (± )-3, 4 and 5 the
splitting patterns of the AB spectra enabled us to distin-
guish between the signals of M and P conformers.
(Fig. 2). However the signals of C-5 methylene protons
of compounds (± )-3, (± )-4 and (± )-5 had a shift pattern
0
0
that was in the following order: (b ) > (a ) > (a) > (b).
The observed chemical shift behaviour on cooling can
be explained on the basis of the proposed solvation
model (Fig. 4). The fact that the aromatic protons of
the naphthyl and the anthryl rings were more shielded
and the hydroxyl proton was more deshielded compared
to the values at 30 ꢁC, pointed that all of the intermole-
cular interactions were getting stronger upon cooling.
This result may be attributed to the slower rotation of
the molecules at lower temperatures. The conformation
might even have been frozen at ꢀ70 ꢁC. For example,
the aromatic groups of the enantiomers and (S)-TFAE
could be positioned parallel to each other enhancing
the intermolecular interactions. In compounds (± )-1
and (± )-2 as the temperature was lowered, the change
The assignments of the four peaks to the four methyl
0
0
groups (Figs. 1 and 4) as a, a , b, b has been done in
0
the following way: Protons (a) and (a ) were both
affected by the shielding zone of the naphthyl group,
giving the signals in the upfield region compared to
0
(
b), (b ). Additionally, proton (a) of the P conformer
was affected by the shielding zone of anthryl (Fig. 4,
3
model D) and anthryl (Fig. 4, model C), whereas the
1
0
0
proton (a ) was shielded by anthryl (Fig. 4, model A).
Since anthryl2 was closer to proton (a ) and the M
conformation could form a tighter complex, this
proton should be more shielded with respect to proton
in the chemical shift values of proton (b ) of the M con-
2
0
former was the most affected one and was strongly
shielded. Based on our model this proton was already
strongly affected by an average anisotropy of the
1
4
(
a) of the P conformer.
anthryl at 30 ꢁC where its rotation was fast around
1
C
2
–C
3
single bond of (S)-TFAE. However, decreas-
sp
sp
0
Protons (b) and (b ) were not affected by the naphthyl
ing the temperature made this rotation slower, making
(
Fig. 4). Proton (b) spends some of the time in the
the anthryl adopt a conformation where the hydroxyl
1
shielding zone of the anthryl (Fig. 4, model C), some
group and the carbinyl hydrogen were aligned on the
2
of the time not (Fig. 4, model D), whereas the proton
same side and –CF positioned nearly orthogonal to
3
0
15
(
b ) was always affected by anthryl (Fig. 4, model A).
the plane of the anthryl group. In this conformation
1
0
0
Since (b ) experiences a stronger shielding effect than
proton (b) of the P conformer it appeared more shielded
than b. Therefore it can be argued that for all the com-
pounds studied (± )-1–5, in the presence of (S)-TFAE
the more deshielded of the diastereotopic C-5 methyl
and the methylene protons (Fig. 1) can be assigned to
the complexed P conformer and the more shielded to
complexed M (Fig. 1). Thus, the second eluted enantio-
mers of (± )-2 and (± )-4 have been assigned to M
conformation.
the anthryl would be closer to the proton (b ) thus
1
shielding it more.
At 30 ꢁC, both of the solvation models (C and D in
Fig. 4) are possible for the P conformer, with model C
being more populated. Proton (b) was only affected by
the anisotropy of anthryl during solvation C, as shown
2
in Figure 4. Therefore, its shielding was only explained
on the basis of Model C. Upon cooling the change in
the solvation model from D to C (Fig. 4) the anisotropy
effect on this proton increased, making it more shielded.
The model also accounts for the difference in chemical
0
shifts between the a and a (upfield, 0.01 ppm) and the
In model A there was a strong p–p interaction between
the aromatic rings, affecting proton (a ). This inter-
molecular interaction became stronger upon cooling,
making this proton more shielded.
0
0
b–b (downfield, 0.02 ppm) pairs (Fig. 1C), which was
6
0
absent for o-iodo derivative. Since the a and a protons
were subjected to similar anisotropy effects, the chemical
shift difference between them was smaller than that of b
0
and b .
