Highly efficient resolutions of 1,4-benzodioxane-2-carboxylic acid with para substituted 1-phenylethylamines

Article abstract of DOI:10.1016/j.tetasy.2005.01.052

The salts of (S)- and (R)-1,4-benzodioxane-2-carboxylic acid with eight (S)-1-arylethylamines were prepared. The determination of their melting points and of their solubilities in alcohol solvents revealed large differences between the diastereomeric benzodioxanecarboxylates of (S)-1-(p-nitrophenyl)ethylamine and of (S)-1-(p-methylphenyl)ethylamine. Therefore, these latter amines were selected to resolve (±)-1,4-benzodioxane-2-carboxylic acid by diastereoselective crystallization finding that both of them display a very high resolution ability for such a substrate, which contrasts with the null efficiency of unsubstituted 1-phenylethylamine. These results are consistent with DSC evidences, which indicated that the two successfully resolved diastereomeric systems are binary mixtures exhibiting a eutectic with a high content of the more soluble diastereomeric salt. The new procedures can advantageously replace the two resolutions we had previously reported, that of the same acid with dehydroabietylamine and that of glycerol acetonide, a precursor of 1,4-benzodioxane-2-carboxylic acid, with 1-phenylethylamine.

Full text of DOI:10.1016/j.tetasy.2005.01.052

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 16 (2005) 1639–1643
Highly efficient resolutions of 1,4-benzodioxane-2-carboxylic
acid with para substituted 1-phenylethylamines
Cristiano Bolchi,* Marco Pallavicini, Laura Fumagalli, Nicoletta Marchini,
Barbara Moroni, Chiara Rusconi and Ermanno Valoti
`
Istituto di Chimica Farmaceutica e Tossicologica, Universita di Milano, viale Abruzzi 42, I-20131 Milano, Italy
Received 27 January 2005; accepted 31 January 2005
Available online 11 March 2005
Abstract—The salts of (S)- and (R)-1,4-benzodioxane-2-carboxylic acid with eight (S)-1-arylethylamines were prepared. The deter-
mination of their melting points and of their solubilities in alcohol solvents revealed large differences between the diastereomeric
benzodioxanecarboxylates of (S)-1-(p-nitrophenyl)ethylamine and of (S)-1-(p-methylphenyl)ethylamine. Therefore, these latter
amines were selected to resolve ( )-1,4-benzodioxane-2-carboxylic acid by diastereoselective crystallization finding that both of them
display a very high resolution ability for such a substrate, which contrasts with the null efficiency of unsubstituted 1-phenylethyl-
amine. These results are consistent with DSC evidences, which indicated that the two successfully resolved diastereomeric systems
are binary mixtures exhibiting a eutectic with a high content of the more soluble diastereomeric salt. The new procedures can advan-
tageously replace the two resolutions we had previously reported, that of the same acid with dehydroabietylamine and that of
glycerol acetonide, a precursor of 1,4-benzodioxane-2-carboxylic acid, with 1-phenylethylamine.
ꢀ 2005 Elsevier Ltd. All rights reserved.
