Enantioseparation of chiral mandelic acid derivatives by supercritical fluid chromatography

Article abstract of DOI:10.1002/chir.23018

Mandelic acid and its derivatives are important chiral analogs which are widely used in the pharmaceutical synthetic industry. The present study investigated the enantiomeric separation of six mandelic acids (mandelic acid, 2-chloromandelic acid, 3-chloromandelic acid, 4-chloromandelic acid, 4-bromomandelic acid, 4-methoxymandelic acid) on the Chiralpak AD-3 column by supercritical fluid chromatography. The influences of volume fraction of trifluoroacetic acid, type and percentage of modifier, column temperature, and backpressure on the separation efficiency were investigated. And the enantiomer elution order was determined. The results show that, for a given modifier, the retention factor, the separation factor, and the separation resolution decreased gradually with increasing the volume ratio of the modifier. At the same volume ratio of modifier, the retention factor of the mandelic acid and its derivatives increased in the order of methanol, ethanol, and isopropanol, except 3-chloromandelic acid. The separation factor and the separation resolution decreased with the increase of column temperature (below the temperature limit). The backpressure affected the enantioseparation process: As the backpressure increased, a corresponding decrease in retention factor was observed. Under the same chiral column conditions, the SFC method exhibited faster and more efficient separation with better enantioselectivity than the HPLC method.

Full text of DOI:10.1002/chir.23018

Received: 30 April 2018
DOI: 10.1002/chir.23018
Revised: 23 August 2018
Accepted: 23 August 2018
R E G U L A R A R T I C L E
Enantioseparation of chiral mandelic acid derivatives by
supercritical fluid chromatography
Jiawei Ding
| Ming Zhang | Huixue Dai | Chunmian Lin
College of Environment, Zhejiang
University of Technology, Hangzhou,
People's Republic of China
Abstract
Mandelic acid and its derivatives are important chiral analogs which are widely
used in the pharmaceutical synthetic industry. The present study investigated
the enantiomeric separation of six mandelic acids (mandelic acid, 2‐
chloromandelic acid, 3‐chloromandelic acid, 4‐chloromandelic acid, 4‐
bromomandelic acid, 4‐methoxymandelic acid) on the Chiralpak AD‐3 column
by supercritical fluid chromatography. The influences of volume fraction of
trifluoroacetic acid, type and percentage of modifier, column temperature,
and backpressure on the separation efficiency were investigated. And the enan-
tiomer elution order was determined. The results show that, for a given modi-
fier, the retention factor, the separation factor, and the separation resolution
decreased gradually with increasing the volume ratio of the modifier. At the
same volume ratio of modifier, the retention factor of the mandelic acid and
its derivatives increased in the order of methanol, ethanol, and isopropanol,
except 3‐chloromandelic acid. The separation factor and the separation resolu-
tion decreased with the increase of column temperature (below the tempera-
ture limit). The backpressure affected the enantioseparation process: As the
backpressure increased, a corresponding decrease in retention factor was
observed. Under the same chiral column conditions, the SFC method exhibited
faster and more efficient separation with better enantioselectivity than the
HPLC method.
Correspondence
Chunmian Lin, College of Environment,
Zhejiang University of Technology,
Hangzhou 310032, People's Republic
of China.
Email: lcm@zjut.edu.cn
Funding information
Natural Science Foundation of Zhejiang
Province, China, Grant/Award Number:
Y506246
KEYWORDS
chiral, enantioseparation, mandelic acid, mobile phase additive, supercritical fluid chromatography
2
1
| INTRODUCTION
tract disinfectant urotropine. Mandelic acid has a chiral
center resulting from the chirality of the carbon directly
As an important structural unit of many natural products
and drug molecules, chiral carboxylic acids play a crucial
role in biochemistry and life sciences. Mandelic acid
attached to the benzene ring, and thus, it usually exists
in the enantiomeric forms of (R)‐MA and (S)‐MA. As
for the mandelic acid compounds, they can be obtained
by replacing the hydrogen of benzene ring in mandelic
acid by different groups, and therefore also show chiral-
ity. It should be noted that the enantiomers of mandelic
acid and the mandelic acid compounds are different in
activity, toxicity, and environmental degradability. For
example, the R‐enantiomer of 2‐chloridic acid can be
1
(MA), known as α‐hydroxyphenylacetic acid, is a kind
of chiral carboxylic acid. Mandelic acid and its derivatives
are important chiral analogs which are widely used in the
pharmaceutical synthetic industry. They are important
raw materials for the synthesis of cephalosporins antibi-
otics, vasodilator drug loop mandelic acid, and urinary
Chirality. 2018;1–12.
wileyonlinelibrary.com/journal/chir
© 2018 Wiley Periodicals, Inc.
1

