HU ET AL.
(400 MHz, DMSO-d6, δ): 7.83-7.85 (m, 4H), 7.30-7.44 (m, 4H+4H), 5.63
study, we first explored a mixed group of D-DBTA, D-DTTA,
and D-DMTA as resolving agents. The mixture of resolving
agents was added to the solution of equimolar amounts of
(s, 2H), 5.10 (s, 1H), 2.36 (s, 6H) ppm.; MS (ESI, m/z): [M + H]+
611.1; found, 610.9.
The complex (R)-1·Ca·4 was suspended in 10% hydrochloric acid
(10 mL) and toluene (5 mL). The mixture was warmed to 80°C for a short
period with vigorous stirring, and then cooled to room temperature. The
precipitate (DTTA) was filtered. The toluene layer was separated and the
aqueous layer was extracted with ethyl acetate (3 × 10 mL). The organic
phases were combined, dried over anhydrous sodium sulfate, and con-
centrated in vacuum. 10 mL of warm toluene was added to the residue
and then the undissolved DTTA was filtered off. The filtrate was concen-
trated to yield 140 mg (60%) of (R)-2-chloromandelic acid with 99% ee; mp:
i
rac-1 and CaO in PrOH-H2O, and the precipitate was ana-
1
lyzed by HPLC and H NMR, which showed that DTTA was
the major component in the diastereomeric salt. (R)-1 was
obtained at 48.5% efficiency and 54.7% ee (Table 1, entry 1).
Dutch resolution is the use of structurally related resolving
agents.27 In the resolution of rac-1, the resolving efficiency
of the Dutch resolution was not as good as that of D-DBTA
or D-DTTA individually (Table 1, entries 1, 2, and 4), and D-
DTTA was the best resolving agent. (R)-1 was obtained at
62.2% efficiency and 54.4% ee in complex formation (Table 1,
entry 4). Thus, an effective resolving agent, DTTA, could be
successfully selected by “Dutch resolution.”
20
116–117°C, ½αꢁD = ꢀ124 (c = 1 in H2O); the (R)-1 reported in Aldrich
(Milwaukee, WI) Chemistry: [α]2D3 = ꢀ126 (c = 3 in H2O). 1H NMR
(400 MHz, DMSO-d6, δ): 7.30-7.52 (m, 4H), 5.34 (s, 1H) ppm. The enantio-
meric purity of (R)-1 was determined by HPLC analyses using an Agilent-
1100 (Palo Alto, CA) instrument with hexane/isopropanol (90/10) and
0.05% trifluoroacetic acid as eluent under detection wavelength of
222 nm. The flow rate was 1.0 mL/min. Retention time: (R)-1 17.3 min,
(S)-1 19.4 min.
In order to improve the efficiency of the resolution, several
kinds of solvents, such as methanol, ethanol, isopropanol, wa-
ter, acetone, and acetonitrile were surveyed and their mix-
tures were examined. The mixed solvent of isopropanol and
water was found to be an excellent solvent system. We ini-
tially utilized tartaric acid derivatives as chiral host reagent
to resolve 2-chloromandelic acid and found the standard mo-
lar ratio of racemic guest and resolving agent was 1:1. But
the efficient molar ratio of tartaric acid derivative neutral cal-
cium salt and racemic compound may be another ratio.21,22
Thus, different ratios were studied and the ratio of 0.8:1 was
found to be the best in the resolving efficiency (Table 1, entry
6). Unfortunately, it was difficult to obtain the qualified single
crystal of less soluble salts for X-ray crystallographic analysis.
The structures of the resolution intermediates were charac-
Racemization of (S)-2-Chloromandelic Acid (S)-1
A typical racemization process in Table 2 is described as follows. The
optically enriched compound (S)-2-chloromandelic acid was obtained
from the mother liquid in the above resolution. A solution of (S)-2-
chloromandelic acid (100 mg, 0.54 mmol, 67% ee) in DMSO (5 mL) and
H2O (0.1 mL) was stirred in a round-bottom flask. Powdered sodium hy-
droxide (64 mg, 1.6 mmol) was added and stirred at 160 °C for 2 h. The
mixture was diluted with 10% hydrochloric acid. The aqueous solution
(pH <2) was extracted with ethyl acetate (3 × 10 mL). The combined ex-
tracts were washed successively with water (10 mL) and brine (10 mL).