1.Introduction
enrichments by recrystallization, preferably of the con-
glomerate forming derivatives of the acid, such as its
methyl ester and the mesylate of 2-hydroxymethyl-1,4-
benzodioxane. However, we were not completely satis-
fied with the two resolution methods, despite their high
efficiency. In fact, the necessity of derivatizing with
phthalic anhydride or of resorting to an unusual resolv-
ing agent like dehydroabietylamine were penalizing lim-
itations. Therefore, we considered the resolution of ( )-1
with 1-phenylethylamine or, alternatively, with its ring
substituted analogues. The fact that no resolutions of
this acid with 1-phenylethylamine have been reported
supposes that such an amine is ineffective, however this
does not exclude the ability of its derivatives of resolving
the same substrate. Saigo and co-workers have recently
demonstrated that the almost null efficiency of the reso-
lution of 1-phenylethylamine by enantiopure 30,40-
methylenedioxymandelic acid is considerably improved
by para substitution on the amine aromatic ring, becom-
ing very high in the case of some 1-phenylethylamines
bearing an electron-withdrawing substituent.4 These re-
sults and the large availability of a number of enantio-
pure 1-arylethylamines, deriving from the resolutions
with isopropylidene glycerol hemiphthalate or 3-car-
boxy-2-naphthoate we had previously developed,5–10
Over the course of our research concerning the synthesis
of chiral benzodioxane derivatives as type-selective a-
adrenoceptor ligands, we developed two different resolu-
tion approaches, which could be conveniently practised
to obtain highly enantiopure 2-substituted benzodiox-
anes. The former method, patented in 1992, is based
on the resolution of racemic glycerol acetonide through
conversion into hemiphthalate and successive diastereo-
selective crystallization of the salt with enantiomerically
pure 1-phenylethylamine.1,2 The latter, recently re-
ported, was a classical resolution as well, not of a rela-
tively remote precursor, such as glycerol acetonide, but
of a chiral 2-substituted benzodioxane, namely ( )-1,4-
benzodioxane-2-carboxylic acid ( )-1, readily available
by condensation of catechol with ethyl 2,3-dibromopro-
pionate and successive saponification of the intermedi-
ate ester.3 The new strategy involved the use of (+)-
dehydroabietylamine as a resolving agent and coupled
the resolution of ( )-1 with subsequent enantiomeric
*
Corresponding author. Tel.: +39 2 50317553; fax: +39 2 50317565;
e-mail: cristiano.bolchi@unimi.it
0957-4166/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2005.01.052

1640
C. Bolchi et al. / Tetrahedron: Asymmetry 16 (2005) 1639–1643
prompted us to consider the possibility of resolving ( )-
1 with enantiopure 1-phenylethylamine and 1-arylethyl-
amines 2–8.
reomeric salt pair can give some reliable predictions of
its resolution process without preparative experiments,
when the laborious determination of the binary phase
diagram is to be spared or cannot be accomplished
due to thermal decomposition.11 In fact, providing that
both diastereomeric salts crystallize separately, efficient
resolution can be expected by preferential crystallization
when they are substantially different in physico-chemical
properties, as simply revealed by a large difference, at
least 20 ꢁC, between their respective melting points. As
shown in Table 1, such a condition was fulfilled by the
diastereomeric salts of (S)-2 and (S)-6, whose fusions
are separated by 29 and 32 ꢁC, respectively. For the
other diastereomeric pairs, we observed significantly
lower or null differences in melting point. These results
promoted (S)-2 and (S)-6 as the best candidates for
the resolution, while rejecting (S)-3 and the two naph-
thylethylamines. In the case of the parent unsubstituted
1-phenylethylamine and of its two p-halo analogues, the
melting point differences were considerable, but not
so promising as those of the p-methyl and p-nitro
derivatives.
Herein, we report (i) a detailed analysis of the melting
points and of the solubilities of all the salts resulting
from the combination of the acid enantiomers with the
S forms of 1-phenylethylamine and 1-arylethylamines
2–8; (ii) some DSC data for the diastereomeric systems
submitted to resolution, namely the (S)-2Æ(S)-1/(S)-
2Æ(R)-1 and (S)-6Æ(S)-1/(S)-6Æ(R)-1 systems and (iii) the
highly efficient resolutions of ( )-1 with (S)-2 and (S)-6.
X = p-Me
= p-OMe
= p-Cl
2
3
4
5
6
7
8
NH2
O
O
COOH
=
p-Br
= p-NO2
=
=
2,3 -CH=CH-CH=CH-
3,4 -CH=CH-CH=CH-
X
1
2.Results and discussion
After determining the melting points, we measured the
solubility of each salt in alcohols at room temperature
(see Table 1). Methanol, ethanol and 2-propanol, the
most frequently used alcohols as solvents in the resolu-
tions of diastereomeric salts, were selected for the exper-
iments. The results listed in Table 1 show that the
solubility (kn) of the n salts, those formed by (R)-1 with
the (S) form of these 1-arylethylamines, is always higher
than the solubility (kp) of the p salts, that is, the salts be-
tween the same (S)-amines and (S)-1. The only excep-
tions are represented by (S)-3Æ(R)-1, less soluble than
(S)-3Æ(S)-1 in methanol, and by (S)-8Æ(R)-1, less soluble
than (S)-8Æ(S)-1 in 2-propanol. As expected, dissolving
ability increases from 2-propanol to methanol. In partic-
ular, this latter alcohol is capable of dissolving the more
soluble salt of each diastereomeric pair enough (5–16
g/dL) to design resolution procedures with acceptable
Among the enantiopure amines obtained by resolution
with carboxyesters of isopropylidene glycerol, we se-
lected the following candidates for the resolution
of ( )-1: (S)-1-phenylethylamine, (S)-1-naphthylethyl-
amine (S)-7, (S)-2-naphthylethylamine (S)-8 and the five
ring substituted analogues of (S)-1-phenylethylamine
(S)-2–(S)-6, bearing a methyl, a methoxyl, a chlorine,
a bromine and a nitro group at the para position, respec-
tively. In order to characterize the resultant eight pairs
of diastereomeric salts with (S)- and (R)-1, all the 16
compounds were individually obtained by concentration
of methanolic solutions containing equivalent amounts
of (S)-amine and (S)- or (R)-acid and successive crystal-
lization of the solid residue from 2-propanol.