2
DING ET AL.
used to synthesize a new, safe, and efficient antiplatelet
aggregation drug clopidogrel, while the S‐enantiomer is
polysaccharide‐based stationary phases (Chiralpak IA,
IB, and IC) in SFC with a set of pharmaceutical race-
mates, including MA. However, the separate study of
the enantiomeric separation of MA and its derivatives
on SFC has not been reported.
3
inactive. The α‐cyclohexylmandelic acid is an important
chiral drug precursor which is used to synthesize a vari-
ety of chiral drugs with biological activity and good effect,
such as oxybutynin; its S‐enantiomer is better than that of
It must be noted that enantioselective separations in
SFC frequently employ temperature below the critical
temperature of CO (T < T ), although the main experi-
4
racemate with fewer side effects.
At present, the ways of obtaining single enantiomers
of MA and its derivatives are mainly asymmetric synthe-
sis, chiral separation, and enantioselective liquid‐liquid
2
c
mental conditions were under its supercritical state. Fur-
thermore, the addition of an organic modifier may lead to
subcritical fluid conditions. In other words, SFC some-
time was run as subcritical fluid chromatography in fact
because of the lower temperature and the addition of
organic modifier. However, usually subcritical or super-
critical state was not distinguished so strictly in
enantioselective separations by SFC or other application
of supercritical fluid technique due to the following rea-
sons: (1) The critical temperature and pressure of mixture
were difficult to determine, (2) the properties of fluid near
critical point changed continuously, and (3) the fluid
retains many of the desirable properties of supercritical
fluids in spite of the little lower of temperature or the
5
6
extraction. For instance, in the study of Tan, poly
(MAH‐β‐CD‐co‐NIPAAm), being synthesized by the
copolymerization of N‐isopropylazylamide (NIPAAm)
with the maleic anhydride (MAH) modified β‐cyclodex-
trin (β‐CD), was used to separate the enantiomer of
MA. The maximum separation factor (α) of MA reached
1
.27 under the optimum conditions. As a new method
of enantiomeric separation, chromatography attracts
much interest and is studied widely. The use of high‐per-
formance liquid chromatography (HPLC) to analyze MA
and its derivatives has been reported many times. High‐
performance liquid chromatography separates efficiently
the chiral organic compounds with macromolecules,
strong polarity, poor thermal stability, and high boiling
17
addition of modifier. The term SFC will be utilized
throughout this paper although some experiment was
carried out at subcritical conditions.
7
8
point. Su achieved completed enantioseparation of α‐
cyclohexylmandelic acid and methyl α‐
In this work, the enantiomeric separation of six MAs
cyclohexylmandelate on an achiral column (ODS3) with
HP‐β‐CD as mobile phase additive by HPLC. The gas
chromatographic separation of MA and its derivatives
has also been reported. Shi et al found that the separa-
tion of enantiomers of methyl mandelate was satisfactory
(Figure 1) was studied on the Chiralpak AD‐3 column
by SFC using supercritical CO as mobile phase. The
2
impacts of volume fraction of trifluoroacetic acid, type
and ratio of modifier, column temperature, and
backpressure on the separation efficiency were investi-
gated. The effect of enantiomeric separation by SFC was
compared with that of traditional HPLC. Besides, the
enantiomeric elution order on the Chiralpak AD‐3 col-
umn was established by analyzing nonracemic samples
enriched by the R‐enantiomers.
9
by means of gas chromatography using β‐cyclodextrin
(CyclodexB, 30 m × 0.25 mm × 0.25 μm). Comparing
supercritical fluid chromatography (SFC) with HPLC,
first, CO is more easily available and less expensive as
2
the mobile phase of SFC; second, only a small amount
of polar modifier is added, which has less impact on the
10
environment and operators ; in addition, SFC can
expand the range of chromatographic analysis by using
2
| MATERIALS AND METHODS
.1 | Materials
11
a variety of detectors, such as UV detector, electron cap-
2
12
ture detector, and so forth. This leads to the comple-
mentarity in the separation of different target
The organic solvents with HPLC grade‐methanol, etha-
nol, and isopropanol were used as modifier. They were
produced by Tianjin Shield Specialty Chemical Ltd. Co
13
compounds. Supercritical fluid chromatography takes
the advantages of short analysis time, fast column equili-
14
bration, and simple mobile phase system.
(Tianjin, China). CO was of dry‐ice grade and purchased
2
Currently, there are not many reports on the separa-
tion of MA and its derivatives by SFC. Medvedovici
from Jingong Specialty Gas Co. Ltd. (Hangzhou, China).
The additive trifluoroacetic acid (TFA) was of HPLC
grade and purchased from Aladdin Company. The six
MAs and their (R)‐MA are the original drugs with the
purity of above 97.0%. The racemic sample was dissolved
in isopropanol to prepare a solution with a mass concen-
tration of about 1000 mg/L for the chromatographic
analysis.
15
et al comparatively studied polysaccharide chiral sta-
tionary phases (CSPs), Chiralcel OD, Chiralpak AD, and
the macrocyclic antibiotic CSPs for the separation of dif-
ferent types of racemic compounds, including MA, by
packed column subcritical fluid chromatography.
16
Mangelings
evaluated three immobilized chiral

