Non-linear effects in the enantiomeric separation of mandelic acid using the mixtures of amphoteric resolving agents

Article abstract of DOI:10.1016/j.tetasy.2015.05.010

Abstract In this study, (S)-phenylalanine, aspartame and (S)-pregabalin, which contain α-, β- and γ-amino acidic moieties were successfully applied as resolving agents for the enantiomeric separation of mandelic acid via the formation of diastereomeric salts. The resolution of mandelic acid was also attempted with mixtures of the corresponding resolving agents and a more efficient enantiomeric separation of the mandelic acid was observed compared with the given sole resolving agents in a few instances. The best results were obtained with a mixture of (S)-phenylalanine and (S)-alanine; the latter resolving agent was insufficient when it was applied alone. The resolution efficiency of mandelic acid could also be improved by applying structurally related amphoteric achiral additives along with the corresponding resolving agents.

Full text of DOI:10.1016/j.tetasy.2015.05.010

Tetrahedron: Asymmetry 26 (2015) 721–731
Contents lists available at ScienceDirect
Tetrahedron: Asymmetry
journal homepage: www.elsevier.com/locate/tetasy
Non-linear effects in the enantiomeric separation of mandelic acid
using the mixtures of amphoteric resolving agents
a
a
a
a
b
b
}
Zsolt Szeleczky , Péter Bagi , Balázs Fodi , Sándor Semsey , Emese Pálovics , Ferenc Faigl ,
Elemér Fogassy a,
⇑
a Department of Organic Chemical Technology, Budapest University of Technology and Economics, Budapest, Hungary
b MTA-BME Organic Chemical Technology Research Group, Hungarian Academy of Sciences, Department of Organic Chemical Technology, Budapest University of Technology
and Economics, Budapest, Hungary
a r t i c l e i n f o
a b s t r a c t
Article history:
In this study, (S)-phenylalanine, aspartame and (S)-pregabalin, which contain a-, b- and c-amino acidic
Received 30 April 2015
Accepted 18 May 2015
Available online 24 June 2015
moieties were successfully applied as resolving agents for the enantiomeric separation of mandelic acid
via the formation of diastereomeric salts. The resolution of mandelic acid was also attempted with mix-
tures of the corresponding resolving agents and a more efficient enantiomeric separation of the mandelic
acid was observed compared with the given sole resolving agents in a few instances. The best results
were obtained with a mixture of (S)-phenylalanine and (S)-alanine; the latter resolving agent was insuf-
ficient when it was applied alone. The resolution efficiency of mandelic acid could also be improved by
applying structurally related amphoteric achiral additives along with the corresponding resolving agents.
Ó 2015 Published by Elsevier Ltd.
1. Introduction
successfully applied in the resolution of 2-chloromandelic acid
which is an intermediate in the synthesis of clopidogrel.4 Kim
Currently, one of the simplest preparations of pure enantiomers
involves the formation and separation of diastereomers. Most of
these methods are carried out by fractional crystallization of the
diastereomeric salts obtained in a corresponding chiral acid—chiral
base reaction.1 In this instance, the racemic acid reacts with a suit-
able homochiral base in an appropriate solvent and the less soluble
diastereomeric salt crystallizes, thus allowing the diastereomers to
be separated by filtration. The corresponding enantiomer or enan-
tiomeric mixture can then be recovered by decomposition of the
diastereomer. The same procedure can also be used when a race-
mic base is reacted with a homochiral acid.2
et al. also applied a-amino acids to separate the enantiomers of
mandelic acid.5 In our previous study, we also used homochiral
amino acids for the partial resolution of mandelic acid and we
observed that crystallization time played a significant role in those
enantiomeric separations.6 In addition to the enantioseparation of
mandelic acid obtained with amino acids, several chiral bases, such
as alkaloids,7a (1R,2S)-ephedrine,7b (1R,2S)-N-methylephedrine,7c
(S)-phenylglycine esters,7d (S)-2aminobutan-1-ol,7e (S)-2-(benzyl-
amino)butan-1-ol,7f
(S)-1-phenylethylamine,7g
(S)-1-amino-
3-phenoxypropan-2-ol,7h (1R,2S)-2-amino-1,2-diphenylethanol7i
and (2S,3S)-2,3-dihydroxy-1,4-bis(hydroxyamino)butane7j have
been used efficiently for the preparation of enantiopure mandelic
acid via the formation of diastereomeric salts. Moreover asymmetric
synthetic8 and enzymatic methods9 are also described in the
literature.
An ideal resolving agent is inexpensive, readily available, non-
toxic and regenerable. Among amphoteric molecules, chiral amino
acids completely fulfil these criteria. For the resolution of racemic
bases, amino acids with acidic side chains, such as aspartic acid
and glutamic acid are suitable. The resolution of racemic
1
-aminoin-
Dutch researchers found that the application of a mixture of
resolving agents with similar characteristics may lead to more effi-
cient enantiomeric separations than the sole application of a given
resolving agent (Dutch resolution).10 The resolution efficiency can
also be improved by using achiral additives with a similar structure
to either the racemic compound or the resolving agent.11
Herein our aim was to extend the Dutch-resolution methodol-
ogy in the sphere of amino acids. We applied mixtures of amino
acids with non-basic side chain and isoelectric point of c.a. 6 as
dane was successfully carried out with half-equivalent of (S)-
aspartic acid. From the corresponding diastereomeric salt, the
(R)-1-aminoindane was separated; this compound is a useful inter-
mediate of rasagiline.3 Amino acids with a basic side chain can also
be employed for the resolution of racemic acids. (S)-Lysine was
⇑
Corresponding author. Tel.: +36 1 4631883; fax: +36 1 4633648.
E-mail address: efogassy@mail.bme.hu (E. Fogassy).
http://dx.doi.org/10.1016/j.tetasy.2015.05.010
0957-4166/Ó 2015 Published by Elsevier Ltd.

722
Z. Szeleczky et al. / Tetrahedron: Asymmetry 26 (2015) 721–731
diastereomer, x is the molar fraction, T is the melting point of a mix-
ture with a molar fraction of x, R the ideal gas constant, and i refers
to the corresponding pure diastereomer.
the resolving agent. (S)-Phenylalanine (S)-2 and (S)-3-(amino-
methyl)-5-methylhexanoic acid [(S)-pregabalin] (S)-4 were chosen
for a- and c-amino acid, respectively. The aspartame (L-aspartyl-L-
The melting point and the enthalpy of fusion of the pure
diastereomers (S)-1ꢀ(S)-2, (R)-1ꢀ(S)-2, (S)-1ꢀ(S)-4 and (R)-1ꢀ(S)-4
were determined by DSC analysis, but the corresponding diastere-
omeric mixtures were not investigated. The results are summa-
rized in Table 1. It is noteworthy that the (S)-1ꢀ(S,S)-3 and
(R)-1ꢀ(S,S)-3 diastereomeric salts decomposed before melting.
Based on the data shown in Table 1, the corresponding binary
melting point diagrams were constructed using Eq. 3 (Fig. 2). The
eutectic compositions were found to be 0.275 and 0.222 for the
(S)-1ꢀ(S)-2/(R)-1ꢀ(S)-2 and the (S)-1ꢀ(S)-4/(R)-1ꢀ(S)-4 systems,
respectively (Table 1). Thereafter, Eq. 2 was used to predict the
maximum resolving capability values (Fmax) for the individual
resolutions using the corresponding eutectic composition values.
For the resolution of mandelic acid 1 with (S)-Phenylalanine (S)-2
or with (S)-pregabalin (S)-4, the Fmax values were 0.63 and 0.70,
respectively (Table 1).
phenylalanine methyl ester) (S,S)-3, which incorporates a b-amino
acid moiety and (S)-alanine (S)-5 were also included in this study.
Mandelic acid 1 was chosen as the racemic compound. The choice
of this model racemic compound was justified by the fact that the
enantiomers of mandelic acid 1 have appropriate chiral recognition
capacity and are widely used resolving agents.12,7i We also wished
to investigate how amphoteric achiral additives, such as glycine 6,
b-alanine 7 and c-aminobutyric acid 8 with similar structures to
the corresponding resolving agent (S)-2–(S)-4 influence the overall
efficiency of the resolution (Fig. 1).