First, the melting point of each of these salts was deter-
mined. It is known that the melting points of a diaste-
Table 1. Melting point (ꢁC) and solubility (g/dL) of the salts of (S)-1-phenylethylamine and the amines (S)-2–(S)-8 with (S)-1 and (R)-1 in alcohols at
25 ꢁC
Salt
Mpa
Solubility
in MeOH
Solubility ratio
SR/SS in MeOH
Solubility in
EtOH
Solubility ratio
SR/SS in EtOH
Solubility
in 2-PrOH
Solubility ratio
SR/SS in 2-PrOH
(S)-1-PhenylethylamineÆ(R)-1
(S)-1-PhenylethylamineÆ(S)-1
(S)-2Æ(R)-1
(S)-2Æ(S)-1
(S)-3Æ(R)-1
(S)-3Æ(S)-1
(S)-4Æ(R)-1
(S)-4Æ(S)-1
(S)-5Æ(R)-1
(S)-5Æ(S)-1
(S)-6Æ(R)-1
(S)-6Æ(S)-1
(S)-7Æ(R)-1
(S)-7Æ(S)-1
(S)-8Æ(R)-1
(S)-8Æ(S)-1
179
15.80
12.13
16.45
3.07
1.3
5.4
0.7
1.6
1.8
6.3
1.1
1.6
3.27
2.34
3.47
0.44
2.33
2.05
2.11
1.15
1.67
1.02
1.51
0.14
1.34
1.18
1.35
0.9
1.4
7.9
1.1
1.8
1.6
10.8
1.1
1.5
0.25
0.20
0.58
0.05
0.35
0.33
0.42
0.13
0.21
0.08
0.17
0.012
0.21
0.15
0.14
0.24
1.2
11
194
190b
219b
185b
178
10.50
14.33
10.16
6.32
1.1
3.2
2.6
14.2
1.4
0.6
182
199b
191b
204b
182b
214b
202b
202b
200b
201b
9.14
5.13
5.73
0.91
6.47
5.89
8.16
5.09
a Determined by a capillary melting point apparatus.
b With decomposition.

C. Bolchi et al. / Tetrahedron: Asymmetry 16 (2005) 1639–1643
1641
volumetric productivity. The most interesting indication
of these solubility tests was that the trend of the solubil-
ity difference between diastereomeric salts parallels that
of the differences in melting point. The p salts (S)-2Æ(S)-1
and (S)-6Æ(S)-1, whose melting point exceed that of the
corresponding n salts of the greatest extent (ꢀ30 ꢁC),
are about sixfold less soluble than those latter in meth-
anol while the solubility ratio (kn/kp) is even higher in
ethanol and in 2-propanol. Conversely, solubility differ-
ences are insignificant for the pairs of diastereomeric
salts formed by the two naphthylethylamines, (S)-7
and (S)-8, having identical melting points. On this scale,
a middle position is occupied by the four pairs of salts of
(S)-3, (S)-4, (S)-5 and (S)-1-phenylethylamine, which
associate moderate melting point differences with appre-
ciable differences in solubility.
of the (S)-6Æ(S)-1/(S)-6Æ(R)-1 mixtures for relatively high
mole fractions (>0.2) of the more soluble salt, while
small amounts of this latter (mole fraction <0.05) were
enough to produce an evident fusion peak of the eutectic
in the DSC curves of the (S)-2Æ(S)-1/(S)-2Æ(R)-1 mix-
tures, as exemplified by the traces depicted in Figure 1.