Proton (a) was affected by the average shielding effect of
both anthryl and anthryl at 30 ꢁC. However, on cool-
1
2
0
The non-resolved (a) and (a ) protons of compounds
ing the P enantiomer may prefer the model C where the
(
± )-2 and (± )-4 (Fig. 1) can be explained on the basis
proton (a) was only affected by anthryl . This proton
1
of the ring size of the chelate structure formed between
the (S)-TFAE and the ring oxygen or sulfur and the
C@O or C@S groups of the heterocycle (Fig. 4), chang-
ing the geometry of the ring formed on chelation. In
although can be expected to be less shielded due to the
same reason noted for proton (b), shielding by the
anthryl had a compensating effect and thus the chemi-
1
cal shift of this proton did not change much on cooling.
compounds (± )-2 and (± )-4, although anthryl was clo-
1
ser to the proton (a) than (± )-1, 3 and 5, anthryl was far
Regarding these results it can be said that the diastereo-
topic methyl signals of the M conformer, forming a tigh-
ter complex with (S)-TFAE, was shielded more
compared to that of the P conformer. These results
3
away from it (model D). This causes a similar shielding
0
effect on protons (a) and (a ), giving the same chemical
shift values for these compounds.
On cooling, it was observed that the upfield chemical
shift differences of the H NMR signals of the C-5
1
à
For compound (± ) 2 and (± ) 4 the chemical shift values of (a) and
0
methyl groups of the compounds (± )-1 and (± )-2
(a ) were the same.

¨
O. Demir-Ordu et al. / Tetrahedron: Asymmetry 16 (2005) 3752–3761
3759
are also consistent with what we observed for 5,5-
dimethyl-3-(o-iodophenyl)-2,4-oxazolidinedione on cool-
ing. For this compound, although the chemical shift of
the signals did not change significantly, the signals of
the P conformer, which formed a tighter complex with
tioresolution of the racemic mixtures was done on Chir-
alpak AD-H column packed with amylose tris-(3,5-
dimethyl)-carbamate as chiral stationary phase and
thermal racemization process was followed by UV at
254nm. To obtain a better resolution of the enantio-
mers, all chromatographic separations were conducted
at 7 ± 2 ꢁC. The same procedure described in our previ-
(
S)-TFAE were the most shielded ones upon cooling.
6
The different shift pattern observed for the compounds
± )-3–5 upon cooling may arise from the difference in
ous work, was followed for thermal racemization. The
(
activation parameters of compounds (± )-2, (± )-3, (± )-4
and (± )-5 are listed in Table 2 and data pertinent to the
chromatographic separations of the enantiomers on
Chiralpak AD-H are shown in Table 3.
the van der Waals radii of C-5 methyl and methylene
protons. Since methyl protons are exposed into space
more, these protons may experience a different aniso-
tropic effect. Even if a different shift pattern was
observed for these compounds on cooling, the
Activation parameters of (± )-1 were found by dynamic
NMR spectroscopy. The two singlets of the diastereo-
topic C-5 methyl protons that were distinguishable at
30 ꢁC coalescenced into one peak at higher temperatures
due to the fast rotation. The kinetic data of the intercon-
version process were determined by the Eyring equa-
0
diastereotopic methyl signals of the M conformer (a -
0
b ) was the most shielded ones as observed in the com-
pounds (± )-1, 2.
2
.4. Activation barriers to hindered rotation
1
6
tion, and it was apparent that the barrier (85.01 kJ/
Activation barriers to the hindered rotation of the com-
pounds studied were determined either by temperature
dependent NMR or by enantioresolution on a chiral
sorbent via HPLC. If the barrier to the restricted rota-
mol in DMSO-d , Table 2) was not high enough to allow
the resolution of compound (± )-1 into the enantiomers
on Chiralpak AD-H.
6
tion about Csp
separation of enantiomeric rotational isomers would
be possible by HPLC.