α
COOH
H2N
COOH
α
NH2
(S)-2
6
Table 1
The melting point and the enthalpy of fusion of the diastereomers formed in the
reaction of mandelic acid 1 and (S)-phenylalanine (S)-2, aspartame (S,S)-3 or (S)-
pregabalin (S)-4
OH
O
β
α
COOH
β
H2N
COOH
α
HOOC
N
H
COOMe
a
Diastereomer Melting point
D
(kJ)
H
Eutectic composition
(—)
Fmax
(—)
NH2
(°C)
7
1
(S,S)-3
(S)-1ꢀ(S)-2
(R)-1ꢀ(S)-2
(S)-1ꢀ(S,S)-3
(R)-1ꢀ(S,S)-3
(S)-1ꢀ(S)-4
(R)-1ꢀ(S)-4.
453
430
Decomposition
Decomposition
411
53.6
31.4
—
—
54.4
31.7
0.275
0.63
α
β
γ
α
—
—
0.222
—
—
0.70
COOH
H2N
COOH
β
γ
NH2
386
(S)-4
8
a
Fmax was calculated using Eq. 2.
α
COOH
NH2
Wynberg et al. proposed another method for predicting the
resolving capability. According to their method (Eq. 4), the
expected resolving capability value (Fexp) can be calculated from
the solubility of the diastereomeric salts.14
(S)-5
Figure 1. Mandelic acid 1, the resolving agents (S)-2–(S)-5 and the achiral additives
6–8 used herein.
Fexp ¼ ðkn ꢁ kpÞ=kn
ð4Þ
where kn and kp are the solubility of the more and the less soluble
salt, respectively.
2. Results and discussion
In order to calculate the expected resolving capability values
(Fexp), the individual diastereomers were prepared by reacting
(R)- or (S)-mandelic acid (R)- or (S)-1 with one equivalent of the
given resolving agent (S)-2, (S,S)-3 or (S)-4 in ethanol. After the
complete removal of the solvent, the corresponding diastereomers
were obtained. The solubility of the given diastereomers was mea-
sured in water at 26 °C, and the expected resolving capability val-
ues (Fexp) were calculated using Eq. 4. The results are summarized
in Table 2. Based on these results, the most efficient enantiomeric
separation of mandelic acid 1 may be expected with (S)-pregabalin
(S)-4 as the resolving agent (Table 2, entry 3).
2.1. Theoretical and experimental values used to describe the
resolution efficiency
The coefficient indicating the efficiency of a particular resolu-
tion is the F-factor, that is, the resolving capability, which is the
product of the optical purity and the yield of the enantiomeric mix-
ture isolated from the corresponding diastereomeric salt (Eq. 1).
F ¼ Yield=100 ꢀ op=100
ð1Þ
Among the methods known for predicting of the efficiency of a
given resolution, one method requires the eutectic composition of
the corresponding diastereomeric salts (xeu) as an input to calcu-
late the maximum resolving capability value (Fmax) (Eq. 2).
2.2. The resolution of mandelic acid 1 with (S)-phenylalanine
(S)-2, aspartame (S,S)-3 or (S)-pregabalin (S)-4
Fmax ¼ ð1 ꢁ 2 ꢀ xeuÞ=ð1 ꢁ xeu
Þ
ð2Þ
After predicting the resolving capability values by different
methods (Eqs. (2)–(4)), we compared these theoretical results to
the experimental resolving capability values obtained with the
application of the corresponding single resolving agents (S)-2–
(S)-4. Therefore the resolution of mandelic acid 1 was carried out
with (S)-phenylalanine (S)-2 or aspartame (S,S)-3 or (S)-pregabalin
(S)-4 in water according to Pasteur’s equivalent,15 the half-equiva-
lent and Pope–Peachey’s half-equivalent method.16 In the latter
instance, Na2CO3 served as an achiral additive.
The eutectic composition of a diastereomeric salt can be deter-
mined by differential scanning calorimetric (DSC) analysis.13 The
liquidus curve of a binary melting point diagram can be calculated
using the simplified Schröeder-van Laar equation (Eq. 3).
lnx ¼
D
Hi=R ꢀ ð1=Ti ꢁ 1=TÞ
ð3Þ
where
D
Hi is the enthalpy of fusion of the corresponding pure
diastereomer, Ti is the melting point of the corresponding pure

Z. Szeleczky et al. / Tetrahedron: Asymmetry 26 (2015) 721–731
723
460
450
440
430
420
410
460
450
440
430
420
410
420
410
400
390
380
370
420
410
400
390
380
370
xeu:0.275
xeu:0.222
0.0 0.2 0.4 0.6 0.8 1.0
x
0.0 0.2 0.4 0.6 0.8 1.0
x
(a)
(b)
Figure 2. The binary melting point diagram of the (S)-1ꢀ(S)-2/(R)-1ꢀ(S)-2 (a) and the (S)-1ꢀ(S)-4/(R)-1ꢀ(S)-4 systems (b).
Table 2
Solubility values of diastereomers obtained in the reaction of (R)- or (S)-1 and the
corresponding resolving agent (S)-2–(S)-4
aspartame (S,S)-3 or (S)-pregabalin (S)-4, we applied the mixtures
of the corresponding resolving agents (S)-2, (S,S)-3 and (S)-4 for the
enantiomeric separation of 1 according to Pasteur’s equivalent, the
half-equivalent and Pope–Peachey’s half-equivalent methods.
In order to fulfil our aim of investigating non-linear interactions
that may occur when resolving agent mixtures are used, a simple
model was implemented which was used previously by our research
group.17 This model comprises of two equations. Eq. 5 was used to
compute the calculated resolving capability value (cF) by weighing
the individual resolution efficiency values obtained with the sole
resolving agents (F) with the molar ratio (n) of the corresponding
resolving agent in the given resolving agent mixture.
a
b
c
Entry
1
Resolving agent
kp (molꢀdmꢁ3
)
kn (molꢀdmꢁ3
)
Fexp
(S)-2
0.0691
0.1599
0.57
0.39
0.73
(S)-1ꢀ(S)-2
(R)-1ꢀ(S)-2
2
3
(S,S)-3
(S)-4
0.0810
0.1328
(S)-1ꢀ(S,S)-3
0.1437
(R)-1ꢀ(S,S)-3
0.5360
(S)-1ꢀ(S)-4
(R)-1ꢀ(S)-4
a
b
c
kp is the solubility of the less soluble diastereomeric salt in water.
kn is the solubility of the more soluble diastereomeric salt in water.
Fexp was calculated using Eq. 4.
In separate experiments, racemic mandelic acid 1 was mixed
cF ¼ nðSÞꢁ2 ꢀ FðSÞꢁ2 þ nðS;SÞꢁ3 ꢀ FðS;SÞꢁ3 þ nðSÞꢁ4 ꢀ F
where nx is the molar ratio of component x in the resolving agent
mixture and Fx is the resolving capability value obtained in the sin-
gle resolution with component x.
Then the quotient (iF) of the measured resolving capability (mF)
and the calculated resolving capability (cF) was computed with Eq.
6. The iF value obtained can be used to quantify a potential non-
linear interaction in the resolving agent mixture. When iF >1, a
positive, synergistic effect is observed. Analogously, an iF value
lower than 1 indicates a negative interaction between the resolving
agents. In Sections 2.4–2.7, the cF and the iF values are used to
describe the non-linear interactions among the corresponding
resolving agents.
ð5Þ
ðSÞꢁ4
with 1.0 equiv (Pasteur’s equivalent method) or 0.5 equiv of the cor-
responding resolving agent (S)-2, (S,S)-3 or (S)-4 (half-equivalent
method), or with 0.5 equiv of the given resolving agent (S)-2, (S,S)-
3 or (S)-4 and 0.25 equiv of Na2CO3 (Pope–Peachey’s half-equivalent
method). The mixture was then dissolved in hot water. The corre-
sponding diastereomeric salt, which appeared upon gradually cool-
ing down the reaction mixture to 26 °C, was filtered off after 15 min
of crystallization. In order to recover the enantiomeric mixture of
mandelic acid 1, the corresponding diastereomer was treated with
hydrochloric acid followed by extraction with ethyl acetate
(Scheme 1).