(S)-2·(R)-1
(S)-2·(S)-1
A detailed DSC analysis, aimed at the construction of
the binary phase diagrams, was hampered by the ther-
mal decomposition of one or both salts of each pair,
which started before the fusion was complete. This hin-
dered the determination of the heats of fusion, allowing
approximate melting points, corresponding to the DSC
melting curve maxima, to be only recorded. This not-
withstanding, we collected the DSC data for a number
of different mixtures of the diastereomeric salts formed
by (S)-2 and (S)-6, hoping that the two diastereomeric
systems (S)-2Æ(S)-1/(S)-2Æ(R)-1 and (S)-6Æ(S)-1/(S)-6Æ(R)-
1 would produce DSC curves typical of conglomerates
with two clearly detectable fusions, the former of the eu-
tectic and the latter of the exceeding pure diastereomer.
Indeed, the DSC data indicated that both the systems
are binary mixtures exhibiting an eutectic and thus sus-
ceptible of resolution by crystallization. In the case of
the (S)-2Æ(S)-1/(S)-2Æ(R)-1 system, we analyzed nine
mixtures with constantly increasing mole fraction of
(S)-2Æ(S)-1 (from 0.1 to 0.9) finding that the fusion of
the eutectic ranges between 175.2 and 177.8 ꢁC and its
composition should be higher than 0.2, but lower than
the 0.3 mole fraction of (S)-2Æ(S)-1. For the (S)-6Æ(S)-1/
(S)-6Æ(R)-1 system, the same investigation showed that
the eutectic melts between 176 and 177 ꢁC, but an even
approximate determination of its composition was
impossible due to the decompositions subsequent to its
fusion and to that of the major diastereomer. However,
we could establish that this eutectic was significantly ri-
cher in the more soluble diastereomeric salt than that of
the (S)-2Æ(S)-1/(S)-2Æ(R)-1 system. In fact, the presence
of the eutectic became appreciable in the DSC traces
(S)-2·(S)-1/ (S)-2·(R)-1
90/10
150
170
190
210
230
250
Temperature (˚C)
Figure 1. DSC traces of the pure salts (S)-2Æ(S)-1 and (S)-2Æ(R)-1
(upper curves) and of their 9:1 mixture (lower curve).
On the basis of all these preliminary results, we were con-
fident that ( )-1 could be successfully resolved by (S)-2
and, even more efficiently, by (S)-6. This amine was tried
first. Equivalent amounts of the racemic acid and (S)-6
were combined in a volume of methanol slightly larger
than the minimum required to dissolve the more soluble
diastereomer at room temperature. After stirring for 5 h
at boiling temperature and then overnight at room tem-
perature, a precipitate of (S)-6Æ(S)-1 was collected in
93.3% yield [46.6% of the starting ( )-1] and with
99.0% ee of (S)-1 (Table 2, entry 1). Such a high resolu-
tion efficiency indicated that the procedure had been
carried out under optimal conditions and was not
susceptible to further significant improvements. In spite
of the more favourable solubility ratios between (S)-
6Æ(S)-1 and (S)-6Æ(R)-1 in ethanol and in 2-propanol,
the use of these two alcohols was not tried due to their
very low dissolving ability of the more soluble (S)-
6Æ(R)-1 salt. Otherwise, in the successive experiments of
resolution with (S)-2, methanol and ethanol were both
tested only rejecting 2-propanol. Again, equivalent
Table 2. Resolution of acid ( )-1 with amines (S)-2 and (S)-6
Yielda (%)
Recovered acid
Eeb (%)
Sexp
Scalcd
c
d
Entry
Resolving agent
Solvent
Precipitate
Precipitate
1
2
3
(S)-6
(S)-2
(S)-2
Methanol
Ethanol
Methanol
(S)-6Æ(S)-1
(S)-2Æ(S)-1
(S)-2Æ(S)-1
93.3
82.7
80.1
93.2
82.6
80.0
99.0
99.8
98.6
0.92
0.82
0.79
0.84
0.87
0.81
a Relative to the theoretical amount, that is, half of the starting racemic acid.
b Enantiomeric excess (determined by chiral HPLC) of the acid liberated from the precipitated salt.
c Experimental resolution efficiency or experimental resolvability calculated from S = chemical yield of the precipitate (%) · ee of the liberated acid
(%)/10,000.
d Maximum theoretical resolvability, calculated from S = (kn À kp)/kn, where kn is the solubility of the more soluble (S)-2Æ(R)-1 or (S)-6Æ(R)-1 salt and
kp the solubility of the less soluble (S)-2Æ(S)-1 or (S)-6Æ(S)-1 salt.