2
–Nsp
2
chiral axis is sufficiently high, the
The barriers of 2,4-oxazolidinediones (± )-1 and (± )-3
(Y = O, Scheme 1) and 2,4-thiazolidinedione (± )-5
(Y = O) derivatives were found to be lower than that
of the rhodanine (± )-4 (Y = S) and 2-thioxo-4-oxazolid-
inone (compound (± )-2, Y = S) derivatives. This differ-
ence was interpreted by means of the higher standard
Dynamic NMR studies of all the compounds, except for
compound (± )-1, indicated no coalescence for the dia-
stereotopic protons up to 110 ꢁC due to the high activa-
tion barrier for restricted rotation. Therefore, thermal
racemization after micropreparative enrichment of these
enantiomers was applied to calculate the barriers. Enan-
˚
bond length of the C@S double bond (1.71 A) than that
˚
of the C@O (1.22 A) and the larger van der Waals radius
˚
˚
of sulfur (1.85 A) than that of the oxygen atom (1.40 A).
In compounds (± )-2 and (± )-4 the repulsion between
Table 2. The kinetic and thermodynamic data for the interconversion process shown in Scheme 1
ꢀ1
DG5, kJ/mol
Compound no.
Solvent
T, K
k, s
a
b
b
c
(
± )-1
DMSO-d
Toluene-d
Ethanol/hexane (20%:80%, v/v)
Ethanol/hexane (60%:40%, v/v)
Ethanol/hexane (50%:50%, v/v)
Ethanol/hexane (70%:30%, v/v)
6
398
378
333
313
351
323
54.52
36.62
85.01 ± 0.05
a
c
8
81.83 ± 0.05
d
ꢀ5 e
ꢀ5 e
ꢀ4 e
ꢀ5 e
f
(
(
(
(
± )-2
± )-3
± )-4
± )-5
4 · 10
3 · 10
2 · 10
5 · 10
109.9 ± 0.07
d
f
103.8 ± 0.07
d
f
111.3 ± 0.07
d
f
105.9 ± 0.07
a
b
c
d
e
f
The coalescence temperature.
Rate constant at coalescence temperature.
Free energy of activation determined by 400 MHz NMR instrument.
The temperature at which the thermal racemization has been done.
The rate constant of interconversion.
Free energy of activation determined by HPLC.
Table 3. Chromatographic parameters for the separation of enantiomers by HPLC on Chiralpak AD-H at 7 ± 2 ꢁC
Compound no. Eluent composition
Retention times, t
1
, t
2
, min Capacity factors, k
1
, k
2
Selectivity, a Flow rate, ml/min
(± )-2
(± )-3
(± )-4
(± )-5
Ethanol/hexane (20%:80%, v/v)
Ethanol/hexane (60%:40%, v/v)
Ethanol/hexane (50%:50%, v/v)
Ethanol/hexane (70%:30%, v/v)
12.96
1.23
1.60
2.40
4.31
1.86
2.94
1.26
1.30
1.79
1.58
0.5
0.5
0.5
1
5.08
19.71
0.78
16.56
2.86
17.542.02
0.542.54
3
2
0.5
2

¨
O. Demir-Ordu et al. / Tetrahedron: Asymmetry 16 (2005) 3752–3761
3760
the thiocarbonyl sulfur atom and the peri hydrogen of
the naphthyl group in the transition state serves as a ste-
ric impediment to the enantiomer interconversion and
makes the passage of the peri hydrogen more difficult,
which in turn, increases the barrier to hindered rotation.
obtain the barriers to rotation, which were found as
109.9, 103.8, 111.3 and 105.9 kJ/mol, respectively. The
difference in activation barriers resulted from the higher
bond length of the C@S double bond than that of C@O,
and the larger van der Waals radius of the sulfur atom
than that of the oxygen atom. The solvent effect of eth-
anol caused an increase in activation barriers by forming
hydrogen bonds with the oxygen atom of C@O groups
making the rotation more hindered.
Compound (± )-5 (X = S) showed a barrier greater than
compounds (± )-1 and (± )-3 (X = O). This difference can
be explained by considering the standard bond length of
C–S (182 pm) and C–O (143 pm) single bonds as has
1
7
been observed before. Owing to the larger bond length
of C–S, the C-2 carbonyl group of compound (± )-5 is
closer to the peri hydrogen of the naphthyl substituent
in the transition state, giving rise to an enantiomeric pair
of M and P atropisomers with higher activation barriers
for interconversion.