Table 3 summarizes the experimental and the predicted resolv-
ing capability values. Based on Fexp, the efficiency of the resolving
agents decreases in the following order: (S)-pregabalin (S)-4, (S)-
phenylalanine (S)-2 and aspartame (S,S)-3. The same relative order
between (S)-2 and (S)-4 was also confirmed by Fmax values calcu-
lated with Eq. 2. Thus according to the predicted resolving capabil-
ity values (Fexp and Fmax), (S)-pregabalin (S)-4 seemed to be the
most suitable resolving agent for the enantiomeric separation of
mandelic acid 1 (Table 3).
iF ¼ mF=cF
ð6Þ
where mF is the measured resolving capability and cF is the calcu-
lated resolving capability.
2.4. Resolution of mandelic acid
1 with mixtures of (S)-
phenylalanine (S)-2, aspartame (S,S)-3 and (S)-pregabalin (S)-4
according to Pasteur’s equivalent method
Using Pasteur’s equivalent and Pope–Peachy’s half-equivalent
methods, the highest resolving capability value (mF) was obtained
with (S)-pregabalin (S)-4. However, (S)-phenylalanine (S)-2 proved
to be the best resolving agent in the case of the half-equivalent
method. It is noteworthy that in all but one instance, the experi-
mental resolving capability values (mF) did not reach the corre-
sponding theoretical results (Fexp and Fmax). The theoretical
resolving capability values (Fexp and Fmax) can only be reached by
setting the optimal resolution parameters and the thorough opti-
mization of these resolutions was not the main aim of this study.
First, the resolution of mandelic acid 1 was carried out with a
resolving agent mixture containing (S)-2, (S,S)-3 or (S)-4 using
Pasteur’s equivalent method (Scheme 2). In these resolution
experiments, the corresponding resolving agents (S)-2, (S,S)-3
or (S)-4 were mixed to obtain a total amount of 1.0 equiv of
the resolving agent mixture (Table 4, entries 4–8), and the reso-
lution of mandelic acid 1 was carried out similarly to the proce-
dure described in Section 2.2. The results are summarized in
Table 4, and the results obtained applying 1.0 equiv of the corre-
sponding sole resolving agent (S)-2–(S)-4 are also shown
(Table 4, entries 1–3).
2.3. The theoretical model describing the non-linear coherence
in the resolving agent mixtures
Applying the resolving agent mixtures containing two resolving
agents in a 1:1 ratio, the resolving capability values (mF) were in
the range of 0.03–0.37 (Table 4, entries 4–6), and the mixture of
After attempting the resolution of mandelic acid 1 with the sole
amphoteric resolving agents, such as the (S)-Phenylalanine (S)-2,

724
Z. Szeleczky et al. / Tetrahedron: Asymmetry 26 (2015) 721–731
Pasteur's equivalent
method
1.0 eq. (S)-2
or
1.0 eq. (S,S)-3
or
1.0 eq. (S)-4
or
half-equivalent
method
OH
COOH
1. water
0.5 eq. (S)-2
+
crystallization
or
DIASTEREOMERIC SALT + MOTHER LIQUOR
2. filtration
1
0.5 eq. (S,S)-3
or
1. HCl
2. extraction
0.5 eq. (S)-4
or
OH
Pope-Peachey's
half-equivalent method
COOH
0.5 eq. (S)-2 + 0.25 Na2CO3
(S)-1
or
enantiomeric mixture
0.5 eq. (S,S)-3 + 0.25 Na2CO3
or
0.5 eq. (S)-4 + 0.25 Na2CO3
Scheme 1. The resolution of mandelic acid 1 with (S)-phenylalanine (S)-2, aspartame (S,S)-3 or (S)-pregabalin (S)-4.
Table 3
Theoretical (Fexp and Fmax) and experimental resolving capability values (mF) obtained in the resolution of mandelic acid 1 with (S)-phenylalanine (S)-2, aspartame (S,S)-3 and (S)-
pregabalin (S)-4
Entry
Resolving agent
Experimental Fa (mF)
Calculated F
Pasteurb
Half-eq.c
Pope–Peacheyd
Fmax
Fexp
e
f
1
2
3
(S)-2
(S,S)-3
(S)-4
0.36
0.39
0.44
0.34
0.11
0.22
0.25
0.11
0.35
0.63
n.d.
0.70
0.57
0.39
0.73
a
b
c
d
e
f
F was calculated using Eq. 1.
See Table 4, entries 1–3 for details.
See Table 5 entries 1–3 for details.
See Table 6 entries 1–3 for details.
Fmax was calculated using Eq. 2. See Table 1.
Fexp was calculated using Eq. 4. See Table 2.
OH
Pasteur's equivalent method
1.0 eq. resolving agent mixture of
(S)-2, (S,S)-3 and (S)-4
1. water
crystallization
COOH
+
DIASTEREOMERIC SALT + MOTHER LIQUOR
2. filtration
1
1. HCl
2. extraction
OH
COOH
(S)-1
enantiomeric mixture
Scheme 2. Resolution of mandelic acid 1 with mixtures of (S)-phenylalanine (S)-2, aspartame (S,S)-3 and (S)-pregabalin (S)-4 according to Pasteur’s equivalent method.

Z. Szeleczky et al. / Tetrahedron: Asymmetry 26 (2015) 721–731
725
Table 4
Resolution of mandelic acid 1 with mixtures of (S)-phenylalanine (S)-2, aspartame (S,S)-3 and (S)-pregabalin (S)-4 according to Pasteur’s equivalent method
Entry
Composition of mixture of resolving agents
(referring to 1 mol mandelic acid)
Measured values
Calculated values
(S)-2 (mol)
(S,S)-3 (mol)
(S)-4 (mol)
opa (%)
Yb (%)
mFc (—)
cFd (—) iFe (—)
1
2
3
4
5
6
7
8
1
0
0
0.5
0
0.5
0.33
0.1
0
1
0
0.5
0.5
0
0.33
0.9
0
0
1
0
0.5
0.5
0.33
0
23
33
76
28
5
48
32
43
158
118
58
116
58
78
91
95
0.36
0.39
0.44
0.33
0.03
0.37
0.29
0.41
—
—
—
0.38
0.42
0.40
0.40
0.39
—
—
—
0.88
0.07
0.93
0.73
1.06
a
b
c
The optical purity was based on the specific rotation.
The yield was based on the half of the racemic mandelic acid 1 which was regarded to be 100% for each antipode.
mF was calculated using Eq. 1.
cF was calculated using Eq. 5.
iF was calculated using Eq. 6.
d
e
(S)-2 and (S)-4 proved to be the most efficient two-component
resolving agent combination leading to an mF value of 0.37
(Table 4, entry 6). The resolving agent mixture containing (S)-2,
(S,S)-3 and (S)-4 in equal quantities led to resolving capability
(mF) of 0.29 (Table 4, entry 7). In all of the above-mentioned
instances, the calculated resolving capability values (cF) were
higher than the corresponding measured values (mF), and the
given iF values were in the range of 0.07–0.93 (Table 4, entries
4–7). It is noteworthy that the iF value obtained with a 1:1 mixture
of (S,S)-3 and (S)-4 was 0.07 indicating a strong negative interac-
tion between the given resolving agents (Table 4, entry 5). The
application of a 9:1 mixture of (S,S)-3 and (S)-2 led to a higher
resolving capability value (mF) than the corresponding 1:1 mixture
(compare Table 4, entries 4 and 8). Moreover, applying (S,S)-3 and
(S)-2 in a 9:1 ratio, a weak but positive interaction (iF = 1.06) was
observed between the resolving agents, whereas only negative
interactions were found in the case of the other resolving agent
combinations (compare Table 4, entries 4–8).