1642
C. Bolchi et al. / Tetrahedron: Asymmetry 16 (2005) 1639–1643
amounts of resolving agent and racemic substrate were
combined to give almost saturated solutions of the more
soluble salt, (S)-2Æ(R)-1, at room temperature. From eth-
anol, (S)-2Æ(S)-1 precipitated in 82.7% yield [41.3% of the
starting ( )-1] and with 99.8% ee of (S)-1 after stirring at
boiling temperature for 20 min, then at 50 ꢁC for further
5 h and, finally, overnight at room temperature (Table 2,
entry 2). Analogous results were obtained by a simple
precipitation using methanol under conditions of quasi-
saturation for (S)-2Æ(R)-1: 80.1% yield of (S)-2Æ(S)-1
and 98.6% ee of (S)-1 (Table 2, entry 3). The recovery
of the S acid from (S)-6Æ(S)-1 and (S)-2Æ(S)-1 precipitates
was always quantitative. Finally, the foreseen inability of
unsubstituted 1-phenylethylamine to resolve ( )-1 was
verified determining negligible enantiomeric excesses
for the acid resultant from precipitated (S)-1-phenethyl-
ammonium salts.
resolving ability of (S)-2, opposite to the null efficiency
of unsubstituted (S)-1-phenylethylamine, surprised us
due to the relatively insignificant effects of methyl substi-
tuent. It is evident that the ability of (S)-2 and (S)-6 to
resolve ( )-1 cannot be associated with a common effect
of the ring substituent, as in the case of the above cited
resolutions with 30,40-methylenedioxymandelic acid,4 or,
generically, with the occupation of the para position of
1-phenylethylamine. If the rationalization of such results
seems impossible at the moment, this is due to the fact
that the resolvability of diastereomers arises from unpre-
dictable differences between their crystal structures. We
think that the X-ray determination of these latter, which
is in progress for some of the present diastereomeric
pairs, will be explicative.
4.Experimental
At the end of the trial resolutions, we were able to com-
pare their experimental efficiencies (SS salt yield · S acid
ee) with the maximum theoretical resolvabilities result-
ing from the relative solubility differences between the
diastereomeric salts, that is, (kn À kp)/kn. As shown in
Table 2 (entries 2 and 3), the experimental efficiencies
(Sexp) are very close to the maximum theoretical resolv-
abilities (Scalcd) for the resolutions with (S)-2 both in
methanol (0.79 vs 0.81) and in ethanol (0.82 vs 0.87),
while, in the case of (S)-6 (Table 2, entry 1), the experi-
mental data (0.92) is even better than the theoretical one
(0.84). The higher resolving ability of (S)-6 with respect
to (S)-2 was consistent with the DSC results, which indi-
cated an eutectic composition closer to the pure more
soluble salt for the (S)-6Æ(S)-1/(S)-6Æ(R)-1 system. In par-
ticular, a 0.17 value of veu is obtained equating the max-
imum theoretical resolvability expression (1 À 2veu)/
(1 À veu) to 0.8, the mean experimental efficiency of
the resolutions with (S)-2. Such an eutectic composition
is not far from the 0.2 to 0.3 range previously suggested
by the inspection of the DSC traces. In the case of (S)-6,
the corresponding equation, that is, (1 À 2veu)/
(1 À veu) = 0.92, is satisfied by veu = 0.07. This value
cannot be demonstrated by DSC evidence. It is signifi-
cantly lower than the eutectic composition inferred
for the (S)-2Æ(S)-1/(S)-2Æ(R)-1 system, thus confirming
what the comparison of the DSC traces of the two con-
glomerates had suggested before the resolution
experiments.