DING ET AL.
3
FIGURE 1 Chemical structures of
6
MA compounds in the present study
(
*denoted chiral center)
2
.2 | Instrumentation
The column temperature and the column pressure were
changed from 29°C to 41°C and from 13 to 17 MPa
respectively. Finally, the optimal separation conditions
The SFC used herein was the Thar SD‐ASFC‐2 model
from Thar Technologies (Pittsburgh, PA, USA) equipped
with a Gilson UV/VIS‐151 detector (Middleton, WI,
USA) and a Rheodyne 7410 injector with a 20 μL loop
volume (Cotati, CA, USA). The system was controlled
by the software of Thar Instruments Superchrom.
of all target compounds were determined. According to
21,22
a large number of preliminary experiments,
the detec-
tion wavelength was set at 220 nm, the flow rate was
2
mL/minute, and injection volume was 10 μL.
The retention factor of ith compounds (k ), the separa-
i
tion factor (α), and the resolution (R ) are calculated as
s
Equations (1), (2), and (3) respectively.
2.3 | Chromatographic conditions
ti − t0
The CSP of the Chiralpak AD‐3 column (250 mm × 4.6 mm
i.d., 5 μm) is made from an amylose tris (3,5‐
dimethylphenylcarbamate); the skeleton of the polymeric
selector consists of D‐glucose units linked by an α‐(1,4)‐
D‐glucose linkage, and it is regularly arranged to form a
ki ¼
(1)
(2)
(3)
t0
kiþ1
ki
α ¼
18
left‐handed
helix
conformation.
The
3,5‐
dimethylphenyl carbamate side chains surround the heli-
cal grooves and delimit the chiral environments. Natu-
rally and chemically modified amylose have highly rigid
and well‐hydrophobic cavity which can serve as a molec-
ular acceptor to recognize a variety of organics, inor-
ganics, biomolecule‐forming host guests as well as
t
− t
iþ1
Rs ¼ 1:18 ×
w þ w
i
iþ1
Here, t is the hold‐up time which is the retention
time of a nonadsorbing component, t is the retention
0
i
time of the component, and w is the peak width at half
i
19,20
supramolecular complexes.
Figure 2 shows the
height of the component.
molecular chemical structure of CSP which was pur-
chased from Daicel Chiral Technologies (Shanghai) Co.
3
| RESULTS AND DISCUSSION
Ltd. The mobile phase was supercritical CO . Different
2
kinds and volume fractions of alcohol modifiers (metha-
nol, ethanol, isopropanol) were added into the mobile
phase in order to explore their influences on the chiral
separation and to optimize the mobile phase composition.
By exploring different influencing factors, the optimum
conditions for the separation of MA compounds could
be determined and the related mechanisms could be rec-
ognized as well, which might provide an insight for the
separation of other enantiomers by SFC.
3
3
.1 | Effect of additive and modifier
.1.1 | Effect of TFA on the
enantioselectivity
Since the MA and its derivatives are rather strong carbox-
ylic acids, they can adsorb onto the silica gel in CSP. This
results in the peak type trailing tail, and the data of the
peak type fitting are then not accurate. An essential
FIGURE 2 Structural formulas of CSPs

4
DING ET AL.
function of basic as well as acidic additives is to suppress
the ionization of strong basic and acidic groups, in order
to achieve elution, enantioseparation, and a satisfying
peak shape. A small amount of TFA can be added to
the mobile phase. As acidic additive, trifluoroacetic acid
improves enantioselectivity by enhancing hydrogen
bonding interactions between the carboxyl groups of the
compounds with carbonyl groups of the carbamate func-
tions of the polysaccharide‐based selectors. Addition of
TFA to the mobile phase reduces the pH and the ioniza-
tion of acidic compounds. Thus, the peak trailing can be
significantly reduced, the peak broadening can be over-
come, the retention time of acidic drugs can be shortened,
and selectivity improved drug molecules.
FIGURE 3 Effects of TFA on the base line fluctuation
Four analytes were selected to explore the effect of
volume fraction of TFA on chiral separation, and the
results are shown in Table 1. With the increase of the vol-
For the same modifier, as can be seen from Figure 4,
k1 gradually decreased with increasing the methanol vol-
ume fraction. It shows that increasing the volume frac-
tion of modifier accelerated the elution speed,
decreasing the retention time. This tendency was pro-
nounced at the initial stage when the modifier volume
fraction increased. Methanol is a protic solvent, which is
also a proton‐donor and proton acceptor. It can occur
hydrogen‐bonding interaction with the compounds and
the chiral stationary phase, thereby competing with com-
pounds for hydrogen bonding sites. Increasing the vol-
ume fraction of methanol enhances this competition,
reduces the interaction between the compound and
CSP, and shortens the retention time.
ume fraction of TFA, the resolution (R ) presented an
s
increasing trend in general while the retention factors
(
k and k ) and the separation factor (α) did not fluctuate
1 2
significantly. Noting that when the TFA concentration
was high, the baseline would become unstable
(Figure 3) and the chiral stationary phase might be dam-
aged. Thus, 0.1% of TFA was determined herein.
3.1.2 | Effect of the modifiers on the
enantioselectivity
Mandelic acid and its derivatives are known to be of
strong molecular polarity whereas the mobile phase CO2
is a nonpolar substance. Thereby, the attractive interac-
tion between the two cannot be strong. Obviously, it is
Figure 5 shows the effect of methanol volume fraction
on α. On the whole, the separation factor decreased with
the increase of the volume fraction of the modifier. This
may be attributed to the fact that the variation degree of
k and k changes with the gradual increase of polar com-
insufficient to elute with only CO as mobile phase with-
2
1
2
out adding any polar modifier. Methanol, ethanol, and
isopropanol used in this study are all proton‐type sol-
vents. They are capable of interacting with the stationary
phase through hydrogen bonding, competing with the
MA compounds for hydrogen bonding sites, and acceler-
ating the elution rate.
ponent percentage in the mobile phase. The α value is the
ratio of two adjacent peak retention factors. Therefore,
the change of k value leads to the change of the α value.
In essence, the α value indicates the thermodynamic
properties of chiral compounds in the distribution equi-
librium between the two phases. It reveals the difference
TABLE 1 Effect of TFA on the enantioselectivity of MA and its derivatives
Compd.
TFA content (%)
k
1
k
2
α
R
s
Compd.
TFA content (%)
k
1
k
2
α
R
s
MA
0%
3.28 4.30 1.31 6.08 3‐ChloroMA
3.31 4.26 1.29 5.96
3.22 4.24 1.32 6.70
0%
0.1%
0.2%
0.3%
3.64 4.20 1.15 2.99
3.64 4.18 1.15 3.10
3.60 4.16 1.15 3.21
3.63 4.20 1.16 3.45
0
0
0
.1%
.2%
.3%
3.27 4.34 1.33 6.69
2‐ChloroMA
0%
.1%
.2%
.3%
4.65 5.20 1.12 2.68 4‐Methoxymandelic acid
4.64 5.16 1.11 2.71
4.55 5.08 1.12 2.84
0%
0.1%
0.2%
0.3%
4.37 4.57 1.05 0.91
4.24 4.50 1.06 1.32
4.36 4.57 1.05 1.15
4.36 4.56 1.05 1.08
0
0
0
4.62 5.19 1.12 3.11
Chromatographic condition: Chiralpak AD‐3 with the modifier of 12% isopropanol at 35°C column temperature and 15 MPa backpressure.

