Resolution of 5-hydroxymethyl-2-oxazolidinone by preferential crystallization and investigations on the nature of the racemates of some 2-oxazolidinone derivatives

Article abstract of DOI:10.1016/j.tetasy.2004.03.038

After ascertaining its conglomerate nature by DSC and solid-state IR analyses, 5-hydroxymethyl-2-oxazolidinone 1, whose enantiomers are very important synthons, was efficiently resolved without chiral auxiliaries by preferential crystallization from a supersaturated isopropanolic solution of the racemate, slightly enriched in one enantiomer (3.7% ee). Favourable conditions to the entrainment were defined utilizing the previously constructed ternary phase diagram {(R)-1, (S)-1, 2-propanol}. Furthermore, the investigations were extended to other chiral 2-oxazolidinones with a functionalized methyl at the 5- or 4-position finding that 5-tosyloxymethyl-2-oxazolidinone is a racemic compound, whereas just the corresponding mesylate is a conglomerate as the parent alcohol 1. Interestingly, 4-hydroxymethyl-2-oxazolidinone 4 proved to be a racemic compound in contrast with its positional isomer 1 demonstrating how a relatively fine variation in the molecular structure can unpredictably influence the crystalline nature of the racemate. The X-ray structure determination carried out on (S)-(+)-1, (±)-4 and (R)-(+)-4 enlightened the importance of the hydrogen bond in determining different supramolecular assembling in the two homochiral compounds with respect to the racemic one and allowed a correlation with the stability of the crystal to be made.

Full text of DOI:10.1016/j.tetasy.2004.03.038

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 15 (2004) 1659–1665
Resolution of 5-hydroxymethyl-2-oxazolidinone by preferential
crystallization1 and investigations on the nature of the racemates of
some 2-oxazolidinone derivatives
Marco Pallavicini,a,* Cristiano Bolchi,a Raffaella Di Pumpo,a Laura Fumagalli,a
Barbara Moroni,a Ermanno Valotia and Francesco Demartinb
aIstituto di Chimica Farmaceutica e Tossicologica, Universitꢀa di Milano, viale Abruzzi 42, I-20131 Milano, Italy
bDipartimento di Chimica Strutturale e Stereochimica Inorganica, Universitꢀa di Milano, via Venezian 21, I-20133 Milano, Italy
Received 18 March 2004; accepted 29 March 2004
Available online 24 April 2004
Abstract—After ascertaining its conglomerate nature by DSC and solid-state IR analyses, 5-hydroxymethyl-2-oxazolidinone 1,
whose enantiomers are very important synthons, was efficiently resolved without chiral auxiliaries by preferential crystallization
from a supersaturated isopropanolic solution of the racemate, slightly enriched in one enantiomer (3.7% ee). Favourable conditions
to the entrainment were defined utilizing the previously constructed ternary phase diagram {(R)-1, (S)-1, 2-propanol}. Furthermore,
the investigations were extended to other chiral 2-oxazolidinones with a functionalized methyl at the 5- or 4-position finding that 5-
tosyloxymethyl-2-oxazolidinone is a racemic compound, whereas just the corresponding mesylate is a conglomerate as the parent
alcohol 1. Interestingly, 4-hydroxymethyl-2-oxazolidinone 4 proved to be a racemic compound in contrast with its positional isomer
1 demonstrating how a relatively fine variation in the molecular structure can unpredictably influence the crystalline nature of the
racemate. The X-ray structure determination carried out on (S)-(+)-1, ( )-4 and (R)-(+)-4 enlightened the importance of the
hydrogen bond in determining different supramolecular assembling in the two homochiral compounds with respect to the racemic
one and allowed a correlation with the stability of the crystal to be made.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
constituted oxazolidinones represent an advantageous
alternative to the previous ones,8;9 which defer con-
struction of the oxazolidinone cycle to the last steps of
the synthetic pathway. In this context, the O-protected
or esterified derivatives of 1, chiefly its (R)-enantiomer,
come to assume the role of key intermediates.