1H NMR spectra were recorded on a Varian Gemini 300
(300 MHz) instrument. Optical rotations were measured
in a 1 dm cell of 1 mL capacity using a Perkin–Elmer
241 polarimeter. HPLC analyses were performed on a
Chiralcel OD column (250 · 4.6 mm i.d.) from Daicel
using a Hitachi 7100 pump, a Hitachi L-7400 UV detec-
tor and a Hitachi D-7000 HPLC System Manager soft-
ware. Melting points were determined by a Buchi
¨
Melting Point B-540 apparatus and by DSC analysis,
taking the temperature of the maximum of the peak.
The DSC curves were recorded and integrated with the
aid of a TA Instruments DSC 2010 apparatus.
Amines (S)-2–(S)-8 were prepared by resolution of the
corresponding racemates as previously reported.5,6,8,10
The racemic acid ( )-1 was synthesized by condensation
of catechol with ethyl 2,3-dibromopropionate and suc-
cessive saponification of the intermediate ester accord-
ing to a literature method.12 (S)-1 and (R)-1 were
available from the previous resolutions of the racemate
with (+)-dehydroabietylamine.3
4.1. Resolution of ( )-1 with (S)-2 in ethanol
A stirred solution of ( )-1 (2.67 g, 14.8 mmol) in ethanol
(75 mL) was added with (S)-2 (2 g, 14.8 mmol), refluxed
for 20 min and then slowly cooled to 50 ꢁC initiating the
crystallization of a white solid. The suspension was stir-
red at this temperature for 5 h and at rt overnight and
finally filtered yielding (S)-2Æ(S)-1 (1.93 g, 82.7% of the
theoretical amount) as a white crystalline solid: mp
219 ꢁC; ee of (S)-1 99.8% (determined by HPLC of the
acid liberated from a sample of the salt on a Chiralcel
OD column; hexane/2-propanol/formic acid 85/13.5/
3.Conclusion
In summary, we have demonstrated that ( )-1 can be
efficiently resolved by two readily available p-substituted
1-phenylethylamines, (S)-2 and (S)-6, via simple proce-
dures. Preliminary investigations on the physical proper-
ties of the salts formed by the 1-arylethylamines (S)-2–
(S)-8 with the enantiomers of 1 allowed (S)-2 and (S)-
6 to be rapidly identified as the best candidates for the
resolution of ( )-1 and their efficiency to be correctly
predicted. Interestingly, two very different substituents,
NO2 and CH3, at the para position of 1-phenylethyl-
amine are both able to make this latter amine an excel-
lent resolving agent of ( )-1. In particular, the high
1
1.5; 0.4 mL/min; (R)-1: k0 = 3.15; (S)-1: k0 = 2.49); H
NMR (DMSO-d6): d 1.41 (d, 3H, J = 8.8 Hz), 2.28 (s,
3H), 4.12 (dd, 1H, J = 11.0, 5.9 Hz), 4.22–4.29 (m,
2H), 4.38 (dd, 1H, J = 2.9, 5.9 Hz), 6.69–6.84 (m, 4H),
7.17 (d, 2H, J = 8.0 Hz), 7.31 (d, 2H, J = 8.0 Hz). The
salt was decomposed by treatment with 1 M NaOH
and dichloromethane. The aqueous phase was sepa-
rated, acidified (pH 1) and extracted with ethyl acetate
four times. The organic extracts were dried with Na2SO4
and concentrated to give (S)-1 (1.1 g, 82.6%) as a white

C. Bolchi et al. / Tetrahedron: Asymmetry 16 (2005) 1639–1643
1643
25
D
solid: mp 97 ꢁC; ½aŠ ¼ À63:9 (c 1, CHCl3); ee identical
sition. The 1:1 (S)-2Æ(S)-1/(S)-2Æ(R)-1 and (S)-6Æ(S)-1/(S)-
6Æ(R)-1 mixtures were prepared by concentration of
solutions containing equivalent amounts of racemic acid
and S amine. The DSC traces of such mixtures exhibited
the melting peak of the eutectic, followed by the antici-
pated fusion of the exceeding SS diastereomer. The
other diastereomeric mixtures were prepared by mixing
(S)-2Æ(S)-1 and (S)-6Æ(S)-1 with increasing quantities of
(S)-2Æ(R)-1 and (S)-6Æ(R)-1, respectively. All the analyses
to that previously determined for the acid liberated from
a sample of the salt with (S)-2; 1H NMR (CDCl3): d 4.38
(dd, 1H, J = 2.9, 11.4 Hz), 4.44 (dd, 1H, J = 4.4,
11.4 Hz), 4.90 (dd, 1H, J = 2.9, 4.4 Hz), 6.88–6.97 (m,
3H), 6.99–7.02 (m, 1H).