The collected organic layers were dried over MgSO4, filtered, and con-
centrated. The product (82 mg, 82% yield) was obtained as a white solid.
1
terized with H NMR and mass spectra. In the experiments,
1H NMR demonstrated that the ratio of 4 and 1 in the less-
soluble salt is close to 1:1. Mass spectra analysis suggested
that the resulting complex was formulated as (R)-1·Ca·4. Af-
ter third being recrystallized from the mixed solvents (iPrOH:
H2O = 2:1), the calcium complex of (R)-1·Ca·4 was suitably
acidified to give (R)-1 by extraction with 99% ee. Thus, utiliz-
ing 4 as a resolving agent, optical pure (R)-1 was prepared
with 99% ee in 60% yield.
RESULTS AND DISCUSSION
Resolution of 2-Chloromandelic Acid
DBTA has been successfully used in calcium salt formation
for the resolution of rac-1, via diastereomeric salt forma-
tions.25 In this method, the best resolution efficiency could
reach 58.2% in the mixture of water and isopropanol (Table 1,
entry 2). Tartaric acid and its diaryl carboxylate derivatives D-
DBTA 3, D-DTTA 4, and D-DMTA 5 are widely used and
nontoxic at chiral resolution agents (Fig. 1). To improve the
resolution efficiency of 1, tartaric acid derivatives (3, 4, and
5) were explored with diastereomeric calcium salts. In this
Racemization of (S)-2-Chloromandelic Acid
While 60% of (R)-1 was obtained from the resolution of the
rac-1 with D-DTTA and Ca2+, undesired (S)-1 remains as
waste. Compound (S)-1, which is a precursor of inactive
clopidogrel enantiomer, could be discarded. We attempted
to find a method to convert this unwanted compound (S)-1
into the useful (R)-1. The conversion could be achieved by
transforming (S)-1 into rac-1 and then into (R)-1. Compound
(S)-1 was thus recycled just by repeating the racemization
and resolution procedure. Herein, the racemization of (S)-1
was extensively studied and many conditions were tried,
which are summarized in Table 2.
Racemization of most carboxylic acids can be carried out
with the coexistence of alkali and dimethyl sulfoxide.28 Thus,
the effects of the category and concentration of base, temper-
ature, and reaction time on rate of racemization are
discussed. As can be seen in Table 2, (S)-1 can be efficiently
racemized into rac-1 with an inorganic base such as sodium
hydroxide, potassium hydroxide, sodium carbonate at 160°
C (Table 2, entries 2–4). However, an organic weak base is
not a suitable base for racemization. Under the same condi-
tion, racemization almost never occurs with Et3N. This means
that strong alkalinity is favorable to racemization of 1. It is
TABLE 1. Resolution of rac-1 by tartaric acid derivatives (D-
DBTA 3, D-DTTA 4 and D-DMTA 5) and CaO
Entry
1: CaO : 3 : 4 : 5a
ee (%)b
Yield (%)c
Eff.(%)d
1
2
3
4
5
6
7
8
3:3:1:1:1
1:1:1:0:0
1:1:0:0:1
54.7
53.8
47.1
54.4
52.3
65.6
60.7
61.6
88.6 (1:3:2)e
108.2
95.8
114.4
114.6
101.8
89.2
76.4
48.5
58.2
45.1
62.2
59.9
66.8
54.1
47.1
1:1:0:1:0
0.9:1:0:1:0
0.8:1:0:1:0
0.7:1:0:1:0
0.6:1:0:1:0
aThe initial molar ratio of rac-1, CaO, 3, 4, and 5.
bIn all experiments, (R)-1 was obtained and the enantiomeric purity was de-
termined by HPLC.
cThe yield of the diastereomeric salt based on half the initial complex.
dResolving efficiency, defined as a product of the yield of the diastereomeric
salt and the ee of the liberated 1.
eThe molar ratio of 3, 4, and 5 in the precipitated salt was determined by 1H
NMR.
Chirality DOI 10.1002/chir