Homochiral 5-hydroxymethyl-2-oxazolidinone
1
is
extensively used to prepare a large number of enantio-
merically pure compounds of great biological interest. It
can be regarded as a protected 3-amino-1,2-propanediol
derivative, which allows, for instance, a variety of
functionalized b-aminoalcohols to be synthesized by
nucleophilic substitution at the exocyclic carbon, fol-
lowed by a ring opening.2;3 Compound 1 is an even more
valuable synthon since many novel pharmacological
agents are N-arylated 2-oxazolidinones bearing a func-
tionalized methyl group at the 5-position. The known
antimicrobial linezolid4 and the antidepressant beflox-
atone5 can be cited as examples. For the synthesis of
such optically active compounds, the recent methods6;7
based on the initial coupling of aryl halides with pre-
OH
OMs
OTs OH
O
O
O
O
HN
HN
HN
HN
O
O
O
O
1
2
3
4
It is therefore no wonder that several preparative
methods for enantiomerically pure 1 have been devel-
oped over the last 20 years, ranging from the enzymatic
resolution of racemic esters of 110 to the carbonylation
of (R)- and (S)-3-amino-1,2-propanediol,11–13 in turn
obtained by classical resolution of the racemate14 or
from chiral pool substances.13 Alternative accesses have
* Corresponding author. Tel.: +39-02-50317524; fax: +39-02-503175-
65; e-mail: marco.pallavicini@unimi.it
0957-4166/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2004.03.038

1660
M. Pallavicini et al. / Tetrahedron: Asymmetry 15 (2004) 1659–1665
been offered by the chromatographic separation of
diastereomeric pairs of N-(a-methylbenzyl)-5-iodo-
methyl-2-oxazolidinone obtained from (R)- or (S)-1-
phenylethylamine2 and, more recently, by the Hofmann
rearrangement of cyclic boronic acid esters of enantio-
merically pure 3,4-dihydroxybutyramide.15 However,
each of these methods has its limitations such as a large
number of steps, the poor availability of the enantio-
merically pure starting material or laborious separation
procedures.
2.6 ratio av between the solubilities, expressed as mole
fraction, of ( )-1 and of (S)-(+)-1, respectively. Such a
value, higher than 2, is typical for conglomerates,17
whose molecules are not dissociable in solution as it is
the case for those of 1.
On the basis of these results, we decided to try the res-
olution by direct crystallization according to the
entrainment procedure, which involves the preferential
precipitation of one enantiomer from a supersaturated
solution of the racemate enriched with a slight excess of
the same enantiomer. Before performing the crystalli-
zation experiments, we determined the binary melting
point phase diagram for the enantiomer system (S)-(+)-
1/(R)-())-1 and the ternary solubility phase diagram for
the system (S)-(+)-1/(R)-())-1/2-propanol at 22 ꢁC. This
solvent was selected as a good candidate for the pref-
erential crystallization considering its reasonable dis-
solving ability for both ( )-1 (8.8 mL/g) and (S)-(+)-1
(23.6 mL/g).
Over the course of our research into new methods for
preparing enantiopure protected C3 synthons, such as
glycerol carbonate and its isostere 1, we noticed that a
71–72 ꢁC melting point had been reported for the race-
mate of 1,11 while the corresponding value for the single
enantiomer ranged from 85 to 90 ꢁC according to
authors.10;13;15 The gap is not so wide as to indicate that
1 crystallizes as a conglomerate and to let us hope for
the resolution of its racemate by a straightforward
procedure of direct crystallization. Nevertheless, we
supposed that the low melting point of the racemate
could have been overestimated and, consequently, its
difference from that of the enantiomer undervalued. On
the basis of this hypothesis, we decided to synthesize
( )-1, (S)-(+)-1 and (R)-())-1 in order to analyze their
physical properties. Herein we report the results of these
analyses, which clearly show that 1 is a conglomerate,
and the successful resolution of ( )-1 by preferential
crystallization according to the entrainment proce-
dure.16 Furthermore, it seemed pertinent to extend the
investigations on the racemate type to other 2-oxazo-
lidinones bearing a functionalized methyl group at the 5-
position, such as the important synthons mesylate 2 and
tosylate 3, and to the 4-substituted isomer of 1, 4-hy-
droxymethyl-2-oxazolidinone 4. The results of such
supplementary researches are also reported herein and,
in particular, the nature of racemic compound of 4 is
compared to that of conglomerate of 1 in the light of
respective X-ray crystallographic data.
The DSC curves for ( )-1, (S)-(+)-1 and (R)-())-1
showed one peak at 56.8, 93.0 and 93.7 ꢁC, respectively,
while two distinct peaks characterized the melting pro-
files of three differently proportioned ( )-1/(R)-())-1
mixtures, the first, between 56 and 57 ꢁC, representing
the fusion of the eutectic, that is the 1/1 mixture of the
two enantiomers, while the second, at temperatures
increasing with the R enrichment, representing the ex-
cess of (R)-())-1. The resultant experimental binary
diagram, depicted in Figure 1 for mole fractions of (R)-
())-1 ranging from 0.5 to 1, is typical of a conglomerate
system. As can be seen in the same figure, the experi-
mental values acceptably fit with the theoretical ones
(solid curve) calculated on the basis of the melting point
of (R)-())-1 and of its heat of fusion (166.1 J/g) by the
€
Schroder–van Laar equation.