DING ET AL.
5
FIGURE 4 Effectofmethanolpercentageontheretentionfactork
1
FIGURE 6 Effect of methanol percentage on the resolution R
s
FIGURE 7 Effect of the modifiers on the retention factor k
1
FIGURE 5 Effect of methanol percentage on the separation
factor α
not be achieved between the enantiomer and CSPs, and
thus, the separation of the two enantiomers was ineffi-
in the interaction force between molecules. That is to say,
with the increase of modifier volume fraction, the differ-
ence in thermodynamic properties and enantioselectivity
of the separated enantiomers turned to be smaller when
the distribution equilibrium was achieved.
As can be seen from Figure 6, the resolution (Rs)
decreased gradually with the increase of methanol vol-
ume fraction since larger volumes of modifier led to the
stronger interaction between methanol and the chiral sta-
tionary phase. The competition for the adsorption sites on
the chiral stationary phase became intensified, and there-
fore, the retention time of the compounds got shortened.
Adequate and effective desorption and adsorption could
cient and R decreased.
s
Different modifiers have different effects on chiral
separation, due to their differences in polar and steric
hindrance. In this paper, we investigated the effect of dif-
ferent modifier species (methanol, ethanol, isopropanol)
on chiral separation of MA and its five derivatives. The
retention factor k , the separation factor α, and the reso-
1
lution R were obtained under the modifier addition of
s
16% (volume fraction) and listed in Figures 7, 8, and 9
respectively. It should be noted that 4‐chloromandelic
acid and 4‐bromomandelic acid cannot be successfully
separated under these isopropanol conditions.

6
DING ET AL.
As shown in Figures 8 and 9, the variation of α and Rs
with the modifier has no obvious regularity. This may be
because the chiral recognition mechanisms of modifiers
used herein to the selected drugs are complex, being
affected not only by the polarity of modifier and the
molecular volume factors but also associating with the
structures of MA compounds. Methanol and ethanol
always performed better in the elution of MA and its
derivatives than isopropanol. The change of retention fac-
tor is basically in accordance with that of the polarity of
modifier and the space resistance. In conclusion, metha-
nol usually provided the largest success rate and specific-
2
3
ity, which agreed well with the report of Syame Khater.
4
| EFFECT OF THE COLUMN
FIGURE 8 Effect of the modifiers on the separation factor α
TEMPERATURE
The temperature of chiral chromatographic column
importantly influences the separation process of chiral
compounds by SFC, changing retention factor, separation
factor, and resolution. Given that the volume fractions of
modifier were distinct under the preferable separation
conditions, thus the separation conditions for this part
of experiments were set to be consistent (0.1% TFA, meth-
anol as modifier, 15 MPa). The column temperature
ranged from 29°C to 41°C (every 3°C for a gradient),
and the impact of column temperature on the separation
of MA and its five derivatives was then investigated. The
results of enantiomeric separation chromatography at dif-
ferent chiral column temperatures were summarized in
Table 2. It can be seen that with the increase of tempera-
ture, the retention factor k and k fluctuated while the
1
2
FIGURE 9 Effect of the modifiers on the resolution R
s
separation factor α and the resolution R decreased.
s
Changing the separating temperature at constant pres-
sure may lead to significant changes of the fluid density.
On increasing temperature, a decrease in fluid density,
and thus an increase in retention, is usually observed.
However, higher temperature can also cause desorption
As shown in Figure 7, except 3‐ChloroMA, the reten-
tion factor of MA and its four derivatives increased in the
order of methanol, ethanol, and isopropanol for the fixed
volume ratio of modifier. The polarity of the three modi-
fiers increased in the order of isopropanol, ethanol, and
methanol. The hydrogen bonding strength of the three
modifiers to CSP obeyed the same sequence. The chain
of isopropanol is more branched than that of methanol
and ethanol. Therefore, the hydrogen bonding of
isopropanol is much weaker, and the elution rate is much
slower. In addition, the molecular volume of the modifier
was in the order of isopropanol, ethanol, and methanol.
The steric hindrance of the modifier with large molecular
volume was larger, and therefore, the attractive interac-
tion between the solvent and the chiral column decreased
and the elution ability of the mobile phase reduced. This
caused the increase of retention time of enantiomers on
the chiral stationary phase.
of both scCO and modifier from the stationary phase
2
24
that leads to a decrease in retention. Temperature and
density are the main factors affecting viscosity. Both the
intermolecular collisions and collisions in the process of
molecular free translation can all cause momentum
transfer. The combined effects of momentum transfer
caused by these two collisions can reflect fluid viscosity.
Temperature and density affect the momentum transfer
mode, which changes the viscosity of the fluid. Typically,
the viscosity of the liquid decreases with increasing tem-
perature; supercritical fluid under high‐density condi-
tions, the viscosity decreases with increasing
temperature; At low‐density conditions, the result is
opposite. The temperature of 35°C was appropriate in this