Applying half-equivalent of the given resolving agent mixture,
the resolving capability values were in the range of 0.06–0.25
(Table 5, entries 4–7), which is somewhat lower than the results
obtained with one equivalent of the corresponding resolving agent
mixtures (compare Table 4, entries 4–8 and Table 5, entries 4–7). In
the case of the half equivalent method, the application of (S)-2 and
(S,S)-3 in a 1:1 ratio proved to be the most efficient (mF = 0.25), and
a weak synergetic effect (iF = 1.11) was also found between (S)-2
and (S,S)-3 (Table 5, entry 4). It is noteworthy that only a negative
interaction could be observed in the cases of the other resolving
agent combinations and the corresponding iF values were in the
range of 0.36–0.64 (Table 5, entries 5–7). Similarly to the equivalent
method, the mF and iF values were the lowest when a 1:1 mixture
of (S,S)-3 and (S)-4 was used (Table 5, entry 5). Using the equivalent
method, the highest resolving capability could be obtained with a
mixture of (S)-2 and (S)-4 (Table 4, entry 6). In case of the half
equivalent method, the application of the same resolving agent
combination [(S)-2 and (S)-4] led to the optical purity of 1
(op = 47%) but the yield remained 38% (Table 5, entry 6).
2.5. Resolution of mandelic acid 1 with mixtures of (S)-
phenylalanine (S)-2, aspartame (S,S)-3 and (S)-pregabalin (S)-4
according to the half-equivalent method
2.6. Resolution of mandelic acid
phenylalanine (S)-2, aspartame (S,S)-3 and (S)-pregabalin (S)-4
1 with mixtures of (S)-
according to Pope–Peachey’s half-equivalent method
The resolution of mandelic acid 1 was also carried out using
half-equivalent of the corresponding resolving agent mixtures con-
taining (S)-2, (S,S)-3 and (S)-4 analogously to the procedure
described in Section 2.4 (Scheme 3). The results are summarized
in Table 5, which also contains the results obtained with half-
equivalent of the corresponding sole resolving agent (S)-2–(S)-4
(Table 5, entries 1–3).
The resolving agent mixtures of (S)-phenylalanine (S)-2,
aspartame (S,S)-3 and (S)-pregabalin (S)-4 were also applied for
the resolution of mandelic acid 1 using Pope–Peachey’s half-
equivalent method. In these experiments, the corresponding
resolving agents (S)-2, (S,S)-3 or (S)-4 were mixed to obtain a total
amount of half-equivalent (Table 6, entries 4–7) and the resolution
OH
half-equivalent method
1. water
crystallization
COOH
+
DIASTEREOMERIC SALT + MOTHER LIQUOR
0.5 eq. resolving agent mixture of
2. filtration
(S)-2, (S,S)-3 and (S)-4
1
1. HCl
2. extraction
OH
COOH
(S)-1
enantiomeric mixture
Scheme 3. Resolution of mandelic acid 1 with mixtures of (S)-phenylalanine (S)-2, aspartame (S,S)-3 and (S)-pregabalin (S)-4 according to the half-equivalent method.

726
Z. Szeleczky et al. / Tetrahedron: Asymmetry 26 (2015) 721–731
Table 5
Resolution of mandelic acid 1 with mixtures of (S)-phenylalanine (S)-2, aspartame (S,S)-3 and (S)-pregabalin (S)-4 according to the half-equivalent method
Entry
Composition of mixture of resolving agents
(referring to 1 mol mandelic acid)
Measured values
Calculated values
(S)-2 (mol)
(S,S)-3 (mol)
(S)-4 (mol)
opa (%)
Yb (%)
mFc (—)
cFd (—) iFe (—)
1
2
3
4
5
6
7
0.5
0
0
0.25
0
0.25
0.17
0
0.5
0
0.25
0.25
0
0
0
35
11
42
38
17
47
37
96
97
53
66
34
38
36
0.34
0.11
0.22
0.25
0.06
0.18
0.13
—
—
—
0.23
0.17
0.28
0.22
—
—
—
1.11
0.36
0.64
0.58
0.5
0
0.25
0.25
0.17
0.17
a
b
c
The optical purity was based on the specific rotation.
The yield was based on the half of the racemic mandelic acid 1 which was regarded to be 100% for each antipode.
mF was calculated using Eq. 1.
cF was calculated using Eq. 5.
iF was calculated using Eq. 6.
d
e
Table 6
Resolution of mandelic acid 1 with mixtures of (S)-phenylalanine (S)-2, aspartame (S,S)-3 and (S)-pregabalin (S)-4 according to Pope–Peachey’s half-equivalent method
Entry
Composition of mixture of resolving agents
(referring to 1 mol mandelic acid)
Measured values
Calculated values
(S)-2 (mol)
(S,S)-3 (mol)
(S)-4 (mol)
opa (%)
Yb (%)
mFc (—)
cFd (—) iFe (—)
1
2
3
4
5
6
7
0.5
0
0
0.25
0
0.25
0.17
0
0.5
0
0.25
0.25
0
0
0
63
9
40
126
64
66
30
0.25
0.11
0.35
0.28
0.15
0.26
0.16
—
—
—
0.18
0.23
0.30
0.24
—
—
—
1.55
0.65
0.74
0.68
0.5
0
0.25
0.25
0.17
55
43
51
49
31
53
51
0.17
a
b
c
The optical purity was based on the specific rotation.
The yield was based on the half of the racemic mandelic acid 1 which was regarded to be 100% for each antipode.
mF was calculated using Eq. 1.
cF was calculated using Eq. 5.
iF was calculated using Eq. 6.
d
e
of mandelic acid 1 was carried out as described in Section 2.2
applying 0.25 equiv of Na2CO3 as the achiral base (Scheme 4).
The results are summarized in Table 6, which also contains the
results obtained with the corresponding sole resolving agents
using Pope–Peachey’s methodology (Table 6, entries 1–3).
The resolving capability values were in the range of 0.15–0.28
using resolving agent mixtures of (S)-2, (S,S)-3 or (S)-4 for the enan-
tiomeric separation of 1 according to Pope–Peachey’s half-equivalent
method (Table 6, entries 4–7). The 1:1 mixture of (S,S)-3 and (S)-4 or
that of (S)-2 and (S)-4 led to higher resolving capability values using
Pope–Peachey’s half-equivalent method than the half-equivalent
method (compare Table 5, entries 5 and 6 and Table 6, entries 5 and
6). The results obtained with the other resolving agent combinations
showed parity in case of both aforementioned methods (compare
Table 5, entries 4 and 7 with Table 6, entries 4 and 7).
2.7. Resolution of mandelic acid 1 with resolving agent mixture
of (S)-phenylalanine (S)-2 and (S)-alanine (S)-5 according to
Pasteur’s equivalent method
Based on the thorough investigation of the resolving agent mix-
tures of (S)-2, (S,S)-3 or (S)-4 detailed in Sections 2.4–2.6, a
synergetic effect could be found between the (S)-phenylalanine
(S)-2 and aspartame (S,S)-3 regardless of the method used
(Table 4, entry 8; Table 5, entry 4 and Table 6, entry 4). However,
only negative interactions were observed for the other resolving
agent combinations.
It was assumed that the structural similarity might be the
underlying phenomenon for the positive interaction between
(S)-2 and (S,S)-3. Therefore, we decided to carry out the resolution
of mandelic acid 1 with (S)-alanine (S)-5, which is another struc-
tural analogue of (S)-phenylalanine (S)-2. The resolution of 1 was
attempted with one equivalent of (S)-alanine (S)-5 similarly to
the procedure described in Section 2.2, but no enantiomeric sepa-
ration was observed (Table 7, entry 2) (Scheme 5).
Thereafter, the resolving agent mixtures containing (S)-
phenylalanine (S)-2 and (S)-alanine (S)-5 were also applied for the
enantiomeric separation of mandelic acid 1. The equivalent method
wasused inthiscase, since thismethodwasmoreefficient inthecase
of the resolving agent mixtures of (S)-2 and (S,S)-3 than the half-
equivalent or Pope–Peachey’s half-equivalent method. The results
are summarized in Table 7 and the resolving capability (mF) values
obtained with different mixtures of (S)-phenylalanine (S)-2 and
(S)-alanine (S)-5 (Table 7, entries 3–7) are also shown in Figure 3.