370
360
350
340
330
320
2. Results and discussion
We synthesized ( )-1 from 3-amino-1,2-propanediol by
treatment with diethylcarbonate according to Dani-
elmeier and Steckhan.13 From the enantiomers of the
aminodiol, we analogously prepared (S)-(+)-1 and (R)-
())-1, which were crystallized from 2-propanol after
chromatographic purification. The melting points of the
racemic oxazolidinone and its enantiomers were accu-
rately determined by DSC under identical conditions
and it was found that ( )-1 melts 36 ꢁC lower than (S)-
(+)-1 and (R)-())-1. Furthermore, IR spectra of ( )-1
and (S)-(+)-1 in the crystalline phase proved perfectly
superimposable. These data demonstrated that 1 is a
member of that minor category of enantiomer systems
(5–10% of the totality of chiral organic solids), which are
mechanical mixtures or conglomerates of crystals, each
of which being made up of homochiral molecules.
Consistently with this evidence, the solubility of the
racemate was approximately twice that of the constitu-
ent enantiomers. In 2-propanol, for instance, we found a
0.5
0.6
0.7
0.8
0.9
1.0
molar ratio of (R)-1
Figure 1. Binary melting-point phase diagram for the enantiomeric
system (R)-1/(S)-1. The solid curve represents the values calculated on
€
the basis of the Schroder–van Laar equation.
The triangular phase diagram (S)-(+)-1/(R)-())-1/2-
propanol was constructed using values of 0.069 and
0.027, that is the solubilities expressed as mole fractions,
which we had previously measured for ( )-1 and (S)-(+)-

M. Pallavicini et al. / Tetrahedron: Asymmetry 15 (2004) 1659–1665
1661
1 in 2-propanol at 22 ꢁC, respectively (see Fig. 2). At this
temperature, the mole fraction of 2-propanol for a sat-
urated solution of ( )-1 is 0.931, while it is 0.973 for a
saturated solution of (S)-(+)-1. The resultant solubility
curve was constituted by two segments (the red lines in
Figure 2) not parallel to the sides of the triangle, which
intersect to form a <60ꢁ angle.
Figure 3 gives a bird’s eye view of the ternary compo-
sitions (points A, B, C and D) assumed by the system
over the course of the first cycle, which was then
reproduced by the successive four.
Figure 2. Solubility diagram of 1 in 2-propanol at 22 ꢁC. The con-
centrations of the components are expressed as mole fractions. The
magnified upper part of the diagram shows the composition of the
saturated solution of racemate (point E: 0:931v2-propanol and
Figure 3. Area (red quadrilateral) of the solubility diagram of ( )-1
usable for the entrainment and, in the magnified part, the first cycle of
resolution of ( )-1. A and C are the ternary compositions of the
supersaturated solutions of ( )-1, R and S enriched, respectively,
where the crystallization takes place; B and D those of mother liquors
remaining from the crystallizations of (R)-())- and (S)-(+)-1, respec-
tively.
ð0:0345 ꢁ 2ÞvðꢂÞꢀ1
Þ
and of single enantiomer (0:973v2-propanol and
0:027vðSÞ-1).
In this triangular phase diagram we defined the region in
which resolution by entrainment is favourable; in other
words, the more suitable conditions of supersaturation
for an efficient resolution by preferential crystallization.
A 0.15 mole fraction of ( )-1 [0.27 g of ( )-1 per milli-
litre of 2-propanol, i.e., more than twice its solubility,
which is 0.11 g/mL] was empirically established as a
concentration corresponding to a sufficiently high but
nevertheless even metastable supersaturation, since,
without seeding, no crystallization took place in a
solution with such a concentration during some days at
room temperature. The area of the ternary diagram
useful for the entrainment resulting from such a limit of
metastable supersaturation is a parallelogram (see Fig.
3) whose sides are formed by the prolongations of the
two segments of the solubility curve and by the respec-
tive parallel lines intersecting at the point corrrespond-
ing to 0.15 mole fraction of ( )-1 (ternary composition:
0:85v2-propanol ꢀ 0:075vðSÞ-1 ꢀ 0:075vðRÞ-1).18 Within this
area, we drew five consecutive crystallization cycles.