DING ET AL.
7
TABLE 2 Separation results of MA and its five derivatives at different column temperatures
Column
Number
MA and its derivatives
Modifier
temperature (°C)
k
1
k
2
α
R
s
1
MA
14% Methanol
29
1.08
1.91
1.68
1.65
1.57
1.50
2.59
2.20
2.15
1.99
1.40
1.35
1.31
1.30
1.26
3.15
3.74
2.94
2.82
2.32
3
3
3
4
2
5
8
1
2
3
4
5
6
2‐ChloroMA
3‐ChloroMA
4‐ChloroMA
4‐BromoMA
4‐MethoxyMA
14% Methanol
10% Methanol
20% Methanol
20% Methanol
20% Methanol
29
2.10
2.86
2.61
2.63
2.58
2.83
3.69
3.24
3.26
3.09
1.35
1.29
1.25
1.24
1.20
3.87
3.98
2.76
3.03
2.38
3
3
3
4
2
5
8
1
29
3.03
3.00
3.39
3.10
3.00
3.67
3.63
4.05
3.70
3.58
1.21
1.21
1.20
1.19
1.19
3.84
3.86
3.79
2.73
3.58
3
3
3
4
2
5
8
1
29
1.18
1.13
1.11
1.06
1.04
1.81
1.72
1.68
1.58
1.53
1.53
1.52
1.51
1.49
1.48
3.55
3.59
3.62
3.60
3.52
3
3
3
4
2
5
8
1
29
1.38
1.52
1.46
1.46
1.34
2.64
2.81
2.61
2.56
2.32
1.91
1.85
1.79
1.75
1.72
7.59
7.44
6.38
5.88
5.28
3
3
3
4
2
5
8
1
29
1.26
1.37
1.32
1.24
1.25
2.05
2.16
2.05
1.90
1.88
1.62
1.58
1.55
1.54
1.50
5.44
5.03
4.26
4.23
3.72
3
3
3
4
2
5
8
1
T = 35°C, 0.1% TFA, on Chiralpak AD‐3.
ꢀ
ꢁ
k
experiment since CO would not reach supercritical state
°
iþ1
°
°
2
−ΔΔG ¼ RT lnα ¼ ln
¼ −ΔΔH þ TΔΔS
ki
at lower temperatures and the chiral separation process
25
could be hindered at higher temperatures. He reported
the enantioseparation of triticonazole by SFC. With the
increase of temperature, retention factors, selectivity, col-
umn efficiency, and resolution kept stable. It meant that
the enantioseparation of triticonazole might be tempera-
ture‐insensitive under the studied experimental condi-
tions. The separation performance may also be related
to the chemicals themselves.
ΔΔH° ΔΔS°
lnα ¼ − RT
þ
R
Here, k is the retention factor, α is the separation fac-
i
tor, R is the universal gas constant, T is the absolute tem-
perature, and Ф is phase ratio. ΔH and ΔS are the free
enthalpy and entropy of the interaction between each
enantiomer and the chiral stationary phase respectively.
ΔΔG° is the Gibbs free energy, while ΔΔH° and ΔΔS°
are the differences of the free enthalpy and free entropy
between two enantiomers respectively. When the
To further investigate the effect of column tempera-
ture, the thermodynamic behavior of enantiomeric sepa-
26,27
ration was studied.
The effect of temperature on the
chiral separation process was interpreted according to
the van't Hoff equation.
−
ΔΔH°/RT plays a dominant role or ΔΔG° is governed
by the enthalpy, the separation process of enantiomers
is called “enthalpy driven.” In other words, the separation
factor decreases as the temperature increases and an
ΔH ΔS
RT
ΔH
*
lnki ¼
þ
þ lnϕ ¼ − RT þ ΔS
R