Using Pope–Peachey’s half-equivalent method, the highest
resolving capability value (mF = 0.28) was obtained with a 1:1
mixture of (S)-2 and (S,S)-3. Moreover, a synergetic effect could
also be observed between the corresponding resolving agents (S)-
2 and (S,S)-3 leading to an iF value of 1.55 (Table 6, entry 4). In
the other instances, the iF values were in the range of 0.65–0.74
indicating
a negative interaction between the corresponding
resolving agents (S)-2, (S,S)-3 or (S)-4 (Table 6, entries 5–7). The
lowest iF value (iF = 0.65) was obtained when (S,S)-3 and (S)-4
were used in a 1:1 ratio (Table 6, entry 5), which is in accordance
with the observations made in the case of the equivalent and half-
equivalent methods (compare Table 4, entry 5; Table 5, entry 5 and
Table 6, entry 5).

Z. Szeleczky et al. / Tetrahedron: Asymmetry 26 (2015) 721–731
727
OH
1. water
Pope-Peachey's half-equivalent method
0.5 eq. resolving agent mixture of
crystallization
COOH
+
DIASTEREOMERIC SALT + MOTHER LIQUOR
2. filtration
(S)-2, (S,S)-3 and (S)-4
+
1
1. HCl
0.25 eq. Na2CO3
2. extraction
OH
COOH
(S)-1
enantiomeric mixture
Scheme 4. Resolution of mandelic acid 1 with mixtures of (S)-phenylalanine (S)-2, aspartame (S,S)-3 and (S)-pregabalin (S)-4 according to Pope-Peachey’s half-equivalent
method.
Table 7
Resolution of mandelic acid 1 with resolving agent mixture of (S)-phenylalanine (S)-2 and (S)-alanine (S)-5 according to Pasteur’s equivalent method
Entry
Composition of mixture of resolving
agents (referring to 1 mol mandelic
acid)
Measured values
Calculated values
(S)-2 (mol)
(S)-5 (mol)
opa (%)
Yb (%)
mFc (—)
cFd (—)
iFe (—)
1f
2
3
4
5
6
7
1
0
0
1
23
0
158
0
26
55
91
0.36
0
—
—
—
—
0.18
0.35
0.5
0.65
0.82
0.82
0.65
0.5
0.35
0.18
53
59
45
40
28
0.14
0.32
0.41
0.48
0.42
0.06
0.13
0.18
0.23
0.30
2.22
2.54
2.28
2.05
1.41
119
150
a
b
c
d
e
f
The optical purity was based on the specific rotation.
The yield was based on the half of the racemic mandelic acid 1 which was regarded to be 100% for each antipode.
mF was calculated using Eq. 1.
cF was calculated using Eq. 5.
iF was calculated using Eq. 6.
Table 4, entry 1.
OH
1. water
Pasteur's equivalent method
crystallization
COOH
1.0 eq. resolving agent mixture of
+
DIASTEREOMERIC SALT + MOTHER LIQUOR
(S)-2 and (S)-5
2. filtration
1
1. HCl
2. extraction
OH
COOH
(S)-1
enantiomeric mixture
Scheme 5. Resolution of mandelic acid 1 with mixtures of (S)-phenylalanine (S)-2 and (S)-alanine (S)-5 according to Pasteur’s equivalent method.
In Figure 3, the dashed line is the corresponding expected
2.8. Resolution of mandelic acid 1 with (S)-phenylalanine (S)-2,
aspartame (S,S)-3 and (S)-pregabalin (S)-4 in the presence of
achiral additives 6–8
resolving capability value (cF), which also means the lack of syner-
gistic effect. Figure 3 shows that the application of every mixture of
(S)-2 and (S)-5 afforded higher measured resolving capability (mF)
than the theoretical values, as also indicated by the iF values that
were in the range of 1.41–2.54 (Table 7, entries 3–7). The highest
resolving capability value (mF = 0.48) was obtained by applying a
mixture of (S)-2 and (S)-5 in a ratio of 0.65:0.35, which result is sig-
nificantly higher than the mF obtained with one equivalent of (S)-2
(compare Table 7, entries 1 and 6). The highest optical purity value
(op = 59%) was achieved by a 0.35:0.65 mixture of (S)-2 and (S)-5
(Table 7, entry 4).
We were also interested in how structurally similar achiral
additives with an amphoteric character 6–8 influence the resolu-
tion of mandelic acid 1 with (S)-phenylalanine (S)-2, aspartame
(S,S)-3 and (S)-pregabalin (S)-4. Glycine 6, b-alanine 7 and
c-
aminobutyric acid 8 were chosen as the structural analogues of
the corresponding resolving agents (S)-2–(S)-4. During the resolu-
tion experiments, the given resolving agent and the achiral addi-
tive [(S)-2 and 6; (S,S)-3 and 7 or (S)-4 and 8] were mixed in a

728
Z. Szeleczky et al. / Tetrahedron: Asymmetry 26 (2015) 721–731
7 led to an even more significant increase both in the optical purity
op = 38%) and the resolving capability ( F = 0.16) compared to
the sole application of half-equivalent of (S,S)-3 (compare
Table 8, entries 3 and 4). The combined application of (S)-4 and 8
led to a 9% of increase in the optical purity, and the mF value
was increased from 0.22 to 0.34 (compare Table 8, entries 5 and 6).
1.0
0.8
0.6
0.4
0.2
0.0
(D
D
3. Conclusion
Herein we have shown that the enantiomers of amino acids or
amino acid derivatives, such as (S)-phenylalanine (S)-2, aspartame
(S,S)-3 and (S)-pregabalin (S)-4 can be used for the partial enan-
tiomeric separation of mandelic acid 1. The mixtures of (S)-2,
(S,S)-3 or (S)-4 were also systematically applied to the enan-
tiomeric separation of mandelic acid 1 and a synergetic effect
was observed between the corresponding resolving agents (S)-2,
(S,S)-3, (S)-4 in a few instances. A significant improvement in the
efficiency of the resolution was observed when the resolution of
mandelic acid 1 was accomplished with a mixture of the struc-
turally related (S)-phenylalanine (S)-2 and (S)-alanine (S)-5, or
with (S)-2, (S,S)-3 or (S)-4 along with the corresponding struc-
turally similar achiral additives 6–8.
However these current methods are not as efficient as the best
resolving agents for the enantiopure mandelic acid, but we have
emphasized that amino acids could be applicable resolving agents
solely or in combinations. Moreover our results underline the
importance of the structural similarity during the selection of the
best resolving agent or resolving agent combination among
amphoteric resolving agents.
0.0
0.2
0.4
0.6
0.8
1.0
x mol (S)-
2
/ mol ((S)-
2
+(S)-5)
Figure 3. The resolving capability values obtained in the resolution of mandelic acid
1 with resolving agent mixtures of (S)-phenylalanine (S)-2 and (S)-alanine (S)-5.
1:1 ratio to obtain a total amount of one equivalent of mixture and
the resolution was carried out similarly to the procedure described
in Section 2.2 (Scheme 6). The results are summarized in Table 8,
which also contains the results obtained with the corresponding
half-equivalent of resolving agents (S)-2–(S)-4.
Applying (S)-2 along with 6 for the resolution of mandelic acid
1, a 15% increase could be observed in the optical purity of 1,
whereas the corresponding resolving capability increased by 0.02
compared to resolution with half-equivalent of (S)-2 (compare
Table 8, entries 1 and 2). In the case of (S,S)-3, the application of
mixture of 0.5 eq. of (S)-2 and 0.5 eq. of 6
OH
COOH
1. water
or
crystallization
mixture of 0.5 eq. of (S,S)-3 and 0.5 eq. of 7
DIASTEREOMERIC SALT + MOTHER LIQUOR
2. filtration
+
or
1. HCl
mixture of 0.5 eq. of (S)-4 and 0.5 eq. of 8
1
2. extraction
OH
COOH
(S)-1
enantiomeric mixture
Scheme 6. Resolution of mandelic acid 1 with (S)-phenylalanine (S)-2, aspartame (S,S)-3 and (S)-pregabalin (S)-4 in the presence of achiral additives 6–8.