In the first cycle, a highly supersaturated solution of ( )-
1 (4.4 g in 20 mL of 2-propanol), enriched with a slight
excess of (R)-())-1 (0.17 g) (see ternary composition A in
Fig. 3), was seeded with crystals of this latter enantio-
mer, vigorously stirred for 1 h and then allowed to stand
over night at room temperature. After this time, we only
decanted the mother liquors isolating a precipitate of the
laevorotatory enantiomer, which was, after drying,
about the triple of the initial enrichment and had a 77%
optical purity. Concomitantly, the sign of the optical
rotation of the mother liquors, now having composition
B, had changed to positive. This S enriched solution
was added with ( )-1 to restore the initial supersatura-
tion (composition C) and the procedure repeated for the
(S)-enantiomer. Again, a crystalline, but now destroro-
tatory precipitate was isolated in analogous yield
with 94% optical purity from mother liquors, which

1662
M. Pallavicini et al. / Tetrahedron: Asymmetry 15 (2004) 1659–1665
had meantime assumed composition D. For each
crystallization, two or three times the amount of the
initial excess of the enantiomerically pure compound
was separated; after five cycles, for both enantio-mers
we obtained a 2.6 g amount with near 90% optical
purity (see Table 1). Finally, these two quantities
were recrystallized from 2-propanol yielding about 2 g of
each enantiomer, more than 11 times the amount of the
initial investment in (R)-())-1, with >99.6% optical
purity.
bilities, expressed as a mole fraction, of ( )-2 and of (R)-
())-2, respectively, in methanol.
Finally, we took into account 4-hydroxymethyl-2-oxa-
zolidinone 4, another important chiral derivative of 2-
oxazolidinone, in order to verify whether it formed a
conglomerate as its 5-hydroxymethyl substituted isomer
1. Similarly to the case of this latter, the nature of the
enantiomeric system has been never examined and the
physical data available from the literature [96–99 and
77–81 ꢁC melting point for (R)-(+)-419 and ( )-4,20
respectively] were not informative. Therefore, we pre-
pared (R)-(+)-4 and ( )-4 from (S)-serine methyl ester
hydrochloride and its racemate, respectively, according
to literature methods:19;21 reaction with triphosgene gave
(S)-())- and ( )-4-methoxycarbonyloxazolidin-2-one,
which were successively reduced to the corresponding
hydroxymethyl derivatives by treatment with sodium
borohydride. On the contrary, an alternative synthetic
route, involving an initial reaction of glycidol with
benzyl isocyanate to give N-benzyl-4-hydroxymethyl-2-
oxazolidinone,22 was left due to the difficulty in debenz-
ylating such an intermediate. DSC analyses showed that
(R)-(+)-4 melts at 103.4 ꢁC, while ( )-4 at 85.6 ꢁC; the
solid-state IR spectra of the single enantiomer is not
superimposable to that of the racemate. On the basis of
these results, we concluded that 4 is not a conglomerate
like 1, but a racemic compound. This means that the
shift of the hydroxymethyl residue from the 5 to the 4
position of the 2-oxazolidinone ring hinders, in the case
of the racemate, the transfer of chirality from molecular
to supramolecular level. It was presumable that such a
dramatic change, due to simple positional isomerism,
resulted from a variation of the hydrogen-bonded
molecular assembly, considering the hydrogen bond
acceptor and donor nature of both the hydroxymethyl
group and the cyclic carbamate. The X-ray analysis of
the solid-state structures of (S)-(+)-1, (R)-(+)-4 and ( )-
4 confirmed this hypothesis revealing that the hydrogen
bond pattern of the racemic compound substantially
Prompted by the successful resolution of ( )-1 by pref-
erential crystallization, we successively considered other
2-oxazolidinones bearing a functionalized methyl group
at the 5-position. In the case of the 5-halomethyl-2-ox-
azolidinones, the literature data indicated that the single
enantiomers melt at the same temperature or at lower
temperatures than the respective racemates, thus
excluding the possibility of forming conglomerates for
such compounds.13 For 5-mesyloxymethyl-2-oxazolidi-
none 2, the melting point of the single enantiomer (118–
120 ꢁC) was reported in the literature,13 but not that
of the racemate; vice versa, for 5-tosyloxymethyl-2-
oxazolidinone 3, the melting point of the racemate
(98–99 ꢁC),11 but not that of the single enantiomer. There-
fore, we prepared ( )-2 and ( )-3 from ( )-1 and (R)-
())-2 and (S)-(+)-3 from (R)-())- and (S)-(+)-1, respec-
tively. The DSC and IR analyses of the tosylate allowed
its nature of racemic compound to be ascertained on the
basis of the melting point of (S)-(+)-3 exceeding that of
( )-3 by only 17 ꢁC (118.5 vs 101.2 ꢁC) and of the
respective solid-state spectra, which are not superim-
posable. On the contrary, the more relevant difference
(30 ꢁC) between the melting point of (R)-())-2 (117.7 ꢁC)
and that of ( )-2 (87.9 ꢁC) and the identity of the
respective IR spectra in the crystalline phase demon-
strated that the mesylate is a conglomerate like its par-
ent alcohol 1 implying the resolvability of the racemic
mixture by direct crystallization. Consistently with these
data, we determined a 2.35 ratio av between the solu-
Table 1. Resolution of ( )-1 by entrainment
Cycle no
1 added (mg)
Recovery of resolved 1 (mg)
a3D0 of the mother
liquors
( )
())
())a
(+)a
1
2
3
4
5
4400
550
500
480
500
610
810
480
460
510
170
524 (77.2%)
+0.397
)0.485
+0.381
)0.427
+0.507
)0.331
+0.391
)0.259
+0.360
)0.280
480 (93.7%)
500 (98.0%)
805 (70.5%)
460 (92.8%)
480 (>99.6%)
610 (87.1%)
480 (92.8%)
510 (77.5%)
370 (>99.6%)
2615c
Total
Residue (mg)d
9300
170
2604b
4140 (R enriched; 4.2% ee)
25
D
a Between brackets, the optical purity resulting from the percent ratio of the observed specific rotation to the specific rotation {½aꢃ ¼ þ30:4, (c 2,
isopropanol)} of pure (S)-1 (>99.6% ee).