8
DING ET AL.
ΔΔH°
ΔΔS°
enthalpy driven occurs. In contrast, an entropy‐driven
separation occurs if ΔΔS°/R governs the value of ΔΔG°
and the separation factor increases as the temperature
increases. When the peaks were coeluted because of the
balance between −ΔΔH°/RT and ΔΔS°/R, the
isoenantioselective temperrature (T ) occurs. T can
Tiso
¼
The results of fitting lnα − 1/T plots were analyzed
using the van't Hoff formula to calculate the thermody-
namic parameters ΔΔH°, ΔΔS°, and Tiso for chiral separa-
tion of enantiomers, which are listed in Table 3.
iso
iso
be calculated through the following equation:
TABLE 3 The thermodynamic parameters of MA and its five derivatives
Number
MA and its derivatives
Fitting equation
ΔΔH (kJ/mol)
ΔΔS (J/mol/K)
Tiso (K)
1
2
3
4
5
6
MA
y = 766.3x − 2.209
y = 892.6x − 2.665
y = 133.9x − 0.2529
y = 298.5x − 0.554
y = 842.3x − 2.145
y = 606.7x − 1.528
−6.37
−7.42
−1.11
−2.48
−7.00
−5.04
−18.37
−22.16
−2.10
346.9
334.9
529.5
538.8
392.7
397.1
2‐ChloroMA
3‐ChloroMA
4‐ChloroMA
4‐BromoMA
4‐MethoxyMA
−4.61
−17.83
−12.70
TABLE 4 Separation results of MA and its five derivatives under different backpressure
Number
MA and its derivatives
Modifier
Pressure (MPa)
k
1
k
2
α
R
s
1
MA
14% Methanol
13
1.90
1.75
1.59
1.47
1.38
2.46
2.23
2.11
2.02
1.96
1.29
1.28
1.32
1.38
1.42
6.04
5.35
5.17
5.13
5.10
1
1
1
1
4
5
6
7
2
3
4
5
6
2‐ChloroMA
3‐ChloroMA
4‐ChloroMA
4‐BromoMA
4‐MethoxyMA
14% Methanol
10% Methanol
20% Methanol
20% Methanol
20% Methanol
13
2.31
2.17
2.03
1.91
1.78
2.71
2.53
2.37
2.24
2.10
1.18
1.16
1.17
1.17
1.18
3.35
2.95
2.81
2.73
2.68
1
1
1
1
4
5
6
7
13
3.54
3.33
3.17
3.04
2.86
4.21
3.94
3.74
3.55
3.35
1.19
1.18
1.18
1.17
1.17
4.36
4.03
3.72
3.36
3.23
1
1
1
1
4
5
6
7
13
1.58
1.38
1.19
1.03
0.90
2.20
1.95
1.74
1.53
1.27
1.39
1.41
1.46
1.48
1.40
5.66
5.26
5.06
4.69
3.08
1
1
1
1
4
5
6
7
13
1.88
1.71
1.55
1.44
1.37
2.67
2.48
2.31
2.18
2.11
1.42
1.45
1.48
1.51
1.54
6.29
6.12
5.98
5.89
5.63
1
1
1
1
4
5
6
7
13
1.41
1.31
1.24
1.18
1.17
1.98
1.83
1.70
1.62
1.54
1.40
1.40
1.37
1.37
1.32
4.64
4.28
3.78
3.63
3.07
1
1
1
1
4
5
6
7
T = 35°C, 0.1% TFA, on Chiralpak AD‐3.

DING ET AL.
9
The highest temperature (41 °C) in this study was
much lower than Tiso of the enantiomers of the MA and
its five derivatives. The chiral recognition process was
mainly enthalpy driven, and the separation factor α
decreased with temperature increasing. This was consis-
tent with the fact that the separation factor of enantio-
mers of chiral drugs varied with temperatures in this
study.
5
| EFFECT OF BACKPRESSURE
The density of mobile phase is proportional to pressure
which may influence the component distribution coeffi-
cient. For the chiral resolution by supercritical CO chro-
2
matography, the pressure range is narrow, basically
between 15 and 20 MPa. Under the optimum separation
condition, the effect of backpressure on MA and its five
derivatives was investigated. Herein, the backpressure of
the SFC system ranged from 13 to 17 MPa and a gradient
was set to be 1 MPa.
It can be seen from Table 4 that, when the
backpressure increased, the retention factors of MA com-
pounds k decreased gradually, and the trend turned to
1
be obvious at low backpressures. It indicates that the
increase of backpressure accelerated the speed of elution
and reduced the retention time. This could be
interpreted by the fact that the increase of backpressure
led to the increase of the overall density of the mixed
mobile phase as well as the enhancement of the solva-
tion ability. The variation of retention factors is similar
to the report of Tarafder. With the increase of the sys-
tem backpressure, the separation factor α remained
within a certain range whereas the resolution Rs
decreased. Considering the factors, including peak speed,
separation degree, and instrument maintenance, herein,
FIGURE 10 The chromatograms of rac‐4‐bromomandelic acid
and R‐4‐bromomandelic acid in standard solutions
elution order, then mark on the chromatogram under
the optimum conditions (Figure 11). The results in this
study on the MA and its five derivatives were compared
28
29
with those reported in literature. Under the same chi-
ral column conditions, SFC resolved all the studied
compounds and the HPLC only succeed for two‐thirds
of the compounds. The SFC method exhibited faster
1
5 MPa was chosen for the chiral separation of MA
and its derivatives.
and
more
efficient
separation
with
better
enantioselectivity than the HPLC method (Table 5). As
3
0
6
| OPTIMUM SEPARATION
a separating technique, SFC has a cost advantage over
RESULTS
normal phase HPLC in terms of (1) the cost of CO ver-
2
sus the cost of normal phase solvents (hexane, heptane,
isooctane, ect), (2) the significant cost reduction in han-
dling waste, (3) the improved efficiency of isolating the
active pharmaceutical ingredient from the mobile phase
after collection, (4) the ease of using SFC as a normal
phase system, and (5) SFC arise from the reduced vis-
cosities and increased diffusivities of supercritical fluids
when compared to liquids. The reduced viscosity
decreases the pressure drop across the column.
Increased diffusivity shifts the optimum linear velocity
to higher values so that higher flow rates can be used
without compromising efficiency.
The resolution results under different conditions show
that all the MA and its five derivatives achieved the
baseline separation within 7 minutes (R > 1.5). The
s
SFC separation method should be optimized based on
the principle of relatively fast separation time and high
tolerance of instruments and CSP chiral column. The
racemic standards and the pure R‐enantiomers were
tested separately (one chromatogram of 4‐bromo
mandelic acid shown in Figure 10), and qualitative
analysis was performed according to the retention times
of the two enantiomers to determine the enantiomer

