Experimental and Model Studies on Continuous Separation of 2-Phenylpropionic Acid Enantiomers by Enantioselective Liquid-Liquid Extraction in Centrifugal Contactor Separators

Article abstract of DOI:10.1002/chir.22567

Multistage enantioselective liquid-liquid extraction (ELLE) of 2-phenylpropionic acid (2-PPA) enantiomers using hydroxypropyl-β-cyclodextrin (HP-β-CD) as extractant was studied experimentally in a counter-current cascade of centrifugal contactor separators (CCSs). Performance of the process was evaluated by purity (enantiomeric excess, ee) and yield (Y). A multistage equilibrium model was established on the basis of single-stage model for chiral extraction of 2-PPA enantiomers and the law of mass conservation. A series of experiments on the extract phase/washing phase ratio (W/O ratio), extractant concentration, the pH value of aqueous phase, and the number of stages was conducted to verify the multistage equilibrium model. It was found that model predictions were in good agreement with the experimental results. The model was applied to predict and optimize the symmetrical separation of 2-PPA enantiomers. The optimal conditions for symmetric separation involves a W/O ratio of 0.6, pH of 2.5, and HP-β-CD concentration of 0.1 mol L^-1 at a temperature of 278 K, where ee_eq (equal enantiomeric excess) can reach up to 37% and Y_eq (equal yield) to 69%. By simulation and optimization, the minimum number of stages was evaluated at 98 and 106 for ee_eq > 95% and ee_eq > 97%. Chirality 28:235-244, 2016.

Full text of DOI:10.1002/chir.22567

CHIRALITY 28:235–244 (2016)
Experimental and Model Studies on Continuous Separation of 2-
Phenylpropionic Acid Enantiomers by Enantioselective Liquid–Liquid
Extraction in Centrifugal Contactor Separators
1
2
2
1
*
*
XIAOFENG FENG, KEWEN TANG, PANGLIANG ZHANG, AND SHUANGFENG YIN
1College of Chemistry and Chemical Engineering, Hunan University, Changsha, People’s Republic of China
2Department of Chemistry and Chemical Engineering, Hunan Institute of Science and Technology, Yueyang, People’s Republic of China
ABSTRACT
Multistage enantioselective liquid–liquid extraction (ELLE) of 2-phenylpropionic
acid (2-PPA) enantiomers using hydroxypropyl-β-cyclodextrin (HP-β-CD) as extractant was
studied experimentally in a counter-current cascade of centrifugal contactor separators (CCSs).
Performance of the process was evaluated by purity (enantiomeric excess, ee) and yield (Y). A
multistage equilibrium model was established on the basis of single-stage model for chiral
extraction of 2-PPA enantiomers and the law of mass conservation. A series of experiments on
the extract phase/washing phase ratio (W/O ratio), extractant concentration, the pH value of
aqueous phase, and the number of stages was conducted to verify the multistage equilibrium
model. It was found that model predictions were in good agreement with the experimental results.
The model was applied to predict and optimize the symmetrical separation of 2-PPA enantiomers.
The optimal conditions for symmetric separation involves a W/O ratio of 0.6, pH of 2.5, and
HP-β-CD concentration of 0.1 mol Lꢀ1 at a temperature of 278 K, where eeeq (equal enantio-
meric excess) can reach up to 37% and Yeq (equal yield) to 69%. By simulation and optimiza-
tion, the minimum number of stages was evaluated at 98 and 106 for eeeq > 95% and eeeq > 97%.
Chirality 28:235–244, 2016. © 2016 Wiley Periodicals, Inc.
Research highlights are as follows:
(1) Multistage enantioselective liquid-liquid extraction is suitable for the separation of
2-phenylpropionic acid enantiomers with HP-β-CD as hydrophilic chiral selector.
(2) Model predictions of multistage countercurrent centrifugal extraction of 2-phenylpropionic
acid enantiomers are in good agreement with the experimental results.
(3) Purity for 2-phenylpropionic acid enantiomers could be remarkably improved by changing
W/O ratio or extractant concentration, and or adding the number of centrifugal contact
separators.
(4) Multistage enantioselective liquid-liquid extraction in centrifugal contactor separators can
be hopeful for separations of various enantiomers at a large-scale.
KEY WORDS: enantioselective liquid-liquid extraction; multistage equilibrium model; symmetrical
separation; simulation and optimization; 2-phenylpropionic acid; hydroxypropyl-β-
cyclodextrin
There is a growing demand for enantiopure compounds in
the pharmaceutical and chemical industry because different
enantiomers often show a large difference in pharmacological
activity and toxicity in living organisms.1,2 (R)-thalidomide, for
example, has the desired sedative effects, while (S)-thalidomide
causes severe birth defects and deformities.3,4 Therefore, it is
very important and necessary to develop technologies for
obtaining optically pure products. Recently, a variety of
methods have become available for separating optically active
compounds. The most used technology on a commercial scale
is the crystallization technique,5 but this method is not always
applicable and the search for an appropriate resolving agent is
rather time-consuming. Many other impressive technologies
such as liquid membrane,6 chromatographic techniques,7,8
kinetic resolution, and capillary electrophoresis,9–11 and so on,
are being developed rapidly and provide very promising ap-
proaches for enantiomeric separation. However, these technol-
ogies have either low versatility or high cost, and they are
difficult to achieve for industrialization. Compared with the
technologies mentioned above, enantioselective liquid–liquid
extraction (ELLE) exhibits several advantages, including
relatively mature theoretical guidance, lower cost in energy
consumption, easier to scale up to commercial scale, and has
a wide range of application.12
ELLE combines the concepts of solvent extraction and chi-
ral recognition in a single technology, which is considered a
promising technology for chiral separation. Although more
Contract grant sponsor: National Basic Research Program of China; Contract
grant number: 2014CB260407.
Contract grant sponsor: Hunan Provincial Natural Science Foundation of
China; Contract grant number: 2015JJ6045.
Contract grant sponsor: Science and Technology Planning Project of Hunan
Province; Contract grant number: 2014GK3108.