b 1905 mg (o.p. > 99.6%) after recrystallization from 2-propanol.
c 2037 mg (o.p. > 99.6%) after recrystallization from 2-propanol.
d Recovered by concentration of the mother liquors remaining from the last crystallization.

M. Pallavicini et al. / Tetrahedron: Asymmetry 15 (2004) 1659–1665
1663
O8
C7
N3
C5
O1
C4
C2
O6
Figure 4. ORTEP partial drawing of the crystal packing for ( )-4,
showing the hydrogen bond pattern.
O6
differs from those of the two homochiral ones, which are
instead quite similar.
C2
C5
N3
O1
A view of a significant portion of the crystal packing of
( )-4 is reported in Figure 4. Here, pairs of enantiomeric
molecules, located about crystallographic inversion
centres, are held together via hydrogen bonds between
C4
O8
the donor OH(8) and the acceptor carbonylic O(6) with
ꢁ
C7
a
O(6)ꢄ ꢄ ꢄO(8) interaction of 2.766(1) A, and
a
O(6)ꢄ ꢄ ꢄH(8)–O(8) angle of 163.9(2)ꢁ. These dimeric units
are stacked to form chains extending along [1 0 0] be-
cause they interact through a further hydrogen bond
between the O(8) hydroxyl atom, which acts, in this
case, as an acceptor, and the donor N(3) atom. The
Figure 5. ORTEP partial drawing of the crystal packing for (S)-(+)-1,
showing the hydrogen bond pattern.
ꢁ
O(8)ꢄ ꢄ ꢄN(3) interaction is 2.882(1) A, while the
O(8)ꢄ ꢄ ꢄH(3)–N(3) angle is 162.15(2)ꢁ. As a result, each
molecule is hydrogen bonded to three other molecules.
The whole crystal is therefore derived by the assembling
of chains among which only weak van der Waals
interactions occur.
In each of the homochiral (S)-(+)-1 and (R)-(+)-4
compounds, the hydrogen bond pattern gives rise to a
tridimensional assemblage of molecules, extending all
over the crystal, where each molecule is hydrogen bond-
ed to four neighbouring molecules (see Figs. 5 and 6). In
all cases OH(8) acts as donor towards O(6) and acceptor
from N(3). The O(8)ꢄ ꢄ ꢄO(6) and O(8)ꢄ ꢄ ꢄN(3) interac-
O1
O8
C7
C5
O6
C2
N3
C4
ꢁ
tions are 2.712(3) and 2.886(4) A for (S)-(+)-1, and
ꢁ
2.735(1) and 2.871(1) A for (R)-(+)-4. The different
pattern of the hydrogen bonds, tridimensionally ex-
tended in (S)-(+)-1 and (R)-(+)-4, but not in ( )-4, is
consistent with the higher stability of the homochiral
crystals with respect to the racemic compound, as indi-
cated by the respective melting points.
Figure 6. ORTEP drawing of a molecule with its neighbours for (R)-
(+)-4.
3. Conclusion
has been first ascertained by us. As proof of the relative
rarity of this type of enantiomeric system, 5-chloro-
methyl-, 5-bromomethyl-, 5-iodomethyl-2-oxazolid-
To the best of our knowledge, the conglomerate nature
of the enantiomeric mixtures of 1 and of its mesylate 2

1664
M. Pallavicini et al. / Tetrahedron: Asymmetry 15 (2004) 1659–1665
inone and 2-oxazolidinones 3 and 4, though structurally
close to 1 and 2, do not form conglomerates. For the
halomethyl derivatives, this is unambiguously indicated
by the literature data, while, in the case of 3 and 4, has
been disclosed by the present investigations, which have
demonstrated their nature of racemic compounds.