Table 8
Resolution of mandelic acid 1 with (S)-phenylalanine (S)-2, aspartame (S,S)-3 and (S)-pregabalin (S)-4 in the presence of achiral additives 6–8
Entry
Resolving agent
Achiral additive
Measured values
Differences
op (%) DF (—)
Referring to 0.5 mol mandelic acid
Referring to 0.5 mol mandelic acid
opa (%)
Yb (%)
mFc (—)
D
1d
2
(S)-2
(S)-2
(S,S)-3
(S,S)-3
(S)-4
—
6
—
7
35
50
11
49
42
51
96
71
97
55
53
66
0.34
0.36
0.11
0.27
0.22
0.34
—
15
—
38
—
9
—
0.02
—
0.16
—
0.12
3e
4
5f
6
(S)-4
8
a
b
c
d
e
f
The optical purity was based on the specific rotation.
The yield was based on the half of the racemic mandelic acid 1, which was regarded to be 100% for each antipode.
mF was calculated using Eq. 1.
See Table 5, entry 1.
See Table 5, entry 2.
See Table 5, entry 3.

Z. Szeleczky et al. / Tetrahedron: Asymmetry 26 (2015) 721–731
729
an op of 35% (Table 5, entry 1). [
The resolution of mandelic acid
a
]
25 = +53.2 (c 1, water); op: 35%.
4. Experimental
4.1. General
D
1 with half-equivalent of
aspartame [(S,S)-3] was carried out according to the representative
procedure C using 2.0 mL of water as the solvent, and 0.37 g (97%)
of (S)-1 was obtained with an op of 11% (Table 5, entry 2). The res-
olution of mandelic acid 1 with half-equivalent of (S)-pregabalin
(S)-4 was carried out according to the representative procedure C
using 2.0 mL of water as the solvent, and 0.20 g (53%) of (S)-1
was obtained with an op of 42% (Table 5, entry 3).
Optical rotations were determined on a Perkin–Elmer 241
polarimeter. Optical purity values (op) were calculated by compar-
ing the specific rotation of the enantiomeric mixture and the corre-
sponding pure enantiomer (the literature data: [a]
25 = +152.0 (c 1,
D
H2O)18,7i). The (S)-3-(aminomethyl)-5-methylhexanoic acid was
synthesized as described earlier.19 The racemic mandelic acid,
the (S)-phenylalanine, the aspartame, the (S)-alanine and the gly-
cine were purchased from Aldrich Chemical Co. The b-alanine
4.5. The resolution of mandelic acid 1 with half equivalent of
(S)-phenylalanine (S)-2 according to Pope–Peachey’s method
(Representative procedure D)
and
c-aminobutyric acid were purchased from Molar Chemical
Ltd. DSC curves were recorded with a Setaram DSC 131 and sam-
A mixture of 0.760 g (5.0 mmol) of racemic mandelic acid 1,
0.410 g (2.5 mmol) of (S)-phenylalanine (S)-2 and 0.133 g Na2CO3
(1.25 mmol) was dissolved in 9.5 mL of hot water. The crystalline
diastereomeric salt, which appeared upon gradually cooling down
the solution to 25 °C was filtered off after 15 min of crystallization
time. The crystals were washed with 1.0 mL of cold water, and then
dried to afford 0.41 g (52%) of the diastereomer (S)-1ꢀ(S)-2. The
decomposition of the diastereomer was accomplished according
to the procedure described in Section 4.3 to afford 0.15 g (40%) of
ples of 1–3 mg were run in hermetically-sealed aluminium pans
with a heating rate of 10 K minꢁ1
.
4.2. Preparation of pure diastereomeric salts incorporating (S)-
or (R)-mandelic acid (S)- or (R)-1 and (S)-phenylalanine (S)-2,
aspartame (S,S)-3 or (S)-pregabalin (S)-4 (Representative
procedure A)
The mixture of 0.077 g (0.5 mmol) of (S)-mandelic acid (S)-1
and 0.080 g (0.5 mmol) of (S)-phenylalanine (S)-2 was dissolved
in 20.0 mL of ethanol and concentrated to give 0.156 g of the
(S)-1ꢀ(S)-2 diastereomeric salt.
(S)-1 with an op of 63% (Table 6, entry 1). [a]
25 = +95.7 (c 1, water);
D
op: 63%. The resolution of mandelic acid 1 with half-equivalent of
aspartame (S,S)-3 and 0.25 equiv Na2CO3 was carried out according
to the representative procedure D using 2.0 mL of water as the sol-
vent, and 0.48 g (126%) of (S)-1 was obtained with an op of 9%
The diastereomeric salts listed in Table 1 were prepared accord-
ing to the representative procedure A from the corresponding pure
enantiomers (S)-1 or (R)-1 and (S)-2, (S,S)-3 or (S)-4.
(Table 6, entry 2). The resolution of mandelic acid
1 with
half-equivalent of (S)-pregabalin (S)-4 and 0.25 equiv Na2CO3 was
carried out according to the representative procedure D using
2.0 mL of water as the solvent, and 0.24 g (64%) of (S)-1 was
obtained with an op of 55% (Table 6, entry 3).
4.3. The resolution of mandelic acid 1 with one equivalent of
(S)-phenylalanine (S)-2 (Representative procedure B)
A mixture of 0.760 g (5.0 mmol) of racemic mandelic acid 1 and
0.820 g (5.0 mmol) of (S)-phenylalanine (S)-2 was dissolved in
8.0 mL of hot water. The crystalline diastereomeric salt which
appeared by gradually cooling down the reaction mixture to
26 °C, was separated from the mother liquor by filtration after
15 min. The crystals were washed with 1.0 mL of cold water, and
then dried to afford 1.34 g (170%) of the diastereomer (S)-1ꢀ(S)-2.
The diastereomeric salt was then dissolved in 1.0 mL of cc. HCl
and 4.0 mL of water, and the resulting mixture was extracted with
ethyl acetate. The combined organic layer was dried (MgSO4) and
concentrated to give 0.60 g (158%) of (S)-1 with an op of 23%
4.6. The resolution of mandelic acid 1 with half equivalent of
(S)-phenylalanine (S)-2 and half equivalent of glycine
6
(Representative procedure E)
The mixture of 0.760 g (5.0 mmol) of racemic mandelic acid 1,
0.410 g (2.5 mmol) of (S)-phenylalanine (S)-2 and 0.190 g
(2.5 mmol) of glycine 6 was dissolved in 4.0 mL of hot water. The
crystalline diastereomeric salt that appeared by gradually cooling
down the solution to 26 °C was filtered off after 15 min of crystal-
lization time. The crystals were washed with 1.0 mL of cold water,
and then they were dried to afford 0.68 g of the diastereomer salt.
The decomposition of the diastereomer was accomplished accord-
ing to the procedure described in Section 4.3 to afford 0.27 g (71%)
(Table 4, entry 1). [a]
25 = +34.9 (c 1, water); op: 23%. The resolution
D
of mandelic acid 1 with one equivalent of aspartame (S,S)-3 was
carried out according to the representative procedure B using
2.0 mL of water as the solvent, and 0.45 g (118%) of (S)-1 was
obtained with an op of 33% (Table 4, entry 2). The resolution of
mandelic acid 1 with one equivalent of (S)-pregabalin (S)-4 was
carried out according to the representative procedure B using
2.0 mL of water as the solvent and 0.22 g (58%) of (S)-1 was
obtained with an op of 76% (Table 4, entry 3).
of (S)-1 with an op of 50% (Table 8, entry 2). [a]
25 = +77.0 (c 1,
D
water); op: 50%. The resolution of mandelic acid 1 with half-equiv-
alent of aspartame (S,S)-3 and half-equivalent of b-alanine 7 was
carried out according to the representative procedure E using
2.0 mL of water as the solvent, and 0.21 g (55%) of (S)-1 was
obtained with an op of 49% (Table 8, entry 4). The resolution of
mandelic acid 1 with half-equivalent of (S)-pregabalin (S)-4 and
4.4. The resolution of mandelic acid 1 with half equivalent of
(S)-phenylalanine (S)-2 (Representative procedure C)
half-equivalent of c-aminobutyric acid 8 was carried out according
to the representative procedure E using 2.0 mL of water as the sol-
vent, and 0.25 g (66%) of (S)-1 was obtained with an op of 51%
(Table 8, entry 6).