4. Experimental
1
13
H and C NMR spectra of all compounds were
recorded on a Varian-Mercury VX-400 MHz-BB (30 ꢁC).
Chromatographic analyses were done using Cecil 1100
pump (P = 3–3.5 MPa), a Rheodyne, 7125 injector
mode with a 20 ll Sample loop, a Cecil UV monitor
(240 nm) and an integrating recorder. The chiral column
used was commercially available Chiralpak AD-H
(4.6 mm I.D. · 250 mm L, 5 lm, Daicel, Tokyo, Japan).
Thermohypersil-Keystone column pocket was used for
the temperature control of the Chiralpak AD-H column.
Elemental analyses were performed on Carlo Erba 1106.
Melting points were recorded using Electrothermal 9100
melting point apparatus.
Comparison of the energy barriers of compounds (± )-1
and (± )-3 revealed the influence of the solvent used
(
Table 2). Compound (± )-3 unexpectedly exerted a
higher barrier than compound (± )-1 although the X
and Y groups were the same (Scheme 1). This result
1
3
can be attributed to the solvent effect, of ethanol that
may form hydrogen bonds with the C@O groups at the
2
- and 4-positions and thus make the rotation around
the Csp
2
–Nsp bond more hindered.
2
4.1. General procedure for the preparation of the
compounds studied (± ±)1ꢀ5
3. Conclusion
The enantiotopic protons of naphthyl compounds (± )-
–5 could be resolved in the presence of the chiral aux-
The compounds were synthesized by the reaction of
0.02–0.03 mol of a-naphthyl isocyanate or a-naphthyl
isothiocyanate and an equimolar amount of the corre-
sponding ethyl esters (ethyl glycolate, ethylthioglycolate
or ethyl a-hydroxy isobutyrate) in the presence of
sodium metal in toluene.
1
iliary (S)-TFAE. The proposed model between the com-
pounds and the chiral auxiliary, which emphasized p–p
interactions between the naphthyl and anthryl groups
as well as the hydrogen bonding interactions enabled
the assignment of the resolved peaks to the M and P
conformations. The strong temperature dependence of
the spectra of the diastereomeric complexes originated
from the adoption of different solvation models at lower
temperatures of the M and P conformers. Thus, at 30 ꢁC
or lower for the M conformation, the most dominant
model was the model A (Fig. 4) while for the P was
the model C (Fig. 4). The population of model C in-
creased even more on cooling upon the change from sol-
vation model D to C.
In a 100-ml three-necked flask, fitted with a thermome-
ter and a reflux condenser, a-naphthyl isocyanate or a-
naphthyl isothiocyanate and corresponding ethyl esters
were mixed in toluene. Sodium metal in small pieces
was added prior to heating. The mixture was heated
for 10 h at about 80 ꢁC after which the temperature
§
was increased to 100–110 ꢁC for 1 h. At the end of the
reflux process, crude products were obtained. After dis-
solving in ethanol, the compounds separated as crystals.
The compounds were further purified by recrystalliza-
1
1
The assignment of the H NMR signals of the M and P
tion from ethanol and identified based on their
NMR spectra and elemental analyses.
H
conformations also enabled us to relate the elution order
in HPLC on Chiralpak AD-H to the absolute conforma-
tions for the compounds (± )-2 and (± )-4. The second
eluted enantiomer had the M conformation on chiral
AD-H column for both compounds.
4.1.1. Synthesis of 5,5)dimethyl)3)(a)naphthyl±)2,4)oxa)
zolidinedione, (± ±)1. Compound (± )-1 was prepared
according to the general procedure using 0.03 mol a-
naphthyl isocyanate, 0.03 mol ethyl a-hydroxy isobuty-
rate, 0.003 mol sodium metal and 25 ml toluene. Yield:
The activation barrier for 5,5-dimethyl-3-(a-naphthyl)-
,4-oxozolidinedione was determined as 85.01 kJ/mol
1
2
5.98 g, 78%. Melting point: 142–143 ꢁC. H NMR
in DMSO-d8 by temperature dependent NMR. The
enantiomers of 5,5-dimethyl-3-(a-naphthyl)-2-thioxo-4-
oxozolidinone, 5,5-dimethyl-3-(a-naphthyl)-2,4-oxozo-
lidinedione, 3-(a-naphthyl)-rhodanine and 3-(a-naph-
thyl)-2,4-thiazolidinedione were resolved micro-
preparatively on Chiralpak AD-H. The thermal racemi-
zation of the enriched enantiomer was followed to
(400 MHz) data in CDCl : Diastereotopic methyl pro-
3
tons at C-5: d = 1.77 ppm (s, 3H), 1.69 ppm (s, 3H). Aro-
1
3
matic protons: d = 7.9–7.16 ppm (m, 7H). C NMR
data in CDCl : Carbonyl carbons in the heterocyclic
3
§
The reaction mixture was refluxed for 7 h for (± )-2.