ride and its racemate were respectively used as starting
materials. The oxazolidinone cycle was formed by
treatment with triphosgene as described in Ref. 21; the
ester function was then reduced to an alcohol with
NaBH4 according to Ref. 19. The only variant was, in
the case of (R)-(+)-4, an additional recrystallization of
the reduction product isolated by silica gel chromato-
graphy and crystallized from methanol in order to
eliminate traces of eutectic (mp 81.7 ꢁC), revealed by the
On the basis of this evidence and considering the
intrinsic value of the enantiomers of 1 as synthetic
intermediates, the resolution of ( )-1 by preferential
crystallization according to the entrainment procedure
was tried and successfully developed. The main advan-
tages of such a new procedure over those described in
the literature (enzymatic or classical resolution and
synthesis from homochiral precursors) can be summa-
rized as follows: (i) the only necessary material, in
addition to the solvent, is ( )-1, which is easily prepared
from largely available racemic 3-amino-1,2-propanediol;
(ii) chiral auxiliaries are not needed, excepting the min-
imum initial enantiomeric enrichment; (iii) the proce-
dure is very simple, not implying laborious working up
and separations, but only consisting of reiterated crys-
tallizations and removals of the mother liquor. Fur-
thermore, compared to the other known resolutions of
nondissociable racemates by entrainment,23 this shows
an excellent efficiency of ca. 20% (chemical yield, relative
to half of the starting racemate, of each crystalliza-
tion · optical purity) and a good productivity of ca.
2.5% (percent ratio of the weight of the precipitate to
that of the mother solution).
DSC curve and resulting in lower specific optical rota-
25
D
tion {+26.2 vs +32.2 ½aꢃ (c 1, MeOH) of recrystallized
(R)-(+)-4} and melting point [102.4 vs 103.4 ꢁC for re-
crystallized (R)-(+)-4]. ( )-4: mp 85.6 ꢁC.
4.2. Typical entrainment procedure
The operation was carried out in a three-necked round-
bottom flask equipped with a condenser, a thermometer
and a magnetic stirrer.
( )-1 (4.4 g, 37.6 mmol) was suspended in 15.7 g of 2-
propanol together with 0.17 g (1.45 mmol) of (R)-())-1.
The stirred suspension was heated at boiling tempera-
ture for 5 min to ensure complete dissolution of the solid
[possible insoluble particles had been previously re-
moved from ( )-1 and (R)-())-1 by filtration of respec-
tive unsatured solutions in 2-propanol on filter paper]
and then slowly cooled to 31 ꢁC. Seeds of (R)-())-1 (5 mg
ca.) were added under stirring. The temperature was let
down to 25 ꢁC in 60–90 min ca. after which stirring was
stopped. The mixture, which still had the appearance of
a clear solution with the previously added seeds settled
to bottom, was allowed to stand overnight at 22 ꢁC.
During this period, a white precipitate of (R)-())-1 was
formed, which adhered to the inner side of the flask and
could be easily recovered by decanting the mother
4. Experimental
4.1. Materials and methods
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. IR spectra were recorded with a FT-IR
Paragon 1000 PC Perkin–Elmer spectrometer. Melting
points were determined 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. For DSC analy-
ses, samples of 2–5 mg were run in crimped aluminium
pans. The enantiomeric mixtures of (S)-(+)-1 and (R)-
())-1 with different molar ratios were prepared by mix-
ing ( )-1 with increasing quantities of (R)-())-1. All the
analyses were performed with a heating rate of
liquor. After drying under vacuum, such a precipitate
20
D
amounted to 524 mg: ½aꢃ ¼ ꢀ23:5 (c 2, 2-propanol); 1H
NMR (DMSO-d6) d 3.21 (pseudo t, 1H, C(4)H,
J ¼ 8:0), 3.39–3.55 (m, 3H, C(4)H and CH2–OH), 4.49
(m, 1H, C(5)H), 5.05 (t, 1H, OH; J ¼ 5:7), 7.38 (br s,
1H, NH). The mother liquor, which showed
a3D0 ¼ þ0:397, was diluted with a little 2-propanol so
that the initial content of solvent (15.7 g) was exactly
restored. ( )-1 (550 mg) was then added. The suspension
was boiled for 5 min and the resultant solution slowly
cooled to 31 ꢁC under stirring and seeded with 5 mg of
(S)-(+)-1. Again, the temperature was let down to 25 ꢁC
in 60–90 min ca. Stirring was stopped and the clear
solution, containing the undissolved crystals used as
seeds, allowed to stand overnight at 22 ꢁC. After this
1 ꢁC minꢀ1
.