*Correspondence to: Kewen Tang, Department of Chemistry and Chemical
Engineering, Hunan Institute of Science and Technology, Yueyang, 414006,
People’s Republic of China. E-mail: tangkewen@sina.com; Shuangfeng Yin,
College of Chemistry and Chemical Engineering, Hunan University,
Changsha, 410082, People’s Republic of China. E-mail: yinsf73@163.com
Received for publication 19 October 2015; Accepted 20 November 2015
DOI: 10.1002/chir.22567
Published online 5 January 2016 in Wiley Online Library
(wileyonlinelibrary.com).
© 2016 Wiley Periodicals, Inc.

236
FENG ET AL.
than half a century has elapsed since the first report of ELLE
and many researchers have attempted to obtain enantiopure
compounds by ELLE, the number of literature studies has
only significantly increased in the past two decades.13–23
Studies cited in the literature focus mainly on the single-stage
extraction equilibrium and the synthesis of new chiral selec-
tors. Only a few studies report the fundamental insights about
the reaction engineering mechanism by combining
experimental investigation and mathematical modeling to pre-
dict the extraction performance and optimize the separation
process.21–23 Recently, studies providing the approaches for
application of ELLE in multistage processes have drawn more
and more attention of researchers.
The centrifugal contactor separator (CCS) is an example of
compact continuous flow equipment that possesses the bene-
ficial properties of large centrifugal forces and excellent mass
transfer. In the CCS device, reaction and separation are com-
bined in a single device, and the two immiscible liquids are in-
tensely mixed and subsequently separated by centrifugal
force. Because of these advantages, CCS is very suitable for
continuous separation of enantiomers by ELLE. CCS has
been successfully used for the synthesis and refining of
biodiesel,24 for separation of ionic liquid dispersions,25 for
oil-water separation,26 and for several other extraction pro-
cesses.27,28 Although the separation of enantiomers by multi-
stage ELLE in CCS has attracted increasing attention in
recent years, studies on experimental and modeling of the
multistage ELLE have been rarely reported.22,29–32
Fig. 2. Chemical structure of HP-β-CD.
hydroxypropyl-β-cyclodextrin (HP-β-CD), was bought from Zhiyuan
Biotechnology (Binzhou, China), and 1,2-dichloroethane with a purity of
99.0 % was obtained from Kermel (Tianjin, China). Sodium dihydrogen
phosphate and phosphoric acid were supplied by Huihong Reagent
(Hunan, China). Solvent for chromatography was of HPLC grade. All other
chemicals used in this work were of analytical reagent grade.
2-Phenylpropionic acid (Figure 1) is an important 2-
arylpropionic acid class which is widely used to synthesize chiral
drugs such as loxoprfen. It has been revealed that 2-arylpropionic
acid enantiomers have their main pharmacological activity on the
(S)-enantiomer, while the (R)-enantiomer is inactive. Several
studies attempting to separate 2-phenylpropionic acid (2-PPA)
enantiomers have been reported, which are regularly achieved
by using capillary electrochromatography,33 synthesis,34 high-
performance liquid chromatography (HPLC) techniques, and
high-speed counter-current chromatography.35 A limitation of
these technologies is their difficulty in large-scale production.
Thus, it is important to explore a high-efficiency technology for
large-scale production of a single 2-PPA enantiomer.
In this work, we report the process optimization of continu-
ous separation of 2-PPA enantiomers using ELLE in CCS de-
vices. HP-β-CD (Figure 2) was used as the chiral selector. It
has been regarded as a very promising extractant and has
been widely investigated because of the unique physical prop-
erties and nontoxic side effects. A multistage counter-current
ELLE equilibrium model for separation of 2-PPA enantiomers
was built. Important process parameters such as the extract
phase/washing phase ratio (W/O), the concentration of chiral
selector, the pH value of aqueous phase, and the number of
stages were investigated. Additionally, the model was further
used to predict and optimize the process.
Experiment Setup
Multistage extraction experiments were performed in a cascade of cen-
trifugal contactor separators, which were connected in line by tubes. The
CINC V02 separators were applied, in which the rotor diameter was
5.1 cm and the height of the annulus was 7 cm. The CCSs were settled
in a thermostatic water bath containing circulated cooling water to main-
tain a stable temperature of 278 K. The rotational speed of the CCSs was
set at a stirring rate of 3000 rpm. Both aqueous phase and organic phase
were transferred to the CCSs by constant flow pumps. Flow rate of the ex-
tract phase (water phase, W) was set at 6 mL minꢀ1 unless otherwise
specified. The effluent was collected in glass containers.
Multistage Extraction Experiments
The chiral extraction system was established by adding HP-β-CD in
aqueous phase as chiral selector, which preferentially recognizes the
(R)-enantiomer. The extract phase (aqueous phase) was prepared by dis-
solving HP-β-CD in 0.1 mol Lꢀ1 NaH2PO4/ H3PO4 buffer solution, race-
mic 2-PPA was dissolved in 1,2-dichloroethane to obtain the feeding
phase, and the raffinate phase (organic phase) was pure 1,2-
dichloroethane. Multistage extraction experiments were performed by
starting the engines of all CCSs, opening the tap of the cooling water flow,
and starting the heavy phase (organic phase) pump. After starting the
heavy pump, the CCSs were filled up in the order from Stage 1 to Stage
N. After the organic phase ran from heavy phase outlet (Stage N), the
pump of the light phase (aqueous phase) was started. Only when the
aqueous phase ran from light phase outlet could the feed pump be started.
As soon as the feed pump started running, samples from the aqueous
phase outlet were taken every 25 min. The concentrations of (S)-2-PPA
and (R)-2-PPA in the aqueous phase outlet were analyzed by HPLC, while
that in organic phase outlet were calculated by the law of the conservation
of mass. When the concentrations of (S)-2-PPA and (R)-2-PPA in the aque-
ous phase outlet were no longer changed, the experiment was stopped
and each CCS acted as a single equilibrium stage at that moment.
MATERIALS AND METHODS
Materials
Racemic 2-phenylpropionic acid was purchased from J&K (Beijing,
China), with
a
purity ≥98%. The hydrophilic extractant,
Analytical Method
The quantification of 2-PPA enantiomers in the aqueous outlet was de-
Fig. 1. Chemical structure of 2-phenylpropionic acid.