¨
O. Demir-Ordu et al. / Tetrahedron: Asymmetry 16 (2005) 3752–3761
3761
ring: 176.6, 153.7 ppm. Aromatic carbons: 134–121 ppm.
Methine carbon (C-5): 84ppm. Diastereotopic methyl
carbons at C-5: 25 and 24ppm. Elemental analysis data:
Calculated for C H NO : C, 70.59; H, 5.09; N, 5.49.
0.02 mol ethylthioglycolate, 0.002 mol sodium metal
and 25 ml toluene. Yield: 2.07 g, 42%. Melting point:
1
158 ꢁC. H NMR (400 MHz) data in CDCl : Diastereo-
3
topic methyl protons at C-5: d = 4.31 and d = 4.27
1
5
13
3
A
B
Found: C, 69.96; H, 4.81; N, 5.22.
(AB, JAB = 5 Hz, 2H), Aromatic protons: d = 7.91–
13
7
.18 (m, 7H). C NMR data in CDCl : Carbonyl car-
3
4
4
.1.2. Synthesis of 5,5)dimethyl)3)(a)naphthyl±)2)thioxo)
)oxazolidinedione, (± ±)2. Compound (± )-2 was pre-
bons in the heterocyclic ring: 170.9, 170.8 ppm, Aro-
matic carbons: 135–122 ppm, Methylene carbon (C-5):
35 ppm. Elemental analysis data: Calculated for
C H NO S: C, 64.19; H, 3.70; N, 5.76; S, 13.17. Found:
pared according to the general procedure using
.025 mol a-naphthyl isothiocyanate, 0.025 mol ethyl
a-hydroxy isobutyrate, 0.0025 mol sodium metal and
0
1
3
9
2
C, 63.11; H, 3.63; N, 5.69; S, 13.46.
2
1
5 ml toluene. Yield: 1.7 g, 24%. Melting point:
1
55 ꢁC. H NMR (400 MHz) data in toluene-d : Diaste-
8
reotopic methyl protons at C-5: d = 1.27 ppm (s, 3H),
Acknowledgement
1
6
.22 ppm (s, 3H1)3, Aromatic protons: d = 7.60–
.90 ppm (m, 7H). C NMR data in CDCl : Carbonyl
This project has been supported by Bo g˘ azi c¸ i University
research fund (BAP) with project number 05B501.
3
carbons in the heterocyclic ring: 194, 181 ppm. Aro-
matic carbons: 142–126 ppm. Methine carbon (C-5):
9
2
2 ppm. Diastereotopic methyl carbons at C-5: 29 and
8 ppm. Elemental analysis data: Calculated for
References
C H NO S: C, 66.42; H, 4.79; N, 5.17; S, 11.80.
1
5
13
2
Found: C, 65.68; H, 4.37; N, 4.95; S, 10.67.
1. Evans, D. A.; Chapman, K. T.; Hung, D. T.; Kawaguchi,
A. T. Angew. Chem., Int. Ed. Engl. 1987, 26, 1184.
4
(
.1.3. Synthesis of 3)(a)naphthyl±)2,4)oxazolidinedione,
± ±)3. Compound (± )-3 was prepared according to the
general procedure using 0.02 mol a-naphthyl isocyanate,
2. Pirkle, W. H.; Pochapsky, T. C. J. Am. Chem. Soc. 1987,
1
09, 5975–5982.