A mixture of 0.760 g (5.0 mmol) of racemic mandelic acid 1 and
0.410 g (2.5 mmol) of (S)-phenylalanine (S)-2 was dissolved in
4.0 mL of hot water. The crystalline diastereomeric salt which
appeared upon gradually cooling down the solution to 26 °C was
filtered off after 15 min of crystallization time. The crystals were
washed with 1.0 mL of cold water, and then they were dried to
afford 0.79 g (100%) of the diastereomer (S)-1ꢀ(S)-2. The decompo-
sition of the diastereomer was accomplished according to the pro-
cedure described in Section 4.3 to afford 0.36 g (96%) of (S)-1 with
4.7. The resolution of mandelic acid 1 with one equivalent of
resolving agent mixture containing (S)-phenylalanine (S)-2 and
aspartame (S,S)-3 in a 1:1 ratio (Representative procedure F)
A mixture of 0.760 g (5.0 mmol) of racemic mandelic acid 1,
0.274 g (1.66 mmol) of (S)-phenylalanine (S)-2, 0.491 g
of (1.66 mmol) aspartame (S,S)-3 and 0.266 g (1.66 mmol) of

730
Z. Szeleczky et al. / Tetrahedron: Asymmetry 26 (2015) 721–731
(S)-pregabalin (S)-4 was dissolved in 4.0 mL of hot water. The crys-
talline diastereomeric salt which appeared upon gradually cooling
down the reaction mixture to 26 °C was separated from the mother
liquor by filtration after 15 min. The crystals were washed with
0.5 mL of cold water, and then were dried to afford 1.92 g of the
diastereomer salt. The decomposition of the diastereomer was
accomplished according to the procedure described in Section 4.3
to afford 0.35 g (91%) of (S)-1 with an op of 32% (Table 4, entry
gradually cooling down the reaction mixture to 26 °C was sepa-
rated from the mother liquor by filtration after 15 min. The crystals
were washed with 0.5 mL of cold water, and then they were dried
to afford 0.73 g of the diastereomer salt. The decomposition of the
diastereomer was accomplished according to the procedure
described in Section 4.3 to afford 0.21 g (55%) of (S)-1 with an op
of 59% (Table 7, entry 4). [a]
25 = +90.0 (c 1, water); op: 59%. The
D
resolution of 1 was also accomplished with the corresponding
mixture of (S)-phenylalanine (S)-2 and (S)-alanine (S)-5 according
to representative procedure I. The results are shown in Table 7,
entries 3–7.
7). [a]
25 = +49.5 (c 1, water); op: 32%. The resolution of 1 was also
D
accomplished with the corresponding resolving agent mixtures of
(S)-2, (S,S)-3 and (S)-4 according to the representative procedure
F. The results are shown in Table 4, entries 4-8.
Acknowledgements
4.8. The resolution of mandelic acid 1 with half equivalent of
resolving agent mixture containing (S)-phenylalanine (S)-2,
The above project was supported by the Hungarian Scientific
and Research Fund (OTKA K104769). The authors are grateful for
the support of Richter Gedeon PhD Scholarship.
aspartame (S,S)-3 and (S)-pregabalin (S)-4 in
a 1:1 ratio
(Representative procedure G)
A mixture of 0.760 g (5.0 mmol) of racemic mandelic acid 1,
0.137 g (0.83 mmol) of (S)-phenylalanine (S)-2, 0.245 g of
(0.83 mmol) aspartame (S,S)-3 and 0.133 g (0.83 mmol) of (S)-pre-
gabalin (S)-4 was dissolved in 4.0 mL of hot water. The crystalline
diastereomeric salt which appeared upon gradually cooling down
the reaction mixture to 26 °C was separated from the mother
liquor by filtration after 15 min. The crystals were washed with
0.5 mL of cold water, and then were dried to afford 0.40 g of the
diastereomer salt. The decomposition of the diastereomer was
accomplished according to the procedure described in Section 4.3
to afford 0.14 g (36%) of (S)-1 with an op of 37% (Table 5, entry
References
1. (a) Jacques, J. C.; Collet, A.; Willen, S. H. Enantiomers, Racemates, and
Resolutions; John Wiley: New York, 1981; (b) Wilen, S. H. Topics in
Stereochemistry; Wiley-Interscience: New York, 1972; (c) Sheldon, R. A.
Chirotechnology, Industrial Synthesis of Optically Active Compounds; Marcel
Dekker: New York, 1993; (d) Kozma, D. Optical Resolutions via Diastereomeric
Salt; CRC Press: London, 2002.
2. (a) Fogassy, E.; Nógrádi, M.; Pálovics, E.; Schindler, J. Synthesis 2005, 2005,
1555–1568; (b) Fogassy, E.; Nógrádi, M.; Kozma, D.; Egri, G.; Pálovics, E.; Kiss,
V. Org. Biomol. Chem. 2006, 4, 3011–3030; (c) Faigl, F.; Fogassy, E.; Nógrádi, M.;
Pálovics, E.; Schindler, J. Tetrahedron: Asymmetry 2008, 19, 519–536; (d)
Siedlecka, R. Tetrahedron 2013, 69, 6331–6363.
3. Dymácek, B. PCT 2011/053211 (Synthon BV) 2012, Chem. Abstr. 157:437939.
4. (a) Bálint, J.; Csatáriné Nagy, M.; Dombrády, Z.; Fogassy, E.; Gajáry, A.; Suba, C.
WO 2003000636 (Sanofi-Synthelabo) 2001, Chem. Abstr. 138:73080. (b)
Bousquet, A.; Musolino, A. WO 9918110 (Sanofi-Synthelabo) 2003, Chem.
Abstr. 130:296510.
5. (a) Pham, X.-H.; Kim, J.-M.; Chang, S.-M.; Kim, I.-H.; Kim, W.-S. J. Mol. Catal. B
Enzym. 2009, 60, 87–92; (b) Kim, J. M.; Chang, S. M.; Kim, I. H.; Pham, X.-H.;
Kim, W.-S. J. Nanosci. Nanotechnol. 2012, 12, 7139.
7). [a]
25 = +57.5 (c 1, water); op: 37%. The resolution of 1 was also
D
accomplished with the corresponding resolving agent mixtures of
(S)-2, (S,S)-3 and (S)-4 according to the representative procedure
G. The results are shown in Table 5, entries 4–7.
4.9. The resolution of mandelic acid 1 with half equivalent of
}
6. (a) Pálovics, E.; Szeleczky, Z.; Fodi, B.; Faigl, F.; Fogassy, E. RSC Adv. 2014, 4,
21254–21261; (b) Szeleczky, Z.; Bagi, P.; Pálovics, E.; Fogassy, E. Tetrahedron:
Asymmetry 2014, 25, 1095–1099; (c) Szeleczky, Z.; Bagi, P.; Pálovics, E.;
Fogassy, E. Tetrahedron: Asymmetry 2015, 26, 377–384; (d) Pálovics, E.;
Szeleczky, Z.; Bagi, P.; Faigl, F.; Fogassy, E. Per. Pol. Chem. Eng. 2015, 59, 26–37.
7. (a) Bergman, S. Arkiv Kemi 1926, 9, 1; (b) Baar, M. R.; Cerrone-Szakal, A. L. J.
Chem. Educ. 2005, 82, 1040–1043; (c) Ebbers, E. J.; Plum, B. J. M.; Ariaans, G. J.
A.; Kaptein, B.; Broxterman, Q. B.; Bruggink, A.; Zwanenburg, B. Tetrahedron:
Asymmetry 1997, 8, 4047; (d) Kesslin, G.; Kelly, K. W. US 4,322,548 (Kay-Fries)
1982, Chem. Abstr. 97:72090; (e) Dannenberg, W.; Schmand, H.: EP 518 (Riedel-
de Haen, A. -G.) 1979, Chem. Abstr. 92:6259. (f) Halmos, I. A.; Forsyth, C. S. US
4,239,912 (American Cyanamid) 1979, Chem. Abstr. 90:203672; (g) Ingersoll,
W.; Babcock, S. H.; Burns, F. B. J. Am. Chem. Soc. 1933, 55, 411–417; (h) Wanga,
P.; Zhanga, E.; Niub, J.-F.; Rena, Q.-H.; Zhaoa, P.; Liu, H.-M. Tetrahedron:
Asymmetry 2012, 23, 1046–1051; (i) Shitara, H.; Shintani, T.; Kodama, K.;
Hirose, T. J. Org. Chem. 2013, 78, 9309–9316; (j) Kobayashi, Y.; Kokubo, Y.;
Aisaka, T.; Saigo, K. Tetrahedron: Asymmetry 2008, 19, 2536–2541.