Chirality DOI 10.1002/chir
termined by HPLC (Waters e2695 Separation Module), and a UV detector

CONTINUOUS SEPARATION OF 2-PPA ENANTIOMERS
237
(Waters 2998 Photodiode Array Detector) with a UV wavelength of
254 nm. The column was Inertsil ODS-3 C18, 5 μm particle size of the
packing material, 250 mm × 4.6 mm I.D. (GL Sciences, Tokyo, Japan).
The mobile phase was glacial acetic acid: methanol: 0.025 mol Lꢀ1 HP-β-
CD aqueous buffer solution (pH = 4.0, adjusted with triethylamine)
(0.5:15:85, v/v/v).35,36 The operating conditions were as follows: flow
rate, 0.8 mL minꢀ1; column temperature, 45 °C; injection volume, 10 μL.
The retention time of (R)-2-PAA was less than that of (S)-2-PPA.
where [AR]aq and [AS]aq are the concentrations of the free (R)-
and (S)-2-PPA in aqueous phase at equilibrium, respectively;
[AR]org and [AS]org are the concentrations of the free (R)-
and (S)-2-PPA in organic phase at equilibrium, respectively.
The subscript j indicates the sequence number of the stages.
The physical partition coefficient of ionic (R)-2-PPA and (S)-
2-PPA is defined as:
Â
Ã
Â
Ã
ARꢀ
ARꢀ
ASꢀ
ASꢀ
aq;j
aq;j
THEORY AND MODEL
Theory of Enantioselective Extraction Mechanism
Â
Ã
Â
Ã
Pi ¼
¼
(2)
org;j
org;j
ELLE is based on the selective recognition of one of the en-
antiomers by a chiral selector. The mechanism of ELLE by
HP-β-CD is schematically shown in Figure 3. In the system
for enantioselective extraction of 2-PPA enantiomers with
HP-β-CD as selective extractor, HP-β-CD was dissolved in
the aqueous phase, which preferentially recognized the (R)-
enantiomer rather than the (S)-enantiomer, indicated by a se-
ries of experiments. Depending on the solubility of 2-PPA en-
antiomers in the organic phase and the aqueous phase, the
location of complexation reaction will either be in the organic
phase, in the aqueous phase, or at the interface. Due to the
fact that HP-β-CD is highly hydrophilic and insoluble in the
organic phase, the possibility of that the reaction will take place
in the organic phase is excluded. Therefore, the complexation re-
actions will take place either in the aqueous phase or at the inter-
face. The solutes of 2-PPA enantiomers can distribute over the
organic and aqueous phases. Thus, the homogeneous aqueous
phase reaction mechanism was applied here and is depicted it
in Figure 3.
where [Aꢀ]aq and [Aꢀ]aq are the concentrations of the ionic (R)-
and (S)-2R-PPA in aqSueous phase at equilibrium, respectively;
[Aꢀ]org and [Aꢀ]org are the concentrations of the ionic (R)-
R
S
and (S)-2-PPA in organic phase at equilibrium, respectively.
The selective recognition takes place in the aqueous phase
depending on the complexation reaction. The complexation
constant KR and KS can be calculated according to Eqs. 3
and 4:
½AR ꢀ HP ꢀ β ꢀ CDꢁaq
KR
¼
(3)
½ARꢁaq½HP ꢀ β ꢀ CDꢁaq
½AS ꢀ HP ꢀ β ꢀ CDꢁaq
½ASꢁaq½HP ꢀ β ꢀ CDꢁaq
KS ¼
(4)
where [AR-HP-β-CD]aq and [AS-HP-β-CD]aq represent the
concentrations of the complexes between (R)-2-PPA and
HP-β-CD and between (S)-2-PPA and HP-β-CD at equilibrium,
respectively.
The dissociation constant of (R)-2-PPA and (S)-2-PPA in the
aqueous phase can be calculated according to Eq. 5:
The physical distribution of two enantiomers between the
two phases is unselective and equal. It is described by the ra-
tio of the two enantiomer concentrations within the aqueous
and organic phase. The physical partition coefficient of molec-
ular (R)-2-PPA and (S)-2-PPA is defined as:
Â
Ã
Â
Ã
Â
Ã
Â
Ã
Hþ
ARꢀ
Hþ
ASꢀ
½ARꢁaq;j
½ASꢁaq;j
aq;j
aq;j
aq;j
aq;j
P0 ¼
¼
(1)
Ka ¼
¼
(5)
½ARꢁorg;j ½ASꢁorg;j
½ARꢁaq;j
½ASꢁaq;j
Fig. 3. Diagram of the mechanism of reactive extraction of 2-PPA enantiomers by HP-β-CD (i = R or S).
Chirality DOI 10.1002/chir

238
Multistage Countercurrent Equilibrium ELLE Model
FENG ET AL.
The overall component mass balances for the enantiomers
Ai is defined as:
A flow diagram of a cascade for the separation of racemic
2-PPA enantiomers into a single enantiomer, (R)-2-PPA and
(S)-2-PPA, is depicted in Figure 4. The racemic 2-PPA in 1,2-
dichloroethane is fed to the CCSs at the feed stage, indicated
with f. The 1,2-dichloroethane (organic phase) was pumped
into the CCSs at the first stage. Dissolving HP-β-CD in
0.1 mol Lꢀ1 NaH2PO4/H3PO4 buffer solution, the aqueous
phase is pumped into the CCSs at the last stage, indicated
with N. The stages f to N form the stripping section, where
(R)-2-PPA (AR) is preferentially extracted to the extract phase
(aqueous phase). The co-extracted (S)-2-PPA (AS) is washed
out of the extract stream in stages 1 to f-1 (wash section). Af-
ter multistage extraction, the AR is primarily left in the aque-
ous phase and, consequently the AS is predominantly
enriched in the organic phase.