10
DING ET AL.
FIGURE 11 SFC chromatograms of MA and its 5 derivatives under optimal separation condition P = 15 MPa, 0.1% TFA, T = 35°C
TABLE 5 Comparison for the enantioseparation of 6 MAs by SFC and HPLC
t
2
α
R
s
MA and its
derivatives
Number
SFC
HPLC
SFC
HPLC
SFC
HPLC
1
2
3
4
5
6
MA
4.6
5.1
7.1
4.1
5.0
4.0
12.2
12.9
19.0
10.1
11.0
46.0
1.28
1.17
1.18
1.46
1.48
1.37
1.20
1.10
1.12
1.00
1.00
1.10
4.50
2.81
3.72
5.06
5.98
3.78
2.86
1.60
1.83
0
2‐ChloroMA
3‐ChloroMA
4‐ChloroMA
4‐BromoMA
4‐MethoxyMA
0
2.07
Chiral columns are Chiralpak AD‐3.

DING ET AL.
11
7
| CONCLUSION
ORCID
Jiawei Ding http://orcid.org/0000-0002-6288-6746
Chunmian Lin http://orcid.org/0000-0003-3173-259X
This study explored the enantiomeric separation of MA
and its five derivatives on the Chiralpak AD‐3 column
by SFC using supercritical CO and various organic mod-
2
ifiers. The effects of volume fraction of TFA, type and per-
centage of mobile phase, column temperature, and
backpressure on the separation efficiency were investi-
gated. The results can be summarized as follows:
REFERENCES
1
2
3
4
. Xiao QB. Synthesis of chiral amino alcohols from L‐
Phenylglycine and their application. Cent South Univ. 2012;1‐47.
. Zhang N. Mandelic compound mandelic acid development pros-
pects. China Chem Ind News. 2002;2–21(3).
1
TFA could effectively solve the problem of peak tailing.
Modifier species and volume fraction greatly affected
the enantioseparation process. Increasing the volume
fraction of modifier facilitated the elution of the compo-
nents and thus decreased the retention time. The ten-
dency was pronounced when the volume fraction
began to increase. As the volume fraction increased,
. Tang T, Wang YQ, Wang HQ, Cai MY. Synthesis of clopidogrel
hydrogen sulfate. Chin J Pharm. 2009;40(50):324‐325.
. Zhou D, Liu JJ, Tang KW. Chiral extraction of racemicα‐
cyclohexyl‐mandelic acid with tartaric acid esters as chiral selec-
tor. J Instr Analyst. 2006;25(6):49‐52.
5
. Lu RC, He Q, Feng C, Peng YF. Enantioselective resolution of 4‐
chloromandelic acid by liquid–liquid extraction using 2‐chloro‐
N‐carbobenzyloxy‐L‐amino acid. Chirality. 2017;29(11):708‐715.
both of α and total R of MA and its five derivatives
s
showed decreasing trend. When the volume fraction
6
. Tan Z, Li F, Zhao C, Teng Y, Liu Y. Chiral separation of
mandelic acid enantiomers using an aqueous two‐phase system
based on a thermo‐sensitive polymer and dextran. Sep Purif
Technol. 2017;172:382‐387.
wasfixedtobe16%,k ofMAanditsderivativeincreased
1
following the order of modifiers: isopropanol > etha-
nol > methanol. The values of α and R did not change
s
significantly with the type of modifier.
7
. Wang Z, Yin DD, Wang XL, Zhan LL. Research progress on
high performance liquid chromatography in chiral drugs analy-
sis. Chin J Pharm Anal. 2015;35(7):1127‐1133.
2
The retention factors k and k fluctuated with tem-
1
2
peratures. The separation factor and resolution
decreased when the column temperature increased.
According to the van't Hoff equation, the chiral sep-
aration process of all MA and its five derivatives
was enthalpy driven. The temperature of 35°C was
8. Su YH. Study on the Chiral Recognition Mechanism of α‐
Cyclohexylmandelic Acid and Methyl α‐Cyclohexylmandelate on
Hydroxypropyl‐β‐Cyclodextrin. Zhejiang University of Technol-
ogy; 2012.
appropriate since CO would not reach supercritical
9. Shi JH, Cheng XW, Yan W. Enantiomeric separation of methy
2
(
mande) ate by gas chromatography and discussion the chiral
recognition mechanism. Chin Pharm Anal.
009;29(10):1681‐1684.
state at lower temperatures and the chiral separa-
tion process could be hindered at higher
temperatures.
The backpressure could affect the enantioseparation
process. As the backpressure increased, a corre-
sponding decrease in retention factor was observed.
The separation factor did not suffer great
fluctuation.
J
2
1
0. Zhao YQ, Pritts WA, Zhang SH. Chiral separation of selected
proline derivatives using a polysaccharide type stationary phase
by supercritical fluid chromatography and comparison with
high‐performance liquid chromatography. J Chromatogr A.
3
4
2008;1189(1–2):245‐253.
1
1. Lou CY, Jiang L, Duan F, et al. Simultaneous determination of
eight allergenous disperse dyes in blended carpet by supercriti-
The desired enantioseparation efficiencies of six
mandelic acids on the AD‐3 column were achieved
by adding 0.1% TFA and methanol as modifier within
cal fluid chromatography‐ultraviolet detection. Chin
Chromatogr. 2017;35(4):453‐457.
J
7
minutes at the column temperature of 35°C and the
1
2. Khater S, Lozach'h MA, Adam I, Francotte E, West C. Compar-
ison of liquid and supercritical fluid chromatography mobile
phases for enantioselective separations on polysaccharide sta-
tionary phases. J Chromatogr A. 2016;1467:463‐472.
backpressure of 15 MPa. In addition, the enantiomer
elution order was determined by comparing the chro-
matograms of MA enantiomers and the pure R‐
enantiomers.
13. Zhang J, Chen XD, Li LQ, He JF, Fan J, Zhang WG. Compara-
tive separation of chiral compounds by supercritical fluid
chromatography and high performance liquid chromatography.
Chin J Chromatogr. 2016;34(3):321‐326.
ACKNOWLEDGMENTS
1
4. Taylor LT. Supercritical fluid chromatography for the 21st cen-
The authors are grateful for financial support from the
Natural Science Foundation of Zhejiang Province, China
tury. J Supercrit Fluid. 2009;47(3):566‐573.
1
5. Medvedovici A, Sandra P, Toribio L, David F. Chiral packed col-
umn subcritical fluid chromatography on polysaccharide and
(Y506246).