8. (a) Sun, Y.; Wan, X.; Wang, J.; Meng, Q.; Zhang, H.; Jiang, L.; Zhang, Z. Org. Lett.
2005, 7, 5425–5427; (b) Schmitt, E.; Schiffers, I.; Bolm, C. Tetrahedron Lett.
2009, 50, 3185–3188; (c) Yan, P.-C.; Xie, J.-H.; Zhang, X.-D.; Chen, K.; Li, Y.-Q.
Chem. Commun. 2014, 15987–15990.
9. (a) Lalonde, J. J.; Govardhan, C.; Khalaf, N.; Martinez, A. G.; Visuri, K.; Margolin,
A. L. J. Am. Chem. Soc. 1995, 117, 6845–6852; (b) Mateo, C.; Chmura, A.; Rustler,
S.; van Rantwijk, F.; Stolz, A.; Sheldon, R. A. Tetrahedron: Asymmetry 2006, 17,
320–323; (c) Cabrera, Z.; Gutarra, M. L. E.; Guisan, J. M.; Palomo, J. M. Catal.
Commun. 2010, 11, 964–967; (d) Gennaro, P. D.; Bernasconi, S.; Orsini, F.;
Corretto, E.; Shello, G. Tetrahedron: Asymmetry 2010, 21, 1885; (e) Chmura, A.;
Rustler, S.; Paravidino, M.; van Rantwijk, F.; Stolz, A.; Sheldon, R. A. Tetrahedron:
Asymmetry 2013, 24, 1225–1232.
resolving agent mixture containing (S)-phenylalanine (S)-2,
aspartame (S,S)-3 and (S)-pregabalin (S)-4 in
a 1:1 ratio
according to Pope–Peachey’s method (Representative
procedure H)
A mixture of 0.760 g (5.0 mmol) of racemic mandelic acid 1,
0.137 g (0.83 mmol) of (S)-phenylalanine (S)-2, 0.245 g of
(0.83 mmol) aspartame (S,S)-3, 0.133 g (0.83 mmol) of (S)-prega-
balin (S)-4 and 0.133 g Na2CO3 (1.25 mmol) was dissolved in
4.0 mL of hot water. The crystalline diastereomeric salt, which
appeared upon gradually cooling down the reaction mixture to
25 °C was separated from the mother liquor by filtration after
15 min. The crystals were washed with 0.5 mL of cold water, and
then they were dried to afford 0.53 g of the diastereomer salt.
The decomposition of the diastereomer was accomplished accord-
ing to the procedure described in Section 4.3 to afford 0.19 g (51%)
of (S)-1 with an op of 31% (Table 6, entry 7). [a]
25 = +47.1 (c 1,
D
water); op: 31%. The resolution of 1 was also accomplished with
the corresponding resolving agent mixtures of (S)-2, (S,S)-3 and
(S)-4 according to representative procedure H. The results are
shown in Table 6, entries 4–7.
10. (a) Vries, T.; Wynberg, H.; van Echten, E.; Koek, J.; ten Hoeve, W.; Kellogg, R. M.;
Broxterman, Q. B.; Minnaard, A.; Kaptein, B.; van der Sluis, S.; Hulshof, L.;
Kooistra, J. Angew. Chem. Int. Ed. 1998, 37, 2349–2354; (b) Nieuwenhuijzen, J.
W.; Grimbergen, R. F. P.; Koopman, C.; Kellogg, R. M.; Vries, T. R.; Pouwer, K.;
van Echten, E.; Kaptein, B.; Hulshof, L. A.; Broxterman, Q. B. Angew. Chem. Int.
Ed. 2002, 41, 4281–4286; (c) Dalmolen, J.; Tiemersma-Wegman, T. D.;
Nieuwenhuijzen, J. W.; van der Sluis, M.; van Echten, E.; Vries, T. R.; Kaptein,
B.; Broxterman, Q. B.; Kellogg, R. M. Chem. Eur. J. 2005, 11, 5619–5624; (d)
Kellogg, R. M.; Nieuwenhuijzen, J. W.; Pouwer, K.; Vries, T. R.; Broxterman, Q.
B.; Grimbergen, R. F. P.; Kaptein, B.; Crois, R. M. L.; de Wever, E.; Zwaagstra, K.;
van der Laan, A. C. Synthesis 2003, 1626–1638.
4.10. The resolution of mandelic acid 1 with mixture of (S)-
phenylalanine (S)-2 and (S)-alanine (S)-5 (Representative
procedure I)
A mixture of 0.760 g (5.0 mmol) of racemic mandelic acid 1,
0.288 g (1.80 mmol) of (S)-phenylalanine (S)-2, 0.289 g
(3.20 mmol) of (S)-alanine (S)-5 was dissolved in 4.0 mL of hot
water. The crystalline diastereomeric salt, which appeared upon

Z. Szeleczky et al. / Tetrahedron: Asymmetry 26 (2015) 721–731
731
11. (a) Pálovics, E.; Fogassy, E.; Schindler, J.; Nógrádi, M. Chirality 2007, 19, 1–4; (b)
Pálovics, E.; Schindler, J.; Faigl, F.; Fogassy, E. Tetrahedron: Asymmetry 2010, 21,
2429–2434.
Bruggink, A. Recl. Trav. Chim. Pays-Bas 1990, 109, 523–528; (c) Bolchi, C.;
Pallavicini, M.; Fumagalli, L.; Marchini, N.; Moroni, B.; Rusconi, C.; Valoti, E.
Tetrahedron: Asymmetry 2005, 16, 1639–1643.
12. (a) Welch, W. M.; Harbert, C. A.; Koe, K. B.; Kraska, A. R. US 4556676 (Pfizer)
1985, Chem. Abstr. 105:12113; (b) Koenig, T. M.; Mitchell, D. Tetrahedron Lett.
1994, 35, 1339–1342; (c) Kinbara, K.; Sakai, K.; Hashimoto, Y.; Nohira, H.;
15. Pasteur, L. C.R. Acad. Sci. 1853, 26, 162.
16. Pope, W. J.; Peachy, S. J. J. Chem. Soc., Trans. 1899, 75, 1066–1093.
17. Markovits, I.; Egri, G.; Fogassy, E. Chirality 2002, 14, 674–676.
18. (a) Huang, H.-R.; Xu, J.-H.; Xu, Y.; Pan, J.; Liu, X. Tetrahedron: Asymmetry 2005,
16, 2113–2117; (b) Shitara, H.; Shintani, T.; Kodama, K.; Hirose, T. J. Org. Chem.
2013, 78, 9309–9316.
Saigo, K. J. Chem. Soc., Perkin Trans.
2 1996, 2615–2622; (d) Bolchi, C.;
Pallavicini, M.; Fumagalli, L.; Ruggeri, P.; Valoti, E. Tetrahedron: Asymmetry
2013, 24, 796–800.
13. Kozma, D.; Pokol, G.; Ács, M. J. Chem. Soc., Perkin Trans. 2 1992, 435–439.
14. (a) van der Haest, A. D. Classical Resolutions; Design of resolving agents and
studies of diastereomeric salts. Ph.D. Thesis, University of Groningen,
Groningen, 1992; (b) van der Haest, A. D.; Wynberg, H.; Leusen, F. J. J.;
19. (a) Grote, T. M.; Huckabee, B. K.; Mulhern, T.; Sobieray, D. M.; Titus, R. D. WO
9640617 A1 (Warner-Lambert) 1996, Chem. Abs. 126:104423; (b) Neu, J.;
Fogassy, E.; Szalma, N.; Kálvin, P.; Schindler, J.; Jakab, G.; Garadnay, S.; Pálovics,
E. WO 1243934 A1 (Richter Gedeon Nyrt.) 2010, Chem. Abstr. 155:535742.