ꢀ
ꢁ
Â
Ã
F½Aiꢁ ¼ W ½Aiꢁaq;N þ Aꢀi aq;N þ ½Ai ꢀ HP ꢀ β ꢀ CDꢁaq;N (9)
0
ꢀ
ꢁ
Â
Ã
þðO þ FÞ ½Aiꢁorg;1 þ Aiꢀ
org;1
The component mass balance for HP-β-CD in aqueous
phase is given by Eq. 10.
½HP ꢀ β ꢀ CDꢁ0 ¼ ½HP ꢀ β ꢀ CDꢁaq;j þ ½AR ꢀ HP ꢀ β ꢀ CDꢁaq;j
þ½AS ꢀ HP ꢀ β ꢀ CDꢁaq;j ¼ ½HP ꢀ β ꢀ CDꢁaq;jþ1
þ½AR ꢀ HP ꢀ β ꢀ CDꢁaq;jþ1 þ ½AS ꢀ HP ꢀ β ꢀ CDꢁaq;jþ1
(10)
The component balances for AR or AS in wash section
(j = 1…f-1), can be written as follows (the subscript, i,
represents R or S):
where [HP-β-CD]0 is the initial concentration of HP-β-CD in
the aqueous phase.
The enantiomeric excess (ee) is used as a measure of the
optical purity of the raffinate and the extract. The ee of
2-PPA in the extract and raffinate can be calculated by:
ꢀ
ꢁ
Â
Ã
Â
Ã
O ½Aiꢁorg;jꢀ1 þ Aꢀi
þ Wð½Aiꢁaq;jþ1 þ Aiꢀ
org;jꢀ1
aq;jþ1
ꢀ
ꢁ
Â
Ã
all
all forms
aq
þ½Ai ꢀ HP-β-CDꢁaq;jþ1Þ ¼ O ½Aiꢁorg;j þ Aiꢀ
½ARꢁ forms-½ASꢁ
aq
org;j
eeR;extract
¼
(11)
(12)
all forms
all forms
ꢀ
ꢁ
Â
Ã
½ARꢁaq
þ½ASꢁaq
-
þW ½Aiꢁaq;j þ Ai aq;j þ ½Ai ꢀ HP ꢀ β ꢀ CDꢁaq;j
all
all forms
½ASꢁ forms-½ARꢁ
org
org
all forms
(6)
eeS;raffinate
¼
all forms
½ASꢁorg
þ½ARꢁorg
For the feed stage, the component balances for Ai is
Besides the ee, another important parameter is the yield, Y.
The yield of (R)-2-PPA and (S)-2-PPA are, respectively,
defined as:
defined as:
ꢀ
ꢁ
Â
Ã
W ½Aiꢁaq;fþ1 þ Aꢀi aq;fþ1 þ ½Ai ꢀ HP ꢀ β ꢀ CDꢁaq;fþ1
ꢀ
ꢁ
Â
Ã
þO ½Aiꢁorg;fꢀ1 þ Aiꢀ
totalARextract½molꢁ
YR;extract
¼
(13)
(14)
org;fꢀ1
totalARfeed ½molꢁ
totalASraffinate½molꢁ
(7)
ꢀ
ꢁ
Â
Ã
þF½Aiꢁ ¼ W ½Aiꢁ þ Aꢀi
þ ½Ai ꢀ HP ꢀ β ꢀ CDꢁaq;f
aq;f
aq;f
0
ꢀ
ꢁ
YS;raffinate
¼
tatalASfeed
Â
Ã
þðO þ FÞ ½Aiꢁorg;f þ Aiꢀ
½molꢁ
org;f
The multistage counter-current ELLE equilibrium model
was programmed in MatLab (MathWorks, Natick, MA).
The influence of several factors, including phase ratio, ex-
tractant concentration, the pH value of aqueous phase, and
the number of stages were modeled.
where [Ai]0 is the initial concentration of (R)- and (S)-2-PPA in
the feed stream.
The component balances for Ai in the stripping section
(j = f+1, f+2…N) can be written as follows:
ꢀ
ꢁ
Â
Ã
Â
Ã
ðO þ FÞ ½Aiꢁorg;jꢀ1 þ Aiꢀ
þ Wð½Aiꢁaq;jþ1 þ Aiꢀ
org;jꢀ1
aq;jþ1
RESULTS AND DISCUSSION
ꢀ
ꢁ
Â
Ã
þ½Ai ꢀ HP ꢀ β ꢀ CDꢁaq;jþ1Þ ¼ ðO þ FÞ ½Aiꢁorg;j þ Aiꢀ
An experimental study of the multistage ELLE of 2-PPA en-
antiomers in CCSs was performed. The racemic feed was
pumped to the cascade at the middle stage. The CCS cascade
was run successfully for 300 min at a temperature of 278 K.
(8)
org;j
ꢀ
ꢁ
Â
Ã
þW ½Aiꢁaq;j þ Aꢀi aq;j þ ½Ai ꢀ HP ꢀ β ꢀ CDꢁaq;j
Fig. 4. Flow diagram of the multistage centrifugal counter-current extraction of 2-PPA enantiomers.
Chirality DOI 10.1002/chir

CONTINUOUS SEPARATION OF 2-PPA ENANTIOMERS
239
The concentrations of (R)- and (S)-2-PPA enantiomers in the
aqueous outlet were detected every 25 min. For exploration
of the equilibrium time, the development of the concentra-
tions of the enantiomers in a particular experiment versus
runtime is depicted in Figure 5. As shown in Figure 5, the
concentrations of enantiomers in aqueous outlet increase
gradually in the first 125 min, and then fluctuated within a nar-
row range. Furthermore, the concentration of the (R)-enantio-
mer remains higher than the (S)-enantiomer. The results
indicate that the time needed for the aqueous concentration
to achieve a steady state is about 125 min. Thus, it can be
suggested that the time needed for chemical and physical
equilibrium is relatively short.
In order to verify the multistage ELLE equilibrium model
and better understand the relationship between process pa-
rameters and extraction performance, the influence of several
parameters on the multistage separation process, including
phase ratio (W/O), extractant concentration, pH value of
aqueous phase, and the number of stages were investigated
by experiment and modeling.