12
DING ET AL.
macrocyclic antibiotic chiral stationary phases. J Chromatogr A.
997;785(1‐2):159‐171.
6. Mangelings D, Klerck KD, Heyden YV. Pharmaceutical‐enantio-
based columns: comparison of normal phase liquid chromatog-
1
raphy
and
supercritical
fluid
chromatography.
Chromatographia. 2017;80(4):547‐557.
1
mers resolution using immobilized polysaccharide‐based chiral
stationary phases in supercritical fluid chromatography.
J Chromatogr A. 2014;1328:85‐97.
25. He J, Fan J, Yan Y, et al. Triticonazole enantiomers: separation
by supercritical fluid chromatography and the effect of the chro-
matographic conditions. J Sep Sci. 2016;39(21):4251‐4257.
1
7. Sun Q, Olesik SV. Chiral separations performed by enhanced‐
fluidity liquid chromatography on a macrocyclic antibiotic chi-
ral stationary phase. Anal Chem. 1999;71(11):2139‐2145.
26. Weng W, Wang QH, Yao BX, Zeng QL. Enantio separation of
amino acid derivatives on an immobilized network polymer
derived from L‐tartaric acid. J Chromatogr A. 2004;1042
(
1–2):81‐87.
1
8. Yamamoto C, Yashima E, Okamoto Y. Structural analysis of
amylose tris(3,5‐dimethylphenylcarbamate) by NMR relevant
to its chiral recognition mechanism in HPLC. J Am Chem Soc.
27. Yang F, Li XC, Zhang W, Pan JB, Chen ZG. A facile light‐emit-
ting‐diode induced fluorescence detector coupled to an
integrated microfluidic device for microchip electrophoresis.
Talanta. 2011;84(4):1080‐1092.
2
002;124(42):12583‐12589.
1
9. Francotte E, Zhang T. Supramolecular effects in the chiral dis-
crimination of meta‐methylbenzoyl cellulose in high‐
28. Tarafder A, Guiochon G. Use of isopycnic plots in designing
operations of supercritical fluid chromatography: I. The critical
role of density in determining the characteristics of the mobile
phase in supercritical fluid chromatography. J Chromatogr A.
performance liquid chromatography.
995;718(2):257‐266.
J
Chromatogr A.
1
2
0. Kucerova G, Kalikova K, Tesarova E. Enantioselective poten-
tial of polysaccharide‐based chiral stationary phases in
supercritical fluid chromatography. Chirality. 2017;29(6):
2011;1218(28):4569‐4585.
29. Wang M. Enantioseparation of mandelic acid compounds with
Chiralpak AD‐H and Chiralcel OJ‐H chiral stationary phases.
Chin J Chromatogr. 2014;32(2):198‐203.
2
39‐246.
2
2
2
1. Miao YL, Xu MX, Zheng XQ, Lin CM. Chiral separation of six
triazole fungicides by supercritical fluid chromatography. Chin
J Pestic Sci. 2015;17(4):432‐438.
30. Webster GK, Holzheuer WB, Wong MM. A comparison of HPLC
and SFC chiral method development screening approaches for
compounds of pharmaceutical interest. Curr Pharm Anal.
2. Dai HX, Yang X, Lin CM. Enantioseparation of 1,4‐
dihydropyridines calcium antagonists by supercritical fluid
chromatography. Chin J Pharm Anal. 2017;8:1513‐1519.
2008;4:101‐105.
3. Khater S, West C. Insights into chiral recognition mechanisms
in supercritical fluid chromatography V. Effect of the nature
and proportion of alcohol mobile phase modifier with amylose
and cellulosetris‐(3,5‐dimethylphenylcarbamate) stationary
phases. J Chromatogr A. 2014;1373:197‐210.
How to cite this article: Ding J, Zhang M, Dai
H, Lin C. Enantioseparation of chiral mandelic acid
derivatives by supercritical fluid chromatography.
Chirality. 2018:1–12. https://doi.org/10.1002/
chir.23018
2
4. Kolderová N, Neveselý T, Šturala J, Kuchař M, Holakovský R,
Kohout M. Enantioseparation of chiral sulfoxides on amylose‐