( )-1, (S)-(+)-1 and (R)-())-1 were prepared from ( )-,
(S)-())- and (R)-(+)-3-amino-1,2-propanediol, respec-
tively, by treatment with diethyl carbonate according to
Ref. 13. ( )-1: mp 56.8 ꢁC; (S)-(+)-1: mp 93.0 ꢁC; (R)-
())-1: mp 93.7 ꢁC. ( )-2 and (R)-())-2 were obtained
from ( )-1 and (R)-())-1 by the method described in
Ref. 13. ( )-2: mp 87.9 ꢁC; (R)-())-2: mp 117.7 ꢁC. ( )-3
and (S)-(+)-3 were synthesized from ( )-1 and (S)-(+)-1
following the procedure reported in Ref. 11. ( )-3: mp
101.2 ꢁC; (S)-(+)-3: mp 118.5 ꢁC. For the preparation of
(R)-(+)-4 and ( )-4, (S)-serine methyl ester hydrochlo-
period, isolation and drying of the precipitate as above
20
D
2, 2-propanol); H NMR was identical to that of (R)-
gave (S)-(+)-1 (480 mg) as a white solid: ½aꢃ ¼ þ28:5 (c
1
())-1. The same cycle of operation was carried out four
more times, yielding a total of 2.615 g of (S)-(+)-1 and
2.604 g of (R)-())-1 with 88% and 87% optical purity,
respectively (see Table 1). These two quantities were
recrystallized from 2-propanol yielding 2.037 g of (S)-
(+)-1 {mp identical to that of the synthesized sample;
25
25
½aꢃ ¼ þ30:4 (c 2, isopropanol); ½aꢃ ¼ þ38:2 (c 1.85,
D
D
25
D
ethanol); lit.15 ½aꢃ ¼ þ38:4 (c 1.35, ethanol) for >99.9%

M. Pallavicini et al. / Tetrahedron: Asymmetry 15 (2004) 1659–1665
1665
ee} and 1.905 g of (R)-())-1 (mp identical to that of the
References and notes
25
synthesized sample; ½aꢃ ¼ ꢀ30:4 (c 2, isopropanol);
D
25
D
1. Pallavicini, M.; Bolchi, C.; Di Pumpo, R.; Fumagalli, L.;
Moroni, B.; Valoti, E. It. Pat. Appl. MI2003A 001788,
2003.
2. Cardillo, G.; Orena, M.; Sandri, S.; Tomasini, C. Tetra-
hedron 1987, 43, 2505.
½aꢃ ¼ ꢀ38:3 (c 1.85, ethanol).
The mother liquor of the last crystallization had
a3D0 ¼ ꢀ0:280 and, after concentration, afforded 4.14 g of
R enriched 1 with ca. 4% enantiomeric excess. These
data were consistent with the amount initially submitted
to resolution (4.57 g) and its optical purity (3.7%).
3. Danielmeier, K.; Schierle, K.; Steckhan, E. Tetrahedron
1996, 52, 9743.
4. The Merck Index, 13th ed.; Merck Research Laboratories:
Whitehouse Station, NJ, 2001; Monograph number 5526.
5. The Merck Index, 13th ed.; Merck Research Laboratories:
Whitehouse Station, NJ, 2001; Monograph number 1022.
6. Madar, D. J.; Kopecka, H.; Pireh, D.; Pease, J.; Pliu-
shchev, M.; Sciotti, R. J.; Wiedeman, P. E.; Djuric, S. W.
Tetrahedron Lett. 2001, 42, 3681.
7. Mallesham, B.; Rajesh, B. M.; Rajamohan Reddy, P.;
Srinivas, D.; Trehan, S. Org. Lett. 2003, 5, 963.
8. Jarreau, F. X.; Rovei, V.; Koenig, J. J.; Schoofs, A. Eur.
Pat. Appl. EP 424244, 1990.
9. Brickner, S. J.; Hutchinson, D. K.; Barbachyn, M. R.;
Manninen, P. R.; Ulanowicz, D. A.; Garmon, S. A.;
Grega, K. C.; Hendges, S. K.; Toops, D. S.; Ford, C. W.;
Zurenko, G. E. J. Med. Chem. 1996, 39, 673.
10. Hamaguchi, S.; Asada, M.; Hasegawa, J.; Watanabe, K.
Agric. Biol. Chem. 1985, 49, 1509.
4.3. X-ray structure determination
Crystal data for (S)-(+)-1: C4H7NO3, f.w. 117.11,
monoclinic, space group P21 (no 4), a ¼ 5:566ð1Þ,
ꢁ
b ¼ 7:900ð1Þ, c ¼ 6:243ð1Þ A, b ¼ 104:98ð1Þꢁ, Z ¼ 2,
dcalcd ¼ 1.467 Mg mꢀ3
,
l ¼ 1:258 cmꢀ1
,
F ð000Þ 124,
R ¼ 0:0304, Rw ¼ 0:0772 for 1287 observed reflections
[I > 2rðIÞ].