Influence of W/O Ratio
In the multistage extraction system, the W/O ratio has a
significant influence on purity and yield. In order to investi-
gate the influence of the W/O ratio on extraction perfor-
mance, a series of extraction experiments were performed
with a W/O ratio in the range from 0.2 to 1.2 at W/F = 6
(F represents flow of feeding phase), pH = 2.5, [HP-β-CD] =
0.1 mol Lꢀ1, [AR,S] = 0.02 mol Lꢀ1, T = 278 K, and N = 10
(N represents number of stages). A comparison of the exper-
imental data with the model predictions of ee and yield is
shown in Figure 6. It is observed from Figure 6 that the exper-
imental data of ee and yield are in good agreement with the
model predictions. The mean relative deviations between
model and experiment were 7.69% for ee and 3.43% for Y. As
shown in Figure 6a, the ee decreases gradually in the extract
phase (aqueous phase) with the increase of W/O ratio, while
an opposite tendency is observed for the ee in the raffinate
phase (organic phase). From Figure 6b, with the increase of
W/O ratio, the yield of (R)-enantiomer in the extract phase
increases while the yield of (S)-enantiomer in the raffinate
phase decreases. The possible reasons for this may be that
with the increase of W/O, more complexes form between
Fig. 6. Influence of W/O ratio on ee and yield for separation of 2-PPA enan-
tiomers. (a) Influence on ee. (b) Influence on yield. Condition: W/F = 6, pH =
8.5, [HP-β-CD] = 0.1 mol Lꢀ1, [AR,S] = 0.02 mol Lꢀ1, T = 278 K, N = 10, feed in
the middle stage.
HP-β-CD and enantiomers in the extract phase, and the
co-extraction of the nonpreferential enantiomer is enhanced,
so the ee in the extract phase and the yield of (S)-enantiomer
in the raffinate phase are decreased. While in the raffinate
phase, with the increase of W/O, fewer enantiomers are left
in the raffinate, and the extracted (R)-enantiomer is always
more than the (S)-enantiomer, so the yield of (R)-enantiomer
in the extract phase and the ee in the raffinate phase are
increased. Furthermore, it is found that the ee and yield in
the extract phase are equal to those in the raffinate phase at
W/O = 0.6. The crosspoint with eeeq (equal enantiomeric
excess) and Yeq (equal yield) is the operating point for
symmetric separation.29
Influence of Extractant Concentration
The extractant concentration is an important parameter in
the multistage extraction process, and it is directly related
to the economic efficiency of enterprises. Therefore, it is vital
to work at a suitable extractant concentration in an industrial
production process. The influence of HP-β-CD concentration
on the purity and yield in both streams was investigated by
varying the concentration from 0 to 0.16 mol Lꢀ1 at W/O =
0.6, W/F = 6, pH = 2.5, [AR,S] = 0.02 mol Lꢀ1, T = 278 K, and
Chirality DOI 10.1002/chir
Fig. 5. Concentration of 2-PPA enantiomers in aqueous outlet.

240
FENG ET AL.
N = 10. A comparison of the experimental data with the model
predictions of purity and yield is shown in Figure 7. The mean
relative deviations between model and experiment were 8.98%
for ee and 10.38% for Y. It is clear from Figure 7a that the ee in
the extract phase increases rapidly when the HP-β-CD
concentration increases from 0 to 0.05 mol Lꢀ1, and then de-
creases with a further increase in HP-β-CD concentration.
The ee in the raffinate phase maintains a rising trend with
the increase of the HP-β-CD concentration. This phenomenon
can be explained as follows: At first, the extractant concentra-
tion in the aqueous phase is very small and the solubility of
2-PPA in the aqueous phase is poor, which results in that
most of the enantiomers are in the raffinate phase (Figure 7b),
thus the ee values in the extract and raffinate phase are very
small. With the concentration of HP-β-CD increasing to
Influence of pH
The pH value is one of the most important factors for con-
sideration in the separation of enantiomers, since it has a sig-
nificant influence on the existing forms of 2-PPA enantiomers.
The ee and yield of multistage ELLE of 2-PPA enantiomers
were performed with the pH value ranging from 2 to 6.5. A
comparison between the experimental results and the model
predictions of the ee and yield in both streams is shown in
Figure 8. The mean relative deviations between model and
experiment were 7.79% for ee and 4.63% for Y.
It can be seen from Figure 8a that the ee in the extract
phase fluctuates within a narrow range at pH value lower than
3.5, and then decreases with a further increase in pH value. A
similar tendency is observed for the ee in the raffinate phase.
It can also be seen from Figure 8b that the yields in both
streams fluctuate within a narrow range with the pH value
increasing from 2 to 3.5. With the further increase of the pH
value, the yield in the extract phase increases gradually, while
an opposite tendency is observed for the yield in the raffinate
phase.
0.05 mol Lꢀ1
, more and more 2-PPA enantiomers are
extracted to the aqueous phase (Figure 7b), and the ee in the
extract phase increases rapidly. This is due to the increase
of enantioselectivity with the increase of HP-β-CD concentra-
tion. With the further increase of HP-β-CD concentration, the
extractant is present in excess with respect to the desired
enantiomer and a considerable amount of undesired enantio-
mers are also extracted to the extract phase, which leads to
the decrease of ee in the extract phase and yield in raffinate.
The possible reasons for this may be that HP-β-CD mainly
has a chiral recognition ability and affinity for molecular
2-PPA, but not for ionic 2-PPA. When the pH value is lower
Fig. 7. Influence of HP-β-CD concentration on ee and yield for separation of
2-PPA enantiomers. (a) Influence on ee. (b) Influence on yield. Condition: W/
O = 0.6, W/F = 6, pH = 2.5, [AR,S] = 0.02 mol Lꢀ1, T = 278 K, N = 10, feed in the
middle stage.
Fig. 8. Influence of pH value on ee and yield for separation of 2-PPA enan-
tiomers. (a) Influence on ee. (b) Influence on yield. Conditions: W/O = 0.6,
W/F = 6, [HP-β-CD] = 0.1 mol Lꢀ1, [AR,S] = 0.02 mol Lꢀ1, T = 278 K, N = 10,
feed in the middle stage.