3
4
5
6
. Cochran, J. E.; Parrot, T. J.; Whitlock, B. H.; Whitlock,
H. W. J. Am. Chem. Soc. 1982, 114, 2269–2270.
. Pirkle, W. H.; Sikkenga, D. L. J. Org. Chem. 1977, 42,
0
1
1
.02 mol ethyl glycolate, 0.002 mol sodium metal and
5 ml toluene. Yield: 0.78 g, 17.30%. Melting point:
1
370–1374.
1
52–154 ꢁC. H NMR (400 MHz) data in CDCl3:
. Jullian, J.-C.; Franck, X.; Latypov, S.; Hocquemiller, R.;
d = 4.56 and d = 4.52 (AB, J = 5 Hz, 2H), Aro-
matic protons: d = 7.50–8.10 (m, 7H). C NMR data
in CDCl : Carbonyl carbons in the heterocyclic ring:
1
A
B
AB
Figad e` re, B. Tetrahedron: Asymmetry 2003, 14, 963–966.
1
3
¨
_
. Demir Ordu, O.; Do g˘ an, I. Tetrahedron: Asymmetry 2004,
15, 925–933.
3
_
_
67, 151 ppm. Aromatic carbons: 134–118 ppm, Methyl-
7. Do g˘ an, I.; Burgemeister, T.; I c¸ li, S.; Mannschreck, A.
ene carbon (C-5): 63 ppm.
Tetrahedron 1992, 48, 7157–7164.
8
9
. Pirkle, W. H.; Beare, S. D.; Muntz, R. L. Tetrahedron
Lett. 1974, 26, 2295–2298.
. Hanna, G. M.; Evans, F. E. J. Pharm. Biomed. Anal. 2000,
4
.1.4. Synthesis of 3)(a)naphthyl±)rhodanine, (± ±)4.
Compound (± )-4 was prepared according to the general
procedure using 0.023 mol a-naphthyl isothiocyanate,
0
and 25 ml toluene. Yield: 2.4g, 41%. Melting point:
1
Diastereotopic methyl protons at C-5: d = 4.38 and
dB = 4.30 (AB, JAB = 5 Hz, 2H). Aromatic protons:
d = 8.01–7.25 (m, 7H). C NMR data in CDCl3:
Carbonyl carbons in the heterocyclic ring: 200, 173
ppm. Aromatic carbons: 134–119 ppm, Methylene
carbon (C-5): 37 ppm. Elemental analysis data: Calcu-
lated for C H NOS : C, 60.23; H, 3.47; N, 5.40; S,
2
4, 189–196.
1
0. Guckian, K. M.; Schweitzer, B. A.; Ren, R. X.-F.; Sheils,
.023 mol ethylthioglycolate, 0.0023 mol sodium metal
C. J.; Tahmassebi, D. C.; Kool, E. T. J. Am. Chem. Soc.
2
000, 122, 2213–2222.
1
71–172 ꢁC. H NMR (400 MHz) data in CDCl3:
11. Rotello, V. M.; Viani, E. A.; Deslongchamps, G.; Murray,
B. A.; Rebek, J. J. Am. Chem. Soc. 1993, 115, 797–
798.
12. Muehldorf, A. V.; Engen, D. V.; Warner, J. C.; Hamilton,
A. D. J. Am. Chem. Soc. 1988, 110, 6561–6562.
A
1
3
¨
_
1
1
3. Demir, O.; Do g˘ an, I. Chirality 2003, 15, 242–250.
4. Beaufour, M.; Merelli, B.; Menguy, L.; Cherton, J. C.
Chirality 2003, 15, 382–390.
1
3
9
2
1
5. Pirkle, W. H.; Finn, J. M. J. Org. Chem. 1981, 46, 2935–
2
4.71. Found: C, 60.13; H, 3.35; N, 5.20; S, 25.08.
2
938.
1
6. Alberty, R. A.; Silbey, R. J. Physical Chemistry; Wiley:
4
.1.5. Synthesis of 3)(a)naphthyl±)2,4)thiazolidinedione,
New York, 1992.
(
± ±)5. Compound (± )-5 was prepared according to the
17. Do g˘ an, I_ .; Pustet, N.; Mannschreck, A. J. Chem. Soc.,
Perkin Trans. 2 1993, 1557–1560.
general procedure using 0.02 mol a-naphthyl isocyanate,