Crystal data for ( )-4: C4H7NO3, f.w. 117.11, mono-
clinic, space group P21/c (no 14), a ¼ 5:131ð2Þ,
ꢁ
b ¼ 8:207ð4Þ, c ¼ 12:246ð4Þ A, b ¼ 92:56ð3Þꢁ, Z ¼ 4,
dcalcd ¼ 1.510 Mg mꢀ3
,
l ¼ 1:300 cmꢀ1
,
F ð000Þ 248,
11. Seneci, P.; Caspani, M.; Ripamonti, F.; Ciabatti, R. J.
Chem. Soc., Perkin Trans. 1 1994, 2345.
12. Warmerdam, E. G. J. C.; Brussee, J.; van der Gen, A.;
Kruse, C. G. Helv. Chim. Acta 1994, 77, 252.
13. Danielmeier, K.; Steckhan, E. Tetrahedron: Asymmetry
1995, 5, 1181.
R ¼ 0:0300, Rw ¼ 0:0818 for 1238 observed reflections
[I > 2rðIÞ].
Crystal data for (R)-(+)-4: C4H7NO3, f.w. 117.11,
orthorhombic, space group P212121 (no 19),
ꢁ
a ¼ 6:909ð3Þ, b ¼ 8:083ð3Þ, c ¼ 9:209ð4Þ A, Z ¼ 4,
ꢀ
14. Pitre, D.; Fedeli, F. Arch. Pharm. 1986, 319, 193.
dcalcd ¼ 1.512 Mg mꢀ3
,
l ¼ 1:302 cmꢀ1
,
F ð000Þ 248,
15. Hollingsworth, R. I.; Wang, G. US 6,288,238 B1, 2001.
16. Jacques, J.; Collet, A.; Wilen, S. Enantiomers, Racemates
and Resolutions; Wiley: New York, 1981; pp 223–241.
17. Ref. 16, p 188.
R ¼ 0:0248, Rw ¼ 0:0675 for 1075 observed reflections
[I > 2rðIÞ].
18. For this empirical manner of determining the region,
which is useful for the entrainment, see Ref. 16, p 236.
19. Sibi, M. P.; Rutherford, D.; Sharma, R. J. Chem. Soc.,
Perkin Trans. 1 1994, 1675.
20. Koellensperger, F. G.; Hartleben, Y.; Kretzschmar, R.;
Neteler, B. Ger. Offen. DE 2538424, 1977.
21. Doyle, M. P.; Dyatkin, A. B.; Protopopova, M. N.; Yang,
C. I.; Miertschin, C. S.; Winchester, W. R.; Simonsen, S.
H.; Lynch, V.; Ghosh, R. Recl. Trav. Chim. Pays-Bas
1995, 114, 163.
Intensity data collection were performed with graphite
ꢁ
monochromatized MoKa radiation (k ¼ 0:71073 A) on
an Enraf-Nonius CAD4 up to 2h ¼ 56ꢁ. Data sets were
corrected for Lorentz-polarization effects and absorp-
tion.24 Structure solution was carried out by direct
25
methods using SHELXS86 and full-matrix least-squares
refinements on F 2 were performed using SHELXL97
26
.
All non-H atoms were refined anisotropically and H
atoms isotropically.
22. Gautun, H. S. H.; Carlsen, P. H. J Tetrahedron: Asym-
metry 1995, 6, 1667.
Crystallographic data (excluding structure factors) for
the structure herein have been deposited with the
Cambridge Crystallographic Data Centre as supple-
mentary publication numbers CCDC 233909-233911.
Copies of the data can be obtained, free of charge, on
application to CCDC, 12 Union Road, Cambridge CB2
1EZ, UK [fax: +44(0)-1223-336033 or e-mail: depos-
it@ccdc.cam.ac.uk].
23. For examples, see the following: (a) Resolution of hydro-
benzoin: Collet, A.; Brienne, M. J.; Jacques, J. Chem. Rev.
1980, 80, 215; (b) Resolution of 2,20-dimethoxy-1,10-
binaphthalene: Gottarelli, G.; Spada, G. P. J. Org. Chem.
1991, 56, 2096; (c) Resolution of intermediates for
diltiazem preparation: Giordano, C.; Tentorio, D.; Casa-
grande, R.; Bertin, P.; Merli, V.; Sagramora, G. US
5,097,059, 1992.
24. North, A. C. T.; Phillips, D. C.; Mathews, F. S. Acta
Crystallogr. Sect. A 1968, 24, 351.
25. Sheldrick, G. M. SHELXS 86, Acta Crystallogr. Sect. A
1990, 46, 467.
Acknowledgements
This work was supported by the Ministero dell’Istruzi-
ꢀ
one, dell’Universita e della Ricerca of Italy.
€
26. Sheldrick, G. M. SHELXS 97, University of Gottingen,
1997.