Chirality DOI 10.1002/chir

CONTINUOUS SEPARATION OF 2-PPA ENANTIOMERS
241
than 3.5, the rising of the pH value has little influence on the
existence state of 2-PPA and the amount of molecular species
remain nearly unchanged, which leads to the ee and the yield
in both streams to fluctuate within a narrow range at pH
values lower than 3.5. With the increase of pH value from
3.5 to 6.5, more and more molecular 2-PPA turned into ionic
2-PPA and entered the extract phase, and the amount of
complexes formed by extractant and enantiomers decreases.
As a result, the ee in both streams and the yield in the
raffinate phase decrease, while the yield in the extract phase
increase. Thus, it should be kept at a low pH value to carry
out the extraction process.
Influence of Number of Stages
In the CCS cascade, the number of stages has a great effect
on ee and yield. High enantiomeric purity and yield are
demanded for the production of a chiral drug. The influence
of the number of stages on ee and yield was modeled at 10
and 20 stages. Experiments were performed by changing
the W/O ratio of 0.4, 0.6, and 0.8 at 10 and 20 stages, respec-
tively. Experimental results and model predictions are shown
in Figure 9. It can be observed from Figure 9 that the eeeq and
Yeq are obviously increased after the number of stages
changed from 10 stages to 20 stages. According to the exper-
imental results, the eeeq can reach 37% and the Yeq can reach
69% at a W/O ratio of 0.6 and N = 20, which are basically in
conformity with the simulation. The mean relative deviations
between model and experiment were 8.19% for ee and 7.55%
for Y. Therefore, higher eeeq and Yeq can be obtained with
more extraction stages.
Modeling and Optimization of the Mutilstage Extraction
System
The results of the above studies show that CCS device is
suitable for multistage counter-current ELLE. Comparison of
the model predictions with the experimental results indicates
that the established multistage equilibrium model is a good
way of predicting the extract efficiency for separation of
2-PPA enantiomers in CCSs over a range of experimental
conditions. Thus, we used the model to explore the influence
of various operating parameters on extraction performance in
the multistage extraction system to predict and optimize the
separation process.
Fig. 9. Influence of number of stages on ee and yield for separation of
2-PPA enantiomers. (a) Influence on ee. (b) Influence on yield. Condition:
W/F = 6, pH = 2.5,[HP-β-CD] = 0.1 mol Lꢀ1, [AR,S] = 0.02 mol Lꢀ1, T = 278 K,
feed in the middle stage.
Location of Feed Stage
The location of the feed stage is an important parameter in
a multistage extraction system. The change of the location of
feed stage could cause a change of the number of stages in
the stripping section and wash section, which could generate
a different extraction efficiency of the symmetric separation.
The eeeq and Yeq in extract phase and raffinate phase with
location of the feed at different stages is shown in Figure 10.
It can be seen from Figure 10 that both the eeeq and the Yeq
reach a maximum when the feed stage is located exactly at
the middle stage. Therefore, higher eeeq and Yeq can be
obtained by using an equal amount of stages in the stripping
section and wash section.
It can be observed from Figure 11 that the eeeq increases rap-
idly with HP-β-CD concentration increasing from 0.02 mol Lꢀ1
to 0.1 mol Lꢀ1, and then it increases slightly. When the pH
value is less than 3.0 and the HP-β-CD concentration is
greater than 0.1 mol Lꢀ1, a plateau appears. From Figure 12,
it is obvious that the eeeq at different HP-β-CD concentrations
follows a similar tendency with the increase of the pH value.
The eeeq rises rapidly with the HP-β-CD concentration
increasing from 0.025 mol Lꢀ1 to 0.10 mol Lꢀ1, and then it in-
creases slowly from 0.10 mol Lꢀ1 to 0.15 mol Lꢀ1. Within the
HP-β-CD concentration range, the eeeq keeps nearly invari-
able at a pH value less than 3.0. When the pH value is greater
than 3.0, the eeeq gradually decreases. Therefore, a pH value
between 2.0 and 3.0 is suitable for separation of 2-PPA
enantiomers. In this work, the pH value of 2.5 is the optimal
pH value during our actual operation. In view of the economic
efficiency and the solution viscosity, the HP-β-CD
Chirality DOI 10.1002/chir
Extractant Concentration and pH Value
The influence of HP-β-CD concentration and pH value on
eeeq for separation of 2-PPA enantiomers has been modeled.
Figures 11 and 12 show the eeeq for separation of 2-PPA enan-
tiomers as a function of HP-β-CD concentration and pH value.

242
FENG ET AL.
concentration is recommended to be set at about 0.1 mol Lꢀ1
,
where the separation efficiency is near its optimum.
Number of Stages and Flow Ratios
According to the experiments above, the phase ratio of
W/O and W/F have great influence on the extraction perfor-
mance. Figures 13 and 14 show the eeeq as a function of the
number of stages at different flow ratios. It can be seen from
Figure 13 that there is a significant effect of the number of
stages and flow ratios on eeeq. The eeeq increases with the in-
crease of W/F at first, and then remains nearly unchanged.
However, the eeeq decreases with the increase of the W/O ra-
tio. When the W/O ratio is lower than 0.58, the eeeq increases
slowly. It can also be seen that the eeeq rises rapidly when the
number of stages is less than 60, and then the change is slight
with a further increase of the number of stages. In order to
further optimize the W/F ratio, the relationship between the
number of stages and eeeq at a different W/F ratio has been
modeled. As shown in Figure 14, the eeeq at a different W/F
ratio follows a similar tendency with the increase of number
Fig. 10. Influence of the location of feed stage on eeeq and Yeq for Separa-
tion of 2-PPA enantiomers. Condition: W/F = 6, pH = 2.5, [HP-β-CD] =
0.1 mol Lꢀ1, [AR,S] = 0.02 mol Lꢀ1, T = 278 K, N = 10.
Fig. 13. Influence of W/O ratio and W/F ratio on eeeq for separation of 2-
PPA enantiomers at different number of stages. Condition: W/O = (0, 5),
W/F = (0, 20), [HP-β-CD] = 0.1 mol Lꢀ1, pH = 2.5, [AR,S] = 0.02 mol Lꢀ1, T =
278 K, feed in middle stage.
Fig. 11. Influence of pH and HP-β-CD concentration on eeeq for separation
of 2-PPA enantiomers. Condition: W/O = (0.01, 5), W/F = 6.0, [AR,S] =
0.02 mol Lꢀ1, T = 278 K, feed in the middle stage, N = 10.