4.2. Resolution of ( )-1 with (S)-2 in methanol
A stirred solution of ( )-1 (2.67 g, 14.8 mmol) in meth-
anol (16 mL) was added with (S)-2 (2 g, 14.8 mmol),
refluxed for 5 h. After stirring overnight, the resultant
suspension was filtered isolating (S)-2Æ(S)-1 (1.87 g,
80.1% of the theoretical amount) as a white crystalline
solid. The enantiomeric excess of the acid liberated from
a sample of the salt was 98.6%. The salt was decom-
posed as described for the resolution in ethanol obtain-
ing (S)-1 (1.07 g, 80.0%) as a white solid: ee identical to
that previously determined for the acid liberated from a
sample of the precipitate.
were performed with a heating rate of 5 ꢁC minÀ1
.
Acknowledgements
This work was supported by the Ministero dellꢀIstruzi-
`
one, dellꢀUniversita e della Ricerca of Italy.
References
1. Piccolo, O.; Pallavicini, M.; Valoti, E.; Villa, L. It. Pat.
Appl. 000232a/92, 1992.
4.3. Resolution of ( )-1 with (S)-6 in methanol
2. Pallavicini, M.; Valoti, E.; Villa, L.; Piccolo, O. Tetrahe-
dron: Asymmetry 1994, 5, 5.
3. Bolchi, C.; Fumagalli, L.; Moroni, B.; Pallavicini, M.;
Valoti, E. Tetrahedron: Asymmetry 2003, 14, 2247.
4. Kinbara, K.; Oishi, K.; Harada, Y.; Saigo, K. Tetrahedron
2000, 56, 6651.
A stirred solution of ( )-1 (5 g, 27.7 mmol) in methanol
(100 mL) was added with (S)-6 (4.61 g, 27.7 mmol), re-
fluxed for 5 h. After stirring overnight, the resultant sus-
pension was filtered isolating (S)-6Æ(S)-1 (4.48 g, 93.3%
of the theoretical amount) as a white crystalline solid:
5. Pallavicini, M.; Valoti, E.; Villa, L.; Piccolo, O. Tetrahe-
dron: Asymmetry 1996, 7, 1117.
1
mp 214 ꢁC; ee of (S)-1 99.0%; H NMR (DMSO-d6): d
1.45 (d, 3H, J = 6.6 Hz), 4.15 (dd, 1H, J = 5.5,
11.0 Hz), 4.25 (dd, 1H, J = 2.9, 11.0 Hz), 4.41–4.46 (m,
2H), 6.72–6.82 (m, 3H), 6.83–6.86 (m, 1H), 7.74 (d,
2H, J = 8.8 Hz), 8.21 (d, 2H, J = 8.8 Hz). The salt was
decomposed as described for the resolution with (S)-2
in ethanol obtaining (S)-1 (2.33 g, 93.2%) as a white
solid: ee identical to that previously determined for the
acid liberated from a sample of the precipitate.
6. Pallavicini, M.; Valoti, E.; Villa, L.; Piccolo, O. Tetrahe-
dron: Asymmetry 1997, 8, 1069.
7. Pallavicini, M.; Valoti, E.; Villa, L.; Piccolo, O. Tetrahe-
dron: Asymmetry 2000, 11, 4017.
8. Pallavicini, M.; Valoti, E.; Villa, L.; Piccolo, O. Tetrahe-
dron: Asymmetry 2001, 12, 2489.
9. Pallavicini, M.; Valoti, E.; Villa, L.; Piccolo, O. Tetrahe-
dron: Asymmetry 2001, 12, 1071.
10. Pallavicini, M.; Bolchi, C.; Fumagalli, L.; Valoti, E.; Villa,
L. Tetrahedron: Asymmetry 2002, 13, 2277.
11. Kozma, D.; Acs, M.; Fogassy, E. Tetrahedron 1994, 50,
6907.
4.4. Thermal analyses
For DSC analyses, samples of 2–3 mg were run in her-
metically sealed aluminium pans to retard the decompo-
12. Koo, J.; Avakian, S.; Martin, G. J. J. Am. Chem. Soc.
1955, 77, 5373.