Fig. 12. Influence of pH on eeeq for separation of 2-PPA enantiomers at dif-
ferent HP-β-CD concentration. Condition: W/O = (0.01, 5), W/F = 6.0, pH = (2,
7), [AR,S] = 0.02 mol Lꢀ1, T = 278 K, feed in the middle stage, N = 10.
Fig. 14. Influence of number of stages on eeeq for separation of 2-PPA en-
antiomers at different W/F ratio. Condition: pH = 2.5, [HP-β-CD] = 0.1 mol Lꢀ1
,
[AR, S] = 0.02 mol Lꢀ1, T = 278 K, feed in the middle stage.
Chirality DOI 10.1002/chir

CONTINUOUS SEPARATION OF 2-PPA ENANTIOMERS
243
TABLE 1. Optimized settings for symmetrical separations with
[AR,S] = 0.02 mol Lꢀ1, pH = 2.5, T = 278 K
and 106 for eeeq > 95% and eeeq > 97% at both stream exits,
respectively.
eeextract and eeraffinate
eeextract and eeraffinate
Variable
> 95% settings
>97% setings
LITERATURE CITED
N
f
98
50
106
54
1. De Camp WH. The FDA perspective on the development of stereoiso-
mers. Chirality 1989;1:2–6.
2. Hutt AJ. Drug chirality: impact on pharmaceutical regulation. Chirality
1991;3:161–164.
[HP-β-CD]
(mol Lꢀ1,)
W/F
0.10
0.10
40
0.54
40
0.55
W/O
3. Strong M. FDA policy and regulation of stereoisomers: paradigm shift and
the future of safer, more effective drugs. Food Drug LJ 1999;54:463–487.
4. Blaschke G, Kraft HP, Fickentscher K, Kohler F. Chromatographic race-
mic separation of thalidomide and teratogenic activity of its enantiomers.
Arzn Forsch 1979;29:1640–1642.
5. Gourlay MD, Kendrick J, Leusen FJJ. Predicting the spontaneous chiral
resolution by crystallization of a pair of flexible nitroxide radicals. Cryst
Growth Des 2008;8:2899–2905.
of stages. The eeeq increases greatly when the W/F ratio is
less than 10, and then the eeeq increases slowly with a further
increase of the W/F ratio. Therefore, a higher W/F ratio, a
lower W/F ratio, and a greater number of stages are required
to achieve higher eeeq.
6. Keurentjes JTF, Nabuurs L, Vegter EA. Liquid membrane technology for
the separation of racemic mixtures. J Membr Sci 1996;113:351–360.
For a symmetrical separation, the minimum number of
stages needed for eeeq >95% and eeeq >97% was calculated
by modeling and optimization at a W/F ratio of 40, pH = 2.5,
[AR,S] = 0.02 mol Lꢀ1, [HP-β-CD] = 0.10 mol Lꢀ1, T = 278 K.
The optimized settings for the two cases are shown in Table 1.
When the eeeq is higher than 95%, a cascade of 98 stages is
needed, whereas for eeeq higher than 97% a minimum of 106
stages is required. Since the enantioselectivity for 2-PPA en-
antiomers with HP-β-CD as chiral selector is moderate, a very
large number of stages are required to obtain higher eeeq.
With the optimal process parameters, the eeeq can be up to
99% with the number of stages of 130. Considering the eco-
nomical exploitation of the principle, the symmetrical separa-
tion of 2-PPA can be replaced by asymmetrical separation, in
which the flow ratios are adjusted to obtain the desired enan-
tiomer with high purity while the undesired enantiomer with
low purity. The asymmetrical separation can make the num-
ber of stages considerably reduced even though the
enantioselectivity is moderate.
7. Schurig V. Separation of enantiomers by gas chromatography.
J Chromatogr A 2001;906:275–299.
8. Schulte M, Strube J. Preparative enantioseparation by simulated moving
bed chromatography. J Chromatogr A 2001;906:5824–5828.
9. Makino K, Goto T, Hiroki Y, Hamada Y. Enantioselectively synthesis of
anti-beta-hydroxy-alpha-amino acids through dynamic kinetic resolution.
Angew Chem Int Ed 2004;43:882–884.
10. Yu T, Du Y, Chen B. Evaluation of clarithromycin lactobionate as a novel
chiral selector for enantiomeric separation of basic drugs in capillary elec-
trophoresis. Electrophoresis 2011;32:1898–1105.
11. Zhou SY, Zuo H, Stobaugh JF, Lunte CE, Lunte SM. Continuous in vivo
monitoring of amino acid neurotransmitters by microdialysis sampling
with online derivatization and capillary electrophoresis separation. Anal
Chem 1995;67:594–599.
12. Schuur B, Verkuijl BJV, Minnaard AJ, de Vries JG, Heeres HJ, Feringa
BL. Chiral separation by enantioselective liquid–liquid extraction. Org
Biomol Chem 2011;9:36–51.
13. Schuur B, Verkuijl BJV, Bokhove J, Minnaard AJ, de Vries JG, Heeres HJ,
Feringa BL. Enantioselective liquid-liquid extraction of (R,S)-
phenylglycinol using a bisnaphthyl phosphoric acid derivative as chiral
extractant. Tetrahedron 2011;67:462–470.
14. Lacour J, Goujon-Ginglinger C, Torche-Haldimann S, Jodry JJ. Efficient
enantioselective extraction of tris (diimine) ruthenium(II) complexes by
chiral, lipophilic TRISPHAT anions. Angew Chem Int Ed
2000;39:3695–3697.
CONCLUSION
15. Viegas RMC, Afonso CAM, Crespo JG, Coelhoso IM. Modelling of the
enantio-selective extraction of propranolol in a biphasic system. Sep Purif
Technol 2007;53:224–234.
16. Tang KW, Chen YY, Huang KL, Liu JJ. Enantioselective resolution of chi-
ral aromatic acids by biphasic recognition chiral extraction. Tetrahedron:
Asymmetry 2007;18:2399–2408.
17. Jiao F, Chen XQ, Hu WG, Ning FR, Huang KL. Enantioselective extrac-
tion of mandelic acid enantiomers by L-dipentyl tartrate and beta-
cyclodextrin as binary chiral selectors. Chemcal Papers 2007;61:326–328.
18. Steensma M, Kuipers NJM, de Haan AB, Kwant G. Identification of
enantioselective extractants for chiral separation of amines and
aminoalcohols. Chirality 2006;18:314–328.
Multistage enantioselective liquid–liquid extraction in a
centrifugal contactor was performed for the separation of
2-PPA enantiomers with HP-β-CD as chiral extractant. The
experimental results show that the performance of extrac-
tion is significantly influenced by the process parameters
such as the extract phase/washing phase ratio, extractant
concentration, the pH value of the aqueous phase, and the
number of stages. A multistage equilibrium model of ELLE
was built on the basis of a single-stage model for chiral
extraction of the 2-PPA enantiomers and the law of mass
conservation. The model has been verified experimentally
and provides a good means for fast optimization of the
operational conditions.
The performance of the multistage separation process was
evaluated by ee and yield. Higher eeeq can be achieved at a
higher W/F ratio, lower W/O ratio, and greater number of
stages. The best process conditions were obtained by model-
ing and optimization at a W/O ratio of 0.6, pH of 2.5, and
HP-β-CD concentration of 0.1 mol Lꢀ1 at 278 K. The optimal
operational eeeq of 37% and Yeq of 69% was close to the model
predictions. By modeling and optimization, the minimum
number of stages for full separation was calculated as 98
19. Dimitrova P, Bart HJ. Extraction of amino acid enantiomers with
microemulsions. Chem Eng Technol 2009;32:1527–1534.
20. Verkuijl BJV, Minnaard AJ, de Vries JG, Feringa BL. Chiral separation of
underivatized amino acids by reactive extraction with palladium-BINAP
complexes. J Org Chem 2009;74:6526–6533.
21. Tang KW, Zhang PL, Pan CY, Li HJ. Equilibrium studies on
enantioselective extraction of oxybutynin enantiomers by hydropholic
β-cyclodextren derivatives. AIChE J 2011;57:3027–3036.
22. Tang KW, Zhang H, Liu Y. Experimental and simulation on
enantioselective extraction in centrifugal contactor separators. AIChE J
2013;59:2594–2602.
23. Koska J, Haynes CA. Modelling multiple chemical equilbria in chiral par-
tition systems. Chem Eng Sci 2001;56:5853–5864.
Chirality DOI 10.1002/chir

244
FENG ET AL.
24. Abduh MY, van Ulden W, van de Bovenkamp HH, Buntara T, Picchioni F,
Manurung R, Heeres HJ. Synthesis and refining of sunflower biodiesel in
a cascade of continuous centrifugal contactor separators. Eur J Lipid Sci
Technol 2015;117:242–254.
31. Tang KW, Wang YQ, Zhang P, Huang Y, Dai G. Process optimization of
continuous liquid-liquid extraction in centrifugal contactor separators
for separation of oxybutynin enantiomers. Sep Purif Technol
2015;150:170–178.
25. Birdwell JF, McFarlane J, Hunt RD, Luo H, DePaoli DW, Schuh DL, Dai S.
Separation of ionic liquid dispersions in centrifugal solvent extraction
contactors. Sep Sci Technol 2006;41:2205–2223.
26. Klasson KT, Taylo PA, Walker JF, Jones SA, Cummins RL, Richardson
SA. Modification of a centrifugal separator for in-well oil-water separation.
Sep Sci Technol 2005;40:453–462.
27. Zhou JZ, Duan WH, Zhou XZ, Zhang CQ. Extraction of hydrocortisone
from the fermentation liquor with annular centrifugal contactors. Sep Sci
Technol 2006;41:573–581.
28. Duan WH, Chen J, Wang JC, Wang SW, Feng XG, Wang XH, Li SW, Xu C.
Application of annular centrifugal contactors in the hot test of the
improved total partitioning process for high level liquid waste. J Hazard
Mater 2014;278:566–571.
32. Zhang PL, Zhang H, Tang KW, Yi JM, Huang Y. Influence of pH on
enantioselective extraction of aromatic acid enantiomers in centrifugal
contactor separators: Experiments and simulation. Sep Purif Technol
2015;141:68–75.
33. Bruggemann O, Freitag R, Whitcombe MJ, Vulfson EN. Comparison of
polymer coatings of capillaries for capillary electrophoresis with respect
to their applicability to molecular imprinting and electrochromatography.
J Chromatogr A 1997;781:43–53.
34. Garcia M, del Campo C, Sanchez-Montero JM, Sinisterra JV.
Chemoenzymatic synthesis of pure enantiomeric 2-aryl propionic acids.
Tetrahedron 1993;49:8433–8440.
35. Tong SQ, Zheng Y, Yan JZ. Application and comparison of high
performance liquid chromatography and high speed counter-current
chromatography in enantioseparation of ( )-2-phenylpropionic acid.
J Chromatogr A 2013;1281:79–86.
29. Schuur B, Winkelman JGM, deVries JG, Heeres HJ. Experimental and
modeling studies on the enantio-separation of 3,5-dinitrobenzoyl-(R),
(S)-leucine by continuous liquidꢀliquid extraction in
a
cascade of
36. Zheng Y, Yan JZ, Tong SQ, Li XN, Chu C. High performance liquid
chromatographic enantioseparation of 2-phenylpropionic acid chiral
mobile phase additive. Chin J Pharm Anal 2013;33:827–830.
centrifugal contactor separators. Chem Eng Sci 2010;65:4682–4690.
30. Tang KW, Wen P, Zhang PL, Huang Y. Fractional reactive extraction for
symmetrical separation of 4-nitro-D, L-phenylalanine in centrifugal
contactor separators: experiments and modeling. Chirality 2015;27:75–81.
Chirality DOI 10